Neighborhood conditions, diabetes, and risk of lower-body functional limitations among middle-aged African Americans : a...

MISSING IMAGE

Material Information

Title:
Neighborhood conditions, diabetes, and risk of lower-body functional limitations among middle-aged African Americans : a cohort study
Abbreviated Title:
BMC Public Health
Physical Description:
Book
Language:
English
Creator:
Schootman, Mario
Andresen, Elena M.
Wolinsky, Fredric D.
Miller, J. Phillip
Yan, Yan
Miller, Douglas K.
Publication Date:

Notes

Abstract:
Background: The relationship between presence of diabetes and adverse neighborhood and housing conditions and their effect on functional decline is unclear. We examined the association of adverse neighborhood (block face) and housing conditions with incidence of lower-body functional limitations among persons with and those without diabetes using a prospective population-based cohort study of 563 African Americans 49-65 years of age at their 2000- 2001 baseline interviews. Methods: Participants were randomly sampled African Americans living in the St. Louis area (response rate: 76%). Physician-diagnosed diabetes was self reported at baseline interview. Lower-body functional limitations were self reported based on the Nagi physical performance scale at baseline and the three-year follow-up interviews. The external appearance of the block the respondent lived on and five housing conditions were rated by study interviewers. All analyses were done using propensity score methods to control for confounders. Results: 109 (19.4%) of subjects experienced incident lower-body functional limitations at three-year follow-up. In adjusted analysis, persons with diabetes who lived on block faces rated as fair-poor on each of the five conditions had higher odds (7.79 95% confidence interval: 1.36-37.55 to 144.6 95% confidence interval: 4.45-775.53) of developing lower-body functional limitations than the referent group of persons without diabetes who lived on block faces rated as good-excellent. At least 80 percent of incident lower-body functional limitations was attributable to the interaction between block face conditions and diabetes status. Conclusions: Adverse neighborhood conditions appear to exacerbate the detrimental effects on lower-body functioning associated with diabetes.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
All rights reserved by the source institution.
Resource Identifier:
doi - 10.1186/1471-2458-10-283
System ID:
AA00012386:00001

Full Text

PAGE 1

1 Additional file 2 Propensity score adjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction co ntrast ratio and attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower body functional limitation at 3 year follow up (weighted n=563) Good excellent block face conditions Fair poor block face conditions Without diabetes With diabetes Without diabetes With diabetes Type of condition OR OR 95% CI OR 95% CI OR 95% CI Interaction contrast ratio 95% CI Attributable proportion 95% CI Housing conditions 1.00* 1. 15 0. 5 7 2. 22 1. 46 0.7 0 3.05 7. 79 1.36 37.55 6.01 0.46, 35.22 0. 80 0.33, 0.96 Noise 1.00* 1. 13 0. 5 6 2.3 0 1. 74 0.87 3.47 144.60 4. 45 775.53 143.0 2. 64, 772.79 0. 99 0.52 1.00 Air quality 1.00* 1.53 0.68 3. 14 2. 84 1.3 4 6.17 23.65 3.73 93.68 19.62 0.00, 89.27 0. 86 0 .00, 0.97 Street & road quality 1.00* 1. 1 4 0. 56 2.39 3.39 1. 24 8.00 79.53 5.85 347.6 7 76.18 2.18, 345.68 0. 95 0. 36, 0.99

PAGE 2

2 Yard & sidewalk quality 1.00* 1.0 1 0.53 2. 00 1. 39 0. 64 3.02 20.79 2.64, 82.37 19.09 1.04, 80.45 0. 93 0 .33, 0.99 Variables included in the propensity score: sampling stratum, age, gender, income, perceived income adequacy, educational attainment, marital status, employment status, number of persons in household, health care insurance, not being able to see a doctor because of c ost, s ocial support, self rated health status, depressive symptoms, a count of the number of chronic conditions, body mass index, risk of alcohol abuse, and physical activity. Referent odds ratio



PAGE 1

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Open Access RESEARCH ARTICLE Bio Med Central 2010 Schootman et al; licensee BioMed Cent ral Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2 .0), which permits un restricted us e, distribution, and reprod uction in any medium, provided the original work is properly cited.Research articleNeighborhood conditions, diabetes, and risk of lower-body functional limitations among middle-aged African Americans: A cohort studyMarioSchootman*1, ElenaMAndresen2, FredricDWolinsky3, J PhilipMiller4, YanYan5 and DouglasKMiller6AbstractBackground: The relationship between presence of diabetes an d adverse neighborhood and housing conditions and their effect on functional decline is unclear. We examined the association of adverse ne ighborhood (block face) and housing conditions with incidence of lower-body functional limitations among persons with and those without diabetes using a prospective population-based cohort study of 563 African Americans 49-65 years of age at their 20002001 baseline interviews. Methods: Participants were randomly sampled African Americans living in the St. Louis area (response rate: 76%). Physician-diagnosed diabetes was self re ported at baseline interview. Lower-body functional limitations were self reported based on the Nagi physical performance scale at baseline and the three-year follow-up interviews. The external appearance of the block the respondent lived on and five housing condit ions were rated by study interviewers. All analyses were done using prop ensity score methods to control for confounders. Results: 109 (19.4%) of subjects experienced incident lower-body functional limitations at three-year follow-up. In adjusted analysis, persons with diabetes who lived on block fa ces rated as fair-poor on each of the five conditions had higher odds (7.79 [95% confidence interval: 1.36-37.55] to 144.6 [95% confidence interval : 4.45-775.53]) of developing lower-body functional limitations than the referent group of persons without diabetes who lived on block faces rated as good-excellent. At least 80 percent of incident lower-body functional limitations was attributable to the interaction between block face conditions and diabetes status. Conclusions: Adverse neighborhood conditions appear to exac erbate the detrimental effects on lower-body functioning associated with diabetes.BackgroundDiabetes has many adverse health effects, but one of the most important with serious consequences for living independently in the community is the impact of diabetes on lower-body function [1]. Poor lower-body function plays a crucial role in the disablement process and has been associated with increased disability days, physician contacts, fear of falling, falls, hip fracture, depression, nursing home placement, and mortality [2,3]. Not surprisingly, diabetes is a major risk factor for the development of lower-body-related functional limitations [4], and a substantial portion of population (adult) disability related to lower-extremity functioning is attributable to diabetes [5]. In the United States lower-body functional limitations are especially high among urban African Americans [6]. For these reasons, the issue of the effect of diabetes on lower-body functioning is particularly important for African Americans. Neighborhood conditions have been shown to predict incident lower-body functional limitations [7,8] and adverse housing conditions have been associated with incident diabetes [9]. There are many potential pathways by which adverse neighborhood conditions might increase the risk of lower-body functional limitations, including increased stress, lower access to medical care, higher social isolation, and lower collective efficacy and social capital [10-12]. Based on the above-described asso* Correspondence: mschootm@im.wustl.edu1 Department of Medicine and Pediatri cs, Washington University School of Medicine, St. Louis, MO, USAFull list of author information is available at the end of the article

PAGE 2

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Page 2 of 7ciations, we hypothesized that adverse neighborhood and housing conditions may exacerbate the detrimental effects on lower-body functioning associated with diabetes. To our knowledge no studies have examined the potential interaction between diabetes and contextual factors such as neighborhood and housing conditions on the development of lower-body functional limitations. Adverse neighborhood and housing conditions may exacerbate the progression toward lower-body functional limitations among persons with diabetes as a result of complications (e.g., neuropathy, amputations), associated conditions (e.g., hypertension, obesity, stroke), low fruit and vegetable consumption and low physical activity, and reduced muscle strength [1], many of which are also associated with adverse neighborhood and housing conditions [13,14]. Therefore, we examined the association of adverse neighborhood (measured at the block face level) and housing conditions with incidence of lowerbody dysfunction among persons with and those without diabetes using a longitudinal study of African Americans.MethodsBaseline sample (wave 1)The sampling design of the African American Health cohort has been described elsewhere [15]. Briefly, the African American Health study is a population-based cohort study of 998 noninstitutionalized African Americans recruited in 2000-2001 using multi-stage probability sampling. All subjects lived in one of two geographic sampling strata: either a poor, inner-city area (St. Louis), Missouri, United States) or more heterogeneous suburbs just northwest of the City of St. Louis. Interviewers (two thirds of whom were African American) with extensive study-specific training screened households for eligibility criteria, which involved self-reported black or African American race, birth date during 1936 through 1950, and Mini-mental Status Examination scores > 16. Subjects were paid volunteers. Sampling proportions were set to recruit approximately equal numbers of subjects from both areas (sampling strata). Each subject was weighted based on the selection probability and the response rate. When these weights are applied, the African American Health study sample represents the noninstitutionalized African American population in the two areas as of the 2000 Census. All subjects received in-home, baseline evaluations (average = 2.5 hours) between September 2000 and July 2001. Baseline response was 76% (998/1320). All procedures were approved by the Institutional Review Boards of Saint Louis University and the University of Michigan.Follow-up sample (wave 4)Follow-up in-home interviews averaging 1.5 hours were conducted 36 months after baseline assessments. Of the 998 persons who participated at baseline, 853 were successfully interviewed at follow-up. Since 51 persons had died between baseline and follow-up, the response for surviving subjects was 90.1% (853/947). No attrition bias during waves 1 through 4 was evident for any of the major variables involved in the current analysis. A total of 290 persons (weighted) had two or more lower body functional limitations at baseline and were excluded from further analysis. Thus, 563 persons (weighted) had one or fewer lower-body functional limitations at baseline and comprised the study sample.Lower-body functional limitationsFive items (0 = no difficulties to 1 = difficulty) from the Nagi physical performance scale assessed lower-body functional limitations, which were summed to form the outcome measure (ranging from 0 to 5) [6]. Items included difficulties in walking a quarter of a mile; walking up and down 10 steps without rest; standing for 2 hours; stooping, crouching, or kneeling; and lifting and carrying 10 pounds [16]. Subjects who expressed any difficulty or inability to perform the functional task at the time of the interview were considered to be limited in that task. Similar to other studies [7], we limited subjects in this study to those with one or fewer lower-body functional limitations at baseline in order to examine the risk of developing two or more lower-body functional limitations three years later. At follow-up, we defined incident lower-body functional limitations as reporting difficulty or being unable to perform at least two of the five physical tasks among those with one or fewer lower-body functional limitations at baseline.Adverse neighborhood and housing conditionsAssessment of neighborhood conditions was comprised of interviewer observations of the block face on which the respondent lived and participants' self-reported neighborhood desirability. An "objective" assessment of the external appearance of the block face (neighborhood) in front of the homes where the participants resided was done by the survey team using a previously published assessment tool [17] during household enumeration, which occurred an average of seven months before the participants were recruited and data were collected. Thus, block face conditions were collected independently from the interview data. Data about the housing conditions were collected at the time of the in-home interview and therefore not independently. On four-point scales (1 = excellent, 4 = poor) observers rated each of five characteristics: condition of houses, amount of noise (from traffic, industry, etc.), air quality, condition of the streets, and condition of the yards and sidewalks in front of homes where the participant lived. Weighted inter-rater Kappa statistics ranged from 0.58 (air quality) to 0.84 (condition of yards and sidewalks [18].

