Patient characteristics as predictors for the psychosocial functioning and quality of life of implantable cardioverter d...

MISSING IMAGE

Material Information

Title:
Patient characteristics as predictors for the psychosocial functioning and quality of life of implantable cardioverter defibrillator recipients
Physical Description:
xi, 112 leaves : ill. ; 29 cm.
Language:
English
Creator:
Lewis, Tara Lynn Saia, 1970-
Publication Date:

Subjects

Subjects / Keywords:
Defibrillators, Implantable   ( mesh )
Social Support   ( mesh )
Quality of Life   ( mesh )
Patients -- psychology   ( mesh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2001.
Bibliography:
Includes bibliographical references (leaves 101-110).
Statement of Responsibility:
by Tara Lynn Saia Lewis.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
oclc - 54019180
ocm54019180
System ID:
AA00011161:00001

Full Text









PATIENT CHARACTERISTICS AS PREDICTORS FOR
THE PSYCHOSOCIAL FUNCTIONING AND QUALITY OF LIFE
OF IMPLANTABLE CARDIOVERTER DEFIBRILLATOR RECIPIENTS














By

TARA LYNN SAIA LEWIS


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

Tara Lynn Saia Lewis












ACKNOWLEDGMENTS


I would like to thank my family for their support and encouragement over the

years during my pursuit of this degree. In particular, I would like to express my deepest

appreciation to my sister Cherie and my husband Carl. Each was instrumental in

providing me with the motivation, support, encouragement, and resources necessary for

me to have completed my dissertation research and degree requirements. Without their

continuous support and assistance, I am certain I would have discontinued my efforts

much earlier along the path toward this goal.

I would also like to express my thanks to the many students and friends within the

department that provided me with endless support over the years. In particular, I would

like to recognize Miles Rogish and Tricia Zawacki-without them I could not have

achieved this degree. In addition, I would like to express a special thanks to the graduate

students in the program who provided me with endless assistance in all phases of this

dissertation research project and without whom there would have been significantly

fewer patients enrolled in this study. In particular, I would like to recognize the valuable

contributions of Guido Urizar, Brian Sirois, Shannon Jackson, and Robyn Wallace. It is

an understatement to say that I owe you, but I will say it anyway.

Within the area of professional development, I would like to express my sincerest

level of admiration to Dr. Lucretia Mann, Dr. Duane Dede, and Dr. Cynthia Belar for

their personal interest in my professional development and willingness to share with me a







part of themselves so that I might learn to become a more talented psychotherapist. They

will always serve in my memory as the examples of clinical excellence that I must strive

to emulate.

I would also like to express my appreciation to each of the members of my

doctoral dissertation committee, Dr. Samuel Sears, Dr. James Rodrigue, Dr. Eileen

Fennell, Dr. Michael Perri, and Dr. Linda Bobroff, for their assistance to me during

various stages of this research endeavor. Special thanks goes to Dr. Rodrigue for his

provision of detailed editorial comments in the final stages of my manuscript preparation.

I found the comments extremely helpful and am fairly certain I would have defended my

dissertation at a much later date if I had not had his assistance regarding the fine-tuning

of my manuscript. My sincere gratitude goes to Dr. Fennell for her continuous support

and encouragement of my personal development and progress over the years. I will be

forever indebted to her for her willingness to make the time to meet with me during my

visits to the department in order to discuss the data and provide me with guidance

regarding statistical analyses procedures and the presentation of the results. Without her

assistance in this area, I believe I would still be in the dark contemplating the optimal

ways to analyze the secondary variables of interest in this study. And finally, I would

like to express my deepest thanks to my dissertation chair, Dr. Samuel Sears, for his

many years of service and friendship that have brightened my experience within the

department and enhanced my training experiences. It has been through our relationship

that my passion for research has developed and my interested in providing psychological

services to medical patients has been strengthened. And while it is true we have had

some differences of opinion in the past, and will likely have more in the future, it is also







true that in the most questionable of times Dr. Sears never let go of his faith in my

potential and his belief in my ability to reach the goals I had set for myself. It is by virtue

of the fact that he trusted and respected me well enough to allow me the latitude to direct

my own course that the fruits of my perseverance are now an even sweeter reward.

While the journey to this point was often arduous, it has not been without recompense

and I am most appreciative of the lessons that I have learned along the way.

And last but not least, I'd like to acknowledge the contribution of my father. It is

he who instilled in me the firm belief that anything is possible through the creation of

one's own luck-the ingredients to which are the combination of intellect, confidence,

determination, and effort. It is in his footsteps I have tried to follow.













TABLE OF CONTENTS

page

ACKNOWLEDGMENTS ................................................ ............................. iii

LIST OF TABLES ......................................................................................... viii

A B STRA CT................................... .............................................................. ......x

LITERATURE REVIEW ................................................... ............................. 1

Introduction to the Implantable Cardioverter Defibrillator (ICD).................
Review of the Psychosocial Functioning of ICD Patients............................12

STUDY PURPOSE AND HYPOTHESES ......................................................26

M ETHOD ................................................................... ................................... 28

Participants.................................................................................................... 28
Procedure ................................................................................................33
M measures ................................................................................................. 35

RESULTS ................................................................... ................................... 42

Initial Psychological Assessment Measures ................................. ..........42
Follow-up Quality of Life Assessment Measures............................... ..43
Follow-up of ICD Specific Post-Implantation Variables............................ 46
Relationships between Predictor and Outcome Variables ..........................51
Regression Models for Predicting Quality of Life Variables .....................53
Predicting Short-term Outcome Variables..................................................54
Predicting Long-term Outcome Variables ............................................63
Exploratory Analyses.................................................................................. 70







DISCUSSION..................... ................................ ..........................................77

Descriptive Analyses of the ICD Patient Sample ....................................... 77
Predictors of Quality of Life in ICD Patients ............................................. 83
Strengths and Limitations.............................. ............................................86
Conclusions............................. ................................................................87
Future Directions ......................................................... ...............................89


APPENDIX A


APPENDIX B


CHRONOLOGICAL REVIEW OF
PSYCHOSOCIAL ICD LITERATURE..................................91

UNPUBLISHED QUESTIONNAIRES ..................................99


REFERENCES ....................................................................................................101

BIOGRAPHICAL SKETCH ................................................. ......................... 111













LIST OF TABLES


Table page

1. General Characteristics of Study Participants........................................29

2. The 10 Most Commonly Diagnosed Medical Conditions .......................31

3. Summary of the Most Frequently Provided Prescription
M medications at Discharge.............................................................32

4. Summary of Mean Quality of Life Questionnaire Scale Scores
At Follow-up Assessment Intervals...............................................44

5. Mean Differences in the Seattle Angina Questionnaire Subscales
Across Samples................................................................ .....47

6. Summary of ICD Firing Experiences at the Short and Long-term
Follow-up Intervals............................................................ .........49

7. Zero-order Partial Correlations Among Predictor and Outcome
Variables Controlling for Age and LVEF....................................52

8. Regression Coefficients for the Predictor Variables
At the Short-term Follow-up Interval ..........................................56

9. Hierarchical Regression Model Results for Predicting Mental Health
Quality of Life at Short-term Follow-up................................ ...58

10. Hierarchical Regression Model Results for Predicting General Health
Quality of Life at Short-term Follow-up............................. ...59

11. Hierarchical Regression Model Results for Predicting Physical Limitations
Quality of Life at Short-term Follow-up.......................................61

12. Hierarchical Regression Model Results for Predicting Disease Perception
Quality of Life at Short-term Follow-up............................. ...62

13. Regression Coefficients for the Predictor Variables
At the Long-term Follow-up Interval ..........................................64







14. Hierarchical Regression Model Results for Predicting Mental Health
Quality of Life at Long-term Follow-up......................................65

15. Hierarchical Regression Model Results for Predicting General Health
Quality of Life at Long-term Follow-up........................................66

16. Hierarchical Regression Model Results for Predicting Physical Limitations
Quality of Life at Long-term Follow-up......................................68

17. Hierarchical Regression Model Results for Predicting Disease Perception
Quality of Life at Long-term Follow-up......................................69

18. Quality of Life Outcomes According to History of Depression
At Initial Assessm ent ................................................................ 72

19. Quality of Life Outcomes According to Trait Anxiety
At Initial Assessment .................................................................73

20. Quality of Life Outcomes According to Dispositional Optimism
At Initial Assessment .................................................................74

21. Quality of Life Outcomes According to History of Social Support
At Initial Assessment ................................................................ 75












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

PATIENT CHARACTERISTICS AS PREDICTORS FOR
THE PSYCHOSOCIAL FUNCTIONING AND QUALITY OF LIFE
OF IMPLANTABLE CARDIOVERTER DEFIBRILLATOR RECIPIENTS

By

Tara Lynn Saia Lewis

August 2001


Chair: Samuel F. Sears
Major Department: Clinical and Health Psychology

Research has shown the implantable cardioverter defibrillator (ICD) may be

associated with psychosocial complications and reduced quality of life. However, few

studies have prospectively examined patient characteristics as predictors of outcome post-

implant. The major objectives of the current study were 1) to obtain descriptive

information regarding pre-implant psychological functioning, experience of ICD firings,

and post-implant quality of life, and 2) to examine the association between pre-implant

psychological patient characteristics, experience of ICD firings, and post-implant quality

of life ratings. This study prospectively assessed 88 first-time implanted ICD patients,

and included both short and long-term follow-up assessments (8 & 14 months,

respectively). Pre-implant psychological functioning was within normal limits for the

ICD study sample, whereas post-implant quality of life was mild to moderately impaired

across all indices. Approximately half of the sample reported experience of ICD firings







at each follow-up. Separate hierarchical regression analyses were conducted to examine

the degree of association between the predictor variables (history of depression, trait

anxiety, dispositional optimism, perceived social support, and ICD firings) and each of

the four dependent variables (mental health, general health, physical limitations, and

disease perception quality of life ratings). All analyses controlled for initial assessment

patient age and ejection fraction. The proposed model accounted for a significant

proportion of the variance observed in patient quality of life ratings for mental health,

general health, and physical limitations at both follow-up intervals (variance accounted

for ranged from 41.7% to 64.5% depending upon outcome assessed). Further, the

proportion of variance accounted for by the psychological variables was equal to or larger

than the variance accounted for by the control variables. To a lesser degree, experience

of ICD firings significantly added to each model. Exploratory ANOVAs were conducted

and ICD patients with lower levels of dispositional optimism had significantly lower

quality of life ratings in all areas assessed at both follow-up intervals. Additionally,

patients with higher levels of trait anxiety had significantly lower quality of life ratings in

all areas assessed at the long-term follow-up interval. This data set emphasizes the

importance of a biopsychosocial model in understanding quality of life outcomes for ICD

patients.












LITERATURE REVIEW


Introduction to the Implantable Cardioverter Defibrillator

The single leading cause of death in the United States is coronary heart disease

(CHD). Approximately one of every five deaths can be attributed to CHD, and according

to the American Heart Association (2001b), CHD was responsible for more than 450,000

deaths in the United States in 1998 alone. Each year, about 220,000 Americans die from

sudden cardiac arrest (a manifestation of CHD) without being hospitalized; more than

80% of these deaths are triggered by life-threatening ventricular arrhythmias, usually

ventricular fibrillation (VF) or ventricular tachycardia (VT). However, sudden cardiac

arrest can be reversed in most victims if it is treated within a few minutes of onset with an

electric shock to the heart to restore a normal heartbeat. Unfortunately, it is estimated

that more than 95 percent of cardiac arrest victims die before reaching the hospital to

receive this treatment (American Heart Association, 2001a). In fact, a victim's chances

of survival are reduced by 7-10 percent with every minute that passes and few attempts at

resuscitation succeed after 10 minutes.

The implantable cardioverter defibrillator (ICD) is a life-saving treatment that has

been developed specifically to treat patients at-risk for sudden cardiac arrest. The ICD is

a device that has been designed to identify, record, and abort life-threatening ventricular

arrhythmias by providing electric shock to the heart muscle when indicated. Given the

lack of efficacious alternatives available to combat the exceptionally high mortality rate







associated with sudden cardiac arrest, the ICD has become commonly utilized, as it has

been proven the most superior treatment option for patients with high risk of sudden

cardiac arrest (Bocker & Breithardt, 2000). Correspondingly, it has been reported that

the number of annual ICD implants rose from approximately 15,000 in 1993 to 50,000 in

1999, an increase of 227% (Winters et al., 2001). Additionally, as reported by Winters

and colleagues, for the year 2000, the implantation of 61,000 ICDs was projected in the

United States, with an additional 20,000 ICD implants projected worldwide; related

expenditure projections ranged from 1.342 and 1.620 billion dollars (United States vs.

worldwide, respectively).

In clinical research, the ICD has been proven highly effective in its treatment and

has greatly reduced mortality in patients at-risk for sudden cardiac arrest (Glikson &

Friedman, 2001), with survival rates from all-cause mortality reported as high as 93% at

one year and 57% at five years (Grimm, Flores, & Marchlinski, 1993). Survival rates

specifically from arrhythmic death have been even more promising, with 98% at one year

and 83% at five years being reported. When considering the survival rate for untreated

individuals experiencing sudden cardiac arrest outside of the hospital is less than 10% in

most circumstances (American Heart Association, 2001a) and between 30-50% of the

individuals with an experience of sudden cardiac arrest have a repeat event within two

years (Gregoratos et al., 1998), the survival rate for individuals implanted with the ICD is

indeed quite remarkable.


Evolution of the Device

The ICD was first developed at Johns Hopkins University in 1980 as a last resort

treatment option for individuals at high risk for sudden cardiac death (Mower, 1994). At







that time, it was approved only for those patients with life-threatening ventricular

arrhythmias who were also refractory to pharmacological therapy. In 1985, it gained

FDA approval for use in the United States as a treatment option for patients diagnosed

with life-threatening ventricular arrhythmias, specifically VT or VF, who were also

refractory to antiarrhythmic drug therapy. In 1991, official class guidelines were

developed and the ICD was recommended as a treatment option for drug refractory

patients diagnosed with life-threatening ventricular arrhythmias, syncope with inducible

ventricular tachycardia episodes, or a history of cardiac arrest not related to myocardial

infarction (Kolettis & Saksena, 1994). However, throughout the early 1990s, despite

these guidelines, ICD therapy was still frequently considered only as a last resort

treatment option, and many physicians still attempted to control their patient's ventricular

arrhythmias primarily with medications.

By 1996, however, numerous trials with the ICD had been conducted which

demonstrated limited surgical complications and low patient mortality. Indeed, in that

first decade of its birth, dramatic technological improvements had occurred with the ICD

(Davidson, VanRiper, Harper, & Wenk, 1994). Along with significant changes in the

implantation procedure (thoracotomy vs. subclavicular), the actual device volume has

been reduced substantially, by more than 50% over the years-without impacting the

longevity of the device (standard volume was 150-200cc in the 1980s compared to the

32-64cc common in the 2000s). Other advances in technology have brought about the

possibility of multiple and tiered therapies, dual chamber devices, and the ability to

record arrhythmic events. With all the new developments taken into consideration, it is







not surprising that the ICD has become a highly desirable treatment option for patients

with life-threatening arrhythmias.

