![ORCID Icon]({"type":"image","src":"/templates/jsp/_style2/_tandf/images/orcid.svg"} "ORCID Icon"){kind=small}
http://orcid.org/0000-0003-4052-1524
ABSTRACT
ABSTRACT
Adherence to pulmonary rehabilitation (PR) is low. Previous studies have focused on clinical predictors of PR completion. We aimed to identify social determinants of adherence to PR. A cross-sectional analysis of a database of COPD patients (N = 455) in an outpatient PR program was performed. Adherence, a ratio of attended-to-prescribed sessions, was coded as low (<35%), moderate (35–85%), and high (>85%). Individual-level measures included age, sex, race, BMI, smoking status, pack-years, baseline 6-minute walk distance (6MWD: <150, 150–249, ≥250), co-morbidities, depression, and prescribed PR sessions (≤20, 21–30, >30). Fifteen area-level measures aggregated to Census tracts were obtained from the U.S. Census after geocoding patients' addresses. Using exploratory factor analysis, a neighborhood socioeconomic disadvantage index was constructed, which included variables with factor loading >0.5: poverty, public assistance, households without vehicles, cost burden, unemployment, and minority population. Multivariate regression models were adjusted for clustering on Census tracts. Twenty-six percent of patients had low adherence, 23% were moderately adherent, 51% were highly adherent. In the best fitted full model, each decile increase in neighborhood socioeconomic disadvantage increased the risk of moderate vs high adherence by 14% (p < 0.01). Smoking tripled the relative risk of low adherence (p < 0.01), while each increase in 6MWD category decreased that risk by 72% (p < 0.01) and 84% (p < 0.001), respectively. These findings show that, relative to high adherence, low adherence is associated with limited functional capacity and current smoking, while moderate adherence is associated with socioeconomic disadvantage. The distinction highlights different pathways to suboptimal adherence and calls for tailored intervention approaches.
Introduction
Completion of pulmonary rehabilitation (PR) for chronic obstructive pulmonary disease (COPD) has been associated with a decrease in exacerbations and hospitalizations (1) and improvement in dyspnea, functional capacity, and quality of life (2). Despite these benefits, many patients with COPD do not complete prescribed PR, with attrition rates up to 60% (3–5). Research on adherence to PR has focused on clinical and demographic characteristics (3, 6). Although the contribution of socioeconomic factors for completion of PR has been acknowledged (7–9), the associations are poorly understood. Few studies have investigated the social determinants affecting the ability of patients to complete prescribed treatments.
The scarcity of social research on adherence to PR stems from a narrow understanding of the ways in which social context impacts adherence to medical recommendations. Studies investigating social predictors of adherence to PR have used measures limited to demographic characteristics, marital and employment status, and educational attainment (10). The social context, however, is a multidimensional construct defined by access to financial, human, and social resources and capital (11). This premise, well-established in the sociological literature (12), has not been applied to research on PR in COPD.
The social, economic, and cultural contexts influence the health-related behaviors of individuals, as evidenced by the social patterning of tobacco and alcohol use, diet and physical activity, or preventive care utilization (13–16). Nevertheless, predictors of PR adherence have seldom been assessed from a sociological perspective. The purpose of this study was to identify predictors of adherence to PR by applying the Health Lifestyle Theory (17, 18), which correlates individual health practices with structural factors such as class circumstances (income, wealth, education, occupational prestige) and living conditions. Based on this framework, we hypothesized that socioeconomic disadvantage is associated with decreased adherence to PR in patients with COPD.
Methods
Study design and population sample
We performed a cross-sectional analysis of a prospectively maintained clinical database of patients (N = 455) who had attended an outpatient PR program at the University of Alabama at Birmingham (UAB) between 1996 and 2013. Due to the large span of the program, analyses controlled for year of PR. The database, described in detail elsewhere (19), included patients with a primary diagnosis of COPD (International Classification of Diseases Codes 491, 492, and 496) at the time of enrollment in PR. PR sessions were prescribed by an exercise physiologist with the number of sessions determined based on medical history review, severity of disease, and walk test distance. Participants attended two to three sessions per week, for a maximum of 36 sessions over 12 weeks. Exercise regimens followed standard PR guidelines (20) and included aerobic exercises such as treadmill walking, cycle, and arm ergometry; resistance training such as machine weights, hand weights, and elastic bands; and breathing training techniques such as diaphragmatic breathing and pursed lips breathing. Regimens were individualized according to patients' exercise tolerance indicated by a baseline 6-minute walk distance test (6MWD) as well as their subjective sense of dyspnea. Those who continued to smoke were allowed to participate in the program.
Demographic and clinical data were obtained prior to the initial PR session. To obtain socioeconomic data, we geocoded patients' residential addresses and assigned a Census tract identifier to each record. Data at the Census tract level were downloaded from the U.S. Census Factfinder website (https://factfinder.census.gov/faces/nav/jsf/pages/index.xhtml) and matched to patients: year 2000 Census data were used for patients with first PR attended between 1996 and 2005, while year 2010 Census data were used for those with first PR attended between 2006 and 2013. The distance from a patient's residence to the outpatient PR facility was calculated using the driving directions provided by Google Maps (https://www.google.com/maps) and selecting the fastest driving route. Ethical approval for this study was obtained from the UAB Institutional Review Board, Protocol X160907008.