PAGE 3

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Page 3 of 7We also obtained a subjective measure of neighborhood conditions from respondents at baseline using a four-item scale of the neighborhood as a place to live, general feelings about the neighborhood, attachment to the neighborhood, and neighborhood safety from crime [19]. Participant responses were dichotomized for each condition, and the scale ranged from 0 to 4. Assessment of housing conditions was an observed five-item scale based on the interviewer's ratings at the baseline interview of the cleanliness inside the building; physical condition of the interior; condition of furnishings; condition of the exterior of the building; and a global rating (all rated as excellent, good, fair, or poor). The testretest reliability was at least 0.68 for each condition [9]. In the present analysis, each block face condition and each housing condition was dichotomized as either fair or poor versus good or excellent.DiabetesThe baseline interview asked respondents about the presence of physician-diagnosed diabetes (test-retest reliability in a subsample of African American Health study participants was 0.94) [20].CovariatesBaseline covariates included in the analysis were patterned after other African American Health cohort research [7]. Socio-demographic variables involved sampling stratum (inner city, suburb), age, gender, income categories, perceived income adequacy (having a comfortable income, having just enough to get by, not having enough to get by), educational attainment, marital status, employment status, number of persons in household, having health care insurance at the time of or during the 12 months prior to interview, and not being able to see a doctor because of cost during 12 months prior to interview. Social support was measured using five items from the Medical Outcomes Study social support instrument. Health at baseline was measured by the self-rated health status question of the Short-form 36, depressive symptoms (score of at least 9 using the 11-item Center for Epidemiology Depressive Symptoms scale), a count of the number of self-reported physician-diagnosed severe chronic conditions ever experienced (asthma, chronic airway obstruction, heart failure, heart attack, angina, stroke, chronic kidney disease, arthritis, and cancer other than a minor skin cancer). Also assessed at baseline were body mass index, current smoking status, risk of alcohol abuse (CAGE), the Yale physical activity instrument, and grip strength. The conceptual model of the relationship among the various types of variables is displayed in Figure 1. The association of interest is the relationship between diabetes and development of lower-body functional limitations as modified by adverse block face/housing conditions. Diabetes and the propensity score are predicting lowerbody functional limitations. The association of diabetes with lower-body functional limitations is modified by adverse block face/housing conditions. The inclusion of the propensity score is aimed at estimating the unbiased association of interest.Statistical analysisFirst, we calculated unadjusted measures of association (odds ratio [OR] and 95% confidence intervals [CI]) and interaction between diabetes and block face/housing conditions for the risk of incident lower-body functional limitation at 3-year follow-up. For each block face and housing condition, a single variable with four categories was created by combining a dichotomous neighborhood/ housing variable (excellent/good vs. fair/poor) with the diabetes variable as part of the additive model [21]. Measures of interaction included the Interaction Contrast Ratio and the Attributable Proportion as measures of departure of additivity. The Interaction Contrast Ratio is the excess risk due to interaction relative to the risk without exposures [22]. If there is no superadditive interaction (null hypothesis), the Interaction Contrast Ratio equals 0. Interaction Contrast Ratio >0 indicates superadditivity, and Interaction Contrast Ratio < 0 indicates subadditivity. Odds ratios were substituted for the relative risks in the Interaction Contrast Ratio calculation. Attributable Proportion refers to the proportion of incident lower-body functional limitations attributable to the interaction among persons who experienced both exposures. If there is no interaction, the Attributable Proportion will equal 0. To adjust for potential confounding from covariates, we used the propensity score method to obtain adjusted estimates of the measures of association and interaction between neighborhood/housing conditions with presence of diabetes on incident lower-body functional limitations [23,24]. All variables were included in the calculation of propensity scores. The propensity score is defined as the conditional probability of a person livingFigure 1 Conceptual Model

PAGE 4

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Page 4 of 7under a certain neighborhood/housing condition and diabetes status, given covariates included. Multivariable logistic regression may be limited in its ability to control for confounders in studies when there are fewer than 10 events per variable analyzed [25]. The use of propensity scores has been proposed as an alternative that may be especially useful when multiple confounders are involved [26,27]. Propensity score methods produce estimates that are more accurate than logistic regression estimates when there are seven or fewer events per confounder, as was the case in the present study [24]. Similar to multivariable methods, unmeasured confounders are not included in the construction of propensity scores. Following previous work [28], we first estimated the probability in each of four categories of the interaction (diabetes Yes/No by block face/housing condition Fair/Poor vs. Good/Excellent) for each individual using a nominal multinomial logistic regression model. Next, we assigned a weight using the inversed predicted probability that the individual was observed in one of the categories. The propensity score-adjusted estimates incorporated these weights in the analyses. Then, we obtained the interaction indices using the log odds ratio estimates from the logistic regression model. Also, we examined the interaction indices for hypertension and arthritis with the contextual variable to challenge the robustness of the findings. Confidence intervals for all effect measures were calculated using the bootstrap percentile methodology. For a 95% confidence interval the 2.5th and 97.5th percentiles of the empirical distribution of Interaction Contrast Ratios were calculated from 1,000 data sets resampled with replacement from the original data. All statistical analyses were conducted using SAS software, version 9.1.ResultsOf 563 subjects with zero or one lower-body functional limitations at baseline, 109 (19.4%) experienced two or more lower-body functional limitations at the 3-year follow-up. Eighteen percent of subjects reported having diabetes at baseline and 82 percent did not. Of those with diabetes, the percentage of subjects who lived on one of the ten block face or housing conditions rated as fair or poor ranged from 10.3% to 23.3%. Of those without diabetes, the percentage of subjects who lived on one of the ten block face or housing conditions rated as fair or poor ranged from 17.2% to 20.2%. There were no statistical differences in block face or housing conditions between persons with and those without diabetes. Baseline characteristics of the study population and factors associated with incident lower-body functional limitations in univariate analysis have been described briefly in Table 1 and more extensively elsewhere [7]. Briefly, persons who were older, unable to visit a doctor because of the cost, scored nine or more on the Center for Epidemiology Depressive Symptoms 11-item scale, experienced greater number of severe chronic conditions, or had one lowerbody functional limitations at baseline were more likely to experience incident lower-body functional limitations at 3-year follow-up. Persons were less likely to have incident lower-body functional limitations at follow-up when they had lived more than five years at the present address or were overweight at baseline.IncidenceOf 563 subjects with zero or one lower-body functional limitations at baseline, the percentage that developed two or more lower-body functional limitations at 3-year follow-up varied according to the participant's diabetes status and each of the five block-face conditions. For example, 65% percent of persons with diabetes who lived on block faces with yards and sidewalks in fair-poor condition developed lower-body functional limitations. In contrast, 19.4 percent of persons without diabetes who lived on block faces with yards and sidewalks in fair-poor condition developed lower-body functional limitations. About 17 percent of persons who lived on block faces with good-excellent conditions (regardless of diabetes status) developed lower-body functional limitations. Similar results were observed for the other four block-face conditions. Little difference was present examining each of the five housing conditions.Interaction between block face conditions and diabetes statusIn unadjusted analysis, (a) persons with diabetes residing on block faces with good-excellent conditions and (b) persons without diabetes living on block faces with fairpoor conditions generally were not significantly more likely to develop lower-body functional limitations three years later than the referent group of persons without diabetes who lived on block faces rated as good or excellent (additional File 1). The only exceptions were for the rating of air quality for both groups and street and road quality for the second group, for which the odds ratios were 2.0 to 2.4. In contrast, persons with diabetes who lived on block faces rated as fair or poor on each of the five conditions had seven to 14 higher odds of developing lower-body functional limitations than the referent group. An interaction existed between block face condition and presence of diabetes for each of the five conditions (all Interaction Contrast Ratio >1.0). At least 75 percent of the incidence of lo wer-body functional limitations was attributable to this interaction for each of the block face conditions (additional File 1). In adjusted analyses, we observed parameter estimates that generally were larger than those in the unadjusted analysis for all conditions (additional File 2). The Attributable Proportion of the incidence of lower-body func-

PAGE 5

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Page 5 of 7tional limitations due to the interaction involving the other four conditions was at least 80 percent. The values of the Interaction Contrast Ratios indicate that the excess risk due to the interaction was large relative to the risk without either exposure.Interaction between housing conditions and diabetes statusIn unadjusted analysis, persons with diabetes who lived under housing conditions rated as good or excellent and those rated as fair or poor had about two times higher odds of developing lower-body functional limitations at 3-year follow-up compared to the reference group (additional File 3). However, based on the Interaction Contrast Ratio and the Attributable Proportion there was no interaction between housing condition and diabetes. In adjusted analysis, there was also no interaction between any of the housing conditions and diabetes, and the associations with the two types of exposures on their own became insignificant (additional File 4). Sensitivity analyses were performed with hypertension and arthritis to examine the specificity of the interaction between diabetes and the contextual variables. However, no interaction was found for hypertension and arthritis (data not shown), suggesting the observed interaction is specific to diabetes and neighborhood conditions.Discussion and ConclusionsThis study has demonstrated the importance of the interaction of diabetes and neighborhood conditions acting in concert on the deterioration of lower-body functional limitations in an urban African American population. This "double jeopardy" suggests that both risk factors combined increased the risk of lower-body functional limitations considerably more than diabetes or adverse block face conditions alone [29]. In contrast, we found noTable 1: Prevalence of selected characteristics at baseline and unadjusted risk of three-year incident lower body functional limitation for subjects in the African-American Health studyUnadjusted risk of incident lower body functional limitation at 3-year follow-up Baseline measure** (weighted n = 563) Odds ratio95% CI Age (mean[s.d])56.1 (4.7)1.061.01 1.11 Gender Women vs. Men54.6%1.460.95 2.24 Length of time at present address More than 5 yrs vs. Less than 5 years73.1%0.830.34 0.82 Objective income < $20,000 vs. > = $50,00017.4%1.680.95 2.94 $20,000 < $50,000 vs. > = $50,00048.8%1.320.29 6.07 Highest level of education < 12 years vs. 12 years or more21.0%0.740.43 1.27 Unable to visit doctor because of cost Yes vs. No6.4%2.351.14 4.83 Center for Epidemiology Depressive Symptoms 11item score 9 Yes vs. No12.5%1.891.08 3.32 No. of severe chronic conditions (per condition)0.8 (1.0)1.561.28 1.90 Lower body limitation at baseline One vs. None29.2%3.562.30 5.49 Body Mass Index >= 30.0 vs. < 25.035.5%0.610.36 1.04 25.0 29.9 vs. < 25.040.7%0.480.28 0.81 CI: confidence interval, s.d.: standard deviation. ** Mean (SD), unless noted.

PAGE 6

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Page 6 of 7evidence of an interaction between diabetes and housing conditions on incident lower-body functional limitations. To our knowledge, ours is the first study that shows a powerful synergy between neighbor conditions and presence of disease on subsequent adverse functional outcome. Adverse neighborhood conditions may exacerbate the progression toward lower-body dysfunction among persons with diabetes as a result of diabetic complications (e.g., heart disease, visual impairment, neuropathy, ulceration, lower extremity amputation), associated conditions (e.g., obesity, hypertension), poor diet and low physical activity, and reduced muscle strength [1]. We adjusted for these risk factors by including them in the propensity score, except for lower-extremity disorders such as peripheral neuropathy, foot ulcers, peripheral arterial disease, and lower-extremity amputation, for which we had little data. These lower-extremity conditions have been associated with numbness in the extremities, and trouble with gait and balance, and lower-body functional limitations [30]. It is possible that adverse neighborhood conditions accelerate the decline toward lower-body functional limitations through lower-extremity conditions in persons with diabetes. However, discerning the meditational pathways was not the purpose of our study. In efforts to prevent future lower-body dysfunction among persons with diabetes by intervening upon lower-extremity disorders, attention to environmental circumstances (especially block face conditions) and the individual's interaction with them needs to be part of the interventional strategy. The findings appear robust with respect to sensitivity analysis. Study limitations include a study sample that involves a single race, a single city, and a restricted age range, all of which may limit generalizability. However, focusing on a single race allows the disentanglement of race and socioeconomic status. Notably, African Americans experience more diabetes and more diabetic complications than does the majority population [31], so the effect of the diabetesneighborhood interaction may be particularly strong in African Americans. We used self-reported diabetes to classify cases of diabetes and thus some cases of prevalent diabetes probably were missed. Despite the very high test-retest reliability of self-reported diabetes in the African American Health data [20], misclassification of diabetes status may still be present, which could lead to biased results. However, unless misclassification of selfreported diabetes was dependent upon block face condition, our results would likely be a conservative estimate of the true relationships. Another limitation is that we had in some instances only 24 persons with diabetes who lived in fair or poor block face conditions at baseline. The relatively small size of this group resulted in wide confidence intervals in some results, but unity was never included. Finally, in most studies of neighborhood effects, multiple study participants are nested within their neighborhood, requiring the use of multilevel statistical techniques. In this study sample, there were 551 block faces, 363 on which only one participant resided (65.9%). Only 3.6 percent of block faces contained five or more participants. We were not able to use multilevel statistical techniques because there was not enough clustering of participants within block faces to support a robust multilevel analytic approach. In a previous study, we randomly selected one subject per block face from the block faces with more than 1 subject and showed that parameter estimates were very similar to our findings using propensity scores [7]. In summary, there appears to be a powerful interaction between adverse block face conditions and the presence of diabetes on decline is a crucial factor for maintaining health and independence (i.e., lower-body physical functioning) in urban-dwelling middle-aged African Americans. Further research is needed to investigate the mediators of this powerful interaction.Additional materialCompeting interests The authors declare that they have no competing interests. Authors' contributions MS conceived of the study and drafted the manuscript, EMA, FDW, JPM participated in drafting of the manuscript, YY performed the statistical analysis, DKM helped conceive of the study and particip ated in drafting of the manuscript. All authors read and approved the final manuscript. Acknowledgements This work was supported by grants from the National Institutes of Health (grant numbers AG10436, DK067172). We thank Mr. James Struthers of the Washington University Diabetes Research and Training Center Prevention and Control Additional file 1 Unadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio and attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower-body functional limitation at 3-ye ar follow-up (weighted n = 563). Additional file 2 Propensity score adjusted measures of association (odds ratio and 95% confidence interv als) and interaction (interaction contrast ratio and attributable pr oportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower-body functional limitatio n at 3-year follow-up (weighted n = 563).* Additional file 3 Unadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio and attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of incident lower-body functional limitation at 3-ye ar follow-up (weighted n = 563). Additional file 4 Propensity score adjusted measures of association (odds ratio and 95% confidence interv als) and interaction (interaction contrast ratio and attributable pr oportion and 95% confidence intervals) between diabetes and housing co nditions for the risk of incident lower-body functi onal limitation at 3-year follow-up (weighted n = 563).

