The Development of a pharmacist model of perceived responsibility for drug therapy outcomes using the triangle model of ...

MISSING IMAGE

Material Information

Title:
The Development of a pharmacist model of perceived responsibility for drug therapy outcomes using the triangle model of responsibility
Physical Description:
Book
Creator:
Planas, Lourdes Gladys
Publication Date:

Record Information

Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 28002824
oclc - 53949625
System ID:
AA00022646:00001

Table of Contents
    Title Page
        Page i
        Page ii
    Dedication
        Page iii
    Acknowledgement
        Page iv
        Page v
    Table of Contents
        Page vi
        Page vii
        Page viii
        Page ix
    List of Tables
        Page x
        Page xi
    List of Figures
        Page xii
    Abstract
        Page xiii
        Page xiv
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
    Chapter 2. Theoretical model
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
    Chapter 3. Review of literature
        Page 32
        Page 33
        Page 34
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
    Chapter 4. Methods
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
    Chapter 5. Results
        Page 99
        Page 100
        Page 101
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
        Page 115
        Page 116
        Page 117
        Page 118
        Page 119
        Page 120
        Page 121
        Page 122
    Chapter 6. Discussion and conclusions
        Page 123
        Page 124
        Page 125
        Page 126
        Page 127
        Page 128
        Page 129
        Page 130
        Page 131
        Page 132
        Page 133
        Page 134
        Page 135
        Page 136
        Page 137
        Page 138
    Appendix A. Pretest pharmacist drug therapy questionnaire
        Page 139
        Page 140
        Page 141
        Page 142
        Page 143
        Page 144
        Page 145
        Page 146
        Page 147
        Page 148
        Page 149
        Page 150
        Page 151
        Page 152
        Page 153
        Page 154
    Appendix B. Text of pretest survey cover letter
        Page 155
    Appendix C. Text of prenotice letter
        Page 156
    Appendix D. Text of screening postcard
        Page 157
    Appendix E. Revised pharmacist drug therapy questionnaire
        Page 158
        Page 159
        Page 160
        Page 161
        Page 162
        Page 163
        Page 164
        Page 165
        Page 166
        Page 167
        Page 168
        Page 169
        Page 170
        Page 171
    Appendix F. Text of survey cover letter
        Page 172
    Appendix G. Text of reminder/thank you postcard
        Page 173
    References
        Page 174
        Page 175
        Page 176
        Page 177
        Page 178
        Page 179
    Biographical sketch
        Page 180
        Page 181
        Page 182
Full Text










THE DEVELOPMENT OF A PHARMACIST MODEL OF PERCEIVED
RESPONSIBILITY FOR DRUG THERAPY OUTCOMES USING THE TRIANGLE
MODEL OF RESPONSIBILITY
















By

LOURDES GLADYS PLANS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001



























Copyright 2001

by

Lourdes Gladys Planas














For my father, Alberto Eduardo Planas














ACKNOWLEDGMENTS

I wish to express my gratitude to the Office of Graduate Minority Programs at the

University of Florida and to the American Foundation for Pharmaceutical Education for

their financial support during my graduate program. I thank the University of Florida

College of Pharmacy Liberty Fund and the Perry A. Foote Program in Health Outcomes

and Pharmacoeconomics for supporting this study. I also extend many thanks to the

pharmacists who participated in this study.

My sincere appreciation goes to my major advisor, Dr. Carole Kimberlin, for her

valuable comments and also for her support and confidence in me; to Dr. David

Brushwood for his enthusiasm and openness; to Dr. Charles Hepler for his inspiring ideas

and exceptional wisdom; to Dr. Richard Segal for his thoughtful guidance and unique

sense of humor; and to Dr. Barry Schlenker for his insightfulness and generosity. I shall

always be grateful to these individuals, as well as to Dr. Donna Berardo, Dr. Earlene

Lipowski, and Dr. L. Douglas Ried for enriching my graduate student experience and for

guiding me to the realization that, as Marcel Proust once wrote, "the real voyage of

discovery consists not in seeking new landscapes but in having new eyes."

I wish to thank my fellow graduate students, past and present, for their collegiality

and friendship, especially Ratanapom Awiphan, Yajaira Bastardo, Barbara Brice, Debbie

Wilson, and Avin Yaldo.

My profound gratitude goes to my father, Alberto Eduardo Planas, for his

examples of perseverance and love for learning; to my mother, Zenaida Planas, for her








unyielding love and support; and to my grandparents, Buenaventura and Candida Planas,

for always believing in me. Finally, my love and appreciation go to my wonderful

husband, Rick Thomrnhill, for his continuous support and devotion. Without him, my

dream would not have become a reality.















TABLE OF CONTENTS

page

A CKN O W LED G M EN TS ................................................................................................. iv

LIST O F TAB LES.............................................................................................................. x

LIST O F FIGU RES .......................................................................................................... xii

ABSTRA CT..................................................................................................................... xiii

CHAPTERS

1 IN TRO D U CTION ........................................................................................................... 1

Problem Statem ent................................................................................ .... ................... 1
Background ..................................................................................................................... 2
Drug-Related M orbidity and M ortality....................................................................... 2
Preventable D rug-Related M orbidity and M ortality................................................... 4
Role of Pharmacists in Preventing Drug-Related Morbidity and Mortality ............... 6
Implementation of pharmaceutical care in community and ambulatory practice... 7
Pharmacist responsibility for drug therapy outcomes............................................. 8
Study Objectives......................................................................................................... 9
Conceptual Foundation................................................................................................... 9
Significance of the Study.............................................................................................. 10

2 TH EO RETICA L M O D EL............................................................................................. 12

Overview ....................................................................................................................... 12
Description of the Triangle Model of Responsibility................................................... 12
A ccountability........................................................................................................... 12
The Elem ents ............................................................................................................ 13
Responsibility ........................................................................................................... 16
The Linkages............................................................................................................. 16
Prescription-event link.......................................................................................... 17
Identity-event link................................................................................................. 17
Prescription-identity link ...................................................................................... 18
Studies U sing the Triangle M odel of Responsibility.................................................... 19





vi








3 REV IEW OF LITERA TURE ........................................................................................ 32

Overview ....................................................................................................................... 32
Responsibility in Social Psychology...................................................................... .. 32
Introduction ............................................................................................................... 32
H eider's Levels of Responsibility ................................... .......................................... 33
Attribution Theory ................................................. ................................................... 35
Limitations of Attribution Theories for Attribution of Responsibility ..................... 39
Social Roles ............................................................... ............................................... 42
D discussion of Responsibility in Social Psychology .................................................. 44
Responsibility in Legal Philosophy .............................................................................. 45
Introduction ............................................................................................................... 45
Hart's Types of Responsibility ...................................... .......................................... 45
Comparison of Legal and Social Psychological Concepts of Responsibility ......... 48
Responsibility in Pharm acy ........................................... ............................................... 50
Introduction ...............................................................................................................50
Conceptualization of Responsibility in Pharm acy .................................................... 50
M easurem ent of Responsibility in Pharm acy ........................................................... 55
D discussion of Responsibility in Pharm acy ............................................................... 59
Application of Triangle Model of Responsibility to Pharmacist Responsibility ...... 61
Research Questions ....................................................................................................... 63

4 M ETH OD S ...................................................................................................................... 65

Overview ....................................................................................................................... 65
Survey Developm ent ......................................................... ............................................ 66
Pharm acist Interview s ............................................. .................................................. 66
Content Validity, Readability, and Face Validity ..................................................... 67
Survey Pretest .................................................................................................. ......... 67
Description of Pharmacist Drug Therapy Questionnaire ...................................... 69
Survey scenarios ......................................................................... .......................... 69
Operationalization of study constructs....................... ........................................... 72
Sum m ary of Survey Developm ent ................................................... ..................... 84
Study Sam ple................................................................................................. ............... 85
D ata Collection ............................................................................................................. 87
Data Analysis ............ ............................................................ ......... ............................... 88
Reliability of M measures ............................................................................................. 88
Convergent and Discrim inant Validity..................................................................... 88
Construct V alidity..................................................................................................... 91
H ypotheses 1 through 4 ......................................... ............................................... 92
Hypotheses 5 through 8 ................... ..................................................................... 96
Sam ple Size D eterm nation ....................................................................................... 97
Sum m ary of M ethods .................................................................................................... 98






vii








5 RESU LTS ......................................................................................................... ............. 99

Response to Survey..................................................................................... .............. 99
Item Responses and Scale Reliability......................................................................... 101
Clarity of Specific Standards.................................................................................. 102
General Clarity of Standards................................................................................... 103
Personal Control...................................................................................................... 104
Professional Duty.................................................................................................... 105
Responsibility for Drug Therapy Outcom es........................................................... 106
Event Im portance.................................................................................................... 107
Pharm aceutical Care Provision............................................................................... 107
Convergent and D iscrim inant Validity....................................................................... 108
Test of H ypotheses...................................................................................................... 111
Hypotheses 1 Through 4......................................................................................... I11
H ypotheses 5 Through 8 ......................................................................................... 119
Sum m ary of Results.................................................................................................... 121

6 D ISCU SSION AND CON CLU SION S ....................................................................... 123

Overview ..................................................................................................................... 123
D discussion of Findings................................................................................................ 123
Prediction of Responsibility for Drug Therapy Outcomes..................................... 123
Prediction of Pharm aceutical Care Provision ......................................................... 126
Differences between Advanced and Non-Advanced Pharmacists.......................... 127
Study Lim itations........................................................................................................ 128
Variable Design and M easurem ent......................................................................... 128
Generalizability....................................................................................................... 129
Im plications of Findings ............................................................................................. 130
Practical Im plications for Pharm acy....................................................................... 130
Theoretical Im plications ......................................................................................... 132
Pharm acy adm inistration .................................................................................... 132
Social psychology............................................................................................... 132
Recom m endations for Future Research...................................................................... 134
Clarity of Standards ................................................................................................ 134
Personal Control...................................................................................................... 135
Professional D uty.................................................................................................... 135
M manipulated Scenarios............................................................................................ 136
Advanced Pharm acists............................................................................................ 136
Prediction of Pharm acist Behavior......................................................................... 137
Conclusion .................................................................................................................. 137

APPENDICES

A PRETEST PHARMACIST DRUG THERAPY QUESTIONNAIRE........................ 139

B TEXT OF PRETEST SURVEY COVER LETTER ................................................... 155








C TEXT OF PRENOTICE LETTER.............................................................................. 156

D TEXT OF SCREENING POSTCARD....................................................................... 157

E REVISED PHARMACIST DRUG THERAPY QUESTIONNAIRE ........................ 158

F TEXT OF SURVEY COVER LETTER...................................................................... 172

G TEXT OF REMINDER/THANK YOU POSTCARD................................................. 173

R EFER EN C E S ................................................................................................................ 174

BIOGRAPH ICA L SKETCH ........................................................................................... 180














LIST OF TABLES


Table page

4-1. Standards for Identifying, Resolving, and Preventing Drug Therapy Problems ............73

4-2. General Clarity of Standards Items from Pretest: Inter-Item Correlations, Corrected
Item-Total Correlations, Means, and Standard Deviations............................................. 75

4-3. Personal Control Items from Pretest: Inter-Item Correlations, Corrected Item-Total
Correlations, Means, and Standard Deviations............................................................... 77

4-4. Professional Duty Items from Pretest: Inter-Item Correlations, Corrected Item-Total
Correlations, Means, and Standard Deviations............................................................... 78

4-5. Responsibility Items from Pretest: Inter-Item Correlations, Corrected Item-Total
Correlations, Means, and Standard Deviations............................................................... 80

4-6. Domains and Items of the Behavioral Pharmaceutical Care Scale-Direct Patient
C are .................................................................................................................................83

5-1. Demographic and Practice Characteristics of Respondents.......................................... 100

5-2. Education, Training, and Certification Characteristics of Respondents....................... 101

5-3. Clarity of Specific Standards ........................................................................................ 102

5-4. General Clarity of Standards Items: Inter-Item Correlations, Corrected Item-Total
Correlations, Means, and Standard Deviations............................................................. 103

5-5. Personal Control Items: Inter-Item Correlations, Corrected Item-Total Correlations,
M eans, and Standard Deviations................................................................................... 104

5-6. Professional Duty Items: Inter-Item Correlations, Corrected Item-Total
Correlations, Means, and Standard Deviations............................................................. 105

5-7. Responsibility Items: Inter-Item Correlations, Corrected Item-Total Correlations,
M eans, and Standard Deviations................................................................................... 106








5-8. Correlations, Means, and Standard Deviations of Summated Scales Used in Path
Model Predicting Responsibility for Drug Therapy Outcomes and BPCS-DPC
Scores............................................................................................................................ 112

5-9. Collinearity Diagnostics for the Regression of Responsibility for Drug Therapy
O utcom es ...................................................................................................................... 112

5-10. Collinearity Diagnostics for the Regression of BPCS-DPC Scores.............................113

5-11. Regression Equations and Path Coefficients for Hypotheses 1 through 4 ................... 116

5-12. Decomposition of Standardized Effects for Path Model Predicting Responsibility
for Drug Therapy Outcomes and BPCS-DPC Scores................................................... 118

5-13. Differences between Advanced and Non-Advanced Pharmacists ............................... 119














LIST OF FIGURES


Figure pRge

1-1. The Triangle M odel of Responsibility............................................................................ 14

1-2. The A accountability Pyram id........................................................................................... 15

1-3. Model of Engagement in Voting as a Mediator of Reported Voting Behavior..............30

4-1. Percent Response to Type of Drug Therapy Problem in Pretest Scenario 1 .................. 70

4-2. Percent Response to Type of Drug Therapy Problem in Pretest Scenario 2 .................. 71

4-3. Percent Response to Standards as Valued Sources for Guidance of Conduct in
Identifying, Resolving, and Preventing Drug Therapy Problems................................... 74

4-4. Proposed Measurement Model for General Clarity of Standards, Personal Control,
and Professional D uty..................................................................................................... 90

4-5. Proposed Path Model of Perceived Responsibility for Drug Therapy Outcomes and
Pharm aceutical Care Provision....................................................................................... 93

5-1. Revised Proposed Measurement Model for General Clarity of Standards, Personal
Control, and Professional D uty..................................................................................... 109

5-2. Measurement Model for General Clarity of Standards, Professional Duty, and
Personal C ontrol............................................................................................................ 110

5-3. Plot of Standardized Residuals versus Standardized Predicted Values in the
Regression of Responsibility for Drug Therapy Outcomes.......................................... 114

5-4. Plot of Standardized Residuals versus Standardized Predicted Values in the
Regression of BPCS Scores .......................................................................................... 115

5-5. Path Model of Perceived Responsibility for Drug Therapy Outcomes and
Pharm aceutical Care Provision..................................................................................... 117

6-1. The Pharmacist Triangle Model of Responsibility....................................................... 133














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE DEVELOPMENT OF A PHARMACIST MODEL OF PERCEIVED
RESPONSIBILITY FOR DRUG THERAPY OUTCOMES USING THE TRIANGLE
MODEL OF RESPONSIBILITY
By

Lourdes Gladys Planas

December 2001


Chairman: Carole L. Kimberlin, Ph.D.
Major Department: Pharmacy Health Care Administration

The primary objective of this study was to develop a pharmacist model of

perceived responsibility for drug therapy outcomes using the Triangle Model of

Responsibility as a theoretical basis. This was accomplished by designing a self-

administered written survey for pharmacists containing measures based on the

operationalization of constructs from the Triangle Model of Responsibility. The validity

of the pharmacist model of perceived responsibility was examined by testing proposed

relationships between the measures using survey responses from community and

ambulatory care pharmacists in Florida.

Subjects' responses to items intended to measure general clarity of standards,

personal control, and professional duty displayed evidence of convergent and

discriminant validity. General clarity of standards, personal control, and professional

duty each had a significant direct effect on responsibility for drug therapy outcomes.








Responsibility for drug therapy outcomes had a significant direct effect on level of self-

reported pharmaceutical care provision, which was measured by a shortened version of

the Behavioral Pharmaceutical Care Scale. Additionally, responsibility for drug therapy

outcomes was found to mediate the effects of general clarity of standards, personal

control, and professional duty on pharmaceutical care provision. General clarity of

standards was also found to have a direct effect on pharmaceutical care provision.

Pharmacists with advanced training or certification reported higher levels of

responsibility for drug therapy outcomes than pharmacists without advanced training or

certification did. There was insufficient evidence to claim differences between advanced

and non-advanced pharmacists' reports of general clarity of standards, personal control,

and professional duty.