Measures
Individual characteristics
Adherence to PR was calculated as a ratio of attended-to-prescribed sessions for the first PR series on record. Adherence rates were not normally distributed, with 44% of participants being 100% adherent. We therefore coded adherence rates as low (<35%), moderate (35–85%), and high (>85%), using cut-offs from previous adherence studies (21–24).
Demographic characteristics included age, sex, and race (White/Black/Other). Clinical characteristics included body mass index (BMI, kg/m2), current smoking (Yes/No based on self-report at the PR enrollment visit), smoking pack-years, and number of co-morbidities (simple count based on physician diagnosis). Depression (Yes/No based on physician diagnosis) was included separately because of its established role in adherence. Baseline functional capacity, assessed with the 6-minute walk distance test (6MWD, in meters), was coded in three categories: <150, 150–249, ≥250 (25). The number of prescribed PR sessions was coded in three categories: ≤20, 21–30, >30.
Neighborhood characteristics
Census tracts, which are small, socio-demographically homogenous geographic units, served as proxies for neighborhoods. Fifteen socioeconomic variables were obtained from the U.S. Census: 1) % Adults age 25+ without high-school diploma; 2) % Adults age 25+ with college degree; 3) Median household income; 4) % Households with interest, dividend, or rental income; 5) Median value of owner-occupied housing; 6) % Households on public assistance or food stamps; 7) % Households in poverty; 8) % Cost-burdened households (with housing costs >30% income); 9) Civilian unemployment rate; 10) % Households without vehicles; 11) Mean number of household vehicles; 12) % Minority population; 13) % Disabled population; 14) % Households with seniors (age 65+) living alone; and 15) % Single-parent households with children. These variables reflect various aspects of income, wealth, education, occupation, housing, and family structure and are comparable to measures used previously to examine the effects of neighborhood disadvantage on health (26–29).
Statistical analysis
Following other studies (27, 30–33), exploratory factor analysis with orthogonal VARIMAX rotation was used to construct an index of neighborhood socioeconomic disadvantage from the 15 candidate variables above. Individual variables were retained if they had a factor loading (i.e., standardized regression coefficient) >0.5. One primary factor emerged, which explained 71% of the total variance and was defined by six variables: public assistance or food stamps, poverty, cost-burden, unemployment, households without a vehicle, and minority population (Table 1). This factor was identified as “socioeconomic disadvantage.”
Table 1. Orthogonal VARIMAX rotated factor patterns and loadings of neighborhood variables.
Next, we constructed a socioeconomic disadvantage index (SDI) by summing the standardized z-scores of the six variables clustered in factor 1 (Cronbach's alpha = 0.91), with higher scores indicating higher disadvantage (range 2.0–59.6, mean = 15.3, SD 11.30). The scores were then categorized into deciles to create a scale of 1 to 10 (mean = 4.52, SD 2.89), which was used continuously. Neighborhoods with SDI scores in the bottom decile were least disadvantaged, while neighborhoods with SDI scores in the top decile were most disadvantaged.
Univariate statistics – means, standard deviations, frequencies, and proportions – were obtained for all non-missing observations. The distribution of individual and neighborhood characteristics was obtained by adherence categories and overall. Bivariate relationships between adherence and every covariate were estimated using simple multinomial logistic regression. Multiple multinomial logistic regression was used to estimate adherence models of individual and neighborhood characteristics. Because the data featured a sparsely populated cluster structure where 49% of clusters represented by Census tracts had a singleton observation, multilevel modeling was not feasible. Instead, we accounted for clustering on Census tracts by adjusting for standard errors and the variance–covariance matrix of estimates to allow for intra-group correlation, relaxing the requirement of independence within clusters (34). Model fit was determined by the difference in Bayesian Information Criterion (BIC) between models. Statistical tests were two-sided and were performed using a 5% significance level (α = 0.05). Analyses were performed using Stata software, version 12.
Results
Study population
The dataset included 455 patients diagnosed with COPD who had participated in a pulmonary rehabilitation program between 1996 and 2013. We excluded patients (n = 40, 8.8%) for whom a Census tract of residence could not be reliably determined due to the use of a P.O. box instead of a street address. The distribution of the excluded patients was random on all variables. The dataset included 211 unique Census tracts: 103 (49%) with only one patient per tract and 108 (51%) with more than one patient per tract: 55 tracts with 2 patients, 29 tracts with 3 patients, and 24 tracts with ≥4 patients (mean = 2, range 1–9).
Neighborhood characteristics differed considerably between the most and least disadvantaged neighborhoods as measured by SDI. For example, compared to the least disadvantaged neighborhoods, the most disadvantaged neighborhoods had more minority population (91.6% vs 6.7%), more poverty (38.5% vs 3.3%), more unemployment (8.4% vs 2.0%), more households without a vehicle (29.1% vs 1.7%), more cost-burdened households (35.6% vs 8.8%), and more households receiving public assistance or food stamps (29.9% vs 2.1%).