PAGE 7

Schootman et al. BMC Public Health 2010, 10 :283 http://www.biomedcentra l.com/1471-2 458/10/283 Page 7 of 7Core (DK20579). We thank Drs. Jack M. Guralnik and Robert B. Wallace for reading an earlier version of this manuscript. Author Details1Department of Medicine and Pediatrics Washington University School of Medicine, St. Louis, MO, USA, 2Department of Epidemiology and Biostatistics, College of Public Health and Health Professions, University of Florida, Gainesville, FL, USA, 3Department of Health Management and Policy, College of Public Health, the University of Iowa, IA, USA, 4Division of Biostatistics, Washington University School of Medicine, St. Louis, MO, USA, 5Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA and 6Indiana University Center for Aging Re search and Regenstrief Institute, Inc., Indiana University School of Medicine, Indianapolis, IN, USA References 1.Bourdel-Marchasson I, Helmer C, Fa got-Campagna A, Dehail P, Joseph PA: Disability and quality of life in elderly people with diabetes Diabetes Metabol 2007, 33(Supplement 1): S66-S74. 2.Andresen EM, Wolinsky FD, Miller JP, Wilson MM, Malmstrom TK, Miller DK: Cross-sectional and longitudinal risk factors for falls, fear of falling, and falls efficacy in a cohort of middle-aged African Americans Gerontologist 2006, 46(2): 249-257. 3.Wolinsky FD, Callahan CM, Fitzgerald JF, Johnson RJ: The risk of nursing home placement and subsequent death among older adults J Gerontol 1992, 47(4): 173-182. 4.Volpato S, Ferrucci L, Blaum C, Osti r G, Cappola A, Fried LP, Fellin R, Guralnik JM: Progression of lower-extremity disability in older women with diabetes: the women's health and aging study Diabetes Care 2003, 26(1): 70-75. 5.Odding E, Valkenburg HA Stam HJ, Hofman A: Determinants of locomotor disability in people aged 55 years and over: The Rotterdam study Eur J Epidemiol 2001, 17(11): 1033-1041. 6.Miller DK, Wolinsky FD, Malmstrom TK, Andresen EM, Miller JP: Inner city, middle-aged African Americans have excess frank and subclinical disability J Gerontol A Biol Sci Med Sci 2005, 60(2): 207-212. 7.Schootman M, Andresen E, Wolinsky F, Malmstrom T, Miller J, Miller D: Neighborhood conditions and risk of incident lower-body functional limitations among middle-aged African Americans Am J Epidemiol 2006, 163: 450-458. 8.Balfour JL, Kaplan GA: Neighborhood environment and loss of physical function in older adults: evidence from the Alameda County Study Am J Epidemiol 2002, 155(6): 507-515. 9.Schootman M, Andresen EM, Wolinsky FD, Malmstrom TK, Miller JP, Yan Y, Miller DK: The effect of adverse housing and neighborhood conditions on the development of diabetes mellitus among middle-aged African Americans Am J Epidemiol 2007, 166: 379-387. 10.Sampson R: Neighborhood-level context and health: Lessons from sociology In Neighborhoods and Health Edited by: Kawachi I, Berkman LF. Oxford: Oxford University Press; 2003:132-146. 11.Glass T, Balfour J: Neighborhoods, aging, and functional limitations In Neighborhoods and Health Edited by: Kawachi I, Berkman L. Oxford: Oxford University Press; 2003:303-334. 12.Mendesde Leon CF, Glass TA, Berkman LF: Social engagement and disability in a community population of older adults: The New Haven EPESE Am J Epidemiol 2003, 157(7): 633-642. 13.Auchincloss AH, Diez Roux AV, Brown DG, Erdmann CA, Bertoni AG: Neighborhood resources for physical activity and healthy foods and their association with insulin resistance Epidemiology 2008, 19(1): 146-157. 14.Stimpson JP, Nash AC Ju H, Eschbach K: Neighborhood deprivation is associated with lower levels of serum carotenoids among adults participating in the Third National Health and Nutrition Examination Survey J Am Diet Assoc 2007, 107(11): 1895-1902. 15.Miller DK, Malmstrom TK, Joshi S, Andresen EM, Morley JE, Wolinsky FD: Clinically relevant levels of depressive symptoms in communitydwelling middle-aged African Americans J Am Geriatr Soc 2004, 52(5): 741-748. 16.Nagi S: An epidemiology of disability among adults in the United States Milbank Q 1976, 54: 439-467. 17.Krause N: Neighborhood deterioration, religious coping, and changes in health during late life Gerontologist 1998, 38(6): 653-664. 18.Andresen E, Malmstrom T, Miller D, Wolinsky F: Reliability and validity of observer ratings of neighborhoods J Aging Health 2006, 18: 28-36. 19.Chandola T: The fear of crime and area differences in health Health Place 2001, 7: 105-116. 20.Andresen EM, Malmstrom TK, Miller DK, Miller JP, Wolinsky FD: Retest reliability of self-reported function, self-care, and disease history Med Care 2005, 43(1): 93-97. 21.Andersson T, Alfredsson L, Kallbe rg H, Zdravkovic S, Ahlbom A: Calculating measures of biological interaction Eur J Epidemiol 2005, 20(7): 575-579. 22.Rothman KJ, Greenland S: Modern Epidemiology 2nd edition. Philadelphia, PA: Lippincott-Raven; 1998. 23.Rubin DB: Estimating causal effects from large data sets using propensity scores Ann Intern Med 1997, 127(8 Pt 2): 757-763. 24.Cepeda MS, Boston R, Farrar JT, Strom BL: Comparison of logistic regression versus propensity score when the number of events is low and there are multiple confounders Am J Epidemiol 2003, 158(3): 280-287. 25.Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR: A simulation study of the number of events per variable in logistic regression analysis J Clin Epidemiol 1996, 49(12): 1373-1379. 26.D'Agostino RB Jr: Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group Stat Med 1998, 17(19): 2265-2281. 27.Diez Roux AV, Borrell LN, Haan M, Jackson SA, Schultz R: Neighbourhood environments and mortality in an elderly cohort: results from the cardiovascular health study J Epidemiol Community Health 2004, 58(11): 917-923. 28.Imbens GW: The role of the propensity score in estimating doseresponse functions Biometrika 2000, 87(3): 706-710. 29.Smith KR, Waitzman NJ: Double jeopardy: interaction effects of marital and poverty status on the risk of mortality Demography 1994, 31(3): 487-507. 30.Ghanassia E, Villon L, Thuan dit Di eudonne J-F, Boegner C, Avignon A, Sultan A: Long term outcome and disability of diabetic patients hospitalised for diabetic foot ulcers: a 6.5 year follow-up study Diabetes Care 2008, 31: 1288-1292. 31.Egede LE, Dagogo-Jack S: Epidemiology of type 2 diabetes: focus on ethnic minorities Med Clin North Am 2005, 89(5): 949-975. viii Pre-publication history The pre-publication history for th is paper can be accessed here: http://www.biomedcentral.com/1471-2458/10/283/prepub doi: 10.1186/1471-2458-10-283 Cite this article as: Schootman et al. Neighborhood conditions, diabetes, and risk of lower-body functional limitations among middle-aged African Americans: A cohort study BMC Public Health 2010, 10 :283Received: 9 November 2009 Accepted: 27 May 2010 Published: 27 May 2010This article is available from: http: //www.biomedcentral.com/1471-2458/10/283 2010 Schootman et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons. org/licenses/by/2.0 ), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. B MC Public Health 2010, 10 :283

PAGE 8

1 Additional file 1 Unadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio an d attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower body functional limitation at 3 year follow up (weighted n=563). Good excellent block face conditions Fair poor block face conditions Without diabetes With diabetes Without diabetes With diabetes Type of condition Prevalence (%) of fair/poor quality OR OR 95% CI OR 95% CI OR 95% CI Interaction c ontrast ratio 95% CI Attributable proportion 95% CI Housing conditions 24.7 1.00* 1.42 0.67 2.97 1.37 0.75 2. 61 7.20 3.98 12.88 5.28 2.57 10.44 0.75 0.52 0.90 Noise 21.0 1.00* 1.48 0.76 3.04 1.39 0.76 2.61 7.91 4.64 14.52 5.92 2.93 5.92 0.76 0.52 0.89 Air quality 18.3 1.00* 2.01 1.09 3.75 2.37 1.30 4.67 13.82 6.74 38.17 10.15 3.65 10.15 0 .75 0.47 0.92 Street & road quality 20.9 1.00* 1.54 0.78 3.12 2.35 1.17 4.56 14.19 7.37 30.56 10.99 5.12 25.99 0.79 0. 60 0.91

PAGE 9

2 Yard & sidewalk quality 24.6 1.00* 1.08 0.53 2.34 1.20 0.70 2.21 9.52 5.34 17.73 8.12 4.40 15.49 0.86 0. 69 0.9 5 *Referent odds ratio

PAGE 10

1 Additional file 2 Propensity score adjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction co ntrast ratio and attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower body functional limitation at 3 year follow up (weighted n=563) Good excellent block face conditions Fair poor block face conditions Without diabetes With diabetes Without diabetes With diabetes Type of condition OR OR 95% CI OR 95% CI OR 95% CI Interaction contrast ratio 95% CI Attributable proportion 95% CI Housing conditions 1.00* 1. 15 0. 5 7 2. 22 1. 46 0.7 0 3.05 7. 79 1.36 37.55 6.01 0.46, 35.22 0. 80 0.33, 0.96 Noise 1.00* 1. 13 0. 5 6 2.3 0 1. 74 0.87 3.47 144.60 4. 45 775.53 143.0 2. 64, 772.79 0. 99 0.52 1.00 Air quality 1.00* 1.53 0.68 3. 14 2. 84 1.3 4 6.17 23.65 3.73 93.68 19.62 0.00, 89.27 0. 86 0 .00, 0.97 Street & road quality 1.00* 1. 1 4 0. 56 2.39 3.39 1. 24 8.00 79.53 5.85 347.6 7 76.18 2.18, 345.68 0. 95 0. 36, 0.99

PAGE 11

2 Yard & sidewalk quality 1.00* 1.0 1 0.53 2. 00 1. 39 0. 64 3.02 20.79 2.64, 82.37 19.09 1.04, 80.45 0. 93 0 .33, 0.99 Variables included in the propensity score: sampling stratum, age, gender, income, perceived income adequacy, educational attainment, marital status, employment status, number of persons in household, health care insurance, not being able to see a doctor because of c ost, s ocial support, self rated health status, depressive symptoms, a count of the number of chronic conditions, body mass index, risk of alcohol abuse, and physical activity. Referent odds ratio