Based on these results, a model called the Pharmacist Triangle Model of

Responsibility was proposed. The model defines community and ambulatory care

pharmacists' perceptions of responsibility for a patient's drug therapy outcome as a

function of the perceived clarity of standards to identify, resolve, or prevent the patients'

drug therapy problems; personal control over the patients' drug therapy outcome; and

professional duty to identify, resolve, or prevent the patient's drug therapy problems.

The findings of this study have practical implications for pharmacy, as well as

theoretical implications for pharmacy administration research. Furthermore, the findings

of this study extend what is known regarding the applicability of the Triangle Model of

Responsibility.













CHAPTER 1
INTRODUCTION




Problem Statement


Occurrences of drug-related morbidity and mortality adversely affect patients'

quality of life and result in substantial costs to society. Therefore, reduction of

preventable cases of drug-related morbidity and mortality would be of great social value.

Pharmacists in community and ambulatory care settings are in unique positions to reduce

preventable drug-related morbidity and mortality by identifying, resolving, and

preventing drug therapy problems. Acceptance of such a mandate by pharmacists

requires an expansion of the traditional dispensing role to a systems view of medication

use in which pharmacists assume responsibility for patients' drug therapy outcomes.

Currently, there is a lack of empirical research related to pharmacists' perceptions of

responsibility for drug therapy outcomes. The development of a theoretical model as a

basis for defining and measuring perceived responsibility for drug therapy outcomes may

provide a better understanding of ways to increase community and ambulatory care

pharmacists' involvement in preventing occurrences of drug-related morbidity and

mortality.








Background


Drug therapy is the most commonly used mode of therapy for illnesses. In the

United States, two-thirds of patient visits to physician offices result in one or more

prescription medication (Cherry, Burt, & Woodwell, 2001). The purpose of drug therapy

is to achieve outcomes that improve the quality of people's lives. In particular, drug

therapy is used (1) to cure, slow, or prevent diseases and (2) to reduce, eliminate, or

prevent symptoms (Hepler & Strand, 1990).

The United States Food and Drug Administration requires that new drugs meet

rigorous standards regarding safety and efficacy before they are approved. Furthermore,

there are federal laws aimed at controlling public access to and use of drug products.

Despite these measures, drug products that are deemed "safe and effective" often fail to

improve the quality of people's lives and result in drug-related morbidity.


Drug-Related Morbidity and Mortality

Drug-related morbidity is defined as a therapeutic malfunction or miscarriage

involving the failure of a therapeutic agent to produce the intended therapeutic outcome.

A drug-related morbidity may manifest as a treatment failure (e.g., failure to control a

disease) or as a new medical problem (e.g., an adverse drug effect) (Hepler & Strand,

1990). Severe cases of drug-related morbidity can ultimately lead to drug-related

mortality.

The prevalence of drug-related morbidity and mortality is substantial. Studies

suggest that up to 25% of hospital admissions are drug-related (Bero, Lipton, & Bird,

1991; Bigby et al., 1987; Dennehy, Kishi, & Louie, 1996; Hallas et al., 1992;








Lakshmanan, Hershey, & Breslau, 1986; Lindley, Tully, Paramsothy, & Tallis, 1992;

Nelson & Talbert, 1996; Nikolaus, Specht-Leible, Kruse, Oster, & Schlierf, 1992).

There are considerable costs associated with drug-related morbidity and mortality.

Ernst and Grizzle (2001) used a cost-of-illness model to estimate direct costs associated

with drug-related morbidity and mortality (i.e., treatment failures, new medical problems)

in the United States ambulatory setting. They estimated that $177.4 billion was spent in

2000 to directly manage drug-related morbidity and mortality. This estimate had more

than doubled since 1995, which was when direct costs associated with drug-related

morbidity and mortality were first estimated at $76.6 billion (Johnson & Bootman, 1995).

To place these figures in perspective, one need only compare them to the estimated direct

costs attributable to diabetes, which totaled $44.1 billion in 1997 (American Diabetes

Association, 1998).

The above researchers' estimates for drug-related morbidity and mortality

included costs associated with additional physician visits, additional prescription

medications, urgent care/emergency department visits, hospital admissions, long-term

care facility admissions, and death (Ernst & Grizzle, 2001; Johnson & Bootman, 1995).

Indirect costs of drug-related morbidity and mortality, such as the economic impact of

lost productivity, were not included in these estimates. In addition, the cost-of-illness

models did not address the cost of drug-related morbidity and mortality due to untreated

indications for drug therapy. Therefore, it is feasible to infer that annual direct costs of

drug-related morbidity and mortality may actually exceed $177.4 billion. Overall, drug-

related morbidity and mortality is extremely costly to patients, insurers, the health care

system, and society.








Preventable Drug-Related Morbidity and Mortality

Instances of drug-related morbidity and mortality are often preceded by one or

more drug therapy problems. A drug therapy problem is a circumstance involving drug

therapy that actually or potentially interferes with a patient experiencing an optimum

therapeutic outcome (Strand, Cipolle, & Morley, 1990). Examples of drug therapy

problems include lack of a necessary drug, ineffective doses, and improper drug

selection. Drug therapy problems that are neither recognized nor resolved may result in

drug-related morbidity or mortality.

While some cases of drug-related morbidity and mortality are not preventable

(e.g., the first time that a patient experiences a medical problem due to a drug allergy),

others are considered preventable. Hepler and Strand (1990) described four elements of a

preventable drug-related morbidity (PDRM). First, the drug-related morbidity must be

preceded by a drug therapy problem that is recognizable. Second, given the drug therapy

problem, the likelihood of a drug-related morbidity must be foreseeable. Finally, the

causes of the drug therapy problem (and the subsequent drug-related morbidity) must be

identifiable and controllable. These four elements have been used as a basis for

operationally defining and validating explicit definitions of PDRM among older persons

(Faris, 2001; MacKinnon, 1999).

Although they vary widely in the literature, prevalence rates of PDRM are

sizeable. For example, several studies of drug-related hospital admissions reported that

47-76% of these admissions were preventable (Bero et al., 1991; Bigby et al., 1987;

Dennehy et al., 1996; Hallas et al., 1992; Lakshmanan et al., 1986; Lindley et al., 1992;

Nelson & Talbert, 1996; Nikolaus et al., 1992). These wide ranges primarily reflect








differences in sampling methods, health systems, and definitions of events and

preventability (Hepler, 2001).

Causes of PDRM have been reported in the literature. Bigby et al. (1987) studied

events leading to 686 emergency hospital admissions of patients. Of the 59 admissions

that were judged as potentially preventable, 19 were due to lack of patient compliance

and 28 were due to inadequate follow-up of abnormal symptoms, signs, or laboratory test

results. A study by Bero and her colleagues (1991) investigated drug therapy problems

associated with 224 hospital readmissions (i.e., admissions within six months of previous

discharge). The 40 readmissions judged as drug-related were characterized according to

the following types of drug therapy problems: adverse drug reaction, patient

noncompliance, overdose, lack of a necessary drug, underdose, problem with over-the-

counter medication, inappropriate drug choice, and drug-drug interaction. In this study,

an overwhelming majority (76%) of the drug-related readmissions were considered

potentially preventable.

The drug-related causes of PDRM mentioned above may have been detected and

dealt with if a health professional had been monitoring the outcomes of these patients'

drug therapies. With therapeutic outcomes monitoring, if a patient had experienced an

abnormal sign or symptom, one that otherwise would have remained unknown to a health

care professional, it could have been detected and addressed. In the next section, the

contribution that community and ambulatory care pharmacists can make to decrease

PDRM is discussed.








Role of Pharmacists in Preventing Drug-Related Morbidity and Mortality

Pharmacists, especially those in community and ambulatory care settings, are in

unique positions to reduce the incidence of PDRM. They are well situated in the

medication use process to influence patients' therapeutic outcomes (Hepler & Grainger-

Rousseau, 1995). They are the last health professional to come in contact with patients

prior to medication consumption. They also are accessible to patients, usually seeing

them several times between routine physician visits. Finally, pharmacists are highly

trusted by the public (Carlson, 2000; McAneny & Saad, 1997).

Pharmacists may reduce PDRM by identifying, resolving, and preventing drug

therapy problems. This particular approach to pharmacy practice shifts the focus from

dispensing pharmaceuticals to optimizing patients' therapeutic outcomes. Hepler and

Strand (1990) proposed that the pharmacy profession adopt this patient-centered,

outcome-oriented approach as its philosophy of practice. This expanded professional

role, known as pharmaceutical care, is defined as "the responsible provision of drug

therapy for the purpose of achieving definite outcomes that improve a patient's quality of

life" (Hepler & Strand, 1990, p. 533).

Pharmaceutical care involves cooperation between patients, their caregivers,

pharmacists, and other health professionals to design, implement, and monitor a

therapeutic plan intended to produce certain therapeutic outcomes for the patient. This

process involves three major functions: (1) identifying potential and actual drug therapy

problems, (2) resolving actual drug therapy problems, and (3) preventing potential drug

therapy problems. Furthermore, pharmaceutical care recognizes the inadequacy of the

traditional medication use process, in which a patient enters with a medical problem and

exits with a prescription to consume. In pharmaceutical care, the medication use process








is changed to a medication use system. Here, there is feedback from the patient at some

point after using a medication. Thus, there is an opportunity for a health professional to

observe the effects of a patients' therapy and to modify the therapeutic plan, if needed

(Hepler & Grainger-Rousseau, 1995).

In the practice of pharmaceutical care, a health care provider accepts

responsibility for a patient's drug therapy outcomes (Hepler & Strand, 1990). This

acceptance includes ensuring that the desired therapeutic goals are achieved and that

drug-induced illness does not occur (Penna, 1990), as well as being held accountable for

meeting patients' drug-related needs (Cipolle, Strand, & Morley, 1998). Further, a health

care provider's acceptance of responsibility for a patient's drug therapy outcomes should

exist in the context of a cooperative health care team, one in which a health care provider

shares responsibility with other members of the health care team (Brushwood & Hepler,

1996). This perspective is relevant because oftentimes the actions of more than one

health care provider influence patients' drug therapy outcomes.

Implementation of pharmaceutical care in community and ambulatory practice

At present, colleges of pharmacy and professional associations (American

Pharmaceutical Association, 1995; American Society of Health-System Pharmacists,

1996) have adopted pharmaceutical care as a standard for pharmacy practice.

Nonetheless, pharmacists, in general, have yet to incorporate pharmaceutical care into

their daily practices.

Barriers to the implementation of pharmaceutical care by community and

ambulatory care pharmacists include lack of organizational support, cooperation of other

health care professionals, financial incentives, access to patient-specific medical

information, and motivation (Penna, 1990). In addition, many community and








ambulatory care pharmacists must reevaluate their professional roles. This often entails

changing their focus from medication distribution to that of medication use and

optimization of patient outcomes.

In the meantime, professional regulations and precedents from legal cases reflect

an expanding role of pharmacists in the health care system. Regulations established by

the Omnibus Budget Reconciliation Act of 1990 (OBRA '90) set forth that pharmacists

must offer to counsel Medicaid patients about their prescription medications and must

engage in prospective drug use review (U.S. Government Printing Office, 1990).

Increasingly, state courts are ruling that pharmacists have a duty to warn patients about

medication safety and usage (Lynn & Ellis, 1998). Despite these trends, however,

community and ambulatory care pharmacists may not see pharmaceutical care practice

standards as either feasible or relevant to their practices. For example, compared with an

expert panel of pharmacists, a random sample of 315 Louisiana pharmacists rated

community-based pharmaceutical care practice standards as less feasible, and they were

significantly less confident about the relevance of these standards to therapeutic

outcomes (Desselle, 1997).

Pharmacist responsibility for drug therapy outcomes

It is apparent from the previous discussion that pharmacists' responsibilities in

patient care may be interpreted by referring to professional regulations and legal

precedents. Nonetheless, the manner in which pharmacists perceive their professional

responsibilities may differ from legal judgments and even from ethical standards (e.g.,

ethical codes adopted by professional associations). With regard to preventing drug-

related morbidity and mortality, the relevant role that pharmacists must fulfill involves

accepting some of the responsibility for patients' drug therapy outcomes. At this time,








there is a lack of empirical research that addresses the conceptualization and

measurement of pharmacists' perceived responsibility for patients' drug therapy

outcomes.




Study Objectives


This study had three objectives:

1. To develop a pharmacist model of perceived responsibility for drug

therapy outcomes using the Triangle Model of Responsibility as a

theoretical basis.

2. To design a self-administered written survey containing measures based

on the pharmacist model of perceived responsibility for drug therapy

outcomes.

3. To examine the validity of the pharmacist model of perceived

responsibility for drug therapy outcomes by testing proposed relationships

between the developed measures using survey responses from community

and ambulatory care pharmacists in Florida.




Conceptual Foundation


The foundation upon which this study was based was the area of social

psychology. This approach was appropriate because the focus of this study was the

definition and measurement of a perception. Perceptions influence how individuals

experience and comprehend their environment, as well as how they socially interact








(Heider, 1958). Social psychology has long been used as a scientific approach to study

people's perceptions of themselves and their social environment (Myers, 1996). The

theoretical model used for this study, the Triangle Model of Responsibility, has its origins

in social psychology. This model is presented in Chapter 2.

In addition, this study utilized the methods of test theory to develop and validate

measures based on a pharmacist model of perceived responsibility for drug therapy

outcomes. Test theory is a general framework for guiding the process of instrument

development (Crocker & Algina, 1986). The process that was used is outlined and

discussed in Chapter 4.

In summary, this study attempted to develop a pharmacist model of

responsibility for drug therapy outcomes by using the Triangle Model of Responsibility

as a theoretical basis for defining and measuring community and ambulatory care

pharmacists' perceptions. Principles of test theory were used to validate measures that

were developed.




Significance of the Study


There are several reasons why this study is needed. The first of these, which was

previously discussed in detail, concerns the prevention of drug-related morbidity and

mortality by pharmacists. The definition and measurement of pharmacists' perceived

responsibility for drug therapy outcomes could provide a better understanding of their

involvement in preventing drug-related morbidity and mortality.

Secondly, the development of theory-based measures of pharmacist responsibility

for drug therapy outcomes makes a unique contribution to pharmacy administration





11


research. Historically, scholarly work in the area of pharmacist responsibility has been of

a legal or philosophical nature. This study was unique because it investigated

pharmacists' perceptions of responsibility for drug therapy outcomes from an empirical

and social psychological perspective. Finally, this study makes a theoretical contribution

to the area of responsibility research in social psychology by furthering what is known

about the application of the Triangle Model of Responsibility.













CHAPTER 2
THEORETICAL MODEL




Overview


The theoretical model used for this study is described in this section. The

Triangle Model of Responsibility attempts to define people's perceptions of

responsibility for an event (Schlenker, Britt, Pennington, Murphy, & Doherty, 1994).

This model clarifies the concept of responsibility, which has remained a vague concept

with multiple meanings, despite its long history in discussions of law, ethics, and social

sanctioning of behavior.

A description of the Triangle Model of Responsibility follows. First, the

relationship between responsibility and accountability is explained. Second, the elements

and linkages of the model are described. Finally, studies using the Triangle Model of

Responsibility are discussed.




Description of the Triangle Model of Responsibility


Accountability

This discussion of responsibility is placed within the context of accountability.

The terms accountability and responsibility are often used synonymously. However, as

we shall see, judgments of responsibility are a necessary, but not sufficient, component of








the process of holding people accountable for events. Accountability is defined as "the

condition of being answerable to audiences, oneself included, for events that are

associated with one's identity and are relevant to salient prescriptions about how things

should be" (Schlenker & Weigold, 1989, p. 23). A component of accountability, separate

from responsibility, is that people must explain and justify their behavior to an audience.

In the accountability process, individuals are judged in an evaluative reckoning

where the evaluators have information about three elements and the linkages between

them.

These elements are (a) the prescriptions that should be guiding the actor's
conduct on the occasion, (b) the event that occurred (or is anticipated) that
is relevant to the prescriptions, and (c) a set of identity images that are
relevant to the event and prescriptions that describe the actor's roles,
qualities, convictions, and aspirations (Schlenker et al., 1994, p. 634).