Characteristics of the sample by adherence status
The characteristics of the final sample (N = 415), overall and by adherence status, are described in Table 2. The patient population, mean age 66.1 years (SD 9.3), was 56% male and 70% White; 11% were current smokers. In terms of adherence to PR, 26% had low adherence, 23% were moderately adherent, and 51% highly adherent.
Table 2. Characteristics of study participants (N = 415): overall and by adherence status.
As seen in Table 2, there were no significant differences between adherence groups in terms of age, sex, race, living arrangements, BMI, smoking pack-years, physician-diagnosed depression, or distance to the PR clinic. However, the three adherence groups differed significantly in baseline functional capacity (6MWD), co-morbidities, smoking status, and number of prescribed PR sessions. They also differed in neighborhood characteristics, specifically % households receiving public assistance or food stamps, % households in poverty, % cost-burdened households, % minority population, and % disabled population. The unadjusted mean SDI scores stratified by adherence levels showed that high adherers experienced the lowest socioeconomic disadvantage (4.2, SD 2.9) while moderate adherers experienced the highest (5.2, SD 3.0), p < 0.01 (Table 2).
Multivariate statistics
In a multiple multinomial logistic regression of individual characteristics (Table 3, Model 1), compared with high adherers, low adherers were more likely to be current smokers (RRR 2.98, p < 0.01) and have lower functional capacity as measured by 6MWD (RRR 0.30, p < 0.01 and RRR 0.17, p < 0.001 for 150–249 m and ≥250 m, respectively; reference category <150 m). Moderate adherers did not differ from high adherers in terms of individual characteristics. After adding neighborhood characteristics (Table 3, Model 2), all differences between low and high adherers were preserved with minimal changes in effect size, while moderate adherers were more likely than high adherers to experience socioeconomic disadvantage as measured by the SDI (RRR 1.16, p < 0.01). In both models, year of PR was positively associated with low adherence (RRR 1.07 and 1.08, respectively; p < 0.05 for both).
Table 3. Multiple multinomial logistic regression of adherence.
The best fitted model of individual and neighborhood characteristics adjusted for clustering into Census tracts is presented in Table 4. By this predictive model, White race doubled and current smoking tripled the relative risk of low vs high adherence (p < 0.05 and p < 0.01, respectively), while each increase in 6MWD category decreased the risk of low adherence by 72% and 84%, respectively (p < 0.001 for both). Year of PR was associated with a 10% increased risk for low adherence (p < 0.001) relative to high adherence. Each decile increase in the neighborhood SDI increased the risk of moderate vs high adherence by 14% (p < 0.01). Thus, the highest socioeconomic disadvantage was associated with a 140% risk of moderate vs high adherence compared to the lowest socioeconomic disadvantage.
Table 4. Individual and neighborhood predictors of adherence.
Discussion
We identified social determinants of adherence to PR by enhancing a clinical database with neighborhood-level socioeconomic measures obtained from the U.S. Census. Findings showed that, relative to high adherence, low adherence was correlated with low adherence as correlated with low functional capacity, current smoking, and White race, while moderate adherence was correlated with socioeconomic disadvantage. The distinction is important because it highlights different pathways to suboptimal adherence and calls for intervention approaches that are tailored to specific types of nonadherence. For example, while adherence-promoting strategies for patients with limited functional capacity may include home-based or inpatient PR, strategies targeting socioeconomically disadvantaged patients may additionally include interventions that address cost- or stress-related barriers.
Socioeconomic disadvantage as a risk factor for adherence to PR is underinvestigated, although it has been acknowledged in the context of adherence to COPD medications (35–38). Our analysis showed that socioeconomic disadvantage is an important correlate of adherence to PR. Similar results have been reported for cardiac rehabilitation, with patients of lower socioeconomic status and current smokers attending three times fewer sessions than patients without such risk factors (39). In our study, patient characteristics predictive of lowest adherence to PR were White race, smoking, and low baseline functional capacity, which may be indicative of health-related barriers to following the exercise regimen. In contrast, patients having moderate adherence were characterized by socioeconomic disadvantage. These findings indicate that in the absence of health-related limitations that drive low adherence, the main barrier to PR is of socioeconomic nature.
Socioeconomically disadvantaged individuals face a variety of chronic stressors in daily living, which trigger unhealthy behaviors (13–16) that help regulate mood (14, 40). The coping functions of such behaviors make them particularly salient and limit the ability of patients to adopt health-beneficial but challenging regimens (15, 41). Future research should test approaches to mitigate the negative impact of socioeconomic disadvantage on adherence to PR. For example, there is evidence that perceived sense of control acts as a buffer against the effect of socioeconomic disadvantage (42). Several studies have reported success in increasing adherence by using positive affect/self-affirmation (43), mindfulness-based stress reduction (44), and financial literacy interventions to reduce financial stress (45).
Traditionally, the adherence behavior has been viewed as a behavioral problem of the patient (46), and adherence-promoting strategies have focused on individual patient psychology, with little consideration of the socio-environmental context in which the patient is situated (47, 48). Meta-analyses show limited success of this approach to behavioral change (49–51), indicating a need for shift toward social factors that impact patients' ability to follow medical recommendations. Understanding adherence to PR as an interaction between the patient and his or her social environment could produce a more effective framework for achieving better adherence (52, 53).