PAGE 12

1 Additional File 3 Unadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio an d attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of i ncident lower body functional limitation at 3 year follow up (weighted n=563). Good excellent housing conditions Fair poor housing conditions Without diabetes With diabetes Without diabetes W ith diabetes Type of condition Prevalence (%) of fair/poor quality OR OR 95% CI OR 95% CI OR 95% CI Interaction c ontrast ratio 95% CI Attributable proportion 95% CI Cleanliness inside building 21.1 1.00* 2.04 1.14, 3.92 0.98 0.50, 2.1 6 1.94 1.01, 4.14 0.06 2.10, 1.66 0.04 1.37, 0.56 Physical conditions 19.5 1.00* 2.07 1.13, 3.81 1.02 0.45, 2.26 2.00 1.03, 3.93 0.06 2.30, 1.57 0.03 1.44, 0.56

PAGE 13

2 interior Furnishings 22.6 1.00* 2.12 1.15, 4.11 1.54 0.75, 3.06 2.88 1.36, 5.6 9 0.12 2.36, 2.44 0.04 1.13, 0.57 Outside condition 20.6 1.00* 2.09 1.12, 3.92 1.20 0.62, 2.40 2.32 1.06, 4.66 0.05 2.14, 1.85 0.02 1.14, 0.55 Overall condition 19.6 1.00* 1.98 1.05, 3.72 1.30 0.70, 2.63 3.04 1.40, 6.72 0.72 1.73, 3.69 0.2 6 0.90, 0.70 *Referent odds ratio

PAGE 14

1 Additional File 4 Propensity score adjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction co ntrast ratio and attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of incident lower body functional limitation at 3 year follow up (weighted n=563). Good excellent housing conditions Fair poor housing conditions Without diabetes With diabetes Without diabetes With diabetes Type of co ndition OR OR 95% CI OR 95% CI OR 95% CI Interaction contrast ratio 95% CI Attributable proportion 95% CI Cleanliness inside building 1.00* 1. 39 0 .71, 2.87 0.85 0.36, 2.14 0.59 0.07, 4.92 0.57 2.65, 3.63 1.08 18.5, 0.82 Physical conditions interior 1.00* 1.32 0.65, 2.47 1.27 0.36, 3.22 0.64 0.08, 5.76 0.76 3.31, 4.23 1.33 22.93, 0.82 Furnishings 1.00* 1.36 0.67, 2.77 1.35 0.47, 3.75 1.38 0.27, 4.51 0.31 3.23, 2.61 0.23 5.84, 0.76

PAGE 15

2 Outside condition 1.00* 1.28 0.59, 2.75 1.28 0.54, 2.91 0. 49 0.08, 5.43 0.95 3.52, 3.42 2.06 20.74, 0.79 Overall condition 1.00* 1.24 0.62, 2.49 1.59 0.66, 3.47 2.46 0.27, 11.67 0.67 2.35, 9.76 0.28 6.59, 0.89 Variables included in the propensity score: sampling stratum, age, gender, income, per ceived income adequacy, educational attainment, marital status, employment status, number of persons in household, health care insurance, not being able to see a doctor because of c ost, s ocial support, self rated health status, depressive symptoms, a count of the number of chronic conditions, body mass index, risk of alcohol abuse, and physical activity Referent odds ratio