When visualized, the three elements and the linkages between them form a two-

dimensional triangle known as the Triangle Model of Responsibility (Figure 1-1). When

an evaluating audience is added to the model in the form of appraising the arrangement of

the elements and linkages, the image is that of a pyramid called the accountability

pyramid (Figure 1-2). An evaluating audience may judge, evaluate, and sanction the

conduct of another person or of itself.


The Elements

The Triangle Model of Responsibility consists of three elements: prescriptions, an

event, and identity images. Prescriptions are rules for conduct. They include laws, moral

codes, regulations, and social norms in groups. Events are "the units of action and their

consequences that actors and observers regard as a unified segment for purposes of some

evaluation" (Schlenker et al., 1994, p. 635). Examples are performance on a single task























IPE link


IE link ]


Figure 1-1. The Triangle Model of Responsibility

(Schlenker et al., 1994)


PI link











































Figure 1-2. The Accountability Pyramid


(Schlenker et al., 1994)








or performance over a defined period of time. Identity images are the actor's roles and

qualities that are applicable in the context of the prescriptions and the event.

In an evaluative reckoning, the elements can vary in their importance to the actor

and judge. Prescriptions are more important when they are seen as valued principles of

conduct or when associated with greater sanctions. Events are more important when they

are associated with important prescriptions or result in significant consequences for the

actor and others. Identity images are more important when they are central and valued

components of the actor's identity (Schlenker et al., 1994).

The combined importance of the three elements is called the potency of the

elements. The potential for grave consequences to an actor increases as the potency of

the elements increases. However, the potency of the elements is only one of two parts of

an evaluative reckoning. Responsibility is the second component of an evaluative

reckoning.


Responsibility

Responsibility is "the psychological adhesive that joins the three elements of all

evaluations" (Schlenker et al., 1994, p. 636). If the three linkages are strong,

responsibility is judged high. As a result, the relevant prescriptions, events, and identity

images are seen as interconnected. The linkages among the three elements are the

prescription-event link, the identity-event link, and the prescription-identity link.


The Linkages

The three links of the Triangle Model of Responsibility are presented below. In

addition to a definition of each link, specific information about antecedents and

consequences of each link are discussed. Antecedents of a strong and weak link are








presented for each of the three links. The effects of strong and weak links on persons are

also mentioned. Finally, attempts by individuals to weaken the links are described.

Prescription-event link

The prescription-event link refers to "the extent to which a clear and salient set of

prescriptions is perceived to exist that should be applied to an event and should govern

conduct" (Schlenker et al., 1994, p. 638). The link is strong when prescriptions for an

event are clearly specified, pertinent to the event, and not in conflict with other applicable

prescriptions. The link is weak when prescriptions for an event are ambiguous, of

questionable pertinence to the event, and in conflict with other applicable prescriptions.

A strong prescription-event link provides the actor with clarity of purpose and

procedure. This clarity may increase the actor's self-confidence with regard to

performance. In contrast, anxiety may result when people feel they are supposed to

behave in a certain manner, but do not know how to do so. Thus, a weak prescription-

event link may result in anxiety and uncertainty. When actors experience or anticipate

task failure, they may attempt to weaken the PE link by claiming that prescriptions were

ambiguous and did not provide adequate guidance for expected behavior. For example,

an actor may say, "The rules were unclear."

Identity-event link

The identity-event link refers to the extent to which the actor is seen as connected

to the event. This link is judged by "perceptions of the extent to which the actor has (or

had) personal control over the event" (Schlenker et al., 1994, p. 639). The link is strong

when the actor is seen as intending to bring about a specific consequence of the event and

is seen as having the ability and freedom to bring about the consequence. The link is

weak when the event is seen as unforeseeable, accidental, or uncontrollable.








A strong identity-event link appears to increase an actor's commitment to a task

and determination for goal accomplishment (Schlenker et al., 1994). In addition, a sense

of personal control over performance of a task is also associated with high self-efficacy

for the task (Bandura, 1989). Actors may attempt to weaken this link when they fail or

anticipate failure on a task in order to detach central components of their identities from

the implications of a threatening event. For example, an actor may attribute failure to

external factors that diminish personal control or to unstable, alterable internal factors.

Prescription-identity link

The prescription-identity link refers to "the extent to which the prescriptions are

perceived as being applicable to the actor by virtue of the actor's characteristics, roles,

and convictions" (Schlenker et al., 1994, p. 638). The link is strong when a set of

prescriptions unambiguously applies to people with certain qualifications and the actor is

seen as possessing those qualifications. The link is weak when there is ambiguity or

conflict about which prescriptions are relevant to the actor's identity or whether the actor

possesses the qualifications that are applicable to the prescriptions.

A strong prescription-identity link provides an actor with purpose and direction.

Insofar as a set of prescriptions is seen as applicable to an actor, this link is analogous to

a sense of duty or obligation. Thus, a strong prescription-identity link is similar to what

Hamilton (1978) describes as role responsibility. As with the prescription-event link, a

weak prescription-identity link can lead to anxiety and uncertainty. In addition, a weak

prescription-identity link can lead to identity crises, role conflicts, and alienation

(Schlenker et al., 1994). On occasions when actors violate a prescription or perform

poorly, they may attempt to weaken the prescription-identity link. For instance, actors

may claim they do not possess the identity-relevant characteristics that are required by








the prescriptions (e.g., skills). Actors may also assert that their identities require that they

prioritize conflicting prescriptions.




Studies Using the Triangle Model of Responsibility


The Triangle Model of Responsibility has been used in several studies (Britt,

1999; Britt, 2001; Schlenker, 1997; Schlenker, 1999; Schlenker et al., 1994). Overall,

these studies have demonstrated that the model's three linkages (i.e., prescription-event,

identity-event, prescription-identity) predicted perceived responsibility, as well as

commitment and determination (referred to as engagement). These studies are discussed

in this section.

Schlenker and his colleagues (1994) conducted two studies to test the Triangle

Model of Responsibility. In the first study, they demonstrated that people used

information about the three linkages as specified by the model when judging the

responsibility of others. One hundred and two introductory psychology students were

randomly assigned to one of eight groups. Each group received a series of four vignettes

describing

(a) an employee in a large urban hospital who was tending to a patient
with a gunshot wound, (b) a student who was taking a course in college,
(c) a mother who was chaperoning her daughter's school trip, and (d) an
employee of an automobile company who was overseeing the assembly of
a prototype vehicle for the press (Schlenker et al., 1994, p. 641).

Each of the four vignettes contained information regarding the prescription-

identity, prescription-event, and identity-event linkages. The eight groups of students

differed on the combinations of information depicting each of the linkages (i.e., strong or








weak). Therefore, each of the linkages was a between-groups variable, while the vignette

type was a within-subjects variable.

The hospital employee vignette is used here to demonstrate how strong and weak

linkages were described. In this vignette, a hospital employee was asked to watch over a

gunshot wound patient whose condition had stabilized but who could possibly experience

complications. The employee was either (a) an emergency room nurse with specialized

training (strong prescription-identity link) or (b) an X-ray laboratory technician with

limited nursing training who was working in the emergency room as a replacement for a

nurse who was sick (weak prescription-identity link). Furthermore, the primary

physician either (a) gave the employee detailed instructions regarding the patient's

medications, possible complications, and key indicators of problems (strong prescription-

event link) or (b) gave only minimal instructions regarding the patient's medications and

no information about possible complications and key indicators of problems (weak

prescription-event link). Finally, the vignette stated that the emergency room was either

(a) relatively quiet with only a few cases and that the staff seemed to have things under

control (strong identity-event link) or (b) extremely busy, with doctors shouting for

assistance and, for the staff, a situation on the edge of chaos (weak identity-event link).

Subjects answered items regarding their perceptions of each vignette's central

character and situation on a nine-point scale. A principal-components analysis, with

varimax rotation, of responses to these items revealed a Responsibility factor and a

Determination factor (with eigenvalues of 4.13 and 3.19, and accounting for 34% and

27% of the variance, respectively).

The Responsibility factor included perceptions of (a) how responsible the
central character was for his or her performance (factor loading=.83), (b)








how much of an obligation or duty the central character had to do well on
the job or task (.71), (c) how blameworthy the central character would be
if he or she performed unsatisfactorily (.85), and (d) how well the central
character would have to perform for his or her work to be considered
satisfactory (.72). The Determination factor included perceptions of (a)
how hard the central character would work on the job or task (.91), (b)
how personally committed the central character was to doing well on the
job or task (.81), (c) how committed the central character was to doing an
excellent job for the intrinsic worth of the job rather than for extrinsic
reasons such as following orders (.78), (d) how well the central character
would do (.83), (e) how confident the central character was that he or she
would do well (.60), and (f) how easy or hard it would be for the central
character to give up on the task in the face of problems (.73) (Schlenker et
al., 1994, p. 642).

Items in each of the two factors were averaged and analyzed as dependent

variables. Mean scores on the items constituting the Responsibility factor ranged from

5.75 (when each linkage was weak) to 7.59 (when each linkage was strong). Analysis of

this factor revealed main effects of each of the three linkages: prescription-identity, F (1,

95)=13.67, p<.001; prescription-event, F (1, 95)=12.36, p<.001; and identity-event, F (1,

95)=9.96, p<01.OL Mean scores on the items that composed the Determination factor

ranged from 5.08 (when each linkage was weak) to 7.46 (when each linkage was strong).

Analysis of this factor resulted in main effects of each linkage: prescription-identity, F (1,

95)=41.31, p<.001; prescription-event, F (1, 95)=3.97, p<.05; and identity-event, F (1,

95)=18.53, p<.001.

Thus, in this first study, judgments of responsibility, as well as determination,

were a function of the strengths of the three linkages (Schlenker et al. 1994).

Furthermore, the results suggested that the central characters in the vignettes were judged

as more responsible and more determined when all linkages were strong, least

responsible and determined when all linkages were weak, and somewhere in between for

intermediate conditions.








The second study by Schlenker et al. (1994) revealed that people sought

information about the three linkages when judging the responsibility of others. A group

of 81 general psychology students was asked to evaluate the responsibility of a business

employee named Matthew regarding his firm's failure to reach its sales goals. Subjects

were given a list of 34 questions to help them with their evaluation. They were told they

could receive answers to as many of the questions as they wished. There were six

questions for each of the linkages.

The questions that reflected the prescription-event link dealt with the clarity and

nonconflicting nature of the relevant sales goals and procedures, as well as the validity of

the sales information as a measure of performance in relation to those goals (e.g., "Were

the steps or procedures that could be used to reach the sales goals clear and well

defined?" "Were there alternative goals or considerations that conflicted with the sales

goals?" "Are the sales indicators of the company valid measures of whether the company

reached its sales goals?"). Questions for the prescription-identity link dealt with the

employee's role in the company (e.g., "Is Matthew's job relevant to sales?" "Does

Matthew possess the abilities and experience required by his job, or did the company

misassign him to a job?"). The questions that reflected the identity-event link concerned

the extent to which circumstances outside of the employee's control would have affected

his ability to influence sales (e.g., "How much control, if any, did Matthew have over the

relevant aspects of sales?" "Did external market conditions exist that would have

affected sales irrespective of what the company did?"). The remaining 16 questions dealt

with consequences of the failure to achieve the sales goals and subject matter that was

irrelevant to the linkages.








The selection rates for the linkage questions were: prescription-event (62%),

identity-event (79%), and prescription-identity (77%). The consequences and linkage-

irrelevant questions were selected 26% and 9% of the time, respectively. An ANOVA of

selection rate based on question type was significant, F (4, 320)=214.28, p<.0001.

Planned comparisons revealed that questions reflecting each of the three linkages were

chosen more often than questions about either consequences or linkage-irrelevant topics

(p<.001).

Subjects also rated the information value of the 34 questions on a five-point scale

(1=valueless or irrelevant.. 5=extremely valuable, necessary for a decision). The

information value of the linkage questions (i.e., prescription-event, prescription-identity,

identity-event) ranged from 3.54 to 3.96, while the consequences and linkage-irrelevant

questions were rated as 2.40 and 1.52, respectively. An ANOVA of information value

based on question type was significant, F (4, 316)=215.62, p<.0001. Planned

comparisons showed that subjects rated each set of linkage questions as more valuable

for the purposes of evaluating responsibility than the consequences and linkage-irrelevant

questions (p<.001).

Finally, subjects were asked to select the 10 questions they believed were most

important for judging responsibility. Collectively, the linkage questions accounted for

94% of the questions selected, with the following percentages for prescription-event,

identity-event, and prescription-identity: 20%, 40%, and 34%. An ANOVA of question

importance based on question type was significant, F (4, 320)=167.59, p<.0001. Planned

comparison revealed that subjects selected linkage questions far more often than

consequences and linkage-irrelevant questions (p<.001).








Hence, subjects in this study sought information about the three linkages when

making judgments about a person's responsibility in a particular situation (Schlenker et

al., 1994). Questions pertaining to the three linkages were selected more often, and were

considered more valuable and important than other types of questions for evaluating

responsibility. This study, as well as the prior one, focused on people's perceptions of

the responsibility of others. The next studies that are outlined dealt with people's

perceptions of their own responsibility.

In another study by Schlenker (1997, 1999), college students' assessments of the

three linkages were used to predict their psychological engagement with regard to

academic achievement, which consisted of responsibility, determination, and

commitment. In turn, psychological engagement predicted the students' grade point

average, controlling for scores on the Scholastic Achievement Test (SAT).

A group of 235 introductory psychology students responded to items assessing

each of the three links and psychological engagement. Examples of the linkage items,

each of which was measured with seven items, were as follows: prescription-event ("The

ways to achieve academic success in college are clear to me"), identity-event ("I have

personal control over my academic success"), and prescription-identity ("At this stage of

my life, being academically successful is my job or duty"). Psychological engagement

was measured with eight items pertaining to perceptions of responsibility for academic

achievement, commitment to academic success, and determination to succeed

academically. Cronbach alpha estimates for these measures ranged from .80 to .84.

A multiple regression revealed that psychological engagement was significantly

predicted by the linkages (overall model R2=.55, p<.001). Each linkage made a unique








contribution to the prediction of psychological engagement (prescription-event: P3=. 13,

p<.02; identity-event: P=.39, p<.001; and prescription-identity: 0=.46, p<.001).

Furthermore, as each of the linkages increased in strength, psychological engagement

increased as well. A second multiple regression showed that grade point average was

significantly predicted by psychological engagement (P=.23, p<.001) and SAT scores

(P=.29, p<.001l) (overall model R2=.15, p<.001).

Thus, these college students reported greater responsibility for their academic

success and determination to be successful academically when they perceived clearer

guidelines for academic success, greater obligation to do well in school, and more

personal control over their academic success (Schlenker, 1999). This study demonstrated

that the Triangle Model of Responsibility might be used when people judge their own

responsibility. The next study that is discussed also focuses on self-assessments of

responsibility, among other things.

The Triangle Model of Responsibility has been tested among United States Army

soldiers (Britt, 1999). The model's three linkages predicted soldier engagement with

regard to a military mission, which was composed of responsibility for and commitment

to the mission. This particular study reported results from three samples of soldiers, each

with slightly different measures. The first sample consisted of 298 soldiers from an Air

Defense Artillery (ADA) unit before their deployment to Saudi Arabia. The second

sample was also from an ADA unit and was composed of 588 soldiers. Unlike the first

sample, however, soldiers from the second sample were surveyed during their

deployment in Saudi Arabia. Finally, the third sample consisted of 132 soldiers who








were part of an Armor unit and were at their home station. For the purposes of this study

description, the measures and results for the three samples are discussed together.

All subjects responded to items concerning the three linkages, responsibility for

their mission, and commitment to their mission. Some examples of these measures are as

follows: prescription-event ("I feel informed about what should happen on the mission,"

"The guidelines for my job are clear," I am confused about what is expected of me");

identity-event ("I will have control over my job on the mission," "Other people have

more control over my job than I do"); prescription-identity ("The mission is relevant to

my role as a soldier," "I am doing what I was trained to do," "My job is not relevant to

what I was trained to do"); and soldier engagement ("I feel responsible for my

performance during the mission," "I am committed to the mission"). In the first sample,

items intended to measure responsibility and commitment were combined into a single

measure of soldier engagement because they were highly correlated (r=.74).

Subsequently, these measures were treated similarly in the other two samples.