The majority of research on adherence to PR has focused on completion of prescribed sessions, dichotomizing patients into completers and non-completers (4, 19, 54, 55). In contrast, we coded adherence rates in three categories – low, moderate, and high – on the basis of a calculated ratio of attended-to-prescribed sessions (<35%, 35–85%, and >85%). We chose this approach after preliminary analysis which showed that non-completers are not a homogenous group in terms of clinical and socioeconomic characteristics, and that discerning a degree of non-completion can reveal different non-adherence typologies. The results confirmed this preliminary observation. The adherence rate of 51%, based on attending at least 85% of prescribed sessions, is consistent with previous reports (3–5).
To our knowledge, this is the first study to investigate social determinants of adherence to PR in patients with COPD. The paucity of such research is partially explained by the lack of socioeconomic data in clinical registries and medical records. Despite an increased scientific interest in the role of socio-environmental factors for health, clinical registries rarely include sufficient socioeconomic data. In this study, we overcame this limitation by using area-level socioeconomic data aggregated to Census tracts, proxy for neighborhoods. Although individual- and area-level measures of wealth, income, education, and occupation are closely associated (26), they are not perfectly correlated, especially for racial/ethnic minorities. Ideally, neighborhood-level social indicators should be used together with, rather than in place of, individual-level indicators, as they provide complementary information on social context and living circumstances. This study shows that neighborhood-level data can be used to assess risk of nonadherence to PR. However, it also highlights the need for expanding health records with socioeconomic and spatial data, which is relevant for clinical decision-making.
Systematic reviews have established that socioeconomic disadvantage negatively impacts COPD outcomes, including morbidity and mortality (56), and results in health disparities (57). Possible mechanisms of that correlation include environmental exposures, occupational risk factors, and smoking, all of which are socially stratified. This study highlights adherence to PR as another potential contributor to COPD disparities by social class, suggesting that for COPD patients in relatively good health, the main barriers to high adherence may be of socioeconomic nature.
Outpatient PR is typically prescribed to patients residing within a reasonable commuting distance to an outpatient PR clinic. Driving distance of >36 miles compared to <6 miles has been reported as a risk factor for adherence (58). In this study, the outpatient PR clinic was located in the City of Birmingham, Alabama, where proximity to the city is correlated with socioeconomic disadvantage. The lack of association between driving distance and adherence in this analysis is in line with the finding that socioeconomic disadvantage is a risk factor for adherence.
Finally, it should be noted that the analysis spanned 17 years, and later year of entry in the PR program was a risk factor for low adherence, most likely reflecting changes in patient characteristics due to program modifications. Specifically, in the initial years of the program, COPD patients were usually referred by a pulmonologist after careful selection in order to maximize the likelihood of PR completion. In later years, referrals to the program increased in number, included more smokers and patients with co-morbidities, and were eventually made automatic at the time of hospital discharge through the implementation of a standard COPD order set. For example, before 2005, 10% of participants were current smokers, compared to 14% after 2005. Similarly, in the first year of the program, patients had an average of 1.1 co-morbidities (SD 1.1), while in the last year they had 3.7 co-morbidities (SD 1.6); the proportion of patients with physician-diagnosed depression increased four-fold. Such changes in patient characteristics have most likely contributed to lower adherence rates over time.
Limitations
This study has several limitations. First, its cross-sectional design limits the ability to make causal inferences about the observed relationships between socioeconomic disadvantage and decreased adherence. The study is also limited to COPD patients in a single outpatient PR clinic and may not be representative of all COPD patients. Third, our data featured a sparsely populated cluster structure that ruled out multilevel modeling. Instead, we accounted for clustering on Census tracts by adjusting for standard errors and the variance-covariance matrix of estimates.
In this analysis, we used physician-reported measure of depression. As depression, an important determinant of COPD outcomes, tends to be clinically underdiagnosed (59, 60), results may have underestimated the true impact of depression on adherence to prescribed PR.
Finally, neighborhood socioeconomic measures, although reasonably correlated with individual-level measures, are not a perfect substitute for individual socioeconomic data. To the extent possible, future adherence studies should collect patient-level socioeconomic indicators.
Conclusions
Socioeconomic disadvantage and health status impact patients' ability to attend PR over extended periods of time. While smoking and low functional capacity are predictors of low adherence, socioeconomic disadvantage is a predictor of moderate adherence. Interventions to improve adherence to PR should be tailored to the specific pathways leading to suboptimal adherence.