!DOCTYPE art SYSTEM 'http:www.biomedcentral.comxmlarticle.dtd'
ui 1471-2458-10-283
ji 1471-2458
fm
dochead Research article
bibl
title
p Neighborhood conditions, diabetes, and risk of lower-body functional limitations among middle-aged African Americans: A cohort study
aug
au ca yes id A1
snm Schootman
fnm Mario
insr iid I1
email mschootm@im.wustl.edu
A2
Andresen
mi M
Elena
I2
andresen@phhp.ufl.edu
A3
Wolinsky
D
Fredric
I3
fredric-wolinsky@uiowa.edu
A4
Miller
J Philip
I4
JPhilipMiller@Wustl.edu
A5
Yan
Yan
I5
yany@wudosis.wustl.edu
A6
Miller
K
Douglas
I6
dokmille@iupui.edu
insg
ins
Department of Medicine and Pediatrics, Washington University School of Medicine, St. Louis, MO, USA
Department of Epidemiology and Biostatistics, College of Public Health and Health Professions, University of Florida, Gainesville, FL, USA
Department of Health Management and Policy, College of Public Health, the University of Iowa, IA, USA
Division of Biostatistics, Washington University School of Medicine, St. Louis, MO, USA
Department of Surgery, Washington University School of Medicine, St. Louis, MO, USA
Indiana University Center for Aging Research and Regenstrief Institute, Inc., Indiana University School of Medicine, Indianapolis, IN, USA
source BMC Public Health
issn 1471-2458
pubdate 2010
volume 10
issue 1
fpage 283
url http://www.biomedcentral.com/1471-2458/10/283
xrefbib
pubidlist
pubid idtype doi 10.1186/1471-2458-10-283
pmpid 20507573
history
rec
date
day 9
month 11
year 2009
acc
27
5
2010
pub
27
5
2010
cpyrt
2010
collab Schootman et al; licensee BioMed Central Ltd.
note This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
abs
sec
st
Abstract
Background
The relationship between presence of diabetes and adverse neighborhood and housing conditions and their effect on functional decline is unclear. We examined the association of adverse neighborhood (block face) and housing conditions with incidence of lower-body functional limitations among persons with and those without diabetes using a prospective population-based cohort study of 563 African Americans 49-65 years of age at their 2000-2001 baseline interviews.
Methods
Participants were randomly sampled African Americans living in the St. Louis area (response rate: 76%). Physician-diagnosed diabetes was self reported at baseline interview. Lower-body functional limitations were self reported based on the Nagi physical performance scale at baseline and the three-year follow-up interviews. The external appearance of the block the respondent lived on and five housing conditions were rated by study interviewers. All analyses were done using propensity score methods to control for confounders.
Results
109 (19.4%) of subjects experienced incident lower-body functional limitations at three-year follow-up. In adjusted analysis, persons with diabetes who lived on block faces rated as fair-poor on each of the five conditions had higher odds (7.79 [95% confidence interval: 1.36-37.55] to 144.6 [95% confidence interval: 4.45-775.53]) of developing lower-body functional limitations than the referent group of persons without diabetes who lived on block faces rated as good-excellent. At least 80 percent of incident lower-body functional limitations was attributable to the interaction between block face conditions and diabetes status.
Conclusions
Adverse neighborhood conditions appear to exacerbate the detrimental effects on lower-body functioning associated with diabetes.
meta
classifications
classification endnote subtype user_supplied_xml type bmc
bdy
Background
Diabetes has many adverse health effects, but one of the most important with serious consequences for living independently in the community is the impact of diabetes on lower-body function abbrgrp abbr bid B1 1. Poor lower-body function plays a crucial role in the disablement process and has been associated with increased disability days, physician contacts, fear of falling, falls, hip fracture, depression, nursing home placement, and mortality B2 2B3 3. Not surprisingly, diabetes is a major risk factor for the development of lower-body-related functional limitations B4 4, and a substantial portion of population (adult) disability related to lower-extremity functioning is attributable to diabetes B5 5. In the United States lower-body functional limitations are especially high among urban African Americans B6 6. For these reasons, the issue of the effect of diabetes on lower-body functioning is particularly important for African Americans.
Neighborhood conditions have been shown to predict incident lower-body functional limitations B7 7B8 8 and adverse housing conditions have been associated with incident diabetes B9 9. There are many potential pathways by which adverse neighborhood conditions might increase the risk of lower-body functional limitations, including increased stress, lower access to medical care, higher social isolation, and lower collective efficacy and social capital B10 10B11 11B12 12. Based on the above-described associations, we hypothesized that adverse neighborhood and housing conditions may exacerbate the detrimental effects on lower-body functioning associated with diabetes. To our knowledge no studies have examined the potential interaction between diabetes and contextual factors such as neighborhood and housing conditions on the development of lower-body functional limitations. Adverse neighborhood and housing conditions may exacerbate the progression toward lower-body functional limitations among persons with diabetes as a result of complications (e.g., neuropathy, amputations), associated conditions (e.g., hypertension, obesity, stroke), low fruit and vegetable consumption and low physical activity, and reduced muscle strength 1, many of which are also associated with adverse neighborhood and housing conditions B13 13B14 14. Therefore, we examined the association of adverse neighborhood (measured at the block face level) and housing conditions with incidence of lower-body dysfunction among persons with and those without diabetes using a longitudinal study of African Americans.
Methods
Baseline sample (wave 1)
The sampling design of the African American Health cohort has been described elsewhere B15 15. Briefly, the African American Health study is a population-based cohort study of 998 noninstitutionalized African Americans recruited in 2000-2001 using multi-stage probability sampling. All subjects lived in one of two geographic sampling strata: either a poor, inner-city area (St. Louis), Missouri, United States) or more heterogeneous suburbs just northwest of the City of St. Louis. Interviewers (two thirds of whom were African American) with extensive study-specific training screened households for eligibility criteria, which involved self-reported black or African American race, birth date during 1936 through 1950, and Mini-mental Status Examination scores 16. Subjects were paid volunteers. Sampling proportions were set to recruit approximately equal numbers of subjects from both areas (sampling strata). Each subject was weighted based on the selection probability and the response rate. When these weights are applied, the African American Health study sample represents the noninstitutionalized African American population in the two areas as of the 2000 Census./p
pAll subjects received in-home, baseline evaluations (average = 2.5 hours) between September 2000 and July 2001. Baseline response was 76% (9981320). All procedures were approved by the Institutional Review Boards of Saint Louis University and the University of Michigan./p
/sec
sec
st
pFollow-up sample (wave 4)/p
/st
pFollow-up in-home interviews averaging 1.5 hours were conducted 36 months after baseline assessments. Of the 998 persons who participated at baseline, 853 were successfully interviewed at follow-up. Since 51 persons had died between baseline and follow-up, the response for surviving subjects was 90.1% (853947). No attrition bias during waves 1 through 4 was evident for any of the major variables involved in the current analysis. A total of 290 persons (weighted) had two or more lower body functional limitations at baseline and were excluded from further analysis. Thus, 563 persons (weighted) had one or fewer lower-body functional limitations at baseline and comprised the study sample./p
/sec
sec
st
pLower-body functional limitations/p
/st
pFive items (0 = no difficulties to 1 = difficulty) from the Nagi physical performance scale assessed lower-body functional limitations, which were summed to form the outcome measure (ranging from 0 to 5) abbrgrpabbr bid="B6"6/abbr/abbrgrp. Items included difficulties in walking a quarter of a mile; walking up and down 10 steps without rest; standing for 2 hours; stooping, crouching, or kneeling; and lifting and carrying 10 pounds abbrgrpabbr bid="B16"16/abbr/abbrgrp. Subjects who expressed any difficulty or inability to perform the functional task at the time of the interview were considered to be limited in that task. Similar to other studies abbrgrpabbr bid="B7"7/abbr/abbrgrp, we limited subjects in this study to those with one or fewer lower-body functional limitations at baseline in order to examine the risk of developing two or more lower-body functional limitations three years later. At follow-up, we defined incident lower-body functional limitations as reporting difficulty or being unable to perform at least two of the five physical tasks among those with one or fewer lower-body functional limitations at baseline./p
/sec
sec
st
pAdverse neighborhood and housing conditions/p
/st
pAssessment of neighborhood conditions was comprised of interviewer observations of the block face on which the respondent lived and participants' self-reported neighborhood desirability. An "objective" assessment of the external appearance of the block face (neighborhood) in front of the homes where the participants resided was done by the survey team using a previously published assessment tool abbrgrpabbr bid="B17"17/abbr/abbrgrp during household enumeration, which occurred an average of seven months before the participants were recruited and data were collected. Thus, block face conditions were collected independently from the interview data. Data about the housing conditions were collected at the time of the in-home interview and therefore not independently. On four-point scales (1 = excellent, 4 = poor) observers rated each of five characteristics: condition of houses, amount of noise (from traffic, industry, etc.), air quality, condition of the streets, and condition of the yards and sidewalks in front of homes where the participant lived. Weighted inter-rater Kappa statistics ranged from 0.58 (air quality) to 0.84 (condition of yards and sidewalks abbrgrpabbr bid="B18"18/abbr/abbrgrp./p
pWe also obtained a subjective measure of neighborhood conditions from respondents at baseline using a four-item scale of the neighborhood as a place to live, general feelings about the neighborhood, attachment to the neighborhood, and neighborhood safety from crime abbrgrpabbr bid="B19"19/abbr/abbrgrp. Participant responses were dichotomized for each condition, and the scale ranged from 0 to 4./p
pAssessment of housing conditions was an observed five-item scale based on the interviewer's ratings at the baseline interview of the cleanliness inside the building; physical condition of the interior; condition of furnishings; condition of the exterior of the building; and a global rating (all rated as excellent, good, fair, or poor). The test-retest reliability was at least 0.68 for each condition abbrgrpabbr bid="B9"9/abbr/abbrgrp. In the present analysis, each block face condition and each housing condition was dichotomized as either fair or poor versus good or excellent./p
/sec
sec
st
pDiabetes/p
/st
pThe baseline interview asked respondents about the presence of physician-diagnosed diabetes (test-retest reliability in a subsample of African American Health study participants was 0.94) abbrgrpabbr bid="B20"20/abbr/abbrgrp./p
/sec
sec
st
pCovariates/p
/st
pBaseline covariates included in the analysis were patterned after other African American Health cohort research abbrgrpabbr bid="B7"7/abbr/abbrgrp. Socio-demographic variables involved sampling stratum (inner city, suburb), age, gender, income categories, perceived income adequacy (having a comfortable income, having just enough to get by, not having enough to get by), educational attainment, marital status, employment status, number of persons in household, having health care insurance at the time of or during the 12 months prior to interview, and not being able to see a doctor because of cost during 12 months prior to interview. Social support was measured using five items from the Medical Outcomes Study social support instrument./p
pHealth at baseline was measured by the self-rated health status question of the Short-form 36, depressive symptoms (score of at least 9 using the 11-item Center for Epidemiology Depressive Symptoms scale), a count of the number of self-reported physician-diagnosed severe chronic conditions ever experienced (asthma, chronic airway obstruction, heart failure, heart attack, angina, stroke, chronic kidney disease, arthritis, and cancer other than a minor skin cancer). Also assessed at baseline were body mass index, current smoking status, risk of alcohol abuse (CAGE), the Yale physical activity instrument, and grip strength./p
pThe conceptual model of the relationship among the various types of variables is displayed in Figure figr fid="F1"1/figr. The association of interest is the relationship between diabetes and development of lower-body functional limitations as modified by adverse block facehousing conditions. Diabetes and the propensity score are predicting lower-body functional limitations. The association of diabetes with lower-body functional limitations is modified by adverse block facehousing conditions. The inclusion of the propensity score is aimed at estimating the unbiased association of interest./p
fig id="F1"
title
pFigure 1/p
/title
caption
pConceptual Model/p
/caption
text
pbConceptual Model/b./p
/text
graphic file="1471-2458-10-283-1" hint_layout="single"/
/fig
/sec
sec
st
pStatistical analysis/p
/st
pFirst, we calculated unadjusted measures of association (odds ratio [OR] and 95% confidence intervals [CI]) and interaction between diabetes and block facehousing conditions for the risk of incident lower-body functional limitation at 3-year follow-up. For each block face and housing condition, a single variable with four categories was created by combining a dichotomous neighborhoodhousing variable (excellentgood vs. fairpoor) with the diabetes variable as part of the additive model abbrgrpabbr bid="B21"21/abbr/abbrgrp. Measures of interaction included the Interaction Contrast Ratio and the Attributable Proportion as measures of departure of additivity. The Interaction Contrast Ratio is the excess risk due to interaction relative to the risk without exposures abbrgrpabbr bid="B22"22/abbr/abbrgrp. If there is no superadditive interaction (null hypothesis), the Interaction Contrast Ratio equals 0. Interaction Contrast Ratio 0 indicates superadditivity, and Interaction Contrast Ratio < 0 indicates subadditivity. Odds ratios were substituted for the relative risks in the Interaction Contrast Ratio calculation. Attributable Proportion refers to the proportion of incident lower-body functional limitations attributable to the interaction among persons who experienced both exposures. If there is no interaction, the Attributable Proportion will equal 0.p