The soldiers scored moderately on the linkage items. For the first sample, means

for the prescription-event, prescription-identity, and identity-event linkages, each of

which was measured with one item using a six-point scale, were 3.38, 3.98, and 3.41,

respectively. The mean engagement score for the first sample, which was measured with

two items (i.e., responsibility and commitment) on a six-point scale, was 4.54. In the

second sample, means for the prescription-event, identity-event, and prescription-identity

linkages, each of which was measured with one positively worded item and one

negatively worded item using a five-point scale, were 3.69, 3.57, and 2.85. The mean

engagement score for the second sample, which was measured with two items (i.e.,








responsibility and commitment) on a five-point scale, was 3.99. For the third sample, in

which each linkage was measured with three items using a five-point scale, the summated

mean scores (and Cronbach alpha estimates) for the prescription-event, identity-event,

and prescription-identity linkages were 10.53 (.76), 10.02 (.76), and 11.50 (.69). In the

third sample, the mean engagement score, which was composed of the sum of five items

using a five-point scale, was 19.72 (.87).

Multiple regression analyses for each of the three samples revealed that, in

general, each of the links independently predicted engagement (overall model R2 from

.17 to .53, p<.05). Standardized regression coefficients for the prescription-event linkage

ranged from .12 to .27; from .10 to .51 for the identity-event linkage; and from .17 to .33

for the prescription-identity linkage. All of these regression coefficients were significant

at the .05 alpha level except for the prescription-event linkage in the third sample, which

did not significantly predict engagement. However, it should be noted that when the

single item assessing responsibility was analyzed instead of engagement, all three

linkages in the third sample accounted for unique variance in perceived responsibility

(Britt, 1999).

The results of this study show that soldiers felt more engaged in their work when

the rules for their performance were clear (prescription-event link), the performance was

seen as relevant to their training (prescription-identity link), and they felt they had

personal control over their performance (identity-event link) (Britt, 1999).

Britt (2001) also examined the usefulness of the Triangle Model of Responsibility

as a predictor of voting engagement and voting behavior in the 2000 presidential election.

A mediational model in which the three linkages predicted voting behavior through








voting engagement was tested. Furthermore, the predictability of the importance of the

event (i.e., outcome of the election) was tested along with that of the three linkages. The

study sample was composed of 117 college students in psychology courses. Subjects

were administered questionnaires in two phases: (1) four days before the election and (2)

one day after the election.

The phase 1 questionnaire included items measuring each of the three linkages:

four items for prescription-event (e.g., "The guidelines for registering to vote in the

election are clear and easy to understand," "I am uncertain about the procedures

regarding voting"); four items for identity-event (e.g., "I will have complete control over

whether I vote on election day," "Factors outside of my control may determine whether I

vote on election day"); and three items for prescription-identity (e.g., "I feel that voting is

relevant to my role as a U.S. citizen," "To me, voting is just one of my many roles as a

citizen, and not one of the most important roles"). The questionnaire also included one

item pertaining to the importance of the election outcome: "How important is it to you

that your candidate win the election?" Lastly, voting engagement was assessed by five

items regarding how responsible and committed one felt for voting (e.g., "I feel

personally responsible for voting on election day," ".... Whether or not I vote on election day

matters a great deal to me"). All of the above items, with one exception, were answered

using 5-point scales (l=strongly disagree.. .5=strongly agree). The scale for importance

of the election ranged from l=not at all important to 5=very important.

The day following the 2000 presidential election, subjects were given phase 2

questionnaires. One of the variables measured in this questionnaire was reported voting.

Subjects were asked, "Did you vote in the presidential election? Yes or No."








The means for the above scales were as follows: prescription-event (3.57),

prescription-identity (3.42), identity-event (3.55), importance of election (3.82), ant

voting engagement (3.51). The coefficient alpha estimates ranged from .63 to .94.

Examination of point-biserial correlations between reported voting and the other

variables revealed a stronger relationship with voting engagement (r=.71, p<.05) than

with the remaining variables (.42 < r < .49, p<.05).

Each of the linkages, as well as importance of the election, made a unique

contribution to the prediction of voting engagement (prescription-event: p=.15, p<.007;

identity-event: P=.24, p<.001; prescription-identity: P=.50, p<.001; importance of

election: P=.27, p<.001). A logistic regression of reported voting was conducted using

voting engagement, importance of the election, and the three linkages as predictors.

Voting engagement was the only significant predictor of reported voting (Wald=5.48,

p<.019). These results suggested that the relationship between reported voting and the

three linkages were mediated by voting engagement (Figure 1-3).

The results of this study expanded the applicability of the Triangle Model of

Responsibility to include voting beliefs and behavior. College students who reported

greater voting engagement (i.e., sense of responsibility and commitment to voting) also

tended to report that voting was relevant to their identities as U.S. citizens, voting

guidelines were clear, they had personal control over being able to vote on election day,

and they felt the outcome of the election was important. The latter finding is valuable

because past studies using the Triangle Model of Responsibility had not tested the

predictive power of event importance. Additionally, the results of this study

demonstrated that voting engagement was predictive of reported voting behavior. The















Clarity of
Guidelines


Personal
Control Over
Voting


Relevance of
Voting to
Identity


Engagement
in Voting


Report of
Voting


Importance
of Election


Figure 1-3. Model of Engagement in Voting as a Mediator of Reported Voting Behavior


(Britt, 2001)








model linkages and event importance were found to have indirect effects on voting

behavior, with voting engagement as a mediator.


In conclusion, several studies have tested the Triangle Model of Responsibility,

particularly the predictive power of the model's linkages. These studies have addressed

peoples' perceptions of other's responsibility, as well as peoples' perceptions of their

own responsibility and psychological engagement.













CHAPTER 3
REVIEW OF LITERATURE




Overview


The concept of responsibility has been studied by researchers and theorists from

various disciplines. This chapter reviews relevant literature concerning responsibility

from social psychology, legal philosophy, and pharmacy administration. Second,

limitations of research in these areas are discussed in order to identify gaps in what has

been studied. Third, similarities between facets of responsibility in these areas are noted.

Finally, research questions based on this literature review and the Triangle Model of

Responsibility are presented.




Responsibility in Social Psychology


Introduction

The area of psychology known as social psychology, which emerged as an

empirical discipline in the late 1880s, is the "scientific study of how people think about,

influence, and relate to one another" (Myers, 1996, p. 3). Social psychology stems from

the disciplines of sociology and psychology, in particular, personality psychology. Like

sociologists, social psychologists are interested in how people behave in groups, albeit on

a smaller scale. In comparison to personality psychologists, social psychologists








concentrate less on differences between individuals and more on commonalities among

individuals as they interact with their social surroundings.

During the 1970s, the focus of social psychology moved toward social cognition,

which is a term to describe how people think about themselves and others (Fiske &

Taylor, 1991). Attribution theory was at the center of this cognitive movement in social

psychology. Coinciding with this movement was the increasing prevalence of studies

dealing with how people judge or attribute responsibility for events. Both of these areas

of research, attribution theory and attribution of responsibility, originated with the work

of Fritz Heider (1958). For many, the use of the term "attribution" to refer to the

assignment of responsibility implies that this process is comparable to the study of

perceived causality as studied in attribution theory (Fincham & Jaspars, 1980). However,

as will be noted later, a distinction between these two areas is justified.

In this section, Heider's (1958) levels for assigning responsibility are presented.

Three of the early attribution theories are described (Heider, 1958; Jones & Davis, 1965;

Kelley, 1967) and the assertion that the assignment of responsibility can be explained by

attribution theory is challenged. In addition, a perspective on responsibility attribution

advocated by Hamilton (1978) is presented. She argued that research on the attribution

of responsibility tends to be more cognitive and less social in nature. Hamilton stated

that the addition of social roles to this area of research would result in a more adequately

social psychological approach to responsibility attribution.


Heider's Levels of Responsibility

In his book, "The psychology of interpersonal relations," Heider (1958) proposes

five levels that people use to judge responsibility. He also states that the attribution of an








action is a central issue in the assignment of responsibility for the action's outcome.

These responsibility levels are ordered according to the relative contribution of personal

(internal) and environmental (external) forces to the outcome of an action. The five

levels are as follows: (1) association, (2) commission, (3) foreseeability, (4)

intentionality, and (5) justification.

First, at the association level, a person is held responsible "for each effect that is

in any way connected with him or that seems in any way to belong to him" (Heider,

1958, p. 113). For example, in a situation of guilt by association, a person in a group

may be perceived as responsible for wrongdoing committed by another group member.

Second, at the commission level, a person is considered responsible for the outcome of

any act that he commits, even if he could not have foreseen the outcome. Here, a "person

is judged not according to his intention but according to the actual results of what he

does" (Heider, 1958, p. 113). Heider likens this level to Piaget's (1932) notion of

objective responsibility, in which an action is judged according to its result and

independently of motive. Third, at the foreseeability level, a person is judged as

responsible "for any aftereffect that he might have foreseen even though it was not a part

of his own goal" (Heider, 1958, p. 113). For instance, a person may injure another due to

negligence. Fourth, at the intentionality level, a person is held responsible for effects that

he intended to occur. Heider compares this level to Piaget's (1932) concept of subjective

responsibility, in which one's motive is the central issue. Finally, at the justification

level, judgments of responsibility are lessened when there are extenuating circumstances

to justify a person's behavior. At this level, "even [a person's] own motives are not

entirely ascribed to him but are seen as having their source in the environment" (Heider,








1958, p. 114). In addition, at this level, responsibility for an event is seen as being shared

by the environment and the actor. The greater the contribution of environmental factors

to an action outcome, the less a person is held responsible.

Although Heider does not define the concept of responsibility in his book, his

levels for judging responsibility have been used extensively in social psychological

studies of attribution of responsibility. Most studies using Heider's levels have

investigated the developmental or cumulative nature of the model, yet a lack of

conceptual clarity regarding Heider's levels remains (Fincham & Jaspars, 1980). Overall,

there is agreement that responsibility for an outcome tends to be attributed when people

are viewed as intentionally bringing about foreseeable outcomes and as acting under

conditions of free choice without justification for their actions (Fiske & Taylor, 1991).


Attribution Theory

The study of how people perceive causes of behavior has been a great source of

interest for social psychologists. Indeed, causal explanations by ordinary people are

central to attribution theory. Attribution theory refers to a collection of theoretical and

empirical work that deals with how ordinary people use information in their social

environment to produce causal explanations for events (Fiske & Taylor, 1991).

There are some basic assumptions of attribution theory. First, people are "naYve

psychologists" who try to make sense of what they observe in the world by forming

attributions, or causal explanations (Heider, 1958). Second, causal attributions are

assumed systematic to the extent that there are patterns to the explanations that people

contrive (Jones, Kanouse, Kelley, Nisbett, Valins, & Weiner, 1972). A third assumption

of attribution theory, which is important in initiating causal analyses, is that people have








an inherent need to predict and control their environment, thus maintaining a sense of

stability around them (Fiske & Taylor, 1991). Finally, it is thought that causal

attributions influence people's behaviors, beliefs, and emotions (Jones et al., 1972).

Attribution theory emerged from Heider's (1958) writings and subsequent

reformulations by Jones and Davis (1965), followed by Kelley (1967). Heider proposes

that people act on the basis of their beliefs. Hence, when social psychologists study

human behavior, people's beliefs should be taken into account, regardless of the validity

of the beliefs.

In his influential writings on the naive analysis of action, Heider (1958) outlines

his views of how people perceive others and how they infer meaning from what occurs

around them. He maintains that people use perceived causes of behavior to anticipate

and influence what will happen to them. According to Heider, people consider locus of

causality when they question why someone behaves in a particular manner. A person

may perceive that a behavior has an internal cause (i.e., in the person), an external cause

(i.e., in the environment), or both. Internal locus of causality consists of personal factors

(e.g., ability, trying, which is effort plus intention), while external locus of causality

consists of situational forces (e.g., task difficulty, luck).

As with the issue of responsibility, Heider places a great deal of emphasis on

intention when he describes the naive analysis of action. He uses the notion of intention

to differentiate between personal and impersonal causality. Intention is seen as the

central factor in the analysis of personal causality, which refers to instances in which a

person intentionally causes an outcome to occur. In contrast, cases of impersonal

causality include, for example, occurrences of unintended consequences. It also is worth








noting that Heider states that responsibility judgment includes the problem of naYve

analysis (attribution) of action. In other words, "once such attribution has been decided

upon, the evaluation of responsibility is possible" (Heider, 1958, p. 114).

Heider's emphasis on the importance of intentionality in assessing personal

causality greatly influenced Jones and Davis (1965) in their formulation of

correspondence inference theory. Their theory is an attempt to explain how individuals

make inferences about a person's intentions and dispositions in the context of a particular

action. Jones and Davis propose that people search for stable and informative

explanations for others' actions. In this sense, a person's action will be most informative

when it is judged as intentional and as a result of an underlying intention that does not

change from situation to situation. According to Jones and Davis (1965), people tend to

make "correspondent inferences" about other people's actions. In essence, an observer is

likely to infer that an actor's action and the intention that produced it correspond to some

underlying, stable quality (i.e., disposition) that the actor possesses. Presumably,

knowing the dispositional attributes of others enables an observer to understand and to

predict another's behavior. Moreover, unexpected or unusual behavior tells us more

about a person than does expected behavior (Jones & Davis, 1965).

Correspondent inference theory is only applicable to intentional behavior (Jones

& Davis, 1965). In order to infer that an actor's behavior is intentional, two conditions

concerning the knowledge and ability of the actor must exist. First, the actor must know

that the act in question has certain effects. Second, the actor must have the ability to

bring about the effects that are observed.








Kelley (1967) is another attribution theorist who was influenced by Heider. His

work in attribution theory built on Heider's (1958) notion that ordinary people gain

understanding of their environment through causal analysis similar to the experimental

method. Kelley (1967) formulated the covariation model for cases in which people have

information from multiple instances of the same or similar events. In the case of

covariation, Kelley drew from empiricist principles in the tradition of John Stuart Mill,

i.e., that "[an] effect is attributed to that condition which is present when the effect is

present and which is absent when the effect is absent" (Kelley, 1967, p. 194).

Covariation is the observed co-occurrence of two events, and is used to infer causal

attributions for events. According to Kelley, the logic of covariation resembles that of

analysis of variance. That is, in order to understand the cause of an effect, one observes

its covariation with different potential causes and attributes the effect to the cause that

most closely covaries with the effect.

Kelley (1967) states that people assess covariation information relevant to an

entity (e.g., a person or thing) across three dimensions: consistency, distinctiveness, and

consensus. Consistency refers to whether a person usually behaves similarly in the

observed situation. Distinctiveness can be described as whether a person behaves

differently in different situations. Finally, consensus refers to whether others behave

similarly in a situation like that which is observed. The strength of each dimension can

be measured as high or low.

An application of the three dimensions would be as follows. Suppose that a boy

named John is seen sleeping during his math class. With regard to consistency, one may

recall whether John usually sleeps during math class (high) or usually stays awake during








math class (low). Second, for distinctiveness, one may consider whether John stays

awake during other classes (high) or sleeps during other classes (low). Finally, with

regard to consensus, one may recall whether other students usually sleep during math

class (high) or usually stay awake during math class (low). In this example, a person will

attribute John's behavior to an internal disposition if there is high consistency, low

distinctiveness, and low consensus. In contrast, if one perceives high consistency, high

distinctiveness, and high consensus in this example, one will attribute John's behavior to

an external, situational cause (e.g., the math teacher is boring).

According to Kelley (1967), people use the combination of the three dimensions

to validate their impressions of events. This validation is sometimes necessary, for

example, when one's impressions are not supported by others or when information about

an event is ambiguous.

The issue of intentionality is neither ruled out nor established in the covariation

model (Kelley, 1967). This exclusion is in contrast to Heider's (1958) levels for judging

responsibility, in which intention is seen as a critical component of personal causality,

and correspondent inference theory, in which intentionality is assumed to exist (Jones &

Davis, 1965).


Limitations of Attribution Theories for Attribution of Responsibility

Although attribution theories are abundantly represented in social psychological

research, few of these studies have focused primarily on responsibility attribution

(Fincham & Jaspars, 1980). Neither Jones and Davis (1965) nor Kelley (1967)

specifically address responsibility attribution, although Kelley (1973) uses this term when

he discusses moral evaluation. Kelley (1971) also states that attribution theory could








provide only a partial understanding of moral evaluation. Moreover, application of

attribution theory to instances involving accidents, omissions, and liability reveals some

limitations of attribution theories in responsibility judgments (Fincham & Jaspars, 1980).

Examples of these limitations are presented in this section.