In 1996 and 1997, two landmark trials originally designed to examine the efficacy

of ICD therapy as compared to more conventional pharmacological methods of treatment

were terminated earlier than planned due to significant decreases found in the all-cause

mortality rates of patients treated with the ICD. The first clinical trial, MADIT

(Multicenter Automatic Defibrillator Implantation Trial), was begun in 1990 and

examined the treatment efficacy of the ICD in post-myocardial infarction patients with

left ventricular ejection fraction (LVEF) dysfunction (LVEF <0.35), nonsustained

ventricular tachycardia, and inducible, sustained ventricular tachycardia. In 1996,

investigators stopped this trial and reported that a 54% decrease in mortality was

observed over a five-year period in patients treated with the ICD as compared to patients

treated with conventional pharmacological therapy (Moss, 1996; Moss et al., 1996). The

researchers noted that observed reductions in mortality in the post-myocardial infarction

patients at high-risk for sudden cardiac death were staggering and suggested that ICD

treatment guidelines should be altered to include post-myocardial infarction patients at

increased risk for sudden cardiac death. After the release of the results, the FDA changed

the indications for the ICD to include patients like those studied in the MADIT trial.

The second clinical trial, AVID (Antiarrhythmics Versus Implantable

Defibrillators), began in 1993 and compared the treatment efficacy of conventional

antiarrhythmic medications (amiodarone or sotalol) with ICD therapy. This trial was

terminated early (in the spring of 1997), having demonstrated a 39% reduction in

mortality at one year for those patients treated with the ICD (AVID Investigators, 1997).







Further, survival rates at two and three years continued to demonstrate significantly

reduced overall mortality in patients treated with the ICD as compared to antiarrhythmic

medications (reductions in mortality by 27 and 31 percent, respectively). AVID was the

first trial to demonstrate the ICD's efficacy in improving overall survival rates

(decreasing both arrhythmic and nonarrhythmic deaths) specifically in patients diagnosed

with life-threatening arrhythmias. The AVID trial effectively demonstrated that

antiarrhythmic medications were an inferior treatment approach to life-threatening

arrhythmias when compared to ICD therapy. As a direct result of this study, ICD

implantation became the indicated treatment plan, rather than a possible treatment option,

for all patients diagnosed with life-threatening VT or VF, and antiarrhythmic medications

became a secondary treatment option in the American Heart Association guidelines.

While the results of the MADIT and AVID trials served to increase the popularity

and acceptance of the ICD as a first line intervention for patients at high risk for life-

threatening ventricular arrhythmias, more recent research protocols have been developed

to examine whether the ICD should be considered a primary prevention method for other

at-risk populations. Large national trials such as the Sudden Cardiac Death in Heart

Failure Trial (SCD-Heft), Defibrillators in Nonischemic Cardiomyopathy Tachycardia

Evaluation, MADIT II, Defibrillators in Acute Myocardial Infarction Trial, Multicenter

Unsustained Tachycardia Trial (MUSTT), and the Coronary Artery Bypass Graft Patch

(CABG-Patch) trial were all designed specifically to determine if primary prevention

with the ICD is indicated in particular cardiac patient populations (Prystowsky & Nisam,

2000; Winters et al., 2001). However, to date only the MUSTT and the CABG-Patch

trials have been concluded. The MUSTT results were remarkably similar to the MADIT







in demonstrating a greater than 50% reduction in mortality for the ICD-treated patients;

the patient populations were also almost identical: previous myocardial infarction,

depressed left ventricular function, nonsustained VT, and inducible sustained VT (Buxton

et al., 1999). In contrast, despite many similarities in the patient populations, the CABG-

Patch trial results conflict with MADIT and MUSTT and do not show a significant

survival advantage to patients given an ICD following coronary artery bypass grafting

(Bigger, 1997). It has been hypothesized that the influence of revascularization on

ischemia and left ventricular function resulted in the lack of ICD benefit in this study

(Block & Breithardt, 1999; Curtis et al., 1997); however, it is also important to note that

the CABG-Patch Trial patients did not have clinically manifest arrhythmias-a key

requirement of patients in MADIT and MUSTT.

As another extension of its use, ICD therapy has been suggested to be a viable

treatment option for many patients awaiting cardiac transplantation (Trappe & Wenzlaff,

1995). Statistics show that approximately one-third of the patients on the waiting list for

heart transplantation die from sudden cardiac arrest; a recent extensive review of the

literature and the use of ICDs prior to heart transplantation by Schmidinger (1999)

concluded that although ICD therapy does not necessarily prolong life in this patient

population at large, its use does not negatively impact heart transplantation survival rates

and may be indicated to prolong survival specifically for patients on the waiting list with

a history of malignant ventricular arrhythmias.

Given the limited number and mixed results of the primary prevention studies

currently completed, it is clear that more research is necessary to determine which cardiac

patient populations may best benefit from prophylactic ICD implantation, particularly







when the patients being considered may be asymptomatic from VT or VF, but are

presumed at higher risk for sudden cardiac arrest (Hohnloser, 1999; Schlapfer,

Kappenberger, & Fromer, 1999). However, as controlled efficacy trials continue to

identify the ICD as a superior and cost-effective, life-saving treatment option for

expanded cardiac patient populations, it can be predicted that there will likely be

widespread usage of the ICD in the decades to come.


Capabilities and Function

The ICD is a sophisticated device that is capable of delivering four types of

therapy: 1) bradycardia pacing, 2) antitachycardia pacing, 3) cardioversion, and 4)

defibrillation. Therapy programs are individually programmed and allow for escalation

from simple pacing functions to the delivery of low-energy synchronous cardioversion or

high-energy defibrillation based upon the patient's needs (Davidson et al., 1994; Vlay,

1996). The strength of the shock is individually programmed, typically ranging from 25

to 42 joules, depending upon the patient's defibrillation threshold (Glikson & Friedman,

2001). If the first shock fails to abort the arrhythmia, the device is designed to recharge

within seconds and deliver between three and seven additional shocks. Each shock that

the patient receives is intended to be a life-saving shock. The initial shock voltage

delivered has been significantly reduced over recent years, reflecting the technological

advances and increased sensitivity of the device, and many shocks are now averted by the

use of low energy cardioversion therapies (usually 1 joule or less). In addition, the ICD

memory is able to store the number of arrhythmic episodes, successful responses, and

cycle length of R-R intervals before and after therapy. Most current devices also record

electrograms to allow for the direct examination of the arrhythmia for which the device







delivered therapy. This information has been extremely useful to practitioners to verify

patient report of incidents and to make adjustments in therapy programming (Knight,

Livingston, Gawlinski, & DeLurgio, 1997).

The ICD device is powered by lithium silver vanadium pentoxide cells and is

capable of delivering approximately 100-200 discharges (Mower, 1994; Vlay, 1996).

Longevity of the device varies depending upon the activated pacing functions, but can be

estimated to be approximately 5 years, ranging from 3-9 years. Battery replacement is

needed at regular intervals. The size of the device has decreased dramatically in the past

decade, and in 1994, the most commonly used active can system had a volume of

approximately 80-cc (a significant decrease from the 150-cc size common just a few

years prior). Today, the smallest device available on the market has a volume of only

32cc, a greater than 75% reduction from the device size available in 1985. For the vast

majority of patients today, the ICD is now placed in the left subclavicular area. The

surgical procedure only requires general anaesthetic and can be conducted within a

catheterization laboratory in an electrophysiological study testing room. Hospital stays

post-ICD implantations are typically only 24-48 hours.


Therapy Description

ICD therapy is programmed to occur whenever an arrhythmia is sensed.

Frequently, patients do not detect ICD therapy; this also holds true for certain types of

arrhythmias. Sometimes, however, unusual sensations may be detected when the ICD is

delivering therapy, even if the patient did not notice the symptoms of the arrhythmia.

This occurrence may sometimes serve as a warning that alarms an unsuspecting patient of

irregularities in the heartbeat.







The types of sensations experienced during ICD therapy will vary depending on

the arrhythmia. There are four main types of ICD therapy described in patient education

materials (Guidant Corporation, 1997). The first type, very low-energy pulses, is

delivered to the heart when the heartbeat slows down too much. This type of ICD

therapy is called bradycardia pacing, and is usually not detected by the patient.

Antitachycardia pacing (ATP) is the most common type of therapy delivered by the ICD.

Many patients do not feel ATP therapy when it is delivered. If it is observed, however,

ATP is typically experienced as a painless fluttering in the chest. Stronger than ATP,

cardioversion is another type of therapy delivered by the ICD. It consists of low-energy

shocks that can feel mildly uncomfortable. The patients typically describe these shocks

as feeling like a thump on the chest. Defibrillation is the strongest type of ICD therapy

that can be delivered. Many patients will become faint or lose consciousness when their

arrhythmia becomes this severe. Because of the physical symptoms of a severe

arrhythmia, patients are unlikely to feel the high-energy shocks of defibrillation. If a

patient is conscious, however, the shock will feel much like a "kick in the chest" and will

occur suddenly. It is described by most patients as painful, but is very brief, usually

lasting only for a second.


Patient Profile

Most ICD recipients are within their fifth to seventh decade of life, have chronic

cardiovascular disease, receive multiple medications, and experience numerous health

problems as well (Pinski & Trohman, 1995). The most typical ICD recipient can be

described as approximately 60 years old, male (80%), and white (90%), based on the

sample of 394 ICD recipients obtained in the AVID Trial (Curtis et al., 1997). In this







study, all of the ICD recipients were diagnosed with life-threatening ventricular

arrhythmias (VT or VF). However, co-morbid cardiac diagnoses were fairly common,

with approximately 81% of the ICD patient population diagnosed with coronary artery

disease, 55% diagnosed with hypertension, 48% diagnosed with angina, and 46%

diagnosed with NYHA (New York Heart Association) Class I or II congestive heart

failure (AVID Investigators, 1997). About two-thirds of the ICD recipients in this

sample had a positive history of myocardial infarction and most were also diagnosed with

depressed left ventricular function (mean LVEF = 0.32).


Evaluation Procedures and Cost

Prior to having ICD therapy recommended, a patient typically receives a

comprehensive medical evaluation and subsequent medical assessments on an outpatient

basis (Block & Breithardt, 1994). When life-threatening cardiac arrhythmias are

suspected, a hospitalization is recommended. During this time, the patient is extensively

evaluated and will usually undergo an electrophysiological assessment (EPS). This study

of the heart is conducted when the cardiologist needs a more precise method of

examination to determine the nature and origin of the cardiac problem; an EPS is almost

always conducted if ventricular arrhythmias are suspected and ICD therapy is being

considered. If the results of the EPS suggest that malignant VT or VF is probable, ICD

therapy is recommended. In most cases, the patient undergoes ICD implantation

concurrent with the EPS procedure or within 24 hours of positive EPS results. The total

cost of the initial hospitalization and ICD implantation has been estimated to range

between $44,000 and $55,000. As predicted, implantation costs did decline with

advances in surgical techniques and device technology, and presently, the ICD itself has







an average cost of $22,000 (Anderson & Camm, 1994; Mushlin et al., 1998;

O'Donoghue, Platia, Brooks-Robinson, & Mispireta, 1990). To address the issue of cost-

effectiveness, MADIT researchers have presented data indicating that providing ICD

implantation and its continued therapy to arrhythmia patients is no more expensive than

providing other approved life-saving procedures to other patient populations (Mushlin et

al., 1998). Other ICD cost analyses support this argument and have demonstrated that the

cost of ICD therapy (approximately $11,300 per year of life saved) is quite similar to or

better than other life-saving treatments and procedures, such as coronary artery bypass

surgery ($7,700 to $44,200), the treatment of hypertension ($11,100 to $23,000), and

kidney dialysis ($57,300 to $59,500) (Gorlin, 1995; Wever et al., 1996). Moreover, a

recent review of the cost-effectiveness ICD literature by Stanton and Bell (2000)

concluded that expenditures related to ICD total therapy might actually break even within

1-3 years post implantation. Further, these authors noted that with anticipated continuing

advances in ICD-related medical technology, cost-effectiveness of the ICD would likely

continue to improve. In summation, researchers in the area of cost-effectiveness have

collectively concluded that early ICD implantation is not only medically indicated, but

significantly more cost-effective as well, as compared to the existing alternative treatment

options, such as late ICD implantation or antiarrhythmic medication therapy and their

associated increased mortality.


Review of the Psychosocial Functioning of ICD Patients

Since the introduction of the ICD as a treatment recommendation for life-

threatening ventricular tachycardia and ventricular fibrillation, researchers have indicated

that many ICD patients do experience a variety of psychological complications following







ICD implantation, including anxiety, depression, adjustment difficulties, excessive fear of

ICD firings, and diminished quality of life. The reported incidence of each of these

patient complaints is significant, but has been known to vary greatly across studies,

primarily due to variances in research design, device technology, type of assessment

measures utilized, and assessment intervals reported. However, despite inconsistencies

within the literature, a common theme has continually emerged-many ICD recipients

experience psychological distress at some point following ICD implantation.

Furthermore, the existence of a relationship between patient experience of ICD firings

and psychological maladjustment has been implicated in a number of studies. The

following section reviews the existing psychosocial ICD literature according to topic, and

provides brief summaries of the major research efforts. At the conclusion of this section,

a methodological critique of the psychosocial ICD literature is provided. In addition, for

ease of reference and review, a table presenting summaries of the primary literature in

chronological order is also provided (see Appendix A).



Anxiety

Within the ICD patient population, anxiety has been the most frequently assessed

psychological variable. The presence of anxiety has been noted in ICD patients since the

earliest of research efforts (Pycha, Gullege, Hutzler, Kadri, & Maloney, 1986), and

anxiety appears to continue to be commonly experienced by ICD patients. Several

researchers have reported that ICD patients experience significantly higher levels of

anxiety than the general population (Keren, Aarons, & Veltri, 1991; Luderitz, Jung,

Deister, & Manz, 1996; Vlay, Olson, Fricchione, & Friedman, 1989). The number of

patients reporting clinically diagnostic levels of anxiety following ICD implantation has







ranged in the literature from 13 to 38 percent (Hegel, Griegel, Black, Goulden, &

Ozahowski, 1997; Herrmann et al., 1997; Konstam, Colbum, & Butts, 1995; Schuster,

Phillips, Dillon, & Tomich, 1998). Recent literature has consistently associated ICD

firings with increased reports of patient anxiety for certain subgroups of ICD patients

(Goodman & Hess, 1999; Heller, Ormont, Lidagoster, Sciacca, & Steinberg, 1998), and

in one study, 88 percent of the patients experiencing ICD firings reported "nervousness"

as a result of ICD firings (Dunbar, Warner, & Purcell, 1993). Further, patients with a

history of ICD firings appear to experience significantly higher levels of anxiety than

patients without a history of ICD firings (Dougherty, 1995; Herrman et al., 1997;

Luderitz et al., 1996). These studies also provide evidence to suggest that patients

experiencing more frequent ICD firings are at greater risk for the development of

clinically significant anxiety as compared to those patients experiencing none or

relatively few ICD firings within the first year (Sears, Todaro, Saia, Sotile, & Conti,

1999).