| Variable | Factor loading | |
|---|---|---|
| 1 | % Adults age 25+ without high-school diploma | |
| 2 | % Adults age 25+ with college degree | |
| 3 | Median household income | |
| 4 | % Households with interest, dividend, or rental income | |
| 5 | Median value of owner-occupied housing | |
| 6 | % Households on public assistance or food stamps | 0.7385 |
| 7 | % Households in poverty | 0.8534 |
| 8 | % Cost-burdened households (housing >30% income) | 0.6083 |
| 9 | Civilian unemployment rate, % | 0.5110 |
| 10 | % Households without vehicles | 0.8745 |
| 11 | Mean number of household vehicles | |
| 12 | % Minority population | 0.7107 |
| 13 | % Disabled population | |
| 14 | % Households with seniors (age 65+) living alone | |
| 15 | % Single-parent households with children |
| Adherence | ||||
|---|---|---|---|---|
| Variables | Overall | Low (<35%) 106 (25.5) | Moderate (35–85%) 98 (23.6) | High (>85%) 211 (50.8) |
| INDIVIDUAL | ||||
| Age, years (range 41–87) | 66.1 (9.3) | 66.3 (9.6) | 64.6 (9.4) | 66.7 (9.0) |
| Sex | ||||
| Male | 233 (56.1) | 63 (59.4) | 46 (46.9) | 124 (58.8) |
| Female | 182 (43.9) | 43 (40.6) | 52 (53.1) | 87 (41.2) |
| Race | ||||
| White | 290 (69.9) | 79 (74.5) | 62 (63.3) | 149 (70.6) |
| Non-White | 125 (30.1) | 27 (25.5) | 36 (36.7) | 62 (29.4) |
| Lives alone | 53 (12.8) | 12 (11.3) | 13 (13.3) | 28 (13.3) |
| BMI, kg/m2 (range 14.2–80.7) | 28.6 (8.5) | 28.4 (8.2) | 29.5 (9.4) | 28.2 (8.2) |
| Co-morbidities (range 0–9)a* | 2.7 (1.9) | 3.0 (1.9) | 2.7 (2.0) | 2.5 (1.8) |
| 6MWD, ma*** | 267.4 (117.0) | 227.0 (121.8) | 270.8 (123.4) | 286.1 (106.5) |
| <150 | 69 (16.6) | 32 (30.2) | 17 (17.4) | 20 (9.5) |
| 150–249 | 110 (26.5) | 29 (27.4) | 24 (24.5) | 57 (27.0) |
| ≥250 | 236 (56.9) | 45 (19.1) | 57 (68.2) | 134 (63.5) |
| Current smoking a** | ||||
| Yes | 46 (11.1) | 19 (17.9) | 13 (13.3) | 14 (6.6) |
| No | 369 (88.9) | 87 (82.1) | 85 (86.7) | 197 (93.4) |
| Smoking pack-years | 46.8 (42.5) | 51.3 (49.0) | 48.8 (54.0) | 43.5 (31.5) |
| Depression | 89 (21.5) | 28 (26.4) | 24 (24.5) | 37 (17.5) |
| Prescribed sessionsa,b* | ||||
| ≤20 | 107 (25.8) | 23 (21.7) | 20 (20.4) | 64 (30.3) |
| 21–30 | 230 (55.4) | 57 (53.8) | 57 (58.2) | 116 (55.0) |
| >30 | 78 (18.8) | 26 (24.5) | 21 (21.4) | 31 (14.7) |
| Distance from home, miles (range 0.7–183) | 18.2 (21.1) | 18.9 (16.1) | 17.5 (26.3) | 18.2 (20.7) |
| NEIGHBORHOOD | ||||
| % College degree | 27.2 (20.7) | 25.8 (19.6) | 30.5 (22.7) | 26.3 (20.2) |
| % No high school diploma | 16.3 (10.5) | 16.7 (10.2) | 15.0 (10.8) | 16.7 (10.4) |
| Median household income, $ | 46,748.1 (22,261.7) | 46,961.5 (22,239.8) | 46,824.7 (24,825.6) | 46,605.3 (21,094.6) |
| % Households with interest, dividend, or rental income | 25.5 (16.5) | 23.3 (14.4) | 24.4 (17.8) | 27.1 (16.8) |
| Civilian unemployment rate, % | 4.5 (3.5) | 4.8 (3.5) | 4.7 (3.3) | 4.2 (3.6) |
| % Households without vehicle | 8.1 (9.4) | 8.5 (10.8) | 9.4 (9.8) | 7.4 (8.2) |
| Mean number of household vehicles | 1.8 (0.4) | 1.8 (0.4) | 1.8 (0.4) | 1.8 (0.4) |
| Median value of owner-occupied housing, $ | 138,194.0 (88,611.9) | 138,741.5 (79,173.59) | 148,136.7 (104,997.9) | 133,300.9 (84,752.2) |
| % Seniors living alone | 9.9 (4.5) | 9.9 (4.3) | 10.1 (5.3) | 9.7 (4.3) |
| % Single parents with children | 19.0 (12.5) | 18.1 (14.6) | 19.2 (13.2) | 19.4 (11.1) |
| % Households with public assistance or food stampsb* | 10.1 (9.6) | 10.6 (9.4) | 11.7 (11.0) | 9.2 (8.9) |
| % Households in povertyb* | 14.2 (11.8) | 14.5 (11.9) | 16.4 (12.3) | 13.0 (11.3) |