First, consider the case of an accident where a man drives his car into a lamp post

and dies (Fincham & Jaspars, 1980). According to Jones and Davis' (1965)

preconditions for the assignment of intention, an act is intentional based on an actor's

knowledge than an action has certain effects and an actor's ability to bring about those

effects. Presumably, the driver in this case knows that driving into a lamp post may

result in death and he possesses the ability to drive into a lamp post. Therefore, using

correspondent inference theory, this incident would be seen as intentional (i.e., the driver

committed suicide), regardless of the fact that the incident was an accident.

With regard to the covariation principle, it is not clear if it would be applicable in

this case because it neither establishes nor rules out intention on the part of the driver

(Kelley, 1967). Moreover, even if an observer uses consistency, distinctiveness, and

consensus information to form a causal attribution to something in the driver or situation,

this conclusion is insufficient to establish responsibility for this occurrence (Fincham &

Jaspars, 1980).

A second type of incident to consider involves a person's failure to commit an act

(Fincham & Jaspars, 1980). For example, a college student may leave her bicycle

unlocked while she attends class, only to return and find that her bicycle has been stolen.

One may say that the student's omission (i.e., not locking her bicycle) simply provided

the occasion for the theft, and that the theft was dependent on something other than her








omission. Assessments of the student's responsibility in this situation may be influenced

by both the dependence of the theft on her failure to lock her bicycle and the

circumstances surrounding other factors) that are involved. Considering the specifics of

this example, "there is little objection to admitting that the theft was a (causal)

consequence of [the student's] act, yet the suggestion that her omission caused the

bicycle to be stolen seems counterintuitive" (Fincham & Jaspars, 1980, p. 84). Hence,

this situation is an example of how attribution theories, which focus on perceived

causality, are limited and somewhat vague in their application to judgments of

responsibility.

Finally, let us consider a situation involving liability of some kind for an outcome

(Fincham & Jaspars, 1980). Suppose that a girl throws a stone through a window. The

girl's mother, although she did not throw the stone, may be held responsible for the

damage to the window. The mother's responsibility in this case does not imply that she

produced the damage. Furthermore, her responsibility hinges upon her relationship to the

person who actually produced the damage. In such a case, attribution theory focuses on

the causative agent (i.e., the girl) rather than on the liability that may be considered as

appropriate.

Therefore, it appears that attribution theory may not be adequate to address issues

related to attribution of responsibility, particularly in cases involving accidents,

omissions, and liability. Attribution theories generally deal with people's perceived

notions of causality, which appear to be only a partial component of the process by which

people attribute responsibility (Fincham & Jaspars, 1980). Further, as was seen in the

last example, the role that a particular person occupies (e.g., mother) may substantially








influence how responsibility is judged. The contribution that roles may play in

responsibility attribution is discussed in the next section.


Social Roles

Hamilton (1978) proposes that the dominant, cognitive approach to responsibility

attribution should be expanded to encompass a more adequately social psychological

approach. According to Hamilton, responsibility refers to "a decision about liability for

sanctions based on a rule" (1978, p. 316). Therefore, an attribution of responsibility is a

function of a rule, an actor's deeds, and others' expectations of what the actor should do.

The expectations of others are defined by the actors' social role and thus, these roles

should be included in responsibility attribution models (Hamilton, 1978).

Hamilton (1978) states that each of Heider's levels has one or more comparable

legal analogs. For example, Heider's first level (association) resembles the legal notion

of vicarious liability, while his third level (foreseeability) is similar to the concept of

negligence seen in both civil and criminal statutes. However, Hamilton notes a

distinction between the manner in which Heider proposes his levels be used and the

manner by which people use legal rules to judge responsibility. According to Heider

(1958), responsibility attribution increases from the first level (association) to the fourth

level intentionalityy), and decreases at the fifth level (justification). In contrast, with the

application of legal rules to a specific situation, a person can be held fully responsible at

each of Heider's levels.

It is clear that the first four Heiderian levels incorporate increasing degrees
of intentionality, yet the law calls forth judgments of guilt from adult
jurors in cases representing all four. And the consequences of these rules
at the lower levels need not be trivial (Hamilton, 1978, p. 318).








Hamilton (1978) argues that, when making responsibility judgments in a court of

law, people consider issues beyond those established by Heider (i.e., causality based on

internal and external forces). People also consider the relation between an actor's deeds

and an actor's role.

Any actor, behaving badly enough and intentionally enough, can be liable
under Intention responsibility rules. But only certain actors in certain
roles and behaving in certain domains may be liable according to
Association responsibility rules. Thus legal Association responsibility
rests on occupancy of specific roles; legal Intention responsibility, on
deeds irrespective of one's social role. The intermediate rule sets reveal
intermediate emphasis on the two (Hamilton, 1978, p. 319).

Hamilton states that a legal analogy is appropriate when evaluating Heider's levels

because legal rules reflect society's normative structure by indicating how instances of

wrongdoing are defined, evaluated, and punished.

Hamilton (1978) also argues that roles should not be treated as external forces but,

rather, as internal manifestations of an external social order. She contrasts her view with

that of Jones and Davis (1965), who regard in-role behavior as uninformative of a

person's dispositions and thus an external force. Hamilton likens her view of roles to

Heider's concept of "ought," which he addresses separately from his responsibility levels

(Heider, 1958). Heider characterizes oughts as "impersonal objective requirements

which are also dispositional in character and intersubjective in validity" (1958, p. 242).

Oughts are objective to the extent that they are acceptable by a consensus, which Heider

refers to as "intersubjective consensus." In addition, oughts are manifested in people as a

sense of obligation or duty to match behavior to a standard of conduct.








Discussion of Responsibility in Social Psychology

Heider's (1958) influential responsibility levels have been used by many social

psychologists to study responsibility attribution. The majority of research using Heider's

levels has focused on the developmental or cumulative nature of the model. However, a

lack of conceptual clarity surrounding the concept of responsibility still exists. This is

partly because "fundamental conceptual issues regarding the notion of responsibility

remain implicit in the literature, or, where articulated, are dealt with only superficially"

(Fincham & Jaspars, 1980, p. 85).

As with his responsibility levels, Heider's writings concerning how people make

attributions about actions influenced many researchers, spearheading the area of

attribution theory. Regardless of its emphasis on perceived causality, attribution theory is

inadequate for explaining responsibility attribution. This revolves around the inability of

attribution theory to address instances where intention, omissions, and liability are

present (Fincham & Jaspars, 1980).

Finally, the contribution that social roles may make to responsibility attribution

was discussed (Hamilton, 1978). According to Hamilton, responsibility for an action

should be considered as a function of a rule, an actor's deeds, and expectations of what

the actor should do (i.e., role). In addition, roles should be regarded as normative

contexts that are internalized and that evoke purposive action.

In her discussion of the contribution of roles to responsibility attribution,

Hamilton refers to Hart's (1968) notion of "role responsibility." This particular type of

responsibility, as well as others that Hart outlines in his legal essays, is one of the major

topics of the next section.








Responsibility in Legal Philosophy


Introduction

The manner in which the concept of responsibility is characterized in the law is of

substantial social relevance. As noted in the previous section, various social

psychologists have referred to legal analogies to describe their concepts of responsibility

(Heider, 1958; Hamilton, 1978). Therefore, this section reviews the concept of

responsibility from a legal perspective.

The legal writings of Hart and Honor6 are the focus of this section. In his essays

on punishment and responsibility, Hart (1968) differentiates between several types of

responsibility. Because causality is an important component of this discussion of

responsibility, a brief description of causation in the law is also included in this section.

Finally, a consideration of some relevant legal principles from a social psychological

perspective is presented.


Hart's Types of Responsibility

Hart (1968) describes several types of responsibility, including role responsibility,

legal liability responsibility, moral responsibility, capacity responsibility, and causal

responsibility. Before outlining these types of responsibility, however, it is worth noting

that Hart excludes the notion of scientific causality from his discussion of responsibility.

His basis for this exclusion is discussed later in this section. Instead of causality, Hart

recognized answerability as the central component of responsibility. Basically,

responsibility exists when one is answerable for one's actions, or for loss or damage that

results from one's actions.








In his description of role responsibility, Hart (1968) states

whenever a person occupies a distinctive place or office in a social
organization, to which specific duties are attached to provide for the
welfare of others or to advance in some specific way the aims or purposes
of the organization, he is properly said to be responsible for the
performance of these duties, or for doing what is necessary to fulfil them
(p.212).

Hart distinguishes "short-lived" duties (i.e., to do or not do some specific act on a

particular occasion) from duties that are singled out as responsibilities. The latter are

relatively complex or extensive duties that require care and attention over a protracted

period, thus defining a "sphere of responsibility." Commonly, a role responsibility is

referred to in the form of "a responsibility" or in the plural. Hart's notion of role

responsibility has been used to describe components of some social psychological models

of responsibility attribution (Hamilton, 1978; Schlenker et al., 1994).

Another type of responsibility that Hart describes, legal liability responsibility, is

considered the central notion of responsibility in legal philosophy. In this form of

responsibility, one is liable to be punished or to pay compensation for an act or a harm if

any of certain criteria are met. The first of these criteria, which has two parts and is

mental/psychological in nature, requires that a person possess (1) normal capacities to

conform to the requirements of law or (2) some element of intention concerning the crime

(i.e., mens rea). The second criterion for legal liability responsibility is a relationship

with a person who engages in illegal action. Here, it is not necessary that the person who

is punished have done what the law forbids. This last criterion is often expressed in cases

of vicarious responsibility. For instance, an innkeeper is liable to punishment if his

employees, without his knowledge and against his orders, sell alcohol after hours (Hart,

1968). It is worth noting that causation is not the only criterion needed for legal liability








responsibility. Therefore, a direct causal connection between a person and an outcome is

neither necessary nor sufficient for legal liability responsibility.

Hart's discussion of moral responsibility is similar to his discussion of legal

liability responsibility. The striking differences between the two are due to "substantive

differences between the content of legal and moral rules and principles rather than to any

variation in meaning of responsibility when conjoined with the word 'moral' rather than

'legal'" (Hart, 1968, pp. 225-6). Therefore, the aforementioned criteria (i.e.,

mental/psychological elements, connections between act and harm, relationship with

perpetrator) are applicable in judging moral responsibility, except that the particular

criteria may be different.

As previously noted, the issue of normal capacity is used as a criterion for legal

liability responsibility and moral responsibility. A person who possesses normal

capacities is considered "responsible for his actions." Hence, Hart considers capacity

responsibility as having a theoretical status distinct from the other types of responsibility.

Causal responsibility is the last type of responsibility described by Hart (1968).

With this type of responsibility, human beings (as well as their actions or omissions),

things, and events may be responsible for outcomes. In an example of causal

responsibility, it is possible to substitute the phrase "was responsible for" with "caused"

(e.g., x "was responsible for" y). It should also be noted that causality responsibility is

customarily used in the past tense.

As seen above, a direct causal link is not necessary for legal liability or moral

responsibility, although it is the most sufficient kind of connection. As noted earlier,

Hart (1968) excluded the scientific causation from his discussion of responsibility. One








reason for this is the difference between the legal and scientific notions of causation. In

their discussion of causation in the law, Hart and Honor6 (1959) note that lawyers make

causal statements about particular situations, whereas scientists formulate universal

statements or generalizations. Furthermore, when dealing with human actions, one

should focus on reasons for action rather than on causes of events. Here, reasons for

action refer to voluntary behaviors with intended effects.

In sum, Hart's (1968) discussion of responsibility explores various types of

responsibility. With regard to role, legal liability, and moral responsibility, the

differences revolve around to whom one is responsible rather than to different meanings

of the concept of responsibility. This is evident in Hart's observation that answerability

is the central component of a legal discussion of responsibility. The legal notion of

causation is a contributory factor in responsibility, while scientific causation is excluded.


Comparison of Legal and Social Psychological Concepts of Responsibility

Legal philosophy and social psychology share some aspects of how responsibility

is attributed. The two areas also have some differences. These similarities and

differences are discussed in this section.

The first of these similarities concerns the concept of causality. Previously, it was

noted that attribution theory is inadequate for addressing attribution of responsibility.

Attribution theory relies almost exclusively on scientific causality as its foundation. This

is evident in Kelley's (1967) covariation model, which is based on analysis of variance.

On the contrary, in attribution of responsibility, as well as in legal philosophy, causality

is only a part of how responsibility is judged. In fact, both Heider (1958) and Hart (1968)

dismiss scientific causality in their writings on responsibility. In his levels of








responsibility, Heider (1958) focuses on the interplay between personal causality (i.e.,

intentionality) and impersonal causality, which is not indicative of an empiricist

approach. In his discussion of types of responsibility, Hart (1968) excludes the scientific

notion of causality because it deals with universal statements, which are not appropriate

for legal purposes. Because of this common sense approach to responsibility attribution

in the law, social psychologists have suggested that responsibility attribution is more a

matter of "man as lawyer" than "man as scientist" (Fincham & Jaspars, 1980).

As mentioned above, intentionality is an important component of how

responsibility is attributed using Heider's levels (1958). The same is true for legal types

of responsibility, in which intention (mens rea) is a criterion for legal liability

responsibility and moral responsibility (Hart, 1968). Thus, it is no surprise that legal

analogies can be made from Heider's levels (Hamilton, 1978). Furthermore,

intentionality is one of the reasons why attribution theories are insufficient for judging

attribution of responsibility in the case of accidents (Fincham & Jaspars, 1980).

Another area in which social psychological and legal notions of responsibility can

be compared is omissions. Heider's (1958) responsibility levels are applicable to events

caused by committed acts, but not by omissions. In legal philosophy, omissions are most

relevant in the context of role responsibility, where a person in a particular situation is

seen as having a duty to behave in a specific manner based on the role that he occupies

(Hart, 1968). The same relationship has been pointed out by Hamilton (1978) with

regard to the attribution of responsibility based on social roles.

In conclusion, social psychological research on attribution of responsibility has

much in common with legal notions of responsibility. These similarities are partly due to








the exclusion of scientific causality, reliance on intention as a major criterion, and the

relevance of omissions to the occupancy of roles and their duties. In the next section of

this literature review, theoretical and empirical work on responsibility in the area of

pharmacy is discussed.




Responsibility in Pharmacy


Introduction

The research field known as pharmacy administration is an eclectic mix of various

disciplines that addresses issues related to pharmacy practice and medication use.

Pharmacy administration researchers draw from areas such as psychology, sociology,

health economics, and consumer behavior. One topic that pharmacy administration

researchers deal with is the expanded professional role of pharmacists in meeting the

needs of society related to medication use. Of particular interest here is reviewing the

pharmacy administration literature to understand how pharmacists' views of

responsibility for drug therapy outcomes have been conceptualized and measured, as well

as gaps in knowledge that exist concerning the conceptualization and measurement of

pharmacists' responsibility for drug therapy outcomes.



Conceptualization of Responsibility in Pharmacy

Most of the literature dealing with the conceptualization of pharmacists'

responsibility for drug therapy outcomes has come from writings about pharmaceutical

care. Hepler and Strand's definition of pharmaceutical care begins with the phrase, "the

responsible provision of drug therapy" (emphasis added) (1990, p. 539). In this context,








responsibility is defined as "an ability to answer for one's conduct and obligations, and as

being morally trustworthy within a specific office, duty, or trust" (Hepler, 1996, p. 36).

This definition of responsibility contains a retrospective and a prospective portion.

The first portion of the definition is related to the concepts of accountability,

liability, answerability and culpability, all of which are retrospective and retributive

(Hepler, 1996). The second portion includes a "prospective sense of moral

trustworthiness, of a private duty yet to be fulfilled" (Hepler, 1996, p. 36). Here,

trustworthiness means "worthy of confidence and reliance on the character, strength,

ability, or truth of someone" (Hepler, 1996, p. 36). In pharmaceutical care, trust between

a provider and a patient is earned and reciprocal (Brushwood & Hepler, 1996). Trust is

inherent in a covenantal relationship where competent care by a provider and acceptance

of professional authority by a patient are exchanged on an equal footing (Hepler &

Angaran, 1996). Thus, responsibility originates in a relationship and is dependent upon a

relationship (Brushwood & Hepler, 1996).