Depression

Symptoms of depression have also been commonly observed within the ICD

patient population, and depressive symptomatology prevalence rates between 24 and 33

percent have been reported at various follow-up intervals (Hegel et al., 1997; Konstam et

al., 1995). In early research, declines in physical ability and functioning were noted to

correspond with negative changes in mood and the development of depression in ICD

patients (Pycha et al., 1986). In more recent research, feelings of loss of control and

helplessness have also been identified, providing evidence to suggest that the learned

helplessness theory of depression may be applicable to ICD patients (Goodman & Hess,







1999; Schuster et al., 1998; Sears, Todaro, et al., 1999). Empirical studies reveal levels

of depression that are clinically diagnostic are present in approximately 10 to 15 percent

of ICD patients following ICD implantation (Herrman et al., 1997; Morris, Badger,

Chielewski, Berger, & Goldberg, 1991). To date, the direct relationship between patient

experience of ICD firings and depression has not been adequately explored, but recent

evidence suggests that multiple ICD firings (>5) are strongly associated with symptoms

of depression (Heller et al., 1998; Sears, Wallace, et al., 2000).


Adjustment Disorder

In the only study of its kind with ICD patients, Morris et al. (1991) conducted

diagnostic interviews with 20 patients approximately eight months following

implantation. Clinical assessment revealed that psychological complications were

common among the ICD patients, as 50 percent of the sample met the criteria of the

DSM-III-R (Diagnostic & Statistical Manual of Mental Disorders, 3rd ed., revised;

American Psychiatric Association, 1987) for a psychological diagnosis. The most

frequent diagnosis was Adjustment Disorder (n=6), although others included Major

Depressive Disorder (n=3), and Panic Disorder (n=1). Analyses revealed that psychiatric

morbidity was significantly related to psychological distress in family members and

perceived inadequacy of social support. In addition, there was a clear trend (p=.055) that

indicated ICD firings were also associated with patient psychological maladjustment and

morbidity. This study was one of the first to document a relationship between ICD

firings and patient psychological adjustment difficulties.







Fear of ICD Firings

Fear of ICD firings has been frequently observed in ICD patients in clinical

practice and has been noted as a common complication throughout much of the literature

examining psychological adjustment to the ICD. In one of the earliest prospective

studies, ICD patients reported experiencing fear and anxiety prior to implantation and this

patient distress remained present following hospital discharge (Pycha et al., 1986). In

another early study, retrospective interviews conducted with patients who had received

ICD firings revealed that 85% of these patients experienced anticipatory fear of ICD

firings (Cooper, Luceri, Thurer, & Myerburg, 1986). In later research, Ahmad,

Bloomstein, Roelke, Bernstein, and Parsonnett (2000) found that approximately one

fourth of their ICD sample reported a "dread" of ICD firings. Further, this is consistent

with other relatively recent studies noting that up to half of all ICD patients, regardless of

ICD firing history, report experiencing fears concerning ICD firings and possible death

(Heller et al., 1998; Schuster et al., 1998; Sneed & Finch, 1992; Vitale & Funk, 1995).

Several researchers have found that patient association of feared ICD firings with

exercise and other physical activities is a common event, and may result in the restriction

of participation in these types of behaviors with many ICD patients (Cooper et al., 1986;

Dunbar et al., 1993; Kuiper & Nyamathi, 1991; Sears, Rauch, Handberg, & Conti, 2001).

Patient fear of ICD firings is of particular concern to healthcare professionals, given that

a recent study by Hegel et al. (1997) reported a negative relationship between

psychological functioning and fear of ICD firings; in this sample, patient fear of ICD

firings was also significantly associated with the increased incidence of depression.

These findings are consistent with more recent research indicating that patients with a







history of ICD firings were more likely to report limitations in leisure activities and have

anxiety about their ICD (Duru et al., 2001).

Although few researchers have examined the course of patient fear of ICD firings,

the literature suggests that ICD firings become less troublesome and fears of ICD firings

decrease over time for the majority of ICD patients (Ahmad et al., 2000; Luderitz et al.,

1996). Even so, fear of ICD firings appears to remain a common complaint in the early

years post-implant, as approximately one third of patients in the ICD studies presented in

this review still reported experiencing fear of ICD firings at follow-up assessment

intervals ranging from as early as 6-months through 4 years post-implant. While research

indicates most patients have a high degree of acceptance of the ICD as a treatment for

managing life-threatening arrhythmias, a significant subset of patients experience distress

related to firings (Herbst, Goodman, Feldstein, & Reilly, 1999; Kohn, Petrucci, Baessler,

Soto, & Movsowitz, 2000; Namerow, Firth, Heywood, Windle, & Parides, 1999; Pauli,

Weidemann, Dengler, Benninghoff, & Kuhlkamp, 1999). Further, many of these patients

have difficulties adjusting psychologically that may persist beyond the first year post-

ICD implant without appropriate psychological intervention.


Quality of Life / Psychological Adjustment

Although the concept of quality of life was long ignored in favor of quantity of

life and survival rates in the ICD literature, following the demonstration of device

efficacy, researchers have recently begun to focus their efforts in this area. However,

there has been little consistency among researchers in how quality of life has been both

defined and measured. In their recent review of quality of life assessment within the

general cardiac literature, Swenson and Clinch (2000) noted that the measurement of







health-related quality of life addresses "the illness experience [as perceived by the

patient] as opposed to the disease" (p. 406). Within the ICD literature, according to

Ahmad et al. (2000), quality of life has been defined "in many different ways, addressing

such issues as awareness of the device, fear of device malfunction, distorted body image,

psychological problems, modified patterns of sexual activity, and concerns about

returning home after device implantation" (p. 937). May et al. (1995) were the first

researchers to attempt to systematically measure quality of life in ICD patients; they

defined quality of life as "the ability of a person to function normally in society as

perceived by that person" (p. 11). According to Namerow et al. (1999), approaches used

within the ICD literature to study quality of life have included literature reviews,

descriptive studies, prospective studies, and comparative studies of patients with and

without ICDs; however, to date the prospective and comparative studies available in the

literature remain notably scarce in this area of research. Therefore, the following section

reviews the quality of life literature in its current state, a blend of studies addressing

quality of life and psychological adjustment issues simultaneously within the ICD patient

population.

Throughout the quality of life literature (Heller et al., 1998; Luderitz et al., 1996;

Pycha et al., 1986; Schuster et al., 1998), numerous patient complaints following ICD

implantation have been identified, including (1) pain, (2) physical discomfort, (3)

constant awareness of the device, (4) sleep disturbances, (5) limited quality of life, (6)

social isolation, (7) declines in sexual activity, and (8) driving restrictions. In addition,

negative feelings have been commonly reported by ICD patients as well, and include

feelings of: (1) insecurity, (2) helplessness, (3) self-doubt, (4) emotional upset, and (5)







general psychological distress. Results of research specifically designed to address the

concerns of younger ICD patients has been similar, with more emphasis on concerns

regarding device appearance, physical activity limitations, sexual relations, social

interactions, and driving restrictions (Dubin, Batsford, Lewis, & Rosenfeld, 1996; Sears,

Burs, Handberg, & Conti, 2001; Vitale & Funk, 1995). However, despite frequent

patient complaints, these researchers also report that ICD acceptance rates are quite high

(approximately 75% of patients report positive feelings), and many patients report having

come to view the ICD as a lifesaver and a symbol of security (Dougherty, 1995; Duru et

al., 2001; Heller et al., 1998). While the findings in the ICD literature vary, most studies

suggest the majority of ICD patients will typically adjust to the device between 4 and 12

months following implantation, with most patients resuming their normal activities

within that time period (May, Smith, Murdock, & Davis, 1995; Vlay et al., 1989).

To more systematically explore the impact of the ICD on patient reported quality

of life, a major prospective examination was conducted by May et al. (1995). These

researchers found that overall quality-of-life and psychosocial functioning appears to

only temporarily decline, with return to pre-ICD implantation levels of functioning

observed as soon as 12 months following ICD implantation. The significant declines

noted at the 6-month assessment interval were in the areas of emotional behavior (e.g.,

increased irritability, negative self-talk, self-blame), alertness behavior (e.g.,

forgetfulness, difficulties concentrating), and social interaction (e.g., irritability and anger

directed at family/friends, less social interaction). These researchers have hypothesized

that it is the decline in ICD firings after the first 6 months that contributes to better

patient adjustment to the ICD at the 12-month interval. However, no analyses were







conducted in this study to address the suspected influence of ICD firings upon patient

psychological functioning.

Other researchers have sought to compare the quality of life experienced by ICD

patients to that of patients in other cardiac populations. In one study, Arteaga and Windle

(1995) compared the quality of life of ICD patients to that of other arrhythmia patients

prescribed pharmacological therapy and a cardiac patient reference group. Results

indicated psychological distress was related to younger patient age and greater cardiac

dysfunction. Further, psychological distress predicted lower quality of life in all three

patient groups. Moreover, patients in both arrhythmia groups reported more distress than

the cardiac reference group. Similarly, recent research conducted by Namerow et al.

(1999) compared coronary artery bypass (CABG) patients to CABG patients with ICDs

6-months post-surgery. Results indicated that the ICD patients had significantly lower

scores on measures of psychological well-being, perceptions of health, and emotional

role functioning. Further, these differences were more marked for those ICD patients

with a positive history of ICD firings, even after controlling for the influence of patient

rehospitalizations. These findings are also consistent with the idea that arrhythmia

patients experience greater impairment that non-arrhythmia patients in terms of quality of

life and psychological adjustment.

In contrast to the above findings, other researchers have found no significant

differences on measures of quality of life and depression between ICD patients and

various cardiac disease reference group patients (Duru et al., 2001; Hermann et al., 1997;

Keren et al., 1991). At this time, although the research has been limited, it appears the

impairment in quality of life and psychosocial functioning that is experienced by ICD







patients is most similar to that experienced by other cardiac arrhythmia patients. While

there is some evidence to suggest ICD patients with a history of firings experience greater

impairment as compared to ICD patients without a history of firings (Heller et al., 1999;

Namerow et al., 1999; Schuster et al., 1998), there is also the suggestion that cardiac

arrhythmia patients, regardless of treatment approach (e.g., ICD vs. medications) may

experience more psychological distress and impairment as compared to other cardiac

patient populations (Herbst et al., 1999), making it essential to carefully consider the

comparison groups used in the research when drawing conclusions about the impact of

the ICD.


Coping Strategies

Relatively few researchers have examined coping strategies used by ICD patients,

and their relationships to physical and psychological adjustment; however, these research

efforts have involved large numbers of ICD recipients. One of the first studies in this

area examined emotion-focused versus task-focused coping styles as they related to the

physical and psychosocial adaptation of patients who underwent ICD implantation at

least two years prior to the assessment (Craney, Mandle, Munro, & Rankin, 1997). In

this study, regression analyses revealed that younger recipients, male sex, and emotion-

focused coping strategies were significantly associated with poorer physical functioning,

and accounted for 25% of the variance. However, for psychosocial functioning only 12%

of the variance was accounted for, with emotion-focused coping strategies being

significantly associated with worse psychosocial functioning.

In a more recent and larger study (N=213), Dunbar, Jenkins, and colleagues

(1999) explored coping on patients assessed pre-ICD implant and at 1 and 3 months post-







implant. The regression analyses performed revealed that the hypothesized predictors

(physical symptoms, illness appraisal, and coping behaviors) significantly explained

additional variance in both functional status and total mood disturbance at follow-up,

above that which was accounted for by the personal-situational variables (sex, age,

optimism, history of sudden cardiac arrest, ejection fraction, comorbid medical

diagnoses, ICD firings, and pre-implant total mood disturbance). For functional status,

appraisal-coping variables accounted for 7% at 1-month follow-up and 5% at 3-months

follow-up (total variance accounted for by the full model was 33% and 36%,

respectively). For total mood disturbance (as measured by the Profile of Mood States),

appraisal-coping variables accounted for 33% of the variance at 1-month follow-up and

20% at 3-months follow-up (total variance accounted for by the full model was 55% and

49%, respectively). The authors concluded that the patterns of change observed in the

appraisal-coping variables and outcomes suggest that the early recovery period is most

important for intervention.


Psychosocial Impact of ICD Firings

Numerous researchers have alluded to the existence of a relationship between

patient distress and ICD firings. Perhaps the most notable research finding regarding

psychological distress and ICD firings was reported by Luderitz et al. (1996). In this

large study (N=95), there was a clear relationship between higher levels of fear/anxiety

and the following two subgroups: 1) patients who experienced more than 5 ICD firings

within the first year, and 2) patients of younger age (<50 years old). In similar but more

recent research, high numbers of ICD firings (210) were related to the presence of

psychological maladjustment (anxiety and/or depression) and reduced quality of life







(Hermann et al., 1997). In this study, more than 50% of the ICD patients experiencing

frequent ICD firings reported impairment in psychological functioning and quality of life.

There was also a relationship between frequent ICD firings and marital distress. In a later

study, the negative impact of ICD firings on patient functioning was again established

(Heller et al., 1998). These researchers reported that experiencing ICD firings was

strongly associated with anxiety, depression, diminished activity, and health concerns in

ICD recipients. Further, this relationship between psychological distress and ICD firings

was exacerbated by a history of multiple ICD firings (15). In a study examining the

psychological reactions of patients and families related to ICD firings, Dougherty (1995)

found that levels of anxiety, depression, anger, and stress were higher for both patients

and family members of ICD recipients who experienced firings. There is also some

evidence to suggest that patients frequently attribute their ICD firings to participation in

recent activities and subsequently restrict the behaviors related to the ICD firings

(Dougherty, 1995; Dunbar et al., 1993). In this way, the psychological distress and

limitations in quality of life often experienced by ICD patients can potentially become

even more severe, especially when coupled with self-imposed restrictions on activity and

social interactions.

The first researchers to conduct a regression analysis specifically designed to

examine the impact of ICD firings on patient psychological distress were Burgess,

Quigley, Moran, Sutton, and Goodman (1997). In this study (N=25), overall

psychological distress was related to diminished physical activity (r-.63) and number of

ICD discharges classified as inappropriate by the patient (r.53). Inappropriate ICD

discharges and diminished physical activity predicted a significant amount of variance







associated with overall psychological distress (C =.41, p<.01), after controlling for age,

psychiatric history, and number of co-morbid medical diagnoses. Appropriate discharges

were related to diminished family responsibilities (r=.48), but not related to overall

psychological distress. Authors concluded that patient beliefs about ICD discharges are

an important influence on patient wellbeing, and noted that improvements in patient

education would be helpful in reducing patient risk of psychological distress. Notably,

this relationship between patient experience of ICD firings, patient age, and the

experience of psychological distress is intriguing and still warrants further study.