| % Cost-burdened householdsb* | 18.1 (12.0) | 18.1 (14.9) | 20.2 (10.5) | 17.2 (10.9) |
| % Minority populationb** | 37.5 (33.4) | 35.0 (31.2) | 45.7 (36.2) | 35.1 (32.7) |
| % Disabled populationa,b** | 8.4 (8.2) | 10.0 (8.4) | 9.5 (8.3) | 7.1 (8.0) |
| SDI, decilesb** | 4.5 (2.9) | 4.5 (2.8) | 5.2 (3.0) | 4.2 (2.9) |
| Model 1 (individual characteristics) | Model 2 (individual + neighborhood characteristics) | |||
|---|---|---|---|---|
| Variables | RRR | [95% Cl] | RRR | [95% Cl] |
| Low Adherencea | ||||
| Individual | ||||
| Age | 0.99 | 0.96–1.02 | 0.99 | 0.96–1.02 |
| Male | 1.14 | 0.65–1.99 | 1.14 | 0.68–1.92 |
| White | 1.50 | 0.81–2.76 | 1.91 | 0.95–3.87 |
| Lives alone | 0.92 | 0.42–2.01 | 0.90 | 0.41–1.96 |
| BMI, kg/m2 | 0.99 | 0.96–1.03 | 0.99 | 0.96–1.03 |
| Co-morbidities | 1.08 | 0.92–1.26 | 1.09 | 0.93–1.27 |
| 6MWD, mb | ||||
| 150–249 | 0.30** | 0.14–0.64 | 0.29** | 0.12–0.67 |
| ≥250 | 0.17*** | 0.08–0.35 | 0.16*** | 0.08–0.35 |
| Current smoking | 2.98** | 1.34–6.63 | 2.78* | 1.22–6.34 |
| Pack-years | 1.00 | 1.00–1.01 | 1.00 | 1.00–1.01 |
| Depression | 1.44 | 0.59–2.19 | 1.20 | 0.61–2.37 |
| Prescribed sessionsc | ||||
| 21–30 | 1.32 | 0.70–2.51 | 1.34 | 0.70–2.56 |
| >30 | 1.48 | 0.65–3.36 | 1.33 | 0.54–3.26 |
| Distance from home, mi. | 1.00 | 0.99–1.02 | 1.01 | 0.99–1.02 |
| Year | 1.07* | 1.01–1.14 | 1.08* | 1.02–1.15 |
| Neighborhood | ||||
| SDI, deciles | — | — | 1.10 | 0.99–1.21 |
| Moderate Adherencea | ||||
| Individual | ||||
| Age | 0.98 | 0.95–1.01 | 0.98 | 0.95–1.01 |
| Male | 0.62 | 0.36–1.06 | 0.58 | 0.33–1.02 |
| White | 0.85 | 0.48–1.51 | 1.26 | 0.66–2.39 |
| Lives alone | 0.97 | 0.46–2.06 | 0.83 | 0.37–1.86 |
| BMI, kg/m2 | 1.00 | 0.97–1.04 | 1.00 | 0.97–1.04 |
| Co-morbidities | 1.00 | 0.85–1.17 | 1.01 | 0.85–1.21 |
| 6MWD, mb | ||||
| 150–249 | 0.51 | 0.22–1.18 | 0.50 | 0.21–1.18 |
| ≥250 | 0.48 | 0.22–1.03 | 0.47 | 0.21–1.06 |
| Current smoking | 1.70 | 0.73–3.97 | 1.48 | 0.59–3.73 |
| Pack-years | 1.01 | 1.00–1.01 | 1.01 | 1.00–1.01 |
| Depression | 1.02 | 0.53–1.97 | 1.06 | 0.54–2.07 |
| Prescribed sessionsc | ||||
| 21–30 | 1.67 | 0.89–3.15 | 1.71 | 0.91–3.20 |
| >30 | 1.87 | 0.82–4.28 | 1.83 | 0.82–4.12 |
| Distance from home, mi. | 1.00 | 0.99–1.01 | 1.00 | 0.99–1.02 |
| Year | 1.04 | 0.99–1.10 | 1.05 | 0.99–1.12 |
| Neighborhood | ||||
| SDI, deciles | — | — | 1.16** | 1.04–1.29 |
| Variables | RRR | 95% CI |
|---|---|---|
| Low Adherence (reference High Adherence) | ||
| Individual | ||
| Age | 0.98 | 0.96–1.01 |
| Male | 1.32 | 0.81–2.13 |
| White | 2.03* | 1.04–3.95 |
| 6MWD, m (base <150) | ||
| 150–249 | 0.28*** | 0.12–0.64 |
| ≥250 | 0.16*** | 0.07–0.34 |
| Current smoking | 3.11** | 1.42–6.81 |
| Prescribed sessions (base ≤20) | ||
| 21–30 | 1.28 | 0.70–2.36 |
| >30 | 1.32 | 0.55–3.16 |
| Year | 1.10*** | 1.04–1.16 |
| Neighborhood | ||
| SDI, deciles | 1.08 | 0.98–1.19 |
| Moderate Adherence (reference High Adherence) | ||
| Individual | ||
| Age | 0.98 | 0.95–1.01 |
| Male | 0.72 | 0.44–1.17 |
| White | 1.33 | 0.71–2.49 |
| 6MWD, m (base <150) | ||
| 150–249 | 0.48 | 0.21–1.10 |
| ≥250 | 0.46 | 0.21–1.02 |
| Current smoking | 1.79 | 0.74–4.32 |
| Prescribed sessions (base ≤20) | ||
| 21–30 | 1.56 | 0.84–2.90 |
| >30 | 1.72 | 0.78–3.79 |
| Year | 1.06 | 1.00–1.12 |
| Neighborhood | ||
| SDI, deciles | 1.14** | 1.03–1.26 |
Declaration of interest
The authors report no conflicts of interest. The content is the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality.