By accepting responsibility, a pharmacist recognizes a duty to respond to

circumstances where, as a result of drug therapy, a patient is at significant risk of harm or

is impeded from attaining therapeutic objectives (i.e., drug therapy problems). This duty

to respond is bound by three factors (Brushwood & Hepler, 1996). First, harm to a

patient from drug therapy must be reasonably foreseeable. Second, activities that a

pharmacist has a duty to perform in response to potential or actual drug therapy problems

are legally permissible and are within a pharmacist's capacity. Finally, a patient's

interests should be placed before the interests of others.








Brushwood and Hepler (1996) define duty as a certain course of conduct that

arises out of a relationship between two parties. They state that the concept of duty has

evolved principally through litigation. This notion of duty as a relationship between two

parties is also present when one defines duty in relation to the concept of rights. Rights

are "justified claims that individuals or groups can make on others or upon society"

(Buerki & Vottero, 1994, p. 38). When one party possesses a right, the party to whom

the claim is made has a duty to behave in a particular manner.

The retrospective aspect of responsibility, the ability to answer for one's conduct

and obligations, implies the existence of an authority to whom accounts are made

(Brushwood & Hepler, 1996). For instance, in the case of a treatment failure, persons in

positions of responsibility may be called upon to account for their actions related to a

patient's care. Yet, in cases of harm due to drug-related morbidity, it is often difficult to

separate responsibilities, in terms of causality, of a drug and of a health care provider.

Brushwood and Hepler (1996) make this distinction by referring to Baier's (1987)

notions of ascribing thing-responsibility and agent-responsibility.

In ascribing thing-responsibility, one identifies the operative factor that caused an

event to occur (Baier, 1987). For example, in a particular case of drug-related injury, one

could say a specific drug was responsible. The ascribing of thing-responsibility is

backward looking. In other words, the "culprit" is identified only after the mishap

occurs.

A major difference between ascribing agent-responsibility and thing-

responsibility is that the former is not merely backward looking (Baier, 1987). When

dealing with agent-responsibility, knowledge of a causal factor in a negative event is








often used to prevent similar occurrences in the future. The consequences of identifying

the "culprit" are more complex when dealing with agents, particularly with human

agents.

Unlike things, human beings cannot only be blamed, they can be found
blameworthy; not only have faults, but be at fault; and not only be due for
repair, but be culpable and deserving of condemnation or punishment or
liable to payment of damages (Baier, 1987, p. 103).

Humans possess the ability of choice in their behavior. In addition, humans, as moral

agents, are capable of understanding guidelines for behavior and obligations by which

they ought to abide (Baier, 1987). Therefore, ascribing agent-responsibility does not

depend on a particular past event. Instead, it is linked to a unique response mechanism

that involves free will and a capacity to recognize duties and requirements.

Pharmacists are agent-responsible in a forward-looking sense for a process
of drug therapy that is free from drug-related diseases. They are agent-
responsible in a backward-looking sense for the problems that drugs
cause. As moral agents, pharmacists have a duty to patients, and this duty
formalizes the pharmacist's responsibility (Brushwood & Hepler, 1996, p.
208).

The forward-looking aspect of agent-responsibility is composed of what Baier

(1987) calls a task-responsibility. A task-responsibility consists of the execution of

whatever is necessary to attain one's agent-responsibility. For example, the duty that

constitutes the forward-looking aspects of a pharmacist's agent-responsibility for drug

therapy outcomes is a task-responsibility to identify and resolve actual drug therapy

problems, and to identify and prevent potential drug therapy problems.

Brushwood and Hepler (1996) outline three types of pharmaceutical care

responsibilities: technical, judgmental, and normative. Technical responsibilities are

based on knowledge and skill. Examples of technical responsibilities are to know the

usual dose of a medication and to provide accurate prescription information to a patient.








Judgmental responsibilities are based on applications of knowledge and skill, such as

making decisions. Examples ofjudgmental responsibilities are to select an appropriate

medication dosage for a specific patient and to decide what prescription information to

emphasize to a patient. Normative responsibilities are based on role obligations and

expectations within relationships. For example, if a patient and a pharmacist agree that

the pharmacist should review the patient's medication profile for possible drug

interactions, then this activity would constitute a normative responsibility.

A health care provider does not function in a vacuum. In today's health care

environment, various health professionals may work with a patient and his family to

optimize health outcomes. In a pharmaceutical care system, health care providers work

with patients and their caregivers as a cooperative team (Hepler, 1993). Brushwood and

Hepler (1996) propose that a pharmaceutical care team, including a patient and family

caregivers, should have written practice guidelines that explain the broad responsibilities

of each team member. Each team member is "held accountable for recognizing the

appropriate course of action and acting appropriately each time a significant opportunity

to improve the outcomes of a patient's drug therapy presents itself' (Brushwood &

Hepler, 1996, p. 205). If a team member does not possess the competence to recognize a

particular opportunity, then the team member is not responsible for its recognition.

However, if an opportunity is noticeable but outside the team member's competence to

resolve, then the team member is responsible for calling it to the attention of the team

member who is competent and thus responsible for resolving the problem. In other

words, the foreseeability of an opportunity to improve the outcomes of a patient's drug

therapy is essential.








Measurement of Responsibility in Pharmacy

The work on pharmacist responsibility in pharmaceutical care that was previously

cited was philosophical in nature. A limited number of studies have measured

perceptions of responsibility among pharmacists using quantitative research methods.

These studies are reviewed in this section.

Hansen and Ranelli (1994) examined relationships between pharmacists'

perceptions of voluntarism as a professional responsibility and benefits of work

requirements outlined in OBRA '90 (U.S. Government Printing Office, 1990). The

study's primary objective was to determine whether pharmacists' views toward

voluntarism as a professional responsibility to society were related to their views

regarding OBRA '90 requirements (i.e., patient counseling, prospective drug use review

(DUR), counseling- and information-recording).

A convenience sample of pharmacists at a continuing education conference was

asked to indicate level of agreement to the following statements regarding OBRA '90

requirements: "counseling requirement is opportunity to establish role, prospective DUR

is necessary for professional survival, counseling and recording information will meet

patient needs" (Hansen & Ranelli, 1994, p. 1022). Opinions regarding voluntarism as a

professional responsibility were assessed by examining pharmacists' level of agreement

to the following statements: "volunteer and community service work is a professional

responsibility, volunteer and community service work go beyond professional

responsibility" (p. 1021). Responses to OBRA '90 and voluntarism items were noted on

a five-point modified Likert scale (l=strongly disagree.. .5=strongly agree). For OBRA

'90 requirements, responses of "strongly agree" and "agree" were noted as favorable








while responses of "neutral," "disagree," and "strongly disagree" were noted as

unfavorable.

Logistic regressions of responding favorably to each of the OBRA '90

requirements were conducted. A one-unit increase in the belief that voluntarism is a

professional responsibility was positively associated with responding favorably to the

OBRA '90 counseling requirement (OR=1.4; p<.00233) and to the counseling- and

information-recording requirement (OR=1.3; p<.0076). Similarly, a one-unit increase in

the belief that voluntarism goes beyond professional responsibility was negatively

associated with responding favorably to the prospective DUR requirement (OR=0.8;

p<.0380). The investigators noted that this study "suggests a possible association

between pharmacists' attitudes toward social responsibility and their acceptance of

expanded professional responsibility" (Hansen & Ranelli, 1994, p. 1020).

Sitkin and Sutcliffe (1991) examined the extent to which professional,

organizational, and legal factors predicted two types of pharmacist behavior: information-

giving and advice-giving. The investigators noted that service orientation and form of

control are two distinguishing characteristics of professional and organizational

influences on pharmacy practice. Professional influence emphasizes service as

commitment to quality patient care, whereas organizational influence emphasizes service

as satisfying customers to retain business. Among professionals, the dominant form of

control is autonomy and self-regulation. However, in organizations, control originates

from the organization's hierarchy or bureaucracy, which often resides outside individual

professionals.








The investigators compared three models (i.e., professional, organizational, and

legal responsibility) in terms of explanatory power for the two pharmacist behaviors

(Sitkin & Sutcliffe, 1991). Ninety-four Texas community pharmacists selected through

stratified random sampling responded to the study survey. The survey included a

scenario involving a patient who was to receive a new prescription medication and an

over-the-counter medication whose concurrent use could result in a drug-drug interaction.

Pharmacists indicated the likelihood that they would explain to the patient why the two

medications should not be taken concurrently (information-giving) as well as the

likelihood that they would instruct the patient as to whether or not to take the two

medications concurrently (advice-giving).

Self-regulation, customer satisfaction, and individual liability were predictors of

information-giving by pharmacists (adjusted R2=. 14, p<.03). The extent to which

pharmacists indicated they were influenced by formal and informal peer evaluation of

their work (standardized P=.29, p_.05), and concerned with malpractice suits and

disciplinary actions by the State Board of Pharmacy (standardized P=.26, p<.05), was

positively related to information-giving. On the contrary, the extent to which pharmacists

were concerned with customer satisfaction/confidence and its relationship to retaining

business for the pharmacy was negatively associated with information-giving

(standardized P3= -.37, p_.05).

For advice-giving, the predictors were customer satisfaction, individual liability,

and centralization of control (adjusted R2=.19, p<.01). The extent to which pharmacists

were concerned with customer satisfaction/confidence and its relationship to retaining

business for the pharmacy (standardized P=.29, p<. 10), and with malpractice suits and








disciplinary actions by the State Board of Pharmacy (standardized 13=.30, p<.05), was

positively associated with advice-giving. Conversely, the degree to which their

organization's policies and practices reflected a concern for avoiding customer

complaints and lawsuits was negatively related to advice-giving (standardized P= -.26,

p<.05).

As noted above, individual liability was a significant predictor of both

information- and advice-giving. This finding was consistent with the researchers'

hypothesis:

what separates a profession from most other occupations is that
responsibility for consequential diagnosis and action is vested in the
individual practitioner. If professionals embrace personal responsibility as
a key part of their professional self-concept, then increased concern with
malpractice or licensing (i.e., individual forms of responsibility), should
be positively associated with valued professional behaviors, such as
advising or informing clients (p. 279).

The researchers' measure of individual liability was a retrospective form of responsibility

because it dealt with sanctioning for consequences of behavior that would have already

occurred.

Using an instrument called the Professional Pharmacists Perspective Profile (P4),

Larson (1994) identified five ethical components of pharmaceutical care: relationships,

access, responsibility, autonomy, and database. Components were obtained by

performing an exploratory factor analysis of the P4 items. The items in the responsibility

factor addressed whether pharmacists, physicians, or patients are ultimately responsible

for identifying, resolving, and preventing drug-related problems. Examples of

responsibility items included:

pharmacists should take responsibility for their patients' drug therapy after
they leave their pharmacy; physicians should take primary responsibility








for monitoring drug therapy since they are the ones who prescribed the
drugs; I like to think of my patients as self-sufficient and not having to
rely on me to solve their drug-related problems (p. 87).

Pharmacists who agreed to participate in a pharmaceutical care demonstration

project responded more positively to the responsibility factor than pharmacists who did

not agree to participate (p<. 10). In addition, staff pharmacists had lower responsibility

scores than managers (p<.0001) and owners (p<.028).


Discussion of Responsibility in Pharmacy

Pharmacists' responsibility for drug therapy outcomes was discussed from a

conceptual perspective. In the philosophy of pharmaceutical care, responsibility is

described in two parts: a retrospective component and a prospective component

(Brushwood & Hepler, 1996). The retrospective component is related to accountability,

where pharmacists are answerable for their conduct and duties. The prospective portion

outlines the duty of pharmacists to respond to drug therapy problems. This duty is bound

by (1) reasonable foreseeability of harm to a patient, (2) the legality of the action to be

taken, (3) the pharmacists' capacity with regard to the action to be taken, and (4) the

notion that a patient's interests come before those of others (Brushwood & Hepler, 1996).

The issue of foreseeability is especially relevant in the context of the current health care

environment, in which patients come in contact with various health care professionals.

Health providers who foresee that a drug-related morbidity is likely to occur, but who do

not have the competency or authority to resolve it, have a duty to warn the appropriate

health provider who may resolve the problem. Brushwood and Hepler (1996) also

advocate the development and use of written practice guidelines that explain the broad

responsibilities of each health care team member.








Brushwood and Hepler use Baier's (1987) concepts of thing-responsibility and

agent-responsibility to note the complexity of determining responsibility, in terms of

causality, when dealing with health care providers and medications. Thing-responsibility

may be used to view a drug as a causative factor for patient harm. Agent-responsibility is

used to judge responsibility of a person. Thing-responsibility is only judged after an

event, while agency-responsibility can be used in both a forward- and backward-looking

manner. With agent-responsibility, people can be punished or held liable for damages.

Further, agent-responsibility takes into account that people are aware of prescriptions for

behavior by which they ought to abide.

This section also reviewed studies in which responsibility was measured from an

empirical perspective. In the study by Hansen and Ranelli (1994), voluntarism was

chosen as an explanatory variable because it denoted an altruistic orientation, which the

investigators stated may manifest itself as a commitment to patient care. However, no

such association between altruism and commitment to patient care was established in this

study. Instead, pharmacists in the study were asked to indicate the extent to which they

agreed or disagreed that volunteering was within the confines of professional

responsibility. In addition, this study assumed that pharmacists' responses to items

pertaining to beliefs about OBRA '90 requirements were representative of their

willingness to assume these requirements and of their attitudes toward these

requirements. Overall, the study lacked clear, conceptual definitions of variables and a

theoretical framework that delineated how these variables were related.

In his study of an ethical foundation of pharmaceutical care, Larson (1994) used

exploratory factor analysis to delineate the components of his model, including the factor








representing responsibility. This factor consisted of five items, three of which addressed

the extent to which the pharmacist or physician should take responsibility for patients'

drug therapy. In view of the complexities involved with the concept of responsibility, as

has been shown in this literature review, it is proposed that a measure of such a concept

should reflect the various characteristics of perceived responsibility.


Application of Triangle Model of Responsibility to Pharmacist Responsibility

Some aspects of the concept of pharmacist responsibility for drug therapy

outcomes, as defined in pharmacy administration literature, are present in the Triangle

Model of Responsibility (Schlenker et al., 1994). The first of these is Hepler's (1996)

definition of pharmacist responsibility in pharmaceutical care. He defines responsibility

as "an ability to answer for one's conduct and obligations, and as being morally

trustworthy within a specific office, duty, or trust" (Hepler, 1996, p. 36). According to

the Triangle Model of Responsibility, responsibility is a necessary component of the

process of being held accountable for events (Schlenker et al., 1994). Therefore, to be

held accountable, one must be in a position of responsibility (as it is defined by the

Triangle Model of Responsibility). This entails being able to answer for one's conduct,

which is not the same as being held accountable. Likewise, the retrospective portion of

the definition of pharmacist responsibility in pharmaceutical care refers to the ability to

answer for one's conduct (Hepler, 1996). Moreover, the prospective portion of the

definition of pharmacist responsibility in pharmaceutical care may be compared to the

prescription-identity link from the Triangle Model of Responsibility. This particular link

refers to a sense of duty or obligation. As such, it is also comparable to Hart's (1968)

notion of role responsibility. In sum, the conceptual definition of pharmacist








responsibility embodies the overall definition of responsibility from the Triangle Model

of Responsibility, as well as one of the model's linkages (i.e., duty).

The prospective portion of Hepler's (1996) definition of pharmacist responsibility

(i.e., the duty to respond to drug therapy problems) is bound by several factors, which

were discussed in previous sections. These factors, which include reasonable

foreseeability of harm to a patient and the legality of the action to be taken (Brushwood

& Hepler, 1996), are similar to two of the responsibility linkages from the Triangle

Model of Responsibility.

The first of these, foreseeability of harm to a patient, may be compared to the

identity-event link. This link is defined as the extent to which a person perceives

personal control over an event. Recall that a drug-related morbidity may be considered

preventable if: (1) the drug-related morbidity is proceeded by a drug therapy problem that

is recognizable; (2) given the drug therapy problem, the likelihood of a drug-related

morbidity is foreseeable; and the causes of the drug therapy problem (and the subsequent

drug-related morbidity) are (3) identifiable and (4) controllable (Hepler & Strand, 1990).

Thus, a pharmacist who is able to foresee harm to a patient may perceive some personal

control over the patient experiencing a drug-related morbidity. A pharmacist may

perceive a greater amount of personal control if he can identify and control the cause of

the drug therapy problem that is resulting (or may potentially result) in a drug-related

morbidity.

The second of these factors that bind a pharmacist's duty to respond to a drug

therapy problem is the legality of the action to be taken (Brushwood & Hepler, 1996).