Methodological Critique

In review, the psychological literature related to ICD patient adjustment is clearly

an emerging science, and appears to have only just begun to develop over the last decade,

with many questions still unanswered. Consequently, numerous methodological flaws

currently exist in the literature. Of primary importance, at the time the current study was

proposed, most of the research conducted had been limited by very small sample sizes-

less than 40 patients (Sears, Todaro, et al., 1999; see also Keren et al., 1991; May et al.,

1995; Morris et al., 1991; Pycha et al., 1986; Vitale & Funk, 1995; Vlay et al., 1989).

Only in the past couple of years have several researchers looked to the future and begun

to report data on large ICD patient samples (Ahmad et al., 2000; Dunbar, Jenkins, et al.,

1999; Dunbar, Kimble, et al., 1999; Duru et al., 2001; Heller et al., 1998; Herbst et al.,

1999; Namerow et al., 1999; Pauli et al., 1999); correspondingly, until recently, the types

of statistical analyses conducted were greatly restricted due to small sample sizes and

power limitations. The literature had also been wrought with the use of non-standardized

assessment measures of psychological functioning and inconsistent operationalization of







psychological constructs in the early years. While most researchers primarily reported

information from such measures in a descriptive fashion (Dougherty, 1995; Dunbar et al.,

1993; Heller et al., 1998; Luderitz et al., 1996), others relied upon survey methodology to

make conclusions regarding patient functioning, rendering study results more difficult to

meaningfully interpret (Konstam, Colburn, Butts, & Estes, 1996). Although number of

ICD firings does appear to be consistently related to psychological adjustment, many of

the earlier studies failed to take this data into consideration when reporting results or

were simply unable to analyze this data due to statistical limitations. In attempts to

compensate for small sample sizes and the use of many non-standardized measures, some

researchers have utilized single item measures to assess patient functioning (May et al.,

1995); however, studies such as this may provide a potentially less complete and less

reliable picture of ICD patient adjustment. Another serious problem with the literature

has been the lack of prospective research designs; to date, a total of seven researcher

groups have collected baseline data in order to adequately address changes in patient

psychological functioning and quality of life over time following implantation (Dunbar,

Jenkins, et al., 1999; Dunbar, Kimble, et al., 1999; Kohn et al., 2000; Luderitz et al.,

1996; May et al., 1995; Pycha et al., 1986; Vitale & Funk, 1995; Vlay et al., 1989). Prior

to the proposal of this research study, even fewer studies had compared the psychological

functioning of ICD patients with that of other cardiac patient populations (Arteaga &

Windle, 1995; Herrmann et al., 1997; Keren et al., 1991); while this has changed

substantially in recent years (Duru et al., 2001; Herbst et al., 1999; Namerow et al.,

1999), the total number is still quite small and the findings have been somewhat mixed.







As such, inferences specifically about the psychosocial functioning and quality of life of

ICD patients as compared to other cardiac patients are limited.

In summary of the literature to date, the following conclusions can be made: (1)

researchers have only recently begun to systematically assess the psychological profile of

cardiac patients recommended for ICD therapy, (2) relatively few prospective studies

have examined psychological functioning and quality of life in ICD recipients, (3) stable

predictors of patient psychosocial functioning and quality of life following ICD implant

have not yet been identified, and (4) the majority of the existing research contains

methodological limitations, such as very small sample sizes, patient selection biases, lack

of prospective assessment, and the use of non-standardized assessment measures.












STUDY PURPOSE AND HYPOTHESES

Not surprisingly, given the nature of potentially life-threatening arrhythmias and

the function of the ICD, the psychosocial ICD literature indicates that ICD recipients do

experience psychosocial distress and may have reduced quality of life following ICD

implantation. However, prospective studies assessing patient characteristics and

subsequent adjustment to the ICD have been limited. Additionally, many of the studies

examining patient functioning post-ICD implantation have not examined the impact of

ICD firings and/or pre-ICD implantation psychological functioning. Further, to date,

very few researchers have examined patient characteristics as predictors of psychosocial

functioning or quality of life in ICD recipients and there have been inconsistent reports in

the literature regarding patient adjustment to the ICD and the possible impact of time and

experience of ICD firings. To address some of the issues within the current literature, the

major objectives of the current study are: 1) to obtain descriptive information regarding

the pre-implant psychological functioning, experience of ICD firings, and post-implant

quality of life, and 2) to examine the association between pre-implant psychological

patient characteristics, experience of ICD firings, and post-implant quality of life ratings

in ICD patients at both short and long-term follow-up intervals (8 & 14 months post-

implant, respectively).

Regarding the proposed multiple regression analyses examining pre-implant

psychological characteristics, experience of ICD firings, and quality of life outcomes at







both short and long-term post-implant, the following working hypotheses were generated

prior to conducting the current study:

1. Dispositional optimism and social support will be positively associated with each
of the quality of life outcome variables of mental health, general health, physical
limitations and disease perception.

2. History of depression, trait anxiety, and ICD firings will be negatively associated
with each of the above referenced quality of life outcome variables.

3. The associations between predictor variables of interest (history of depression,
trait anxiety, dispositional optimism, social support, and ICD firings) and the
quality of life outcome variables will be statistically significant.

4. The percentage of variance in the outcome variables accounted for by the
predictor variables of interest (as defined in #3) will be equal to or greater than
the variance accounted for by the control variables of patient age and left
ventricular ejection fraction.












METHOD


Participants

The sample consisted of a total of 88 patients who received an ICD implant for

the first time as their primary treatment for a diagnosis of life-threatening ventricular

arrhythmia during the time period of August 1997 through February 1999. All subjects

were recruited from the following two locations: 1) the Electrophysiology Clinic at the

Division of Cardiovascular Medicine, Shands Hospital, University of Florida,

Gainesville, FL, and 2) the Electrophysiology Clinic at the Division of Cardiovascular

Medicine, Vanderbilt University Medical Center, Vanderbilt University, Nashville, TN.

Participants were voluntary and no monetary compensation was offered. Individuals

were excluded from participation in this study if they were: 1) less than 18 years of age,

2) unable to read and write in English, 3) cognitively impaired, or 4) indicated they

would not be available for follow-up assessment. Informed consent procedures as

required and approved by the Institutional Review Boards (IRB) at both the University of

Florida Health Science Center and the Vanderbilt University Medical Center were

followed.

Table 1 provides a summary of the general demographic characteristics of the

study sample. Of the 88 patients who participated in this study, the majority was male,

with a mean age of 65.3 years (range: 22 to 89). The sample was almost exclusively







Table 1

General Characteristics of Study Participants


Characteristic n Percent


Patient Age
Under 50 10 11.4%
50-59 16 18.2%
60-69 22 25.0%
70-79 34 38.6%
80 and above 6 6.8%

LV Ejection Fraction
15% or less 19 21.6%
16-25% 23 26.1%
26-35% 17 19.3%
36-45% 8 9.1%
46-55% 7 8.0%
Greater than 55% 11 12.5%
Unknown 3 3.4%

Sex
Male 73 83.0%
Female 15 17.0%

Ethnicity
Caucasian 81 92.0%
African American 3 3.4%
Hispanic 1 1.1%
Asian 1 1.1%
Unknown 2 2.3%

Marital Status
Married 67 76.1%
Widowed 7 8.0%
Single 6 6.8%
Divorced 5 5.7%
Cohabiting 1 1.1%
Unknown 2 2.3%







Caucasian and most patients were married; approximately half were Protestant and

indicated an educational background at or above the high school graduate level.

The most frequently occurring medical diagnoses in this sample were

predominantly cardiovascular in nature and are presented in Table 2. The vast majority

of patients had primary diagnoses of ventricular arrhythmia (95.4%) and coronary artery

disease (69.0%), and the reported mean left ventricular ejection fraction (LVEF) was

30.6% (range: 5 to 70). However, other less common comorbid diagnoses included

chronic obstructive pulmonary disease (14.9%), peripheral vascular disease (12.6%),

cerebral vascular events (11.5%), and renal insufficiency (10.3%). In addition,

psychiatric disorders, such as substance abuse, depression, and anxiety, were diagnosed

infrequently (eight patients: 4.6%, 3.4%, and 1.1%, respectively). Given the serious

medical conditions diagnosed in this patient sample, all prescribed medications at time of

hospital discharge were assessed as well. Those medications most frequently prescribed

are presented according to drug classification in Table 3, and consisted of common

cardiac medications such as ACE inhibitors, diuretics, aspirin, digoxin, and beta-

blockers. In addition, other drugs, such as nutritional supplements (46.4%), psychotropic

medications (16.0%), pain medications (14.3%), and breathing medications (13.1%),

were prescribed, albeit with less frequency, to smaller groups of patients in this sample.

On average, the majority of patients (n = 58, 66.0%) were prescribed between five to nine

different classes of medications, with a mean of 6.54 different medication classes

prescribed per patient (SD=2.55).







Table 2

The 10 Most Commonly Diagnosed Medical Conditions


Medical Diagnosis Percentage (n)

Ventricular Arrhythmia (VT/VF) 95.4% (83)
Coronary Artery Disease (CAD) 69.0% (60)
Hypertension 49.4% (43)
Cardiomyopathy 42.5% (37)
Atrial Fibrillation 40.2% (35)
Past Myocardial Infarction (MI) 39.1% (34)
Diabetes Mellitus 35.6% (31)
Congestive Heart Failure (CHF) 31.0% (27)
Hyperlipidemia 27.6% (24)
Past Sudden Cardiac Death (SCD) 26.4% (23)

Note. The primary medical diagnosis in this sample was ventricular arrhythmia;
however all subjects received at least two medical diagnoses (M = 5.82, SD = 2.19).
As comorbidity was high, the percentages of the other most commonly diagnosed
conditions are provided. Total n = 87 as this data was not available for 1 subject.







Table 3

Summary of the Most Frequently Provided Prescription Medications at Discharge


Class of Drug Prescribed (example) Percentage (n)

ACE inhibitors (vasotec, lisinopril) 70.2% (59)
Diuretics (lasix) 66.9% (57)
Aspirin (enteric coated ASA) 66.7% (56)
Digitalis glycoside (digoxin) 52.4% (44)
Beta-blockers (lopressor) 44.0% (37)
Vasodilators (nitroglycerin, ismo) 34.5% (29)
Antidiabetics (insulin, glucophage) 34.5% (29)
Antilipidemics (lipitor) 33.3% (28)
Gastrointestinals (prilosec, axid) 33.3% (28)
Antiarrhythmics (amiodarone, sotalol) 32.1% (27)
Blood modifiers (coumadin, ticlid) 29.8% (25)

Note. Total n = 84 as this data was not available for 4 subjects. All patients in the
sample were prescribed at least one medication; the majority was receiving
prescriptions from a combination of five or more different drug classes.







Procedure


Initial Assessment

All patients were recruited to participate in the study during their inpatient

hospitalization for ICD implantation. Following informed consent procedures,

arrangements were made for patients to complete their participation prior to hospital

discharge. The initial assessment phase consisted of a brief history of depression

interview and the administration of several self-report, paper-and-pencil questionnaires

designed to provide information regarding patient demographics, trait anxiety,

dispositional optimism, and perceived social support. This protocol was part of a larger

assessment battery that typically required approximately 15-20 minutes for patients to

complete.


Follow-up assessment: Short and long-term intervals

Short and long-term follow-up assessments for all participants were conducted by

telephone and targeted two time periods: 1) 6-9 months post-ICD implantation and 2) 12-

15 months post-ICD implantation. At these intervals, the quality of life measures of

mental health, general health, physical limitations, and disease perception were

administered. In addition, several questions were asked regarding patient ICD

knowledge, concerns, driving practices, and experience of firings (e.g., number, episodes,

storms). All patients were contacted prior to the follow-up assessment in order to

schedule their telephone interview appointment in advance. Upon scheduling an

appointment time, patients were mailed a response sheet to use during the telephone

interview. This response sheet served as an answer key for the questionnaire instruments







administered and allowed the battery to be completed by the patient in a more time

efficient manner. The telephone interviews were approximately 30-45 minutes in

duration.

The mean follow-up time for the short-term assessment interval was 8.2 months

post ICD-implantation (SD = 1.8), with a total of 42 patients (47.7%) from those initially

recruited participating in the telephone assessment. Of the original sample, 19 (21.6%)

could not be contacted, 10 (11.4%) were confirmed deceased, 7 (8.0%) were excluded

due to severe health impairment (e.g., stroke), 6 (6.8%) were excluded due to insufficient

initial assessment data, 1 (1.1%) was excluded due to ICD extraction, and 3 (3.4%)

refused to participate. After the exclusion of ineligible subjects, the remaining eligible

sample size consisted of 64 subjects, 42 of who (65.6%) participated at the short-term

follow-up interval.

At the long-term follow-up assessment interval, the mean follow-up time was

14.3 months post ICD-implantation (SD = 1.8), with 49 of the initially recruited patients

(55.7%) participating in the telephone assessment. Of the original sample, 12 (13.6%)

could not be contacted, 15 (17.0%) were confirmed deceased, 3 (3.4%) were excluded

due to severe health impairment (e.g., stroke), 6 (6.8%) were excluded due to insufficient

initial assessment data, 1 (1.1%) was excluded due to ICD extraction, and 2 (2.3%)

refused to participate. After the exclusion of ineligible subjects, the remaining eligible

sample size consisted of 63 subjects, 49 of who (77.8%) participated at the long-term

follow-up interval.

In all, a total of 60 (68.2%) of the original 88 patients recruited participated in at

least one follow-up interval, with 31 (35.2%) participating in both the short and long-







term follow-up intervals. In contrast, 28 (31.8%) of the original sample could not be

assessed at either follow-up interval, due to factors such as death, severe health

impairment, relocation, or refusal. Analyses were conducted to determine if differences

existed between the follow-up participants and non-participants. Variables specifically

examined for between groups differences were gender, ethnicity, marital status, education

level, patient age, ejection fraction, history of depression, trait anxiety, dispositional

optimism, and perceived social support. However, the comparisons between the follow-

up participants and the subjects unable to be assessed did not reveal any significant

differences between groups on any of the demographic, medical, or psychological

variables measured at the time of ICD implant. While the unable to be assessed group is

sizeable, the exceptionally low refusal rate and the lack of initial assessment differences

between the participating and non-participating subjects in follow-up suggest that the

chance of a systematic self-selection bias is relatively low in this sample.