Funding
This project was supported by grant number K12HS023009 from the Agency for Healthcare Research and Quality.
References
- Puhan MA, Gimeno-Santos E, Cates CJ, Troosters T. Pulmonary rehabilitation following exacerbations of chronic obstructive pulmonary disease. Cochrane Database Syst Rev. 2016;12:CD005305. [PubMed], [Google Scholar]
- Garvey C, Bayles MP, Hamm LF, et al. Pulmonary rehabilitation exercise prescription in chronic obstructive pulmonary disease: review of selected guidelines: an official statement from the American association of cardiovascular and pulmonary rehabilitation. J Cardiopulm Rehabil Prev. 2016;36(2):75–83. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Hayton C, Clark A, Olive S, et al. Barriers to pulmonary rehabilitation: characteristics that predict patient attendance and adherence. Respir Med. 2013;107(3):401–407. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Boutou AK, Tanner RJ, Lord VM et al. An evaluation of factors associated with completion and benefit from pulmonary rehabilitation in COPD. BMJ Open Respir Res. 2014;1(1):e000051. [Crossref], [PubMed], [Google Scholar]
- Hogg L, Garrod R, Thornton H, McDonnell L, Bellas H, White P. Effectiveness, attendance, and completion of an integrated, system-wide pulmonary rehabilitation service for COPD: prospective observational study. COPD. 2012;9(5):546–554. [Taylor & Francis Online], [Google Scholar]
- Heerema-Poelman A, Stuive I, Wempe JB. Adherence to a maintenance exercise program 1 year after pulmonary rehabilitation: what are the predictors of dropout? J Cardiopulm Rehabil Prev. 2013;33(6):419–426. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Blackstock FC, ZuWallack R, Nici L, Lareau SC. Why don't our patients with chronic obstructive pulmonary disease listen to us? The enigma of nonadherence. Ann Am Thorac Soc. 2016;13(3):317–323. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Keating A, Lee A, Holland AE. What prevents people with chronic obstructive pulmonary disease from attending pulmonary rehabilitation? A systematic review. Chron Respir Dis. 2011;8(2):89–99. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Spruit MA, Franssen FM. What to do before pulmonary rehabilitation to improve adherence? Chron Respir Dis. 2010;7(3):131–133. [Crossref], [PubMed], [Google Scholar]
- Fischer MJ, Scharloo M, Abbink JJ, et al. Drop-out and attendance in pulmonary rehabilitation: the role of clinical and psychosocial variables. Respir Med. 2009;103(10):1564–1571. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Braveman PA, Cubbin C, Egerter S, et al. Socioeconomic status in health research: one size does not fit all. JAMA. 2005;294(22):2879–2888. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Adler NE, Rehkopf DH. U.S. disparities in health: descriptions, causes, and mechanisms. Annu Rev Public Health. 2008;29:235–252. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Lynch JW, Kaplan GA, Salonen JT. Why do poor people behave poorly? Variation in adult health behaviours and psychosocial characteristics by stages of the socioeconomic lifecourse. Soc Sci Med. 1997;44(6):809–819. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Lantz PM, House JS, Mero RP, Williams DR. Stress, life events, and socioeconomic disparities in health: results from the Americans' Changing Lives Study. J Health Soc Behav. 2005;46(3):274–288. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Lutfey K, Freese J. Toward some fundamentals of fundamental causality: socioeconomic status and health in the routine clinic visit for diabetes. American Journal of Sociology. 2005;110(5):1326–172. [Crossref], [Web of Science ®], [Google Scholar]
- Pampel FC, Krueger PM, Denney JT. Socioeconomic disparities in health behaviors. Annu Rev Sociol. 2010(36):349–370. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Cockerham WC. Health lifestyle theory and the convergence of agency and structure. J Health Soc Behav. 2005;46(1):51–67. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Cockerham WC. Social Causes of Health and Disease. 2nd ed. Cambridge, UK: Polity 2013. vii, 255. [Google Scholar]
- Brown AT, Hitchcock J, Schumann C, Wells JM, Dransfield MT, Bhatt SP. Determinants of successful completion of pulmonary rehabilitation in COPD. Int J Chron Obstruct Pulmon Dis. 2016;11:391–397. [PubMed], [Web of Science ®], [Google Scholar]