Pharmacy laws are one of many types of prescriptions or standards for pharmacist








conduct. The prescription-event link from the Triangle Model of Responsibility is

defined as the extent to which a clear and salient set of prescriptions is perceived to exist

that should be applied to an event and should govern conduct (Schlenker et al., 1994).

Hence, this particular factor may be related to prescription clarity.

The above discussion suggests that pharmacist responsibility for drug therapy

outcomes may be composed of three components, namely the three linkages from the

Triangle Model of Responsibility. This assertion points to the inadequacy of studies that

define pharmacist responsibility as unidimensional constructs and to the need for a

multidimensional measure of pharmacists' perceived responsibility for drug therapy

outcomes.




Research Questions


As outlined in Chapter 1, the objectives of this study were to (1) develop a

pharmacist model of perceived responsibility for drug therapy outcomes using the

Triangle Model of Responsibility as a theoretical basis, (2) design a self-administered

written survey containing measures based on this pharmacist model, and (3) examine the

validity of the pharmacist model of perceived responsibility for drug therapy outcomes by

testing proposed relationships between the developed measures using survey responses

from community and ambulatory care pharmacists in Florida. The following three

research questions were posed for the purpose of fulfilling the third study objective:

1. Will pharmacist measures of the (a) prescription-event link (prescription

clarity), (b) identity-event link (personal control), and (c) prescription-








identity link (duty) predict levels of perceived responsibility for drug

therapy outcomes?

2. Will perceived responsibility for drug therapy outcomes predict levels of

self-reported pharmaceutical care activities?

3. Will pharmacists with advanced training or certification report greater

levels of pharmacist measures of (a) the prescription-event link

(prescription clarity), (b) the identity-event link (personal control), (c) the

prescription-identity link (duty), and (d) perceived responsibility for drug

therapy outcomes than will pharmacists without advanced training or

certification?













CHAPTER 4
METHODS




Overview


The primary objective of this study was to develop a pharmacist model of

perceived responsibility for drug therapy outcomes using the Triangle Model of

Responsibility (Schlenker et al., 1994) as a theoretical basis. This was accomplished by

designing a self-administered written survey for pharmacists containing measures based

on the operationalization of constructs from the Triangle Model of Responsibility. The

validity of the pharmacist model of perceived responsibility was then examined by testing

proposed relationships between the measures using survey responses from community

and ambulatory care pharmacists in Florida. This chapter describes the methods that

were used to achieve the study objectives.

The process by which the written survey was designed included operationalization

of study constructs, a pretest, and a subsequent revision for the validation study. All of

these are discussed in the next section. The sections that follow describe the sampling

procedure used to identify the pool of potential subjects for the validation study, the data

collection methods, data analysis techniques, and study hypotheses.








Survey Development


A self-administered written survey called the Pharmacist Drug Therapy

Questionnaire (PDTQ) was developed for this study. The process by which the PDTQ

was developed is described in this section. First, a convenience sample of community

and ambulatory care pharmacists were interviewed regarding their experiences with

patients and drug therapy problems. Their comments were used to aid in developing

scenarios and items for the PDTQ. Once the PDTQ was designed and reviewed by an

expert panel for content validity, it was pretested among a sample of pharmacist members

of a pharmacy association. This section describes how the PDTQ was developed,

pretested, and revised in preparation for the validation study.


Pharmacist Interviews

A convenience sample of eight community and ambulatory care pharmacists from

central Florida were interviewed regarding their experiences with patients and drug

therapy problems. Interviewees were asked open-ended questions about drug-related

events (e.g., a patient experiences a treatment failure, new medical problem). They were

asked to comment on examples of standards (prescriptions) they thought they should

have used as guides for conduct. In addition, they were asked to describe the clarity of

these standards (e.g., applicability, ambiguity, conflicts with other standards) in relation

to these events, whether they felt bound by these standards as pharmacists, and their

opinion about their personal control over these adverse patient outcomes.

Comments from the interviewees were used to facilitate the designing of the

PDTQ. Two patient care scenarios were written, as well as items intended to measure

constructs from the Triangle Model of Responsibility. The operationalization of these








constructs is presented in the next section. But first, the various steps that led to the

development of the PDTQ are discussed.


Content Validity. Readability, and Face Validity

The purpose of content validation is to assess whether an instrument's items

adequately represent a specific domain or construct (Crocker & Algina, 1986). Once the

PDTQ was designed, an expert panel consisting of four pharmacy administration faculty

members and one social psychology faculty member assessed it for content validity.

Members of the panel were provided with definitions of scale domains and instructions

for evaluation of the instrument.

To assess the readability and face validity of the PDTQ, the survey was

administered to two pharmacists. Each took 20 minutes to complete the survey.

Afterward, each pharmacist was asked to comment on parts of the survey that he found to

be confusing or unclear.


Survey Pretest

Following review of the PDTQ by the above individuals, the survey instrument

was pretested among a sample of pharmacists who were members of a state pharmacy

association. Some of the objectives of the pretest were to obtain a conservative return

rate estimate for the PDTQ and to determine how pharmacists viewed the importance of

the selected standards, as well as how they viewed the intended drug therapy problems in

the patient care scenarios. Additionally, the pretest allowed for preliminary reliability

assessments of the Triangle Model of Responsibility measures. This was carried out by

estimating internal consistency using coefficient alpha, which provides a measure of how

consistently subjects perform across items intended to measure the same construct








(Crocker & Algina, 1986). An alpha estimate greater than or equal to .65 was considered

acceptable (Nunnally, 1978). Furthermore, item analysis was carried out by examining

the corrected item-total correlation for each item, which is the correlation between the

item score and the total for the remaining items in the corresponding domain (Crocker &

Algina, 1986). Corrected item-total correlation estimates greater than or equal to .50 and

inter-item correlations greater than or equal to .30 were considered acceptable.

For the pretest, 300 names were randomly selected from a community pharmacist

member list of a state pharmacy association. A cover letter accompanying the survey

stated the purpose of the study, provided instructions, and assured the respondents of

confidentiality (Appendix B). Business reply envelopes were also included for returning

surveys. Eighty-one completed surveys (27%) were returned. Of these, 58 were from

pharmacists who indicated they worked in community pharmacy settings and were thus

used for the pretest analyses.

Seventy-six percent of the pretest sample was male. The number of years that

these pharmacists had practiced pharmacy ranged from 2 to 46 with a mean of 21.5 years.

The percentage employed in independent, chain, and grocery/discount pharmacies was

36.2%, 32.8%, and 17.2%, respectively. The remainder of the pretest sample worked in

outpatient clinics (6.9%) and other types of community pharmacies (6.9%). Almost half

were pharmacy managers (48.3%). Approximately one-fifth were either staff

pharmacists (20.7%) or pharmacy owners (19.0%).

Based on analyses from the pretest, several revisions were made to the PDTQ

before testing it in the validation study. The next section describes the revised PDTQ,

including the revisions made from the pretest version.








Description of Pharmacist Drug Therapy Questionnaire

Survey scenarios

The PDTQ consisted of two scenarios in which a patient's experience with a

drug-related morbidity was described (i.e., treatment failure, potential adverse effect).

Each scenario consisted of specific information about a patient, including diagnosis,

prescription medication, and refill history. The scenarios were designed to simulate

realistic situations in which pharmacists would find themselves and in which they would

be faced with decisions regarding identifying, resolving, or preventing a drug therapy

problem. The primary purpose for including two scenarios in the PDTQ was to avoid

limiting responses to only one hypothetical situation.

Scenario 1. The first scenario described a 23-year old asthmatic patient named

Ms. Smith who was requesting a refill of her beta-agonist metered dose inhaler

(albuterol). The patient's refill history for albuterol showed that her last refill was 14

days earlier, indicating a possible pattern of use considered too frequent for a medication

that is prescribed to alleviate (as opposed to prevent) asthma symptoms. Furthermore,

the patient complains of symptoms that are usually associated with worsening of asthma.

The scenario stated that albuterol was the only medication that had been prescribed for

the patient's asthma. The facts of this scenario were intended to imply that Ms. Smith

was currently experiencing a treatment failure due to the lack of a necessary drug (i.e., a

corticosteroid to treat her worsening asthma).

In the pretest PDTQ, respondents were asked to select what type(s) of drug

therapy problems) they thought Ms. Smith was currently experiencing (item 7, Appendix

A). Over half (54.4%) of the pretest respondents agreed that Ms. Smith had a medical









condition that required drug therapy, but was not receiving a drug for that indication

(Figure 4-1).




100
R
P
e 80
S
r 60 -No
p
0 40Q EYes
o04
n
n s 20
t

e 0 '-- --




Type of Drug Therapy Problem


Figure 4-1. Percent Response to Type of Drug Therapy Problem in Pretest Scenario 1



Additionally, the percentage of respondents who agreed that Ms. Smith was being

treated with too much of the correct drug and with the wrong drug was 36.8% and 28.1%,

respectively. Very few or none of the respondents believed Ms. Smith was taking a drug

for no medically valid indication (0%), was being treated with too little of the correct

drug (10.5%), had a medical condition that was the result of an adverse effect (5.3%), or

had a medical condition that was the result of a drug interaction (0%).

Scenario 2. The second scenario described a 60-year old patient named Mr.

Wilson who was presenting a new prescription for a sulfonylurea (glipizide) to treat his

recently diagnosed diabetes. The patient in this scenario mentioned that he rarely ate

breakfast. This piece of information was intended to have respondents infer that Mr.

Wilson was at a high risk to experience an adverse effect due to the excessive lowering of








his blood sugar by glipizide. In the pretest scenario, the glipizide was prescribed to be

taken every morning.

In the pretest PDTQ, respondents were asked to select what type(s) of drug

therapy problems) they thought Mr. Wilson was at a high risk to experience (item 26,

Appendix A). Forty-four percent of the pretest respondents agreed that Mr. Wilson was

at a high risk to experience a medical condition resulting from an adverse effect (Figure

4-2).





pR H100
e 80-
es
S
r P 60 No
CO
0 40 .Yes
n
n
ns 20 -
0

66\o6 e\, l ,6 0" 60" W6 ON
0 0 4vvool S
00

'$O 00o

Type of Potential Drug Therapy Problem


Figure 4-2. Percent Response to Type of Drug Therapy Problem in Pretest Scenario 2



Very few or none of the respondents believed Mr. Wilson was at a high risk to

experience a drug therapy problem due to taking a drug for no medically valid indication

(1.8%), taking the wrong drug (5.6%), taking too much of the correct drug (7.4%), or a

drug interaction (9.2%). However, approximately one-third of respondents believed that

Mr. Wilson was at a high risk to experience a medical condition for which additional








drug therapy would be indicated (35.2%) or because too little of the correct drug was

being taken (31.5%).

Respondents who selected these last two types of drug therapy problems may

have been reacting to the once-a-day dosing ofglipizide. It is common for physicians to

begin newly diagnosed diabetics on once-a-day dosing and then increase the dose until

adequate blood sugar control occurs. After the titration period, the dose that is typically

seen in clinical practice is twice-a-day dosing. In light of this, a revision was made to

this scenario for the validation study. The revised scenario had Mr. Wilson presenting a

prescription for glipizide that was dosed twice a day. The scenario described how Mr.

Wilson's physician had previously given him free samples ofglipizide and started him on

once a day dosing, but had increased the dosing frequency to twice a day for better

glucose control.

Each scenario was followed by items intended to measure clarity of six particular

standards, general clarity of the set of standards, personal control, professional duty,

responsibility for drug therapy outcomes, and importance of the event. All of these

measures, as well as demographic variables, are described next.

Operationalization of study constructs

According to the Triangle Model of Responsibility, responsibility is a function of

three linkages that join the three elements of the model (i.e., prescriptions, identity, and

event). The three linkages are the prescription-event link (prescription clarity), the

identity-event link (personal control), and the prescription-identity link (duty) (Schlenker

et al., 1994). For the purposes of this study, the model elements and linkages were

defined and operationalized as described below. The survey instrument also contained a








measure of pharmacist provision of pharmaceutical care services and various

demographic variables.

Clarity of standards. This variable was defined as the extent to which a clear set of

prescriptions was perceived to exist that should be applied to identifying, resolving, or

preventing drug therapy problems in a particular drug-related event. Recall that

prescriptions are rules for conduct, which may include laws, moral codes, regulations,

and social norms in groups (Schlenker et al., 1994). For this study, clarity of six specific

prescriptions (standards) was measured, as well as general clarity of the set of standards

(Table 4-1).




Table 4-1. Standards for Identifying, Resolving, and Preventing Drug Therapy Problems
Standards
Pharmacy laws and regulations
Practice standards advocated by pharmacy associations
Policies at my workplace
Clinical or disease management guidelines
Professional code of ethics
My knowledge of usual practices by fellow pharmacists



The six specific standards were selected based on pharmacists' comments from

earlier interviews, pharmacy administration literature, and the Triangle Model of

Responsibility. To confirm that pharmacists viewed these six standards as valued

principles of conduct, respondents in the pretest were asked, "To what extent do you

agree or disagree that each of these are valued sources for guidance of conduct?" (items 1

- 6, Appendix A). A total of 68 90 % of respondents either strongly agreed or

somewhat agreed with the questions (Figure 4-3).









100%
R
P 80%
P e 80% Strongly Disagree
es
r 60% OMl Somewhat Disagree
c 0- INeither
e 40% Somewhat Agree
ns 20% | Strongly Agree
t 20%
0%0-

0% 0



Standards


Figure 4-3. Percent Response to Standards as Valued Sources for Guidance of Conduct in
Identifying, Resolving, and Preventing Drug Therapy Problems


Clarity of the six specific standards was measured in with the following question

in Scenario 1, "To what extent do you agree or disagree that each of these standards

provides clear goals and rules to resolve Ms. Smith's drug therapy problemss)" (items 8

- 13, Appendix A) and the following question in Scenario 2, "To what extent do you

agree or disagree that each of these standards provides clear goals and rules to prevent

potential drug therapy problems that Mr. Wilson may experience?" (items 27 32,

Appendix A). Responses were measured on a 5-point scale: l=strongly agree,

2=somewhat agree, 3=neither agree nor disagree, 4=somewhat disagree, and 5=strongly

disagree. Responses were reverse coded so that higher scores were indicative of greater

levels of clarity.








In the pretest PDTQ, general clarity of standards was measured with two items in

Scenario 1: "Overall, I think there is a clear set of standards for helping to resolve Ms.

Smith's drug therapy problemss" and "In general, the set of standards for helping to

resolve Ms. Smith's drug therapy problems) are very ambiguous" (items 14 and 15,

Appendix A). Two items were used to measure this construct in Scenario 2: "Overall, I

think there is a clear set of standards for helping to prevent Mr. Wilson's potential drug

therapy problemss" and "In general, the set of standards for helping to prevent Mr.

Wilson's potential drug therapy problems) is very ambiguous" (items 33 and 34,

Appendix A). Responses were measured on a 5-point scale: 1 =strongly agree,

2=somewhat agree, 3=neither agree nor disagree, 4=somewhat disagree, and 5=strongly

disagree. Positively worded items were reverse coded so that higher scores were

indicative of greater levels of clarity. These four items had a coefficient alpha of .81 in

the pretest. All corrected item-total correlations were .50 or higher and all inter-item

correlations were .30 or higher (Table 4-2).




Table 4-2. General Clarity of Standards Items from Pretest: Inter-Item Correlations,
Corrected Item-Total Correlations, Means, and Standard Deviations
R14* 15 R33* 34
R14 ---
15 .66 ---
R33 .69 .49 ---
34 .37 .39 .55 ---
Corrected item-total .71 .62 .71 .50
correlation
Mean 4.20 3.35 4.15 3.55
Standard Deviation 0.83 1.27 0.74 1.00
Note: N=57
*Positively worded items are reverse coded (R) so that higher scores are indicative of
greater levels of clarity (possible range = I 5).








Because clarity may also be influenced by lack of conflicting standards or

prescriptions (Schlenker et al., 1994), an extra item was added to each scenario in the

revised PDTQ: "In my opinion, none of the standards for helping to resolve Ms. Smith's

drug therapy problems) conflict with one another" and "In my opinion, none of the

standards for helping to prevent Mr. Wilson's drug therapy problems) conflict with one

another" (items 9 and 28, Appendix E).