Measures

Patient assessment instruments were administered at each of three intervals: initial

assessment, short-term follow-up, and long-term follow-up. At the initial assessment, the

following instruments were administered: (1) demographic questionnaire (see Appendix

B), (2) Interpersonal Support Evaluation List, (3) Life Orientation Test, (4) Schedule for

Affective Disorders and Schizophrenia: History of Depression Scale, and (5) State-Trait

Anxiety Inventory: Trait Scale. At each of the follow-up intervals, the following

instruments were administered: (1) Seattle Angina Questionnaire: Physical Limitations

and Disease Perception Scales, (2) Short-Form 36 Health Survey: General Health

Perceptions and Mental Health Scales, and (3) Survey of ICD Information, Firings,







Concerns, and Driving (see Appendix B). Presented below, in alphabetical order, is a

detailed description of each of the assessment measures administered.


Interpersonal Support Evaluation List Short-form (ISEL)

The short-form of the ISEL is a 16-item, self-report questionnaire that was

developed to measure perceived availability of supportive social resources that might

facilitate coping with stressful situations (Cohen, Mermelstein, Kamarck, & Hoberman,

1985). The items are rated on the likelihood that the type of support described would be

available if needed (i.e., probably true, probably false). Desirability of the items is

reverse balanced, in that about half of the items are phrased in the positive direction (7 vs.

9). The ISEL has been shown to have strong internal consistency and to moderately

correlate with other measures of social support, such as perceived availability of social

support (Inventory of Socially Supportive Behaviors; r=.46), involvement and emotional

support (Moos University Residence Environment Scale; r=.62), number of close friends

(r=.46), and network size (r-.39) (Cohen et al., 1985). The four scales of this measure

assess the following domains: (1) perceived availability of someone to talk to about one's

problems (Appraisal), (2) perceived availability of people with whom to do things

(Belonging), (3) perceived availability of positive social comparison (Self-esteem), and

(4) perceived availability of instrumental assistance (Tangible). Although each of these

scales can be interpreted separately, the literature indicates the scores on the ISEL scales

can also be summed to derive a single, valid indicator of perceived availability of social

support (Bennett, 1993; Brookings & Bolton, 1988; Cohen et al., 1985). For the purposes

of this study, this approach was taken and only the ISEL total score was examined in

analyses. The Cronbach's alpha for the ISEL total score was .79 in the current sample.







Life Orientation Test (LOT)

The LOT is an 8-item, self-report questionnaire (with 4 additional filler items)

that assesses generalized expectancies for positive versus negative outcomes. Subjects

rate the extent to which they agree or disagree with each item using a 5-point scale. Half

of the items are phrased in the positive direction (e.g., "In uncertain times, I usually

expect the best") (Scheier & Carver, 1985). The scores for the negative items are

reversed, and then all items are summed to yield an overall dispositional optimism score.

The LOT has a reported reliability alpha of 0.76 and test-retest reliability of 0.79 (Scheier

& Carver, 1985). Although researchers have suggested the measurement of dispositional

optimism (using the LOT) overlaps with neuroticism or negative affectivity (see Smith,

Pope, Rhodewalt, & Poulton, 1989), several recent analyses conducted with the data from

4,309 subjects have demonstrated that associations between optimism and outcome

variables remain significant even when the effects of neuroticism, trait anxiety, self-

mastery, and self-esteem are statistically controlled (Scheier, Carver, & Bridges, 1994).

The authors conclude that overall, the LOT has good predictive validity, and dispositional

optimism (as measured by the LOT) is quite distinguishable as an independent construct,

as compared to the constructs of neuroticism and negative affectivity. The Cronbach's

alpha for the LOT was .74 in this sample.


Schedule for Affective Disorders and Schizophrenia (SADS)

The SADS is a structured interview developed to assess the presence of

psychological disorders according to research diagnostic criteria (Spitzer, Endicott, &

Robins, 1978). For the purposes of this study, only the history of depression module was

administered. This module is a brief diagnostic interview that allows the examiner to







assess for the history of a depressive episode that lasted at least one week or more, and

provides information regarding depressed mood, anhedonia, and associated symptoms of

depression. On this measure, a criterion of at least four depressive symptoms (in addition

to anhedonia or depressed mood) must be endorsed for a diagnosis of past major

depressive episode to be given. The depressive symptoms on this measure are consistent

with those currently listed in the DSM-IV (4h ed.; American Psychiatric Association,

1994).


Seattle Angina Questionnaire (SAQ)

The SAQ is a 19-item, self-report questionnaire that was designed to measure

cardiac functioning and health-related quality of life in patients diagnosed with coronary

artery disease (Spertus et al., 1995). The SAQ includes five clinically relevant scales

including: physical limitations, disease perception, treatment satisfaction, anginal

frequency, and anginal stability. For the purposes of this study, only the Physical

Limitations and Disease Perception scales were examined. The Physical Limitations

scale is a measure of how daily activities are limited specifically by the patient's cardiac

disease. The Disease Perception scale is a measure of the perceived degree of burden that

cardiac disease has upon the patient's quality of life. In a sample of 117 cardiac patients

assessed at 3-month intervals, the test-retest reliabilities of the Physical Limitations and

Disease Perception scales were 0.83 and 0.78, respectively (Spertus et al., 1995).

Further, this study established that the Physical Limitations scale of the SAQ possessed

adequate concurrent validity as a measure of quality of life after cardiac illness, and was

comparable to that of the Duke Activity Status Index (r=0.43, p < 0.001) and the Specific

Activity Scale (r=0.84, p < 0.001). The Disease Perception scale was also reported to







possess adequate concurrent validity with the General Health Perceptions scale of the SF-

36 Health Survey (r=0.60, p < 0.0001). In addition, the SAQ manual provides a

prescribed strategy for handling missing data, which allows for mean substitution of

missing items, given that the majority of items of the scale are completed. This method

was utilized to reduce the incidence of missing scale data in the analyses. For the SAQ

scales examined in this sample at the short-term follow-up interval, the Cronbach alphas

were .93 (Physical Limitations) and .67 (Disease Perception). At the long-term follow-up

interval, the Cronbach alphas were .95 (Physical Limitations) and .65 (Disease

Perception).


Short-Form 36 Health Survey (SF-36)

The SF-36 is a 36-item, self-report questionnaire that measures health-related

quality of life (Ware, Snow, Kosinski, & Gandek, 1993). The SF-36 was both rationally

and empirically developed as part of the Medical Outcomes Study and consists of eight

health domain scales including: physical functioning, role limitations due to physical

problems, role limitations due to emotional problems, social functioning, bodily pain,

mental health, vitality, and general health perceptions. For the purposes of this study,

only the general health perceptions and mental health scales were examined. The

General Health Perceptions scale (5 items) measures personal evaluation of health, health

outlook, perceived resiliency to illness, and perceived changes in health from one year

ago. The Mental Health scale (5 items) measures general mental health functioning

including depression, anxiety, behavioral-emotional control, and positive affect. Despite

having only a few items on each scale, the internal reliabilities of the scales were found to

range from 0.77 to 0.92 in a sample of 3,053 adults (Stewart et al., 1992). Additionally,







the SF-36 manual provides a prescribed strategy for handling missing data, which allows

for mean substitution of missing items, given that the majority of items of the scale are

completed (Ware et al., 1993). This method was utilized to reduce the incidence of

missing scale data in the analyses.


State-Trait Anxiety Inventory (STAI)

The STAI is a 40-item, self-report questionnaire designed to measure both state

and trait anxiety (Spielberger, 1977; Spielberger, Gorsuch, Lushene, Vagg, & Jacobs,

1983). For the purposes of this study, only the 20-item trait scale of this questionnaire

was administered. Trait anxiety is defined as a relatively enduring personality

characteristic, or more specifically, as anxiety proneness. The internal reliability of both

the state and trait anxiety scales has been shown to be uniformly high across samples of

adults ranging from 0.89 to 0.96. Test-retest stability coefficients for multiple samples of

college students ranged from 0.73 to 0.86, with test-retest validity specifically for the trait

scale being reported at 0.73 for males and 0.77 for females. Concurrent validity between

the STAI and IPAT Anxiety Scale and the Taylor Manifest Anxiety Scale ranged from

0.83 to 0.73 (Spielberger et al., 1983). The Cronbach's alpha derived for the trait anxiety

scale in the current sample was .89.



Survey of ICD Information, Firings, Concerns, and Driving (SIFCAD)

As standardized and validated measures designed to assess the level of ICD

information provided to patients, experience of ICD firings, current concerns, and driving

practices were not available in the literature, a brief survey to address these areas of

interest was specifically developed for use in this study. The SIFCAD is a 4-item




41


questionnaire designed for administration in an interview format. Patient responses to the

SIFCAD questions provide data regarding the information the patient has received about

their ICD, experience of ICD firings (frequency and number of episodes), current

concerns about the ICD, and typical driving practices (days/week, distances, ICD-related

adverse events).













RESULTS


Initial Psychological Assessment Measures

A summary of the descriptive results obtained for the measures of history of

depression, trait anxiety, dispositional optimism, and perceived social support at the time

of the initial assessment is presented below.

On the measure of history of depression (SADS: History of Depression module),

29.5% of this sample (n = 26) reported having experienced a period of feeling

significantly depressed at least one week or more during their lifetime. Further, 22.7% of

the patients (n = 20) met the established criteria for diagnosis of a past episode of major

depression, and reported experiencing a mean of 5.52 (SD = 1.97) depressive symptoms

in addition to depressed mood during their most severe episode. The mean number of

reported past major depressive episodes was 3.64 (SD = 3.79) for this sample; however,

54.5% of those patients with a positive history of depression reported a total of only one

or two major depressive episodes in the past.

Regarding trait anxiety, in this sample a total of 14 patients (18.7%) reported trait

anxiety levels more than one standard deviation higher than the mean established for a

large heart transplant sample (N = 207, M = 34.90, SD = 9.14; Sears, Rodrigue, Sirois,

Urizar, & Perri, 1999). The overall mean for the ICD study sample was 36.29 (SD =

9.25), which was not statistically different than the heart transplant sample mean, but was

significantly lower (p < .001) than the mean score published for a sample of general







medical patients by the authors of this measure (N = 161, M = 41.91, SD = 12.70;

Spielberger et al., 1983).

The mean score derived from the LOT for this sample was 21.49 (S = 5.13),

indicating a normal level of dispositional optimism in this ICD patient sample. This

mean did not differ significantly from the published norms available for this measure a

large college undergraduate sample (N = 624, M = 21.8, SD = 4.8; Scheier et al., 1985).

In terms of perceived social support, a mean of 13.72 (SD = 2.66) was derived for

the composite score of the ISEL, and indicated relatively high degrees of social support

across the four dimensions assessed. As short-form means have not been published in the

literature, comparative means were calculated using the means and standard deviations

for the full general population scale from a community sample (N = 64, M = 13.46, SD =

2.19; Cohen et al., 1985). A t-test analysis revealed no significant differences between

the two samples.


Follow-up Quality of Life Assessment Measures

At both the short and long-term follow-up assessment intervals, patients

completed the Mental Health and General Health subscales of the SF-36 and the Physical

Limitations and Disease Perceptions subscales of the SAQ. A brief summary is presented

in Table 4.

On the SF-36, short and long-term mean scores for the Mental Health subscale

were calculated at 76.48 (SD = 21.11) and 75.92 (SD = 20.84), respectively. For the

General Health subscale, the short and long-term mean scores were calculated at 55.41

(SD = 25.51) and 48.14 (SD = 20.84), respectively. Paired samples t-test analyses for




































0Nc^
- oo









0
men







mO10




00'


r
0 0
o








IU


U


Ch


tis
.- a








o
o o






0 C
aa




5 o
so



c






.-


a "












' 0
t o


U.t













--0


yr|
o-


^ ^
Cs 00
00 0












000
odod
- -













-CN








'r10
tr *t







(1\(1
*-< v-



V- h
Vri
v (-s
^ o
*^ ^

>r> ^

in (-


C
S
$C,


E"~i-;a
~C
~dq~Q)
P~ut:
3:
\o
m

v,







both the Mental Health and General Health subscales indicated that the mean scores did

not differ significantly from each other over time. To provide a reference for the

magnitude of these scores, additional paired comparisons were made. Since these were

exploratory analyses, multiple comparison corrections were not performed. Results

indicated there were no significant differences between the Mental Health subscale means

in this sample and the published norms for this subscale in three somewhat similar

cardiac patient samples (hypertension, congestive heart failure, acute MI; Ware et al.,

1993). In contrast, for each of the comparative cardiac patient groups, there were

significant differences between the established norms and the General Health subscale

mean obtained in this sample. More specifically, the General Health subscale mean at the

short-term follow-up interval was significantly higher (p < .05) than the mean reported in

the congestive heart failure sample (N = 216, M = 47.05), although there was no

significant difference found between the two groups at the long-term follow-up interval.

In contrast, the General Health subscale mean at the long-term follow-up interval was

significantly lower than the means obtained in the hypertension (p < .001, N = 2089, M =

63.30) and recent MI samples (p < .01, N = 107, M = 59.17), but differences were not

significant at the short-term interval.

As measured by the SAQ, the short and long-term mean scores obtained for the

Physical Limitations subscale were 71.49 (SD = 26.87) and 65.34 (SD = 29.98),

respectively. For the Disease Perception subscale, the short and long-term mean scores

were calculated at 68.25 (SD = 24.36) and 63.78 (SD = 23.79), respectively. Paired

samples t-test analyses indicated that the mean scores for each of these subscales did not

differ significantly from each other over time. To provide a reference for the magnitude







of these scores, additional paired comparisons were made. Interestingly, the study

patients evidenced significantly higher mean scores (and thus indicated higher levels of

quality of life) at each follow-up interval as compared to the coronary artery disease

patients sampled in two different published studies in the literature (Dougherty,

Dewhurst, Nichol, & Spertus, 1998; Spertus et al., 1995). These results indicate that this

sample of cardiac arrhythmia patients is distinctly different and higher functioning than a

typical group of coronary artery disease patients. Table 5 provides more specific details

about the mean scores and significant differences between these patient groups.


Follow-up of ICD Specific Post-Implantation Variables

At each of the follow-up intervals, patients also responded to questions about their

perception of the ICD information received, firings experienced, current concerns, and

driving practices. The information obtained is described below.


ICD Information Received

At the time of ICD implantation, a large number of patients responded "mostly

yes" (58.5%) or "definitely yes" (22.6%) to the question of whether or not they received

enough information regarding their ICD, while relatively few responded "mostly no"

(7.5%) or "definitely no" (3.8%). Comparatively, the information patients received

regarding their ICD increased significantly by the time of the short-term follow-up

interval, such that almost all the patients in the sample responded, "mostly yes" (39.0%)

or "definitely yes" (53.7%) to the same question. There was little change in this pattern

at the time of the long-term follow-up interval (32.7% and 59.2%, respectively).








Table 5

Mean Differences in the Seattle Angina Questionnaire Subscales Across Samples

Coronary Artery Disease Patients
Sample 1 Sample 2
Arrhythmia Patients Mean P Mean P
Seattle Angina Questionnaire Subscale Difference Value Difference Value

Short-term Follow-up Interval
Physical Limitations 21.29 .000 18.39 .000
Disease Perception 11.55 .004 17.85 .000

Long-term Follow-up Interval
Physical Limitations 15.14 .002 12.24 .011
Disease Perception 7.08 .043 13.38 .000

Note. For Sample 1, N = 117 (Dougherty et al., 1998). For Sample 2, N = 107 (Spertus
et al., 1995). All patients in the comparative samples had a primary medical diagnosis of
coronary artery disease.