pTo adjust for potential confounding from covariates, we used the propensity score method to obtain adjusted estimates of the measures of association and interaction between neighborhoodhousing conditions with presence of diabetes on incident lower-body functional limitations abbrgrpabbr bid="B23"23abbrabbr bid="B24"24abbrabbrgrp. All variables were included in the calculation of propensity scores. The propensity score is defined as the conditional probability of a person living under a certain neighborhoodhousing condition and diabetes status, given covariates included. Multivariable logistic regression may be limited in its ability to control for confounders in studies when there are fewer than 10 events per variable analyzed abbrgrpabbr bid="B25"25abbrabbrgrp. The use of propensity scores has been proposed as an alternative that may be especially useful when multiple confounders are involved abbrgrpabbr bid="B26"26abbrabbr bid="B27"27abbrabbrgrp. Propensity score methods produce estimates that are more accurate than logistic regression estimates when there are seven or fewer events per confounder, as was the case in the present study abbrgrpabbr bid="B24"24abbrabbrgrp. Similar to multivariable methods, unmeasured confounders are not included in the construction of propensity scores. Following previous work abbrgrpabbr bid="B28"28abbrabbrgrp, we first estimated the probability in each of four categories of the interaction (diabetes YesNo by block facehousing condition FairPoor vs. GoodExcellent) for each individual using a nominal multinomial logistic regression model. Next, we assigned a weight using the inversed predicted probability that the individual was observed in one of the categories. The propensity score-adjusted estimates incorporated these weights in the analyses. Then, we obtained the interaction indices using the log odds ratio estimates from the logistic regression model. Also, we examined the interaction indices for hypertension and arthritis with the contextual variable to challenge the robustness of the findings.p
pConfidence intervals for all effect measures were calculated using the bootstrap percentile methodology. For a 95% confidence interval the 2.5supth supand 97.5supth suppercentiles of the empirical distribution of Interaction Contrast Ratios were calculated from 1,000 data sets resampled with replacement from the original data. All statistical analyses were conducted using SAS software, version 9.1.p
sec
sec
sec
st
pResultsp
st
pOf 563 subjects with zero or one lower-body functional limitations at baseline, 109 (19.4%) experienced two or more lower-body functional limitations at the 3-year follow-up. Eighteen percent of subjects reported having diabetes at baseline and 82 percent did not. Of those with diabetes, the percentage of subjects who lived on one of the ten block face or housing conditions rated as fair or poor ranged from 10.3% to 23.3%. Of those without diabetes, the percentage of subjects who lived on one of the ten block face or housing conditions rated as fair or poor ranged from 17.2% to 20.2%. There were no statistical differences in block face or housing conditions between persons with and those without diabetes. Baseline characteristics of the study population and factors associated with incident lower-body functional limitations in univariate analysis have been described briefly in Table tblr tid="T1"1tblr and more extensively elsewhere abbrgrpabbr bid="B7"7abbrabbrgrp. Briefly, persons who were older, unable to visit a doctor because of the cost, scored nine or more on the Center for Epidemiology Depressive Symptoms 11-item scale, experienced greater number of severe chronic conditions, or had one lower-body functional limitations at baseline were more likely to experience incident lower-body functional limitations at 3-year follow-up. Persons were less likely to have incident lower-body functional limitations at follow-up when they had lived more than five years at the present address or were overweight at baseline.p
tbl id="T1"
title
pTable 1p
title
caption
pPrevalence of selected characteristics at baseline and unadjusted risk of three-year incident lower body functional limitation for subjects in the African-American Health studyp
caption
tblbdy cols="4"
r
c
p
c
c
p
c
c cspan="2" ca="center"
p
bUnadjusted risk of incident lower body functional limitation at 3-year follow-upb
p
c
r
r
c cspan="4"
hr
c
r
r
c
p
c
c ca="left"
p
bBaseline measure**b
p
p
b(weighted n = 563)b
p
c
c ca="center"
p
bOdds ratiob
p
c
c ca="center"
p
b95% CIb
p
c
r
r
c cspan="4"
hr
c
r
r
c ca="left"
pAge (mean[s.d])p
c
c ca="left"
p56.1 (4.7)p
c
c ca="center"
p1.06p
c
c ca="center"
p1.01 1.11p
c
r
r
c ca="left"
pGenderp
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
pWomen vs. Menp
c
c ca="left"
p54.6%p
c
c ca="center"
p1.46p
c
c ca="center"
p0.95 2.24p
c
r
r
c ca="left"
pLength of time at present addressp
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
pMore than 5 yrs vs. Less than 5 yearsp
c
c ca="left"
p73.1%p
c
c ca="center"
p0.83p
c
c ca="center"
p0.34 0.82p
c
r
r
c ca="left"
pObjective incomep
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
p< $20,000 vs. = $50,000p
c
c ca="left"
p17.4%p
c
c ca="center"
p1.68p
c
c ca="center"
p0.95 2.94p
c
r
r
c indent="1" ca="left"
p$20,000 < $50,000 vs. = $50,000p
c
c ca="left"
p48.8%p
c
c ca="center"
p1.32p
c
c ca="center"
p0.29 6.07p
c
r
r
c ca="left"
pHighest level of educationp
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
p< 12 years vs. 12 years or morep
c
c ca="left"
p21.0%p
c
c ca="center"
p0.74p
c
c ca="center"
p0.43 1.27p
c
r
r
c ca="left"
pUnable to visit doctor because of costp
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
pYes vs. Nop
c
c ca="left"
p6.4%p
c
c ca="center"
p2.35p
c
c ca="center"
p1.14 4.83p
c
r
r
c ca="left"
pCenter for Epidemiology Depressive Symptoms 11-item score ≥9p
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
pYes vs. Nop
c
c ca="left"
p12.5%p
c
c ca="center"
p1.89p
c
c ca="center"
p1.08 3.32p
c
r
r
c ca="left"
pNo. of severe chronic conditions (per condition)p
c
c ca="left"
p0.8 (1.0)p
c
c ca="center"
p1.56p
c
c ca="center"
p1.28 1.90p
c
r
r
c ca="left"
pLower body limitation at baselinep
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
pOne vs. Nonep
c
c ca="left"
p29.2%p
c
c ca="center"
p3.56p
c
c ca="center"
p2.30 5.49p
c
r
r
c ca="left"
pBody Mass Indexp
c
c
p
c
c
p
c
c
p
c
r
r
c indent="1" ca="left"
p= 30.0 vs. < 25.0p
c
c ca="left"
p35.5%p
c
c ca="center"
p0.61p
c
c ca="center"
p0.36 1.04p
c
r
r
c indent="1" ca="left"
p25.0 29.9 vs. < 25.0p
c
c ca="left"
p40.7%p
c
c ca="center"
p0.48p
c
c ca="center"
p0.28 0.81p
c
r
tblbdy
tblfn
p* CI: confidence interval, s.d.: standard deviation.p
p** Mean (SD), unless noted.p
tblfn
tbl
sec
st
pIncidencep
st
pOf 563 subjects with zero or one lower-body functional limitations at baseline, the percentage that developed two or more lower-body functional limitations at 3-year follow-up varied according to the participant's diabetes status and each of the five block-face conditions. For example, 65% percent of persons with diabetes who lived on block faces with yards and sidewalks in fair-poor condition developed lower-body functional limitations. In contrast, 19.4 percent of persons without diabetes who lived on block faces with yards and sidewalks in fair-poor condition developed lower-body functional limitations. About 17 percent of persons who lived on block faces with good-excellent conditions (regardless of diabetes status) developed lower-body functional limitations. Similar results were observed for the other four block-face conditions. Little difference was present examining each of the five housing conditions.p
sec
sec
st
pInteraction between block face conditions and diabetes statusp
st
pIn unadjusted analysis, (a) persons with diabetes residing on block faces with good-excellent conditions and (b) persons without diabetes living on block faces with fair-poor conditions generally were not significantly more likely to develop lower-body functional limitations three years later than the referent group of persons without diabetes who lived on block faces rated as good or excellent (additional File supplr sid="S1"1supplr). The only exceptions were for the rating of air quality for both groups and street and road quality for the second group, for which the odds ratios were 2.0 to 2.4. In contrast, persons with diabetes who lived on block faces rated as fair or poor on each of the five conditions had seven to 14 higher odds of developing lower-body functional limitations than the referent group. An interaction existed between block face condition and presence of diabetes for each of the five conditions (all Interaction Contrast Ratio 1.0). At least 75 percent of the incidence of lower-body functional limitations was attributable to this interaction for each of the block face conditions (additional File supplr sid="S1"1supplr).p
suppl id="S1"
title
pAdditional file 1p
title
text
p
bUnadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio and attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower-body functional limitation at 3-year follow-up (weighted n = 563).b
p
text
file name="1471-2458-10-283-S1.DOC"
pClick here for filep
file
suppl
pIn adjusted analyses, we observed parameter estimates that generally were larger than those in the unadjusted analysis for all conditions (additional File supplr sid="S2"2supplr). The Attributable Proportion of the incidence of lower-body functional limitations due to the interaction involving the other four conditions was at least 80 percent. The values of the Interaction Contrast Ratios indicate that the excess risk due to the interaction was large relative to the risk without either exposure.p
suppl id="S2"
title
pAdditional file 2p
title
text
p
bPropensity score adjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio and attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower-body functional limitation at 3-year follow-up (weighted n = 563).*b
p
text
file name="1471-2458-10-283-S2.DOC"
pClick here for filep
file
suppl
sec
sec
st
pInteraction between housing conditions and diabetes statusp
st
pIn unadjusted analysis, persons with diabetes who lived under housing conditions rated as good or excellent and those rated as fair or poor had about two times higher odds of developing lower-body functional limitations at 3-year follow-up compared to the reference group (additional File supplr sid="S3"3supplr). However, based on the Interaction Contrast Ratio and the Attributable Proportion there was no interaction between housing condition and diabetes. In adjusted analysis, there was also no interaction between any of the housing conditions and diabetes, and the associations with the two types of exposures on their own became insignificant (additional File supplr sid="S4"4supplr).p
suppl id="S3"
title
pAdditional file 3p
title
text
p
bUnadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio and attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of incident lower-body functional limitation at 3-year follow-up (weighted n = 563).b
p
text
file name="1471-2458-10-283-S3.DOC"
pClick here for filep
file
suppl
suppl id="S4"
title
pAdditional file 4p
title
text
p
bPropensity score adjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio and attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of incident lower-body functional limitation at 3-year follow-up (weighted n = 563).b
p
text
file name="1471-2458-10-283-S4.DOC"
pClick here for filep
file
suppl
pSensitivity analyses were performed with hypertension and arthritis to examine the specificity of the interaction between diabetes and the contextual variables. However, no interaction was found for hypertension and arthritis (data not shown), suggesting the observed interaction is specific to diabetes and neighborhood conditions.p
sec
sec
sec
st
pDiscussion and Conclusionsp
st
pThis study has demonstrated the importance of the interaction of diabetes and neighborhood conditions acting in concert on the deterioration of lower-body functional limitations in an urban African American population. This "double jeopardy" suggests that both risk factors combined increased the risk of lower-body functional limitations considerably more than diabetes or adverse block face conditions alone abbrgrpabbr bid="B29"29abbrabbrgrp. In contrast, we found no evidence of an interaction between diabetes and housing conditions on incident lower-body functional limitations. To our knowledge, ours is the first study that shows a powerful synergy between neighbor conditions and presence of disease on subsequent adverse functional outcome.p
pAdverse neighborhood conditions may exacerbate the progression toward lower-body dysfunction among persons with diabetes as a result of diabetic complications (e.g., heart disease, visual impairment, neuropathy, ulceration, lower extremity amputation), associated conditions (e.g., obesity, hypertension), poor diet and low physical activity, and reduced muscle strength abbrgrpabbr bid="B1"1abbrabbrgrp. We adjusted for these risk factors by including them in the propensity score, except for lower-extremity disorders such as peripheral neuropathy, foot ulcers, peripheral arterial disease, and lower-extremity amputation, for which we had little data. These lower-extremity conditions have been associated with numbness in the extremities, and trouble with gait and balance, and lower-body functional limitations abbrgrpabbr bid="B30"30abbrabbrgrp. It is possible that adverse neighborhood conditions accelerate the decline toward lower-body functional limitations through lower-extremity conditions in persons with diabetes. However, discerning the meditational pathways was not the purpose of our study. In efforts to prevent future lower-body dysfunction among persons with diabetes by intervening upon lower-extremity disorders, attention to environmental circumstances (especially block face conditions) and the individual's interaction with them needs to be part of the interventional strategy. The findings appear robust with respect to sensitivity analysis.p
pStudy limitations include a study sample that involves a single race, a single city, and a restricted age range, all of which may limit generalizability. However, focusing on a single race allows the disentanglement of race and socioeconomic status. Notably, African Americans experience more diabetes and more diabetic complications than does the majority population abbrgrpabbr bid="B31"31abbrabbrgrp, so the effect of the diabetes-neighborhood interaction may be particularly strong in African Americans. We used self-reported diabetes to classify cases of diabetes and thus some cases of prevalent diabetes probably were missed. Despite the very high test-retest reliability of self-reported diabetes in the African American Health data abbrgrpabbr bid="B20"20abbrabbrgrp, misclassification of diabetes status may still be present, which could lead to biased results. However, unless misclassification of self-reported diabetes was dependent upon block face condition, our results would likely be a conservative estimate of the true relationships. Another limitation is that we had in some instances only 24 persons with diabetes who lived in fair or poor block face conditions at baseline. The relatively small size of this group resulted in wide confidence intervals in some results, but unity was never included.p