Personal control. This variable was defined as the extent to which a pharmacist

perceived s/he had personal control over identifying, resolving, or preventing a drug

therapy problem from becoming an adverse drug therapy outcome. In the pretest PDTQ,

personal control was measured with three items in Scenario 1: "Factors outside of my

control interfere with my ability to help resolve Ms. Smith's drug therapy problemss)" "I.... I

have plenty of control over helping Ms. Smith experience better outcomes from her

asthma therapy," and "There is very little I can do to help Ms. Smith achieve better

control of her asthma" (items 16, 19, 22; Appendix A). Three items were also used to

measure this construct in Scenario 2: "Factors outside of my control interfere with my

ability to help prevent Mr. Wilson from experiencing drug therapy problems," "I have

plenty of control over helping Mr. Wilson avoid adverse outcomes from his diabetes

therapy," and "There is very little I can do to help Mr. Wilson achieve stable control of

his diabetes" (items 35, 38, 41; Appendix A). Responses were measured on a 5-point

scale: 1 =strongly agree, 2=somewhat agree, 3=neither agree nor disagree, 4=somewhat

disagree, and 5=strongly disagree. Positively worded items were reverse coded so that

higher scores were indicative of greater levels of personal control.








These six items had a coefficient alpha of .87. All corrected item-total

correlations were .50 or higher and all inter-item correlations were .30 or higher (Table 4-

3). These items remained the same in the revised PDTQ (items 10, 13, 16, 29, 32, 35;

Appendix E).




Table 4-3. Personal Control Items from Pretest: Inter-Item Correlations, Corrected Item-
Total Correlations, Means, and Standard Deviations
16 R19* 22 35 R38* 41
16 ---
R19 .56 ---
22 .54 .51 ---
35 .71 .57 .44 ---
R38 .54 .65 .32 .68 ---
41 .62 .54 .71 .53 .47 ---
Corrected item-total .75 .70 .60 .76 .69 .69
correlation
Mean 3.35 3.46 4.48 3.33 3.48 4.24
Standard Deviation 1.56 1.34 0.69 1.44 1.22 0.91
Note: N=54
*Positively worded items are reverse coded (R) so that higher scores are indicative of
greater levels of personal control (possible range = 1 5).



Professional duty. This variable was defined as the extent to which a pharmacist

perceived s/he had a professional duty as a pharmacist to identify, resolve, or prevent a

drug therapy problem. In the pretest PDTQ, professional duty was measured with three

items in Scenario 1: "As a pharmacist, I have a professional duty to help resolve Ms.

Smith's drug therapy problemss)" "Helping to resolve Ms. Smith's drug therapy

problems) is outside of my professional duty as a pharmacist," and "As a pharmacist, I

am professionally obligated to ensure that Ms. Smith's drug therapy problems) is(are)

resolved" (items 17, 20, 23; Appendix A). Three items were also used to measure this

construct in Scenario 2: "As a pharmacist, I have a professional duty to help prevent Mr.








Wilson from experiencing potential drug therapy problems," "Helping to prevent Mr.

Wilson from experiencing drug therapy problems is outside of my professional duty as a

pharmacist," and "As a pharmacist, I am professionally obligated to ensure that Mr.

Wilson does not experience any drug therapy problems" (items 36, 39, 42; Appendix A).

Responses were measured on a 5-point scale: I =strongly agree, 2=somewhat agree,

3=neither agree nor disagree, 4=somewhat disagree, and 5=strongly disagree. Positively

worded items were reverse coded so that higher scores were indicative of greater levels of

professional duty.

The professional duty items had a coefficient alpha of .80. The corrected item-

total correlations for these six items ranged from .42 to .74, with items 20 and 36 having

corrected item-total correlations below .50 (Table 4-4).




Table 4-4. Professional Duty Items from Pretest: Inter-Item Correlations, Corrected Item-
Total Correlations, Means, and Standard Deviations
R17* 20 R23* R36* 39 R42*
R17 ---
20 .37 ---
R23 .60 .38 ---
R36 .25 .25 .32 ---
39 .55 .51 .52 .29 ---
R42 .31 .13 .67 .43 .45 ---
Corrected item-total .58 .42 .74 .42 .66 .56
correlation
Mean 4.53 4.42 3.78 4.45 4.40 3.98
Standard Deviation 0.88 0.96 1.27 0.74 0.78 1.13
Note: N=55; corrected item-total correlations <50 and inter-item correlations <.30 are
underlined.
*Positively worded items are reverse coded (R) so that higher scores are indicative of
greater levels of professional duty (possible range = I 5).








Three of the four inter-item correlations below .30 were from items across scenarios.

Because of the apparent lack of clarity regarding the drug therapy problem in Scenario 2,

it was expected that revision of this scenario in the revised survey instrument would

improve the performance of the professional duty items. Therefore, these items remained

the same in the revised PDTQ (items 11, 14, 17, 30, 33, 36; Appendix E).

Responsibility for drug therapy outcomes. This variable was defined as the extent

to which a pharmacist perceived s/he was responsible for a patient's drug therapy

outcomes. In the pretest PDTQ, responsibility for drug therapy outcomes was measured

with three items in Scenario 1: "I will feel responsible if Ms. Smith's drug therapy

problems) interfere(s) with her asthma therapy outcomes," "I won't feel responsible if

Ms. Smith's drug therapy problems) remains) unresolved," and "I will feel responsible

if Ms. Smith does not experience her desired asthma therapy outcomes" (items 18, 21,

24; Appendix A). Three items were also used to measure this construct in Scenario 2: "I

will feel responsible if Mr. Wilson experiences a drug therapy problem that interferes

with his diabetes therapy outcomes," ".... I won't feel responsible if Mr. Wilson experiences

a drug therapy problem with his diabetes treatment," and "I will feel responsible if Mr.

Wilson does not experience his desired diabetes therapy outcomes" (items 37, 40, 43;

Appendix A). Responses were measured on a 5-point scale: 1 =strongly agree,

2=somewhat agree, 3=neither agree nor disagree, 4=somewhat disagree, and 5=strongly

disagree. Positively worded items were reverse coded so that higher scores were

indicative of greater levels of responsibility for drug therapy outcomes.

These six items had a coefficient alpha of .94. All corrected item-total

correlations were .50 or higher and all inter-item correlations were .30 or higher (Table 4-








5). Therefore, these items remained the same in the revised PDTQ (items 12, 15, 18, 31,

34,37; Appendix E).




Table 4-5. Responsibility Items from Pretest: Inter-Item Correlations, Corrected Item-
Total Correlations, Means, and Standard Deviations
R18* 21 R24* R37* 40 R43*
R18 ---
21 .69 ---
R24 .78 .63 ---
R37 .69 .55 .73 ---
40 .75 .76 .64 .73 ---
R43 .75 .63 .90 .80 .64 ---
Corrected item-total .84 .73 .84 .79 .80 .86
correlation
Mean 3.96 3.89 3.58 3.69 3.78 3.64
Standard Deviation 0.98 1.12 1.26 1.15 1.07 1.04
Note: N=55
*Positively worded items are reverse coded (R) so that higher scores are indicative of
greater levels of responsibility for drug therapy outcomes (possible range = 1 5).



Event importance. The Triangle Model of Responsibility defines an event as a

unit of action and its consequences (Schlenker et al., 1994). For the purposes of this

study, an event was defined as the consequences that a patient would experience if his or

her drug therapy problems were not identified, resolved, or prevented. This variable was

measured with one item in Scenario 1: "I think Ms. Smith will experience severe

consequences if her drug therapy problems) is(are) not resolved" and one item in

Scenario 2: "I think Mr. Wilson will experience severe consequences if his potential drug

therapy problems) is(are) not prevented" (items 25 and 44, Appendix A). Responses

were measured on a 5-point scale: 1 =strongly agree, 2=somewhat agree, 3=neither agree

nor disagree, 4=somewhat disagree, and 5=strongly disagree. Both items were reverse








coded so that higher scores were indicative of greater importance (i.e., consequences) of

the event.

The means (standard deviations) for items 25 and 44, respectively, were 4.22

(0.85) and 3.87 (1.00). The means (standard deviations) for items 25 and 44,

respectively, were 4.22 (0.85) and 3.87 (1.00). The difference in pharmacists' ratings of

the severity of the two events was statistically significant (paired t = 2.34, p<.023). The

paired samples correlation for these two items was .32 (p<.O 19).

Identity importance. This variable was defined as the extent to which pharmacists

believed their professional identity to be important. Identity importance was measured

with two items: "Being a pharmacist is a crucial part of who I am" and "There are more

important parts of my personal identity than being a pharmacist" (items 71 and 72,

Appendix A). Responses were measured on a 5-point scale: l=strongly agree,

2=somewhat agree, 3-=neither agree nor disagree, 4=somewhat disagree, and 5=strongly

disagree. Item 71 was reverse coded so that a higher score was indicative of a greater

level of identity importance.

The means (standard deviations) for items 71 and 72, respectively, were 4.50

(0.68) and 1.96 (1.01). The inter-item correlation between these two items was .10 and

the coefficient alpha was .17. Item 72 was intended to be interpreted as a negatively

worded item about identity importance, where pharmacists who felt their professional

identity to be important would have responded with a high number (i.e., 4=somewhat

disagree, 5=strongly disagree); however, only 3 out of 58 (5.2%) pharmacists responded

in this manner. It is apparent that most of the respondents felt they had more important








parts of their personal identity than being a pharmacist (e.g., parent, spouse). Having

these views does not negate believing one's professional identity to be important.

Both of these items were excluded from the revised PDTQ. One reason for the

exclusion was the problem associated with item 72. Secondly, responses to item 71 were

highly skewed with a mean of 4.50 out of a possible 5. In addition, these items were not

crucial to the study hypotheses, which are discussed later in this chapter. Finally, these

items were good candidates to be excluded in order to shorten the PDTQ for the

validation study. Due to the low response rate observed for the pretest (27%), decreasing

the estimated administration time of 20 minutes seemed indicated.

Pharmaceutical care provision. Pharmacists' self-reports of pharmaceutical care

activities were measured using the Direct Patient Care component of the Behavioral

Pharmaceutical Care Scale (BPCS-DPC) (Odedina & Segal, 1996). In the BPCS-DPC,

respondents are asked to indicate to how many of their last five patients with chronic

conditions (e.g., asthma, diabetes) they provided certain services. This instrument is

composed of 18 items intended to measure pharmacists' performance of activities

associated with six pharmaceutical care domains (Table 4-6).

Coefficient alpha estimates from the pretest for the multiple-item domains were as

follows: patient assessment (.91), documentation (.79), and implementation of therapeutic

objectives and monitoring plans (.83). The coefficient alpha for the entire BPCS-DPC

was .92.

Due to the low response rate to the pretest (27%), this instrument was shortened

for the revised PDTQ. Items from the Documentation domain were excluded, resulting

in a total of 12 items with a possible score range of zero to 60 (items 39 50, Appendix











Table 4-6. Domains and Items of the Behavioral Pharmaceutical Care Scale-Direct
Patient Care
Patient Assessment Domain
Ask the patient to describe his/her medical condition, including a description of
medical problems and symptomatology.
Ask the patient what he/she wants to achieve from the drug therapy.
Ask the patient questions to assess actual patterns of use of the medication.
Ask the patient questions to find out if he/she might be experiencing drug therapy
problems.
Ask the patient questions to find out about the perceived effectiveness of drugs
he/she is taking.
Ask the patient questions to ascertain whether the therapeutic objectives) is(are)
being reached.
Documentation Domain
Document information about the patient's medical conditions on written records,
computerized notes, or by other formal mechanisms in a form that could be read and
interpreted by another health care practitioner in my absence.
Document all medications currently being taken by the patient on written records,
computerized notes, or by other formal mechanisms in a form that could be read and
interpreted by another health care practitioner in my absence.
Document the desired therapeutic objectives for the patient.
Document the drug therapy problems, potential or actual, on written notes.
Document the desired therapeutic objectives) for each of the drug therapy problems
identified.
Document any intervention made on the patient's file, prescription, report, or
medical order in a form that could be read and interpreted by another health care
professional.
Implementation of Therapeutic Objectives and Monitoring Plans Domain
Implement a strategy to resolve (or prevent) the drug therapy problems.
Establish follow-up plans to evaluate the patient's progress toward his/her drug
therapy objectives.
Carry out the follow-up plans established for the patient's progress toward his/her
drug therapy objectives.
Patient Record Screening Domain
* Check the patient's records for potential drug therapy problems.
Patient Advising and Counseling Domain
* Discuss the patient's drug therapy with him or her.
Verification of Patient Understanding Domain
* Verify that the patient understands the information I present to him or her.
(Odedina & Segal, 1996)








E). The documentation domain was selected for exclusion because it was not as directly

involved in identifying and resolving drug therapy problems. In addition, this subscale

had the lowest coefficient alpha, had items that were lengthy and time consuming to read,

and tied with the patient assessment subscale for having the highest number of items (i.e.,

6).

Demographic variables. Several demographic variables were measured.

Respondents were asked to indicate how many years they had practiced pharmacy, which

educational degrees they had received, which (if any) advanced certifications, licenses or

residencies they had received/completed, their current practice setting, their current

position at their job, their age, and their gender. For the sake of brevity, two

demographic items that had been included in the pretest were removed from the revised

PDTQ: year graduated from pharmacy school and number of pharmacy association

memberships.


Summary of Survey Development

This section described how the PDTQ was developed for this study. Following

the design of the survey instrument, it was pretested among a sample of community

pharmacists. Results from the pretest indicated that these pharmacists were generally

able to identify the relevant drug therapy problem in the asthma scenario (Scenario 1).

This issue was not as clear with the diabetes scenario (Scenario 2), but this scenario was

revised in preparation for the validation study.

The length of the PDTQ may have been an issue in the low pretest response rate

(27%). Revisions were made to the PDTQ with the intent of decreasing the

administration time. These revisions included removing two items measuring identity








importance, shortening the BPCS-DPC by excluding the six Documentation items and

removing two demographic items. Furthermore, the revised PDTQ did not contain items

measuring importance of standards or identification of drug therapy problems; these

items had been included in the pretest version for preliminary analyses to establish the

appropriateness of the standards selected and the patient cases that were used as stimuli in

the instrument. Neither scale was measuring a study construct or a variable in a test of

hypotheses. Overall, the changes in the revised PDTQ lowered the administration time

from approximately 20 to 15 minutes.

The revised PDTQ was then tested among a sample of Florida community and

ambulatory care pharmacists in order to validate the measures that were developed. The

next section describes how the measures were validated.




Study Sample


The sample for the validation study consisted of community and ambulatory care

pharmacists in Florida. This section describes the procedure that was used to select the

study sample.

The 2000 Florida Board of Pharmacy mailing list was used as the sampling frame

for the validation study. The list contained names and addresses of 13,994 active

members of the Florida Board. The names of the 300 pharmacists who had been mailed

surveys for the pretest, as well as the 8 pharmacists who had been interviewed, were

removed from this initial list. A systematic sampling of the remaining 13,686

pharmacists was undertaken by selecting every other name on the list. Of the 6,843

names selected, 6,798 had complete usable addresses. These remaining pharmacists








constituted the sample that was contacted regarding the study. Each pharmacist was

assigned a subject identification number for study purposes.

Because the Florida Board of Pharmacy list did not include information regarding

pharmacists' practice settings, a screening process was used to identify pharmacists who

practiced in community and ambulatory care settings. On July 17, 2001, a prenotice

letter and screening postcard (see Appendices C and D, respectively) were mailed to each

of the 6,798 selected pharmacists.

The prenotice letter was printed on letterhead stationery, and included

individualized salutations for each addressee (e.g., "Dear John Doe") and an original

signature in contrasting ink. These elements of correspondence personalization have

been shown to increase response rates between 5% and 11% in general public surveys

(Dillman, 2000). An ancillary endorsement statement, "address service requested," was

printed on the envelopes in which the letters and postcards were mailed. The purpose of

this endorsement was to alert the United States Post Office to send the new address of

each forwarded mail piece to the principal investigator. Consequently, further

correspondence with subjects was facilitated by having their most recent addresses.

The prenotice letter informed pharmacists about the study and requested that they

complete and return the screening postcard. The postcard contained two questions: (1)

"Are you currently practicing as a pharmacist?" and (2) "What is your practice setting?"

The following practice settings were listed as choices: community pharmacy, hospital

pharmacy, ambulatory/outpatient clinic, and other. Subject identification numbers were

printed below the two questions on each postcard. The reverse side of the screening

postcard contained a business reply address and postage.