ICD Firing Experience

For a general overview of the information obtained regarding experience of ICD

firings at each follow-up interval, Table 6 is provided.

At the short-term follow-up interval, approximately half of the sample providing

information indicated they had not received any ICD firings (54.2%, n = 32), while the

other half indicated they had received at least one to date (45.8%, n = 27). Further, for

the latter subgroup of patients, the mean number of ICD firings was 4.6 per patient

(range: 1-19); however, the vast majority of patients received far fewer, with 40.7%

reporting 1-2 ICD firings, 33.3% reporting 3-4 ICD firings, 14.8% reporting 5-10 ICD

firings, and 11.1% reporting more than 10 ICD firings. Regarding the number of ICD

firing episodes experienced, 11 patients (40.7%) reported only one episode; while 10

patients (37.0%) reported 2-3 episodes and 6 patients (22.2%) reported 4-6 episodes. Of

the total sample, 16.9% reported experiencing at least one ICD storm (23 firings in a

single episode), whereas within the subgroup of patients reporting a positive history of

ICD firing experience, 37.0% noted experiencing at least one ICD storm (n = 10).

At the long-term follow-up interval, results were similar, with again

approximately half of the sample providing information indicated they had not received

any ICD firings (44.9%, n = 22), while the other half indicated they had received at least

one to date (55.1%, n = 27). For the latter subgroup of patients, the mean number of ICD

firings was 5.6 per patient (range: 1-28); however, the vast majority of patients received

far fewer, with 37.0% reporting 1-2 ICD firings, 29.6% reporting 3-4 ICD firings, 22.2%







Table 6

Summary of ICD Firing Experiences at the Short and Long-term Follow-up Intervals


Follow-up Interval


ICD Firing History


Short-term


Long-term


Positive for ICD Firing 45.8% (n = 27 of 59) 55.1% (n= 27 of 49)
Positive for ICD Stormn 16.9% (n = 10 of 59) 18.4% (n = 9 of 49)

For the Subgroup of Patients with a Positive ICD Firing History

Mean Number (Range) 4.6 (1-19) 5.6 (1-28)

ICD Storm Experience 37.0% 33.3%

Number of Firings
1-2 40.7% 37.0%
3-4 33.3% 29.6%
5-10 14.8% 22.2%
>10 11.1% 11.1%

Number of Episodes
1 40.7% 33.3%
2-3 37.0% 37.0%
4-6 22.2% 25.9%
>6 0.0% 3.7%

Note. IaCD Storm is defined as an experience of 3 or more firings within a 24-hour time
period. Multiple ICD firings (2 or more) in a single episode have been associated with
significantly decreased survival over time (Pacifico et al., 1999).







reporting 5-10 ICD firings, and 11.1% reporting more than 10 ICD firings. Regarding the

number of CD firing episodes experienced, nine patients (33.3%) reported only one;

while 10 patients (37.0%) reported 2-3, 7 patients (25.9%) reported 4-6, and 1 patient

reported 10 separate episodes. Of the total sample, 18.4% reported experiencing at least

one ICD storm (>3 firings in a single episode), whereas within the subgroup of patients

reporting a positive history of ICD firing experience, 33.3% noted at least one ICD storm

= 9).


Current Concerns

About half of the patients in this sample did not express any concerns related to

the ICD and its functions at either the short or long-term follow-up interval (48.8% and

49.0%, respectively). Further, the total number of concerns was small at both follow-up

intervals (range 1-4), with relatively few patients volunteering concerns and the majority

noting a concern (43.9% at short-term and 40.8% at long-term) stating only 1-2 specific

areas of concern (e.g., travel limitations, size, things to avoid to prevent triggering a

firing, fear of firings).


Driving Practices

At follow-up, the vast majority of patients (80.5% at short-term and 85.7% at

long-term) indicated they were driving fairly regularly, with approximately 1/3 of the

patients driving 1-3 days per week, 1/3 driving 4-6 days per week, and 1/3 driving 7 days

per week at each of the follow-up intervals. When they chose to drive, the vast majority

of patients in this sample (84.8% at short-term and 82.9% at long-term) indicated they







drove 30 miles or less during a typical round-trip. There were no significant changes in

driving patterns over time. Further, no patient experienced an ICD firing while in the act

of driving, although a few patients did report either before or after driving ICD firing

experiences.


Relationships between Predictor and Outcome Variables

For the short and long-term follow-up intervals, Table 7 presents the zero-order

partial correlations between the predictor and outcome variables, while controlling for

age and LVEF. As expected, there were several significant correlations observed

between the psychological measures, ICD firing experiences, and the quality of life

measures. More specifically, at short-term follow-up, dispositional optimism was

significantly and moderately correlated with the quality of life outcome measures of

mental health, general health, and physical limitations (r = .58, .57, and .51, respectively;

p < .01), such that having higher levels of dispositional optimism significantly related to

higher quality of life ratings. Further, at this interval, history of depression was

significantly and moderately correlated with lower levels of general health (r = -.59, p <

.01) and, similarly, experience of ICD firings significantly related to increased physical

limitations (r = -.43, p < .05).

At the long-term follow-up, trait anxiety was significantly and moderately

correlated with the quality of life outcome measures of mental health, general health, and

physical limitations (r = -.59, -.44, and -.45, respectively; p < .01), such that having

higher levels of trait anxiety significantly related to lower quality of life ratings. In

addition, at this interval, history of depression was significantly and moderately




































0o \ a ,
enI I o


,-.
I I


*
*







CN N
I I


**
*




*






*


i*
as


N 00 o -


rl~
I I




I I


i I





I I


*




* *
* *
* *
I I
an s
+ *


0 -ICl Cl


*


0\e0





Ir,.





Cl w


"o M





Oe co


*~


II





0




0
oi









I,
8



















0-
aC















II















=-



t o
* V





co II*
go












sz



II
C II





H a *












dev3 .


)O







correlated with a lower rating of mental health quality of life (r = -.51, p < .001) and,

similarly, higher levels of dispositional optimism significantly related to higher ratings of

both mental health and general health quality of life (r = -.56 and .43, respectively; 1 <

.01).


Regression Models for Predicting Quality of Life Variables

In order to determine the relative importance of the selected baseline medical and

psychological variables and ICD firings in predicting the short and long-term quality of

life outcome measures, separate blocked hierarchical regression analyses were planned.

Prior to performing the analyses, the distributional characteristics of the predictor

variables were examined via residual, partial regression, and normal probability plots for

possible violations of the underlying assumptions for multiple regression analyses (i.e.,

the assumptions of linearity, constant variance, independence, and normality). All

variables were within established guidelines for inclusion in the regression analyses, with

the exception of sex, which demonstrated an exceptionally skewed distribution.

Although sex was originally planned to serve as a control variable, the low number of

females in this sample resulted in insufficient power to detect group differences.

Therefore, this variable was removed from consideration in the regression model.

Otherwise, all of the multiple regression analyses were conducted as originally planned.

In this study, four different dimensions of quality of life (mental health, general

health, physical limitations, and disease perception) were predicted through hierarchical

multiple regression analyses for both the short-term and long-term follow-up intervals. In

all, eight separate hierarchical multiple regression analyses were conducted to examine







the degree of association between selected predictor (independent) variables and the

outcome (dependent) variable at each of the follow-up intervals. Figure 1 illustrates the

model utilized in the analyses. The dependent variables were identified as: (1) mental

health, (2) general health, (3) physical limitations, and (4) disease perception.

Independent variables were selected from both the medical and psychological domains,

and included (1) history of depression, (2) trait anxiety, (3) trait optimism, (4) social

support, and (5) ICD firings. All analyses controlled for the biological variables of

patient age and ejection fraction by entering these variables into the model first (Step I).

Following Step I, the psychological variables were entered as a block into the analysis

(Step II). Then, history of ICD firings (positive or negative) was then entered into the

analysis as Step III. This blocked hierarchical technique was used to allow for the

determination of the relative importance of each type of the independent variables in the

prediction of each dependent variable and to provide an overall measure of the variance

predicted by the full model.


Predicting Short-term Outcome Variables

In each of the regressions performed at the short-term follow-up interval, the

relationships between the predictor variables and quality of life outcome were in the

anticipated directions. Table 8 provides more detailed information regarding the beta

coefficients derived for the predictor variables in Step 3 of each of these analyses.







Independent Variables


STEP I


STEP II


Psychological Variables

History of Depression
Trait Anxiety
Dispositional Optimism
Perceived Social Support


STEP III


Dependent Variables


Regression 1

Mental Health

Regression 2

General Health

Regression 3

Physical Limitations

Regression 4

Disease Perception


Figure 1. Model Utilized in the Hierarchical Multiple Regression Analyses


Control Variables

Age
Ejection Fraction


Medical Variable

ICD Firing History







Table 8

Regression Coefficients for the Predictor Variables at the Short-term Follow-up Interval


Unstandardized Standardized
Coefficients Coefficients
Predictor Variables B Std. Error Beta t Sig.
Mental Health
Patient Age** .714 .235 .394 3.032 .006
Ejection Fraction .372 .190 .255 1.960 .061
History of Depression** -24.539 8.157 -.393 -3.008 .006
Trait Anxiety .371 .453 .116 .819 .420
Dispositional Optimism** 2.495 .755 .481 3.304 .003
Perceived Social Support .516 1.783 .038 .290 .774
History of ICD Firings -9.326 5.986 -.207 -1.558 .132
General Health
Patient Age .511 .365 .233 1.400 .174
Ejection Fraction .586 .294 .331 1.991 .058
History of Depression -9.589 12.642 -.127 -.759 .455
Trait Anxiety .540 .702 .140 .769 .449
Dispositional Optimism* 3.167 1.170 .505 2.705 .012
Perceived Social Support 1.111 2.763 .068 .402 .691
History of ICD Firings -5.072 9.277 -.093 -.547 .589
Physical Limitations
Patient Age .516 .324 .274 1.592 .128
Ejection Fraction** .872 .294 .514 2.963 .008
History of Depression -12.675 16.538 -.131 -.766 .453
Trait Anxiety .780 .663 .220 1.177 .254
Dispositional Optimism 2.354 1.296 .365 1.816 .085
Perceived Social Support 3.016 2.580 .191 1.169 .257
History of ICD Firings -15.518 8.830 -.307 -1.757 .095
Disease Perception
Patient Age .677 .341 .341 1.985 .058
Ejection Fraction .533 .275 .333 1.938 .064
History of Depression -24.275 11.816 -.354 -2.054 .051
Trait Anxiety .123 .656 .035 .188 .853
Dispositional Optimism 1.412 1.094 .249 1.290 .209
Perceived Social Support -.921 2.582 -.062 -.357 .724
History of ICD Firings -9.620 8.671 -.195 -1.110 .278
p <.05. ** < .01.







Mental Health Quality of Life

In this regression (see Table 9), the control variables of age and LVEF

significantly accounted for 21.1% of the variance of scores on the SF-36: Mental Health

Scale. As expected, however, the psychological variables of history of depression, trait

anxiety, trait optimism, and social support were most important in predicting mental

health quality of life, and significantly accounted for 39.9% of the variance, almost twice

that of the control variables. When ICD firing history was added to the model, it was a

relatively small but significant contributor, and accounted for 3.5% of the variance. In

all, the total variance in short-term mental health quality of life that was accounted for by

this significant model was 64.5% (p < .001).


General Health Quality of Life

In the second regression (see Table 10), the control variables of age and LVEF

significantly accounted for 13.7% of the variance of scores on the SF-36: General Health

Scale. As anticipated, the psychological variables of history of depression, trait anxiety,

trait optimism, and social support were highly important in predicting general health

quality of life, and significantly accounted for 27.4% of the variance, exactly twice that of

the control variables. When ICD firing history was added to the model, it was a relatively

small contributor, although still significant, and accounted for 0.7% of the variance. In

all, the total variance in short-term general health quality of life that was accounted for by

this significant model was 41.7% (p = .039).

































a,

ab~

VIU
Lr,


a,
M


U
L4


dU


0 I

WU'


o
0


0



u













00
EE
-


00










0t 0
O O



w a













































000
; '





1
'5 S a
dj a ^
*- oy o o





> f [A


(U >
- --3 i-
M Ws


I


hhh
--c~lm
vvv































4)

u









GO


U
M 6





0 3

(U
I~ I

od
a S


O
-o

r:


0


to




u








o o









o o










oo
00








U V



r- 0 0



> n
E E
I-
















eg


hhh
~Nrc,
vvv







Physical Quality of Life

In the third regression (see Table 11), as expected, the control variables of age and

LVEF were relevant, significantly accounting for 23.4% of the variance of scores on the

SAQ: Physical Limitations Subscale. However, the psychological variables of history of

depression, trait anxiety, trait optimism, and social support were relatively equal in

importance in predicting the physical quality of life ratings, and significantly accounted

for 24.1% of the variance, slightly more than that of the control variables. When ICD

firing history was added to the model, it represented a slightly smaller, but still significant

contribution, accounting for 7.3% of the variance. In all, the total variance in short-term

physical quality of life that was accounted for by this significant model was 54.8% (p =

.018).


Disease Perception Quality of Life

In the fourth regression (see Table 12), although the full model was not significant

in predicting disease perception scores in this sample (p = .071), it did approach

significance, with ICD firing history accounting for 3.1% of the total variance. When the

third block (ICD firing history) was removed from the model, the control and

psychological variables continued to demonstrate a strong trend in predicting the outcome

(p = .063), accounting for a total of 34.9% of the variance (16.5% and 18.4%,

respectively).






























a E






u
5






0

* I
0
O m


IL
-0 (U
5Z'


o
C


- cJ c


0





as
o











o o
000






>E>
3 E

00

a 4 C







CEA
Co0






00
W, '




. N *-i
0-0














o a-.
a, ,
*S -
h !!
u y
' c <* *-


















C/3


-a
U





33
oU
L
h)




U








d0^


0 E

d
U
*<3


-o
0


0 r-~














00
oo










0N eo
- 0


















00N 0












en

Cil C14
-!
- Cl







ooc m







o o







0 0


- C r
'L '- r-


0
4-






Q


t: t
o o








00






W 3









o
w i
o >
OT













00



S8






02iii
L00










*- Cr 93


UQQir
i- i l -


I I







Predicting Long-term Outcome Variables

In each of the regressions performed at the long-term follow-up interval, the

relationships between the predictor variables and quality of life outcome were in the

anticipated directions. Table 13 provides more detailed information regarding the beta

coefficients derived for the predictor variables in Step 3 of each of these analyses.