pFinally, in most studies of neighborhood effects, multiple study participants are nested within their neighborhood, requiring the use of multilevel statistical techniques. In this study sample, there were 551 block faces, 363 on which only one participant resided (65.9%). Only 3.6 percent of block faces contained five or more participants. We were not able to use multilevel statistical techniques because there was not enough clustering of participants within block faces to support a robust multi-level analytic approach. In a previous study, we randomly selected one subject per block face from the block faces with more than 1 subject and showed that parameter estimates were very similar to our findings using propensity scores abbrgrpabbr bid="B7"7abbrabbrgrp.p
pIn summary, there appears to be a powerful interaction between adverse block face conditions and the presence of diabetes on decline is a crucial factor for maintaining health and independence (i.e., lower-body physical functioning) in urban-dwelling middle-aged African Americans. Further research is needed to investigate the mediators of this powerful interaction.p
sec
sec
st
pCompeting interestsp
st
pThe authors declare that they have no competing interests.p
sec
sec
st
pAuthors' contributionsp
st
pMS conceived of the study and drafted the manuscript, EMA, FDW, JPM participated in drafting of the manuscript, YY performed the statistical analysis, DKM helped conceive of the study and participated in drafting of the manuscript. All authors read and approved the final manuscript.p
sec
bdy
bm
ack
sec
st
pAcknowledgementsp
st
pThis work was supported by grants from the National Institutes of Health (grant numbers AG10436, DK067172). We thank Mr. James Struthers of the Washington University Diabetes Research and Training Center Prevention and Control Core (DK20579). We thank Drs. Jack M. Guralnik and Robert B. Wallace for reading an earlier version of this manuscript.p
sec
ack
refgrp
bibl id="B1"
title
pDisability and quality of life in elderly people with diabetesp
title
aug
au
snmBourdel-Marchassonsnm
fnmIfnm
au
au
snmHelmersnm
fnmCfnm
au
au
snmFagot-Campagnasnm
fnmAfnm
au
au
snmDehailsnm
fnmPfnm
au
au
snmJosephsnm
fnmPAfnm
au
aug
sourceDiabetes Metabolsource
pubdate2007pubdate
volume33volume
issueSupplement 1issue
fpageS66fpage
lpageS74lpage
xrefbib
pubid idtype="doi"10.1016S1262-3636(07)80058-9pubid
xrefbib
bibl
bibl id="B2"
title
pCross-sectional and longitudinal risk factors for falls, fear of falling, and falls efficacy in a cohort of middle-aged African Americansp
title
aug
au
snmAndresensnm
fnmEMfnm
au
au
snmWolinskysnm
fnmFDfnm
au
au
snmMillersnm
fnmJPfnm
au
au
snmWilsonsnm
fnmMMfnm
au
au
snmMalmstromsnm
fnmTKfnm
au
au
snmMillersnm
fnmDKfnm
au
aug
sourceGerontologistsource
pubdate2006pubdate
volume46volume
issue2issue
fpage249fpage
lpage257lpage
xrefbib
pubid idtype="pmpid" link="fulltext"16581889pubid
xrefbib
bibl
bibl id="B3"
title
pThe risk of nursing home placement and subsequent death among older adultsp
title
aug
au
snmWolinskysnm
fnmFDfnm
au
au
snmCallahansnm
fnmCMfnm
au
au
snmFitzgeraldsnm
fnmJFfnm
au
au
snmJohnsonsnm
fnmRJfnm
au
aug
sourceJ Gerontolsource
pubdate1992pubdate
volume47volume
issue4issue
fpage173fpage
lpage182lpage
bibl
bibl id="B4"
title
pProgression of lower-extremity disability in older women with diabetes: the women's health and aging studyp
title
aug
au
snmVolpatosnm
fnmSfnm
au
au
snmFerruccisnm
fnmLfnm
au
au
snmBlaumsnm
fnmCfnm
au
au
snmOstirsnm
fnmGfnm
au
au
snmCappolasnm
fnmAfnm
au
au
snmFriedsnm
fnmLPfnm
au
au
snmFellinsnm
fnmRfnm
au
au
snmGuralniksnm
fnmJMfnm
au
aug
sourceDiabetes Caresource
pubdate2003pubdate
volume26volume
issue1issue
fpage70fpage
lpage75lpage
xrefbib
pubidlist
pubid idtype="doi"10.2337diacare.26.1.70pubid
pubid idtype="pmpid" link="fulltext"12502660pubid
pubidlist
xrefbib
bibl
bibl id="B5"
title
pDeterminants of locomotor disability in people aged 55 years and over: The Rotterdam studyp
title
aug
au
snmOddingsnm
fnmEfnm
au
au
snmValkenburgsnm
fnmHAfnm
au
au
snmStamsnm
fnmHJfnm
au
au
snmHofmansnm
fnmAfnm
au
aug
sourceEur J Epidemiolsource
pubdate2001pubdate
volume17volume
issue11issue
fpage1033fpage
lpage1041lpage
xrefbib
pubidlist
pubid idtype="doi"10.1023A:1020006909285pubid
pubid idtype="pmpid" link="fulltext"12380718pubid
pubidlist
xrefbib
bibl
bibl id="B6"
title
pInner city, middle-aged African Americans have excess frank and subclinical disabilityp
title
aug
au
snmMillersnm
fnmDKfnm
au
au
snmWolinskysnm
fnmFDfnm
au
au
snmMalmstromsnm
fnmTKfnm
au
au
snmAndresensnm
fnmEMfnm
au
au
snmMillersnm
fnmJPfnm
au
aug
sourceJ Gerontol A Biol Sci Med Scisource
pubdate2005pubdate
volume60volume
issue2issue
fpage207fpage
lpage212lpage
xrefbib
pubid idtype="pmpid"15814864pubid
xrefbib
bibl
bibl id="B7"
title
pNeighborhood conditions and risk of incident lower-body functional limitations among middle-aged African Americansp
title
aug
au
snmSchootmansnm
fnmMfnm
au
au
snmAndresensnm
fnmEfnm
au
au
snmWolinskysnm
fnmFfnm
au
au
snmMalmstromsnm
fnmTfnm
au
au
snmMillersnm
fnmJfnm
au
au
snmMillersnm
fnmDfnm
au
aug
sourceAm J Epidemiolsource
pubdate2006pubdate
volume163volume
fpage450fpage
lpage458lpage
xrefbib
pubidlist
pubid idtype="doi"10.1093ajekwj054pubid
pubid idtype="pmpid" link="fulltext"16421245pubid
pubidlist
xrefbib
bibl
bibl id="B8"
title
pNeighborhood environment and loss of physical function in older adults: evidence from the Alameda County Studyp
title
aug
au
snmBalfoursnm
fnmJLfnm
au
au
snmKaplansnm
fnmGAfnm
au
aug
sourceAm J Epidemiolsource
pubdate2002pubdate
volume155volume
issue6issue
fpage507fpage
lpage515lpage
xrefbib
pubidlist
pubid idtype="doi"10.1093aje155.6.507pubid
pubid idtype="pmpid" link="fulltext"11882524pubid
pubidlist
xrefbib
bibl
bibl id="B9"
title
pThe effect of adverse housing and neighborhood conditions on the development of diabetes mellitus among middle-aged African Americansp
title
aug
au
snmSchootmansnm
fnmMfnm
au
au
snmAndresensnm
fnmEMfnm
au
au
snmWolinskysnm
fnmFDfnm
au
au
snmMalmstromsnm
fnmTKfnm
au
au
snmMillersnm
fnmJPfnm
au
au
snmYansnm
fnmYfnm
au
au
snmMillersnm
fnmDKfnm
au
aug
sourceAm J Epidemiolsource
pubdate2007pubdate
volume166volume
fpage379fpage
lpage387lpage
xrefbib
pubidlist
pubid idtype="doi"10.1093ajekwm190pubid
pubid idtype="pmpid" link="fulltext"17625220pubid
pubidlist
xrefbib
bibl
bibl id="B10"
title
pNeighborhood-level context and health: Lessons from sociologyp
title
aug
au
snmSampsonsnm
fnmRfnm
au
aug
sourceNeighborhoods and Healthsource
publisherOxford: Oxford University Presspublisher
editorKawachi I, Berkman LFeditor
pubdate2003pubdate
fpage132fpage
lpage146lpage
xrefbib
pubid idtype="doi"full_textpubid
xrefbib
bibl
bibl id="B11"
title
pNeighborhoods, aging, and functional limitationsp
title
aug
au
snmGlasssnm
fnmTfnm
au
au
snmBalfoursnm
fnmJfnm
au
aug
sourceNeighborhoods and Healthsource
publisherOxford: Oxford University Presspublisher
editorKawachi I, Berkman Leditor
pubdate2003pubdate
fpage303fpage
lpage334lpage
xrefbib
pubid idtype="doi"full_textpubid
xrefbib
bibl
bibl id="B12"
title
pSocial engagement and disability in a community population of older adults: The New Haven EPESEp
title
aug
au
snmMendesde Leonsnm
fnmCFfnm
au
au
snmGlasssnm
fnmTAfnm
au
au
snmBerkmansnm
fnmLFfnm
au
aug
sourceAm J Epidemiolsource
pubdate2003pubdate
volume157volume
issue7issue
fpage633fpage
lpage642lpage
xrefbib
pubidlist
pubid idtype="doi"10.1093ajekwg028pubid
pubid idtype="pmpid" link="fulltext"12672683pubid
pubidlist
xrefbib
bibl
bibl id="B13"
title
pNeighborhood resources for physical activity and healthy foods and their association with insulin resistancep
title
aug
au
snmAuchinclosssnm
fnmAHfnm
au
au
snmDiez Rouxsnm
fnmAVfnm
au
au
snmBrownsnm
fnmDGfnm
au
au
snmErdmannsnm
fnmCAfnm
au
au
snmBertonisnm
fnmAGfnm
au
aug
sourceEpidemiologysource
pubdate2008pubdate
volume19volume
issue1issue
fpage146fpage
lpage157lpage
xrefbib
pubidlist
pubid idtype="doi"10.1097EDE.0b013e31815c480pubid
pubid idtype="pmpid" link="fulltext"18091002pubid
pubidlist
xrefbib
bibl
bibl id="B14"
title
pNeighborhood deprivation is associated with lower levels of serum carotenoids among adults participating in the Third National Health and Nutrition Examination Surveyp
title
aug
au
snmStimpsonsnm
fnmJPfnm
au
au
snmNashsnm
fnmACfnm
au
au
snmJusnm
fnmHfnm
au
au
snmEschbachsnm
fnmKfnm
au
aug
sourceJ Am Diet Assocsource
pubdate2007pubdate
volume107volume
issue11issue
fpage1895fpage
lpage1902lpage
xrefbib
pubidlist
pubid idtype="doi"10.1016j.jada.2007.08.016pubid
pubid idtype="pmpid" link="fulltext"17964308pubid
pubidlist
xrefbib
bibl
bibl id="B15"
title
pClinically relevant levels of depressive symptoms in community-dwelling middle-aged African Americansp
title
aug
au
snmMillersnm
fnmDKfnm
au
au
snmMalmstromsnm
fnmTKfnm
au
au
snmJoshisnm
fnmSfnm
au
au
snmAndresensnm
fnmEMfnm
au
au
snmMorleysnm
fnmJEfnm
au
au
snmWolinskysnm
fnmFDfnm
au
aug
sourceJ Am Geriatr Socsource
pubdate2004pubdate
volume52volume
issue5issue
fpage741fpage
lpage748lpage
xrefbib
pubidlist
pubid idtype="doi"10.1111j.1532-5415.2004.52211.xpubid
pubid idtype="pmpid" link="fulltext"15086655pubid
pubidlist
xrefbib
bibl
bibl id="B16"
title
pAn epidemiology of disability among adults in the United Statesp
title
aug
au
snmNagisnm
fnmSfnm
au
aug
sourceMilbank Qsource
pubdate1976pubdate
volume54volume
fpage439fpage
lpage467lpage
bibl
bibl id="B17"
title
pNeighborhood deterioration, religious coping, and changes in health during late lifep
title
aug
au
snmKrausesnm
fnmNfnm
au
aug
sourceGerontologistsource
pubdate1998pubdate
volume38volume
issue6issue
fpage653fpage
lpage664lpage
xrefbib
pubid idtype="pmpid"9868846pubid
xrefbib
bibl
bibl id="B18"
title
pReliability and validity of observer ratings of neighborhoodsp
title
aug
au
snmAndresensnm
fnmEfnm
au
au
snmMalmstromsnm
fnmTfnm
au
au
snmMillersnm
fnmDfnm
au
au
snmWolinskysnm
fnmFfnm
au
aug
sourceJ Aging Healthsource
pubdate2006pubdate
volume18volume
fpage28fpage
lpage36lpage
xrefbib
pubidlist
pubid idtype="doi"10.11770898264305281001pubid
pubid idtype="pmpid" link="fulltext"16468180pubid
pubidlist
xrefbib
bibl
bibl id="B19"
title
pThe fear of crime and area differences in healthp
title
aug
au
snmChandolasnm
fnmTfnm
au
aug
sourceHealth Placesource
pubdate2001pubdate
volume7volume
fpage105fpage
lpage116lpage
xrefbib
pubidlist
pubid idtype="doi"10.1016S1353-8292(01)00002-8pubid
pubid idtype="pmpid" link="fulltext"11470224pubid
pubidlist
xrefbib
bibl
bibl id="B20"
title
pRetest reliability of self-reported function, self-care, and disease historyp
title
aug
au
snmAndresensnm
fnmEMfnm
au
au
snmMalmstromsnm
fnmTKfnm
au
au
snmMillersnm
fnmDKfnm
au
au
snmMillersnm
fnmJPfnm
au
au
snmWolinskysnm
fnmFDfnm
au
aug
sourceMed Caresource
pubdate2005pubdate
volume43volume
issue1issue
fpage93fpage
lpage97lpage
xrefbib
pubid idtype="pmpid" link="fulltext"15626939pubid
xrefbib
bibl
bibl id="B21"
title
pCalculating measures of biological interactionp
title
aug
au
snmAnderssonsnm
fnmTfnm
au
au
snmAlfredssonsnm
fnmLfnm
au
au
snmKallbergsnm
fnmHfnm
au
au
snmZdravkovicsnm
fnmSfnm
au
au
snmAhlbomsnm
fnmAfnm
au
aug
sourceEur J Epidemiolsource
pubdate2005pubdate
volume20volume
issue7issue
fpage575fpage
lpage579lpage
xrefbib
pubidlist
pubid idtype="doi"10.1007s10654-005-7835-xpubid
pubid idtype="pmpid" link="fulltext"16119429pubid
pubidlist
xrefbib
bibl
bibl id="B22"
aug
au
snmRothmansnm
fnmKJfnm
au
au
snmGreenlandsnm
fnmSfnm
au
aug
sourceModern Epidemiologysource
publisherPhiladelphia, PA: Lippincott-Ravenpublisher
edition2edition
pubdate1998pubdate
bibl
bibl id="B23"
title
pEstimating causal effects from large data sets using propensity scoresp
title
aug
au
snmRubinsnm
fnmDBfnm
au
aug
sourceAnn Intern Medsource
pubdate1997pubdate
volume127volume
issue8 Pt 2issue
fpage757fpage
lpage763lpage
xrefbib
pubid idtype="pmpid"9382394pubid
xrefbib
bibl
bibl id="B24"
title
pComparison of logistic regression versus propensity score when the number of events is low and there are multiple confoundersp
title
aug
au
snmCepedasnm
fnmMSfnm
au
au
snmBostonsnm
fnmRfnm
au
au
snmFarrarsnm
fnmJTfnm
au
au
snmStromsnm
fnmBLfnm
au
aug
sourceAm J Epidemiolsource
pubdate2003pubdate
volume158volume
issue3issue
fpage280fpage
lpage287lpage
xrefbib
pubidlist
pubid idtype="doi"10.1093ajekwg115pubid
pubid idtype="pmpid" link="fulltext"12882951pubid
pubidlist
xrefbib
bibl
bibl id="B25"
title
pA simulation study of the number of events per variable in logistic regression analysisp
title
aug
au
snmPeduzzisnm
fnmPfnm
au
au
snmConcatosnm
fnmJfnm
au
au
snmKempersnm
fnmEfnm
au
au
snmHolfordsnm
fnmTRfnm
au
au
snmFeinsteinsnm
fnmARfnm
au
aug
sourceJ Clin Epidemiolsource
pubdate1996pubdate
volume49volume
issue12issue
fpage1373fpage
lpage1379lpage
xrefbib
pubidlist
pubid idtype="doi"10.1016S0895-4356(96)00236-3pubid
pubid idtype="pmpid" link="fulltext"8970487pubid
pubidlist
xrefbib
bibl
bibl id="B26"
title
pPropensity score methods for bias reduction in the comparison of a treatment to a non-randomized control groupp
title
aug
au
snmD'Agostinosnm
fnmRBfnm
sufJrsuf
au
aug
sourceStat Medsource
pubdate1998pubdate
volume17volume
issue19issue
fpage2265fpage
lpage2281lpage
xrefbib
pubidlist
pubid idtype="doi"10.1002(SICI)1097-0258(19981015)17:19<2265::AID-SIM9183.0.CO;2-Bpubid
pubid idtype="pmpid" link="fulltext"9802183pubid
pubidlist
xrefbib
bibl
bibl id="B27"
title
pNeighbourhood environments and mortality in an elderly cohort: results from the cardiovascular health studyp
title
aug
au
snmDiez Rouxsnm
fnmAVfnm
au
au
snmBorrellsnm
fnmLNfnm
au
au
snmHaansnm
fnmMfnm
au
au
snmJacksonsnm
fnmSAfnm
au
au
snmSchultzsnm
fnmRfnm
au
aug
sourceJ Epidemiol Community Healthsource
pubdate2004pubdate
volume58volume
issue11issue
fpage917fpage
lpage923lpage
xrefbib
pubidlist
pubid idtype="doi"10.1136jech.2003.019596pubid
pubid idtype="pmcid"1732601pubid
pubid idtype="pmpid" link="fulltext"15483307pubid
pubidlist
xrefbib
bibl
bibl id="B28"
title
pThe role of the propensity score in estimating dose-response functionsp
title
aug
au
snmImbenssnm
fnmGWfnm
au
aug
sourceBiometrikasource
pubdate2000pubdate
volume87volume
issue3issue
fpage706fpage
lpage710lpage
xrefbib
pubid idtype="doi"10.1093biomet87.3.706pubid
xrefbib
bibl
bibl id="B29"
title
pDouble jeopardy: interaction effects of marital and poverty status on the risk of mortalityp
title
aug
au
snmSmithsnm
fnmKRfnm
au
au
snmWaitzmansnm
fnmNJfnm
au
aug
sourceDemographysource
pubdate1994pubdate
volume31volume
issue3issue
fpage487fpage
lpage507lpage
xrefbib
pubidlist
pubid idtype="doi"10.23072061754pubid
pubid idtype="pmpid"7828768pubid
pubidlist
xrefbib
bibl
bibl id="B30"
title
pLong term outcome and disability of diabetic patients hospitalised for diabetic foot ulcers: a 6.5 year follow-up studyp
title
aug
au
snmGhanassiasnm
fnmEfnm
au
au
snmVillonsnm
fnmLfnm
au
au
snmThuan dit Dieudonnesnm
fnmJ-Ffnm
au
au
snmBoegnersnm
fnmCfnm
au
au
snmAvignonsnm
fnmAfnm
au
au
snmSultansnm
fnmAfnm
au
aug
sourceDiabetes Caresource
pubdate2008pubdate
volume31volume
fpage1288fpage
lpage1292lpage
xrefbib
pubidlist
pubid idtype="doi"10.2337dc07-2145pubid
pubid idtype="pmcid"2453665pubid
pubid idtype="pmpid" link="fulltext"18390801pubid
pubidlist
xrefbib
bibl
bibl id="B31"
title
pEpidemiology of type 2 diabetes: focus on ethnic minoritiesp
title
aug
au
snmEgedesnm
fnmLEfnm
au
au
snmDagogo-Jacksnm
fnmSfnm
au
aug
sourceMed Clin North Amsource
pubdate2005pubdate
volume89volume
issue5issue
fpage949fpage
lpage975lpage
noteviiinote
xrefbib
pubidlist
pubid idtype="doi"10.1016j.mcna.2005.03.004pubid
pubid idtype="pmpid" link="fulltext"16129107pubid
pubidlist
xrefbib
bibl
refgrp
sec
st
pPre-publication historyp
st
pThe pre-publication history for this paper can be accessed here:p
p
urlhttp:www.biomedcentral.com1471-245810283prepuburl
p
sec
bm
art