Mental Health Quality of Life

In this regression (see Table 14), the control variables of age and LVEF

significantly accounted for 27.1% of the variance of scores on the SF-36: Mental Health

Scale. As anticipated, however, the psychological variables of history of depression, trait

anxiety, trait optimism, and social support were of relative equal importance in predicting

mental health quality of life, and significantly accounted for 25.4% of the variance,

similar to that of the control variables. When ICD firing history was added to the model,

it was a relatively small contributor, although still significant, and accounted for 1.3% of

the variance. In all, the total variance in long-term mental health quality of life that was

accounted for by this significant model was 53.8% (p < .001).


General Health Quality of Life

In the second regression (see Table 15), the control variables of age and LVEF

significantly accounted for 27.3% of the variance of scores on the SF-36: General Health

Scale. As anticipated, the psychological variables of history of depression, trait anxiety,

trait optimism, and social support were still important in predicting general health quality

of life, and significantly accounted for 17.6% of the variance, approximately two-thirds

that accounted for by the control variables. When ICD firing history was added to the







Table 13

Regression Coefficients for the Predictor Variables at the Long-term Follow-up Interval


Unstandardized Standardized
Coefficients Coefficients
Predictor Variables B Std. Error Beta t Sig.
Mental Health
Patient Age .416 .242 .264 1.720 .095
Ejection Fraction .374 .194 .263 1.928 .063
History of Depression -10.543 7.371 -.196 -1.430 .162
Trait Anxiety -.747 .465 -.301 -1.606 .118
Dispositional Optimism 1.130 .776 .260 1.456 .155
Perceived Social Support -.915 1.212 -.122 -.755 .456
History ofICD Firings -5.447 5.751 -.133 -.947 .351
General Health
Patient Age .600 .365 .272 1.644 .110
Ejection Fraction* .660 .292 .332 2.257 .031
History of Depression -4.884 11.110 -.065 -.440 .663
Trait Anxiety -.830 .701 -.239 -1.184 .245
Dispositional Optimism 1.939 1.170 .318 1.657 .107
Perceived Social Support -2.178 1.827 -.207 -1.192 .242
History ofICD Firings -8.514 8.668 -.148 -.982 .333
Physical Limitations
Patient Age .435 .449 .187 .969 .341
Ejection Fraction .454 .338 .222 1.343 .190
History of Depression 2.714 13.042 .035 .208 .837
Trait Anxiety* -2.302 .898 -.623 -2.563 .016
Dispositional Optimism .844 1.406 .135 .600 .553
Perceived Social Support -4.022 2.111 -.372 -1.905 .067
History of ICD Firings 11.047 10.199 .180 1.083 .288
Disease Perception
Patient Age .476 .364 .255 1.309 .200
Ejection Fraction .286 .292 .170 .980 .334
History of Depression -9.808 11.090 -.153 -.884 .383
Trait Anxiety -.654 .699 -.222 -.935 .357
Dispositional Optimism .777 1.168 .150 .665 .511
Perceived Social Support -1.315 1.824 -.147 -.721 .476
History of ICD Firings -3.190 8.653 -.065 -.369 .715
p <.05.





65







C o o


4-
Sooo -





o00o
00 e

Co m "4


0o o0
*





-5
00 M0-
C0 0 C>

ao o oQ






0 .00


a m




U
a 0 3










c
.c o s oo o








U o
o o
0 0 r' r






0 |A 0 -S-


> I I I> m > C
0c l cq m "













0 I (n M
0a U
- U UU-oo







( ( r s





66








0 00
ooo o

Ci)








000
oa



!2 t oo C- C
o- 00



0 0 0 e

0
0 t C0
-o o


t2 .- 0 *


0 0
f... C














O Q\on ---
-~4




om o





oo o


I l en en v
a d s






t as-
> >-


rj2






73 -0
0a C/ C en






SU

U -i r- Cc ^r(r^
t- Il) -







model, it was again a relatively small contributor, although still significant, and

accounted for 1.6% of the variance. In all, the total variance in long-term general health

quality of life that was accounted for by this significant model was 46.5% (p = .003).


Physical Quality of Life

In the third regression (see Table 16), as expected, the control variables of age and

LVEF significantly accounted for some of the variance of scores on the SAQ: Physical

Limitations Subscale (14.6%). However, not surprisingly, the psychological variables of

history of depression, trait anxiety, trait optimism, and social support were again more

important in predicting the physical limitations score at this interval, and significantly

accounted for 26.4% of the variance, almost twice that accounted for by the control

variables. When ICD firing history was added to the model, it was again a relatively

small contributor, although still significant, and accounted for 2.4% of the variance. In

all, the total variance in long-term physical quality of life that was accounted for by this

significant model was 43.4% (p = .016).


Disease Perception Quality of Life

In the fourth regression (see Table 17), the full model was again not significant in

predicting disease perception scores in this sample (p = .173). However, at the long-term

follow-up interval, when only the first block of control variables were considered, age

and LVEF were significant in predicting outcome (p = .047), and accounted for a small

percent of the overall variance (15.3%).




68











o* 00


oo m 0
NM# Q


0 0
=3 0o


g -
0 coo


0 0

W N- 00






t0 -o *
- o aa L -
aw 0 *







U N U m 0 o4o
r 1 00
4-4










a ) 1 s 0


0 I- ON 00 0 0
- coco
0 c O -
.3 .,-! .- .-

4) 4--oo
C0 --- -

- | *a m-
&" C4 o C^ C
0 4)



'32

( 5 5U U rs d


























































































- ch'
v- '


co




















b,




U
L4


0-
4-


tt
oo

0





00




co






.
S^
'3 "3








o o


. .n .-


, 's






coo
L00

















0~ 3 r>
D aj M













000
S (U
UOU
E"" 0 O
>* M

s- Q &


'U
o
WcU'
O S








'U
aa

*faS ^


0)







Exploratory Analyses


Predictor Variables

Although not originally planned in the proposal of this dissertation, additional

analyses were conducted to examine the differences, if any, in the short and long-term

quality of life outcome variables according to classification group for each of the medical

and psychological variables measured at the time of the initial ICD implantation. This

approach would allow for more information about the individual importance of each of

the variables to be determined. The medical and psychological predictor variables of

patient age, ejection fraction, trait anxiety, dispositional optimism, and perceived social

support were dichotomized into low and high groups based on the median noted for each

variable, while history of depression was dichotomized into negative and positive history

groups based on patient report. Afterward, ANOVAs were conducted to test for

significant differences between groups in the short and long-term quality of life measures

for each predictor variable. Analyses revealed no significant differences between groups

for the medical variables of patient age and ejection fraction on the indices of short and

long-term quality of life. In contrast, however, there were significant differences noted

between groups for each of the psychological variables in at least one quality of life

outcome variable.

For history of depression, there were significant group differences noted for long-

term rating of disease perception and for both the short and long-term mental health

ratings (p < .05). The differences were in the expected direction, with positive history of

depression being associated with poorer quality of life outcomes, and are presented in







detail in Table 18. For trait anxiety, while there were no significant between group

differences at the short-term follow-up interval, there were significant differences

between level of trait anxiety and quality of life ratings for each outcome measured at the

long-term follow-up interval such that higher anxiety related to poorer quality of life in

each instance (p < .05; see Table 19). For dispositional optimism, there were significant

differences between levels of dispositional optimism and quality of life ratings for each of

the outcomes measured at both the short and long-term follow-up intervals (p < .05).

Again, differences were in the expected direction such that less optimism related to

poorer quality of life in each instance (see Table 20). For social support, as is shown in

Table 21, while there was a significant difference in the expected direction between

groups for the mental health quality of life rating at the long-term follow-up interval (p <

.05), all other group differences were non-significant.

The results of the ANOVAs further substantiate the previous regression analyses

results in that they provide additional information regarding the significant association

between the pre-ICD implant psychological variables and indices of quality of life at both

the short and long-term follow-up intervals. Further, these analyses also highlight the

psychological variable of dispositional optimism as being particularly important (relative

to the other variables measured) in the prediction of both short and long-term quality of

life outcomes.







Table 18

Quality of Life Outcomes According to History of Depression at Initial Assessment


History of
Depression


General
Health


Short-Term Follow-up Interval
Mental Physical
Health' Limitations


Negative Mean 55.000 79.314 71.157 70.714
N 35 35 30 35
Positive Mean 57.429 62.286 73.958 55.952
N 7 7 4 7

Long-Term Follow-up Interval
History of General Mental Physical Disease
Depression Health Healthb Limitations Perceptionc
Negative Mean 52.000 80.947 68.316 68.202
N 38 38 32 38
Positive Mean 34.818 58.545 56.692 48.485
N 11 11 11 11

"Group mean differences were significant (p<.05).
bGroup mean differences were significant (p<.001).
cGroup mean differences were significant (p<.01).


Disease
Perception


--~----







Table 19

Quality of Life Outcomes According to Trait Anxiety at Initial Assessment


Trait
Anxiety


General
Health


Short-Term Follow-up Interval'
Mental Physical
Health Limitations


Disease
Perception


Low Mean 56.545 80.182 70.614 72.727
N 22 22 19 22
High Mean 53.667 71.467 70.960 60.556
N 15 15 11 15

Long-Term Follow-up Interval
Trait General Mental Physical Disease
Anxiety Healthb Healthc Limitationsb Perceptiond
Low Mean 56.200 85.400 75.000 73.333
N 20 20 17 20
High Mean 39.560 67.360 53.977 55.000
N 25 25 22 25

"At the short-term follow-up interval, the ANOVA revealed no significant differences.
bGroup mean differences were significant (p<.05).
'Group mean differences were significant (p<.005).
dGroup mean differences were significant (p<.01).







Table 20

Quality of Life Outcomes According to Dispositional Optimism at Initial Assessment


Dispositional
Optimism


General
Health


Short-Term Follow-up Interval
Mental Physical
Health Limitations'


Disease
Perception


Low Mean 40.933 58.400 56.597 53.889
N 15 15 12 15
High Mean 66.550 87.600 77.369 75.833
N 20 20 17 20

Long-Term Follow-up Interval
Dispositional General Mental Physical Disease
Optimism Healtha Healthb Limitationse Perceptiond
Low Mean 34.857 61.714 51.879 55.556
N 21 21 17 21
High Mean 59.714 88.381 72.778 69.841
N 21 21 20 21

"Group mean differences were significant (pS.005).
bGroup mean differences were significant (p<.001).
'Group mean differences were significant (p<.05).
dGroup mean differences were significant (p<.01 for short-term, p<.05 for long-term).







Table 21

Quality of Life Outcomes According to Social Support at Initial Assessment


Social Support


General
Health


Short-Term Follow-up Interval
Mental Physical
Health Limitations


Low Mean 52.769 70.154 65.972 66.026
N 13 13 8 13
High Mean 56.083 79.500 68.813 67.014
N 24 24 22 24

Long-Term Follow-up Interval
Social Support General Mental Physical Disease
Health Health" Limitations Perception
Low Mean 39.875 67.000 64.187 62.500
N 16 16 14 16
High Mean 50.821 80.000 62.847 63.691
N 28 28 24 28


Disease
Perception


'The group mean difference was significant (p<.05). The ANOVA revealed no other
significant group mean differences for quality of life outcomes at either follow-up
interval.







ICD Firing Experience History

Although not originally planned in the dissertation proposal, additional analyses

were conducted in order to more closely examine the impact of experiencing ICD firings

upon quality of life outcomes and the relationship between the psychological predictor

variables and patient experience of ICD firings. First, ICD firing experience was

classified into three groups: 1) no firing experience, 2) at least one firing and no storms

experienced, and 3) at least one ICD storm experienced. Second, given the observed

variability in patient experience of ICD firings, the conservative non-parametric Kruskal-

Wallis test was utilized. However, at both the short and long-term follow-up intervals,

this analysis did not reveal any significant differences between ICD firing groups for

either the quality of life outcome or psychological predictor variables. Further, while the

medical predictor variables did not significantly distinguish between ICD firing groups at

the short-term follow-up interval, the Kruskal-Wallis test did identify significant

differences among the ICD firing groups for the medical predictor variables at the long-

term follow-up interval. These differences, according to patient age and ejection fraction,

were such that older individuals and those patients with lower ejection fractions were

more likely to have experienced ICD firings and subsequently ICD storms at the long-

term follow-up interval as compared to the younger patients and those patients with more

healthy or higher ejection fractions (p < .01 for each medical variable).












DISCUSSION

The major objectives of the current study were: 1) to obtain descriptive

information regarding the pre-implant psychological functioning, experience of ICD

firings, and post-implant quality of life, and 2) to examine the association between pre-

implant psychological patient characteristics, experience of ICD firings, and post-implant

quality of life ratings in ICD patients at both short and long-term follow-up intervals.

Specifically, this study investigated through hierarchical regression analyses the strength

of the associations between the predictor variables of patient history of depression, trait

anxiety, dispositional optimism, social support, and ICD firings and the quality of life

outcome variables of mental health, general health, physical limitations, and disease

perception, while controlling for patient age and left ventricular ejection fraction (LVEF),

at two separate follow-up intervals.


Descriptive Analyses of the ICD Patient Sample


Demographics

The sample consisted of 88 patients receiving an ICD implant for the first time,

the majority of whom were male, Caucasian, and approximately 65 years old. Moderate

to severe impairment characterized the cardiac function of the participants, with

approximately two thirds of the patients having a documented LVEF < 35%. The vast

majority had primary medical diagnoses of ventricular arrhythmia and coronary artery

disease, and prescriptions for multiple classes of medication. While the current study







sample was smaller, the patient characteristics found in this ICD patient sample were

similar to those of the ICD patients documented in the much larger medical trials such as

AVID, MADIT, MUSTT, and the CABG-Patch Trial (AVID Investigators, 1997; Bigger,

1997; Buxton et al., 1999; Curtis et al., 1997; Namerow et al., 1999).


Psychological Indices

With regard to the psychological characteristics of the ICD patients at the time of

implantation, 22.7% of the sample (n = 20) reported a past major depressive episode and

18.7% (n = 14) reported current levels of trait anxiety significantly higher than the

average derived in a comparable cardiac sample of patients awaiting heart

transplantation. While these figures may initially appear high (representing

approximately 1 in 5 patients with a past or present psychiatric disturbance), these rates

are actually somewhat conservative compared with those previously documented in the

ICD literature (Sears, Todaro, Saia, Sotile, & Conti, 1999) and in other cardiac patient

samples (Carney et al., 1997; Lesperance & Frasure-Smith, 2000). In terms of general

psychological functioning at the time of ICD implantation, not surprisingly the ICD

patient sample reported levels of trait anxiety, dispositional optimism, and social support

that were not significantly different than community sample norms published for these

measures within the literature.


Quality of Life Indices

Regarding mental health quality of life, at both the short and long-term follow-up

intervals, ICD patients reported relatively high levels of functioning for a medical

sample-similar to that of patients in other cardiac samples such as hypertension (HTN),







congestive heart failure (CHF), and acute myocardial infarction (MI). In contrast, at the

short-term follow-up interval while ICD patients were similar to HTN and MI samples in

their report of moderately impaired general health quality of life, their ratings were

significantly higher than those of the more comparable CHF patients. Over time, though,

this situation reversed as ICD patient general health ratings declined further, becoming

more similar to that of CHF patients and significantly more impaired as compared to the

functioning of HTN and MI patients at the long-term follow-up interval. While these

changes were significant in analyses comparing the ICD patients to other cardiac

samples, the 7-point change in rating in ICD patients from the short to the long-term

follow-up interval was not sufficiently large enough to reach statistical significance for

within sample comparisons.

With regard to disease-specific quality of life, although scores for the Physical

Limitations and Disease Perceptions subscales of the Seattle Angina Questionnaire did

decrease over time, the differences between the two follow-up intervals were not

statistically significant. Interestingly, however, the ICD patients evidenced significantly

higher levels of cardiac-specific quality of life at each follow-up interval as compared to

two otherwise similar samples of coronary artery disease (CAD) patients; these analyses

suggest that ICD patients may enjoy a higher disease-specific quality of life than typical

CAD patients. This can likely be attributed to the physiological differences apparent

between the two medical diagnoses and the resulting illness imposed restrictions.


ICD Specific Knowledge and Adjustment

In terms of general knowledge about the ICD itself, at each follow-up interval the

vast majority of patients sampled indicated they had "mostly" or "definitely" received







enough information regarding their ICD (short-term: 81%, long-term: 92%). This finding

suggests that providing more intensive patient education for all ICD recipients as a

routine practice is likely to be unnecessary, given the level of current patient satisfaction

in this area. However, there were still small groups of patients at both the short and long-

term follow-up intervals that felt they had not received enough information regarding

their ICD; it is plausible these patients may be at higher risk for psychological

maladjustment to the ICD following an ICD firing experience. As such, it is this group of

patients that would most likely benefit from additional psycho-educational interventions,

such as "debriefing" sessions following an ICD firing, support group involvement, and/or

proactive informational discussions. Also in this sample, patients overwhelmingly

reported relatively few concerns related to their ICD and noted these issues as minor

(complaints such as device size and travel restrictions), suggesting good overall

acceptance of the device and providing additional indication of adequate patient

education for most ICD recipients. This high degree of acceptance is consistent with

previous findings in the literature (Sears, Todaro, et al., 2000). However, there remained

a quantity of patients at both follow-up intervals that reported fear of ICD firings as a

significant concern that impacted their activities; some of these patients had not received

an ICD firing in more than 6 months, while others had not yet received an ICD firing

even at the long-term follow-up interval. For this group of patients, similar to those

reporting inadequate ICD knowledge, it appears that additional psycho-educational

intervention focused on teaching patients to resume the healthy aspects of their previous

lifestyles is clearly indicated and would facilitate increased patient adjustment and

preparedness related to ICD firings. Indeed, such interventions have proven themselves







to be effective in the recent literature (Kohn et al., 2000). In terms of returning to normal

levels of activity and independence in one area of their lives, consistent with research by

Jung and Luderitz (1996) and the recently relaxed driving restrictions for ICD recipients,

the majority of sample at both follow-up intervals revealed they were driving regularly

(range: 1-7 days/week) without experience of adverse incidents related to the ICD.


ICD Firings

In this sample, approximately half of the ICD patients assessed at each follow-up

interval indicated they had experienced at least one ICD firing. While this percentage

was lower than the expected two-thirds reported in various other medical trials, this

difference could be attributed to the recent marked advances in the device industry's

development of technology. At present, for many patients, some ICD firings can be

successfully averted through the use of interventional techniques, more advanced pacing

technologies, dual-function devices, and sophisticated device programs which can

respond to abnormal heart rhythms with less invasive treatment approaches such as

cardioversion (multiple low level pulse firings) prior to the delivery of a major shock. It

appears that these medical advances are not only reducing the incidence of high voltage

shocks in the ICD patient population, but hold the potential to warn the patient and

possibly reduce the severity of the ICD firing when it is experienced. However, despite

these advances, there are still a large number of patients experiencing a phenomenon of

three or more ICD firings in a 24-hour period, termed an "ICD storm." It is these patients

who have been identified to be at the most risk for psychological maladjustment in the

ICD literature (Sears, Wallace, et al., 2000). In the present sample, at each of the follow-

up intervals, approximately one third of the patients who had experienced a firing noted







an ICD storm experience; this represents 16.9% and 18.4% of the total patient sample at

the short and long-term follow-up intervals, respectively. Interestingly, the percentage

obtained in the current study was significantly higher than the 10% previously reported in

the ICD literature (Credner, Klingenheben, Mauss, Sticherling, & Hohnloser, 1998).

When coupled with the evidence from the CABG-Patch Trial indicating that ICD patients

who have experienced firings report a significantly worse quality of life than ICD

patients without experience of firings (Namerow et al., 1999) and the documentation that

increased risk of mortality is associated with multiple firings in a single episode (Pacifico

et al., 1999), the high percentages of storms in this patient sample are concerning.

Further, of the subgroup of patients with an ICD firing experience, at the short-term

follow-up interval approximately one fourth of the patients noted five or more ICD

firings while approximately one third met this criterion at the long-term follow-up

interval.

In the current ICD patient sample, slightly lower levels of quality of life were

observed for patients with a history of ICD firings as compared to those without;

however, the differences between the two groups were not statistically significant. It is

possible that differences between the two groups were not found in the patients of this

sample due to the increased awareness of and attention to psychological issues on the part

of the cardiac specialty physicians caring for these patients. The referring physicians

were well educated about the need for psychological intervention in ICD patients prior to

their involvement in referring patients to this study. Therefore, they may have

inadvertently provided ICD patients with increased psycho-educational intervention at

their medical visits as compared to providers at other institutions and may have referred







patients for additional psychological treatment at a differentially higher rate. However,

this hypothesis remains unsubstantiated, as the receipt of increased attention from

medical providers and the provision of referrals for subsequent psychological

interventions were not a primary focus of this study; consequently these occurrences were

not tracked in the current study.


Predictors of Quality of Life in ICD Patients

The second major objective of this study was to examine the utility of using a

model including pre-implant variables and experience of ICD firings to predict quality of

life outcomes post-ICD implantation. The hierarchical regression analyses conducted

revealed that the proposed model accounted for a significant proportion of the variance

observed in patient quality of life ratings of mental health, general health, and physical

limitations at both the short and long-term follow-up intervals. Furthermore, it was noted

that in predicting quality of life outcomes, the proportion of unique variance accounted

for by the pre-implant psychological variables (history of depression, trait anxiety,

dispositional optimism, and social support) was relatively equal to, if not larger than, the

variance accounted for by the medical variables of age and ejection fraction. In all cases,

experience of ICD firings significantly added to the model, although to a lesser degree

than the medical or psychological variables. In contrast to the findings reported above, at

neither follow-up interval were the analyses significant in demonstrating an association

between the predictor variables and disease perception quality of life. While it might be

true that the effect size associated with this variable required a larger sample size to

detect an effect, it is also plausible that a model containing different predictor variables

might have been more useful. According to the "sickness scoreboard" theory previously







hypothesized in the literature (Sears et al., 1999), it is likely that a model which included

such variables as number of medical appointments post-implant, health

beliefs/expectations, physical limitations, and experience of ICD firings may have

accounted for more of the variance in patient rating of disease perception quality of life.

This theory may warrant further testing as this study was designed before it was

postulated.

Prior to the implementation of this study, it had been hypothesized that experience

of ICD firings would account for a significantly higher proportion of the unique variance

as compared to the psychological variables, ultimately supporting the recommendation

that enhanced patient education programs regarding ICD firings and subsequent

"debriefing" sessions would be indicated for the ICD patient population at large.

However, results have instead indicated that, while experience of ICD firings is relevant,

the pre-implant psychological variables are actually more important to consider and

account for a significantly larger proportion of the variance in outcome. Further, the lack

of a stronger association between ICD firings and the outcome variables in this study

provides support to suggest that patient experience of ICD firings does not necessarily

warrant immediate psychological intervention and may be of less independent importance

than the literature has previously suggested. The degree to which psychological variables

are significantly associated with the patient outcome variables can assist clinicians in

better understanding the typical characteristics of presenting ICD recipients and the

impact these psychological variables may have upon psychosocial functioning and

quality of life. With regard to specific impact on quality of life associated with each pre-







implant psychological variable, the ad-hoc analyses conducted revealed the following

information:

1. ICD patients with a positive history ofdepression at pre-implant were
significantly more likely to report lower ratings of mental health quality of life
at both the short and long-term follow-up intervals and lower ratings of
disease perception quality of life at the long-term follow-up interval;

2. ICD patients with higher levels of trait anxiety at pre-implant were
significantly more likely to report lower ratings of mental health, general
health, physical limitations, and disease perception quality of life at the long-
term follow-up interval;

3. ICD patients with lower levels of dispositional optimism at pre-implant were
significantly more likely to report lower levels of mental health, general
health, physical limitations, and disease perception quality of life at both the
short and long-term follow-up intervals; and

4. ICD patients with lower levels of social support at pre-implant were
significantly more likely to report lower levels of mental health quality of life
at the long-term follow-up interval.

Given the significant relationships between positive history of depression, high trait

anxiety, low dispositional optimism, and low social support and the quality of life

outcome variables, it appears that the previously proposed provision of a psychological

intervention for all ICD recipients experiencing firings are better indicated for those

patients presenting with identified risk-factor characteristics for maladjustment and/or

psychological difficulties. Further, there is evidence to suggest that targeting specific

interventions toward those ICD recipients reporting high levels of trait anxiety and low

levels of dispositional optimism would provide the most positive impact on long-term

quality of life outcomes post-ICD implantation.







Strengths and Limitations

The current study has several methodological strengths that allow for more

conclusive interpretations of the findings than many of the studies previously published

in the ICD literature. First, the study design involved prospective data collection and

included the examination of a variety of patient psychological characteristics prior to an

opportunity for patient experience with the ICD to alter these variables. Second,

assessments were restricted to the use of standardized and validated measures that were

chosen specifically to address previous findings in the literature and for their established

reliability and validity in measuring the constructs of interest. Third, this study included

the collection of follow-up data at two separate endpoints-both short and long-term

intervals post-ICD implantation-to address previously discrepant findings in the

literature and provide descriptive information about changes over time in patient

functioning and quality of life. Fourth, the total number of patients recruited to

participate in the study was designed to be sufficiently large such that, even after

suffering the expected significant levels of attrition, the minimum number of subjects

required for adequate power given the planned regression analyses was still available at

each of the follow-up intervals. Fifth, this study involved the collection of data from a

sample which was very similar in demographic and medical characteristics to those

patients described in large national ICD medical samples, thereby increasing the

likelihood that these findings should generalize well to other ICD patient populations.

Given the methodological weaknesses currently existing in the psychosocial ICD

literature to date, this study represents a significant contribution in terms of progress and

can serve as a useful aid to clinicians and researchers alike striving to understand the







differential impact of medical factors, psychological factors, and experience of ICD

firings upon patient functioning and quality of life.

Despite the many strengths of the current study, several weaknesses should be

considered as they may limit the interpretation of the findings. First, it is important to

note that while the study sample mimics the typical ICD patient population, it is

demographically restricted to older, male, Caucasian patients; consequently, the study

findings may not be as generalizable to patient samples that are more youthful or include

large percentages of female and/or non-white ICD patients. Second, both the short and

long-term follow-up intervals were marked by sizeable attrition; while participants and

non-participants at follow-up did not significantly differ from each other at the time of

initial ICD implantation on medical or psychological variables, it is possible that they

may have differed in terms of quality of life outcomes. Third, although very few study

patients were prescribed psychotropic medications and/or carried a psychological

diagnosis at the time of initial ICD implantation, patient participation in psychological

intervention by private providers or support groups, etc., post-ICD implantation was not

tracked; as differences in quality of life outcomes potentially exist between patients in

treatment versus those not, this represents a possible limitation in the interpretation of the

results.


Conclusions

The present study provides greater insight into the psychosocial functioning and

quality of life of ICD recipients, a population in need of additional assessment and

interventional research given the high incidence of psychological disturbance and

impaired quality of life in these patients. Study results indicate that ICD recipients with a







positive history of depression, high trait anxiety, low dispositional optimism, and low

social support at the time of ICD implantation are particularly "at-risk" for exhibiting

lower levels of functioning on mental health, general health, and physical health quality

of life indices post-implant, even after controlling for patient age or ejection fraction. In

addition, experience of ICD firings was found to further complicate patient adjustment to

the ICD, even after accounting for the influence of both the medical and psychological

variables. Furthermore, it is essential to note that pre-implant levels of high trait anxiety

and low dispositional optimism were the two most critical variables in predicting lower

functioning in terms of mental health, general health, physical limitations, and disease

perception quality of life at the long-term follow-up interval. Collectively, these findings

identify subgroups of ICD patients that are at higher risk for experiencing poor quality of

life outcomes and are subsequently in need ofpsychosocial interventions directed at

facilitating overall patient adjustment to the ICD and improving patient preparedness for

an ICD firing experience. To date, however, there have been very few studies that have

examined the impact of such interventions on ICD patient adjustment and psychosocial

functioning (Kohn et al., 2000).

Considering the evidence that the integration of psychosocial interventions within

rehabilitation programs leads to reduced cardiovascular morbidity and mortality in

comparison to those programs without this intervention component (Linden, 2000), it

naturally follows that ICD patients, particularly those patients with identified risk-factors

for psychological maladjustment, would benefit from such interventions. The findings of

the current study are particularly important because of evidence that unaddressed

psychological factors in medical patients frequently lead to poorer health outcomes and







increased medical utilization, thereby increasing health care costs (Friedman et al., 1995).

Results from the current study demonstrate the importance of assessing pre-implant

psychological factors in intended ICD recipients, particularly as these variables were

found to be at least equally as important of medical variables in predicting several

dimensions of quality of life post-implant. Clinicians can utilize this information to

improve quality of life outcomes in ICD recipients by providing patients with increased

attention to their psychological needs and referrals for psycho-educational interventions

when indicated.


Future Directions

Given the findings of the current study and other studies documenting various

degrees of psychological maladjustment to the ICD, future research needs to next address

the impact of psychological interventions, such as support group involvement, proactive

informational discussions, and individual "debriefing" sessions when indicated, on the

psychosocial functioning and quality of life outcomes of ICD patients post-implant.

Additionally, given the paucity of research that has been conducted to date examining

individual differences influencing the adjustment of females to the ICD, future studies

would be beneficial in providing useful information to clinicians about the potential

differences between male and female ICD recipients. And finally, additional research

identifying and further scrutinizing potential risk factors for psychological maladjustment

to the ICD is indicated. While the current study provides useful information in this

regard, future studies could more specifically address the independent value of each of

the psychological predictor variables by determining the differential risk associated with

each of these factors. While this type of analysis was beyond the original scope of the