PAGE 1

1 Additional File 4 Propensity score adjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction co ntrast ratio and attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of incident lower body functional limitation at 3 year follow up (weighted n=563). Good excellent housing conditions Fair poor housing conditions Without diabetes With diabetes Without diabetes With diabetes Type of co ndition OR OR 95% CI OR 95% CI OR 95% CI Interaction contrast ratio 95% CI Attributable proportion 95% CI Cleanliness inside building 1.00* 1. 39 0 .71, 2.87 0.85 0.36, 2.14 0.59 0.07, 4.92 0.57 2.65, 3.63 1.08 18.5, 0.82 Physical conditions interior 1.00* 1.32 0.65, 2.47 1.27 0.36, 3.22 0.64 0.08, 5.76 0.76 3.31, 4.23 1.33 22.93, 0.82 Furnishings 1.00* 1.36 0.67, 2.77 1.35 0.47, 3.75 1.38 0.27, 4.51 0.31 3.23, 2.61 0.23 5.84, 0.76

PAGE 2

2 Outside condition 1.00* 1.28 0.59, 2.75 1.28 0.54, 2.91 0. 49 0.08, 5.43 0.95 3.52, 3.42 2.06 20.74, 0.79 Overall condition 1.00* 1.24 0.62, 2.49 1.59 0.66, 3.47 2.46 0.27, 11.67 0.67 2.35, 9.76 0.28 6.59, 0.89 Variables included in the propensity score: sampling stratum, age, gender, income, per ceived income adequacy, educational attainment, marital status, employment status, number of persons in household, health care insurance, not being able to see a doctor because of c ost, s ocial support, self rated health status, depressive symptoms, a count of the number of chronic conditions, body mass index, risk of alcohol abuse, and physical activity Referent odds ratio


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ERHNOJXOH_ZPOSZI INGEST_TIME 2012-10-02T14:27:52Z PACKAGE AA00012386_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


xml version 1.0 encoding utf-8 standalone no
mets ID sort-mets_mets OBJID sword-mets LABEL DSpace SWORD Item PROFILE METS SIP Profile xmlns http:www.loc.govMETS
xmlns:xlink http:www.w3.org1999xlink xmlns:xsi http:www.w3.org2001XMLSchema-instance
xsi:schemaLocation http:www.loc.govstandardsmetsmets.xsd
metsHdr CREATEDATE 2012-07-26T00:05:51
agent ROLE CUSTODIAN TYPE ORGANIZATION
name BioMed Central
dmdSec sword-mets-dmd-1 GROUPID sword-mets-dmd-1_group-1
mdWrap SWAP Metadata MDTYPE OTHER OTHERMDTYPE EPDCX MIMETYPE textxml
xmlData
epdcx:descriptionSet xmlns:epdcx http:purl.orgeprintepdcx2006-11-16 xmlns:MIOJAVI
http:purl.orgeprintepdcxxsd2006-11-16epdcx.xsd
epdcx:description epdcx:resourceId sword-mets-epdcx-1
epdcx:statement epdcx:propertyURI http:purl.orgdcelements1.1type epdcx:valueURI http:purl.orgeprintentityTypeScholarlyWork
http:purl.orgdcelements1.1title
epdcx:valueString Neighborhood conditions, diabetes, and risk of lower-body functional limitations among middle-aged African Americans: A cohort study
http:purl.orgdctermsabstract
Abstract
Background
The relationship between presence of diabetes and adverse neighborhood and housing conditions and their effect on functional decline is unclear. We examined the association of adverse neighborhood (block face) and housing conditions with incidence of lower-body functional limitations among persons with and those without diabetes using a prospective population-based cohort study of 563 African Americans 49-65 years of age at their 2000-2001 baseline interviews.
Methods
Participants were randomly sampled African Americans living in the St. Louis area (response rate: 76%). Physician-diagnosed diabetes was self reported at baseline interview. Lower-body functional limitations were self reported based on the Nagi physical performance scale at baseline and the three-year follow-up interviews. The external appearance of the block the respondent lived on and five housing conditions were rated by study interviewers. All analyses were done using propensity score methods to control for confounders.
Results
109 (19.4%) of subjects experienced incident lower-body functional limitations at three-year follow-up. In adjusted analysis, persons with diabetes who lived on block faces rated as fair-poor on each of the five conditions had higher odds (7.79 [95% confidence interval: 1.36-37.55] to 144.6 [95% confidence interval: 4.45-775.53]) of developing lower-body functional limitations than the referent group of persons without diabetes who lived on block faces rated as good-excellent. At least 80 percent of incident lower-body functional limitations was attributable to the interaction between block face conditions and diabetes status.
Conclusions
Adverse neighborhood conditions appear to exacerbate the detrimental effects on lower-body functioning associated with diabetes.
http:purl.orgdcelements1.1creator
Schootman, Mario
Andresen, Elena M
Wolinsky, Fredric D
Miller, J Philip
Yan, Yan
Miller, Douglas K
http:purl.orgeprinttermsisExpressedAs epdcx:valueRef sword-mets-expr-1
http:purl.orgeprintentityTypeExpression
http:purl.orgdcelements1.1language epdcx:vesURI http:purl.orgdctermsRFC3066
en
http:purl.orgeprinttermsType
http:purl.orgeprinttypeJournalArticle
http:purl.orgdctermsavailable
epdcx:sesURI http:purl.orgdctermsW3CDTF 2010-05-27
http:purl.orgdcelements1.1publisher
BioMed Central Ltd
http:purl.orgeprinttermsstatus http:purl.orgeprinttermsStatus
http:purl.orgeprintstatusPeerReviewed
http:purl.orgeprinttermscopyrightHolder
Mario Schootman et al.; licensee BioMed Central Ltd.
http:purl.orgdctermslicense
http://creativecommons.org/licenses/by/2.0
http:purl.orgdctermsaccessRights http:purl.orgeprinttermsAccessRights
http:purl.orgeprintaccessRightsOpenAccess
http:purl.orgeprinttermsbibliographicCitation
BMC Public Health. 2010 May 27;10(1):283
http:purl.orgdcelements1.1identifier
http:purl.orgdctermsURI http://dx.doi.org/10.1186/1471-2458-10-283
fileSec
fileGrp sword-mets-fgrp-1 USE CONTENT
file sword-mets-fgid-0 sword-mets-file-1
FLocat LOCTYPE URL xlink:href 1471-2458-10-283.xml
sword-mets-fgid-1 sword-mets-file-2 applicationpdf
1471-2458-10-283.pdf
sword-mets-fgid-3 sword-mets-file-3 applicationmsword
1471-2458-10-283-S1.DOC
sword-mets-fgid-4 sword-mets-file-4
1471-2458-10-283-S2.DOC
sword-mets-fgid-5 sword-mets-file-5
1471-2458-10-283-S3.DOC
sword-mets-fgid-6 sword-mets-file-6
1471-2458-10-283-S4.DOC
structMap sword-mets-struct-1 structure LOGICAL
div sword-mets-div-1 DMDID Object
sword-mets-div-2 File
fptr FILEID
sword-mets-div-3
sword-mets-div-4
sword-mets-div-5
sword-mets-div-6
sword-mets-div-7



PAGE 1

1 Additional file 1 Unadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio an d attributable proportion and 95% confidence intervals) between diabetes and block face conditions for the risk of incident lower body functional limitation at 3 year follow up (weighted n=563). Good excellent block face conditions Fair poor block face conditions Without diabetes With diabetes Without diabetes With diabetes Type of condition Prevalence (%) of fair/poor quality OR OR 95% CI OR 95% CI OR 95% CI Interaction c ontrast ratio 95% CI Attributable proportion 95% CI Housing conditions 24.7 1.00* 1.42 0.67 2.97 1.37 0.75 2. 61 7.20 3.98 12.88 5.28 2.57 10.44 0.75 0.52 0.90 Noise 21.0 1.00* 1.48 0.76 3.04 1.39 0.76 2.61 7.91 4.64 14.52 5.92 2.93 5.92 0.76 0.52 0.89 Air quality 18.3 1.00* 2.01 1.09 3.75 2.37 1.30 4.67 13.82 6.74 38.17 10.15 3.65 10.15 0 .75 0.47 0.92 Street & road quality 20.9 1.00* 1.54 0.78 3.12 2.35 1.17 4.56 14.19 7.37 30.56 10.99 5.12 25.99 0.79 0. 60 0.91

PAGE 2

2 Yard & sidewalk quality 24.6 1.00* 1.08 0.53 2.34 1.20 0.70 2.21 9.52 5.34 17.73 8.12 4.40 15.49 0.86 0. 69 0.9 5 *Referent odds ratio



PAGE 1

1 Additional File 3 Unadjusted measures of association (odds ratio and 95% confidence intervals) and interaction (interaction contrast ratio an d attributable proportion and 95% confidence intervals) between diabetes and housing conditions for the risk of i ncident lower body functional limitation at 3 year follow up (weighted n=563). Good excellent housing conditions Fair poor housing conditions Without diabetes With diabetes Without diabetes W ith diabetes Type of condition Prevalence (%) of fair/poor quality OR OR 95% CI OR 95% CI OR 95% CI Interaction c ontrast ratio 95% CI Attributable proportion 95% CI Cleanliness inside building 21.1 1.00* 2.04 1.14, 3.92 0.98 0.50, 2.1 6 1.94 1.01, 4.14 0.06 2.10, 1.66 0.04 1.37, 0.56 Physical conditions 19.5 1.00* 2.07 1.13, 3.81 1.02 0.45, 2.26 2.00 1.03, 3.93 0.06 2.30, 1.57 0.03 1.44, 0.56

PAGE 2

2 interior Furnishings 22.6 1.00* 2.12 1.15, 4.11 1.54 0.75, 3.06 2.88 1.36, 5.6 9 0.12 2.36, 2.44 0.04 1.13, 0.57 Outside condition 20.6 1.00* 2.09 1.12, 3.92 1.20 0.62, 2.40 2.32 1.06, 4.66 0.05 2.14, 1.85 0.02 1.14, 0.55 Overall condition 19.6 1.00* 1.98 1.05, 3.72 1.30 0.70, 2.63 3.04 1.40, 6.72 0.72 1.73, 3.69 0.2 6 0.90, 0.70 *Referent odds ratio