- Nici L, Donner C, Wouters E, et al. American thoracic society/European respiratory society statement on pulmonary rehabilitation. Am J Respir Crit Care Med. 2006;173(12):1390–1413. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Modi AC, Cassedy AE, Quittner AL, et al. Trajectories of adherence to airway clearance therapy for patients with cystic fibrosis. J Pediatr Psychol. 2010;35(9):1028–1037. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Oates GR, Stepanikova I, Gamble S, Gutierrez HH, Harris WT. Adherence to airway clearance therapy in pediatric cystic fibrosis: socioeconomic factors and respiratory outcomes. Pediatr Pulmonol. 2015;50(12):1244–1252. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Wilbur J, Chandler P, Miller AM. Measuring adherence to a women's walking program. West J Nurs Res. 2001;23(1):8–24; discussion -32. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Donesky-Cuenco D, Janson S, Neuhaus J, Neilands TB, Carrieri-Kohlman V. Adherence to a home-walking prescription in patients with chronic obstructive pulmonary disease. Heart Lung. 2007;36(5):348–363. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Dajczman E, Wardini R, Kasymjanova G, Prefontaine D, Baltzan MA, Wolkove N. Six minute walk distance is a predictor of survival in patients with chronic obstructive pulmonary disease undergoing pulmonary rehabilitation. Can Respir J. 2015;22(4):225–229. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Sampson R, Morenoff J, Gannon-Rowley T. Assessing “Neighborhood Effects”: social processes and new directions in research. Annu Rev Sociol. 2002(28):443–478. [Crossref], [Web of Science ®], [Google Scholar]
- Diez-Roux AV, Kiefe CI, Jacobs DR Jr., et al. Area characteristics and individual-level socioeconomic position indicators in three population-based epidemiologic studies. Ann Epidemiol. 2001;11(6):395–405. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Barber S, Hickson DA, Kawachi I, Subramanian SV, Earls F. Neighborhood disadvantage and cumulative biological risk among a socioeconomically diverse sample of African American adults: an examination in the Jackson Heart Study. J Racial Ethn Health Disparities. 2016;3(3):444–456. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Blanc PD, Yen IH, Chen H, et al. Area-level socio-economic status and health status among adults with asthma and rhinitis. Eur Respir J. 2006;27(1):85–94. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Diez Roux AV, Merkin SS, Arnett D, et al. Neighborhood of residence and incidence of coronary heart disease. N Engl J Med. 2001;345(2):99–106. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Bird CE, Seeman T, Escarce JJ, et al. Neighbourhood socioeconomic status and biological ‘wear and tear’ in a nationally representative sample of US adults. J Epidemiol Commun Health. 2010;64(10):860–865. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- King KE, Morenoff JD, House JS. Neighborhood context and social disparities in cumulative biological risk factors. Psychosom Med. 2011;73(7):572–579. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Clark CR, Ommerborn MJ, Hickson DA, et al. Neighborhood disadvantage, neighborhood safety and cardiometabolic risk factors in African Americans: biosocial associations in the Jackson Heart study. PLoS One. 2013;8(5):e63254. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Rogers WH. Regression standard errors in clustered samples. Stata Technical Bulletin Reprints. 3. College Station, TX: Stata Press; 1993. p. 88–94. [Google Scholar]
- Mackey K, Parchman ML, Leykum LK, Lanham HJ, Noel PH, Zeber JE. Impact of the chronic care model on medication adherence when patients perceive cost as a barrier. Prim Care Diabetes. 2012;6(2):137–142. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Krauskopf K, Federman AD, Kale MS, et al. Chronic obstructive pulmonary disease illness and medication beliefs are associated with medication adherence. COPD. 2015;12(2):151–164. [Taylor & Francis Online], [Google Scholar]
- Tottenborg SS, Lange P, Johnsen SP, Nielsen H, Ingebrigtsen TS, Thomsen RW. Socioeconomic inequalities in adherence to inhaled maintenance medications and clinical prognosis of COPD. Respir Med. 2016;119:160–167. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Kale MS, Federman AD, Krauskopf K, et al. The association of health literacy with illness and medication beliefs among patients with chronic obstructive pulmonary disease. PLoS One. 2015;10(4):e0123937. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Gaalema DE, Savage PD, Rengo JL, et al. Patient characteristics predictive of cardiac rehabilitation adherence. J Cardiopulm Rehabil Prev. 2017;37(2):103–110. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Williams DR. Race/ethnicity and socioeconomic status: measurement and methodological issues. Int J Health Serv. 1996;26(3):483–505. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Pearlin L. The sociological study of stress. J Health Soc Behav. 1989;30(3):241–256. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Zilioli S, Imami L, Slatcher RB. Socioeconomic status, perceived control, diurnal cortisol, and physical symptoms: a moderated mediation model. Psychoneuroendocrinology. 2017;75:36–43. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Charlson ME, Wells MT, Peterson JC, et al. Mediators and moderators of behavior change in patients with chronic cardiopulmonary disease: the impact of positive affect and self-affirmation. Transl Behav Med. 2014;4(1):7–17. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Sgalla G, Cerri S, Ferrari R, et al. Mindfulness-based stress reduction in patients with interstitial lung diseases: a pilot, single-centre observational study on safety and efficacy. BMJ Open Respir Res. 2015;2(1):e000065. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Patel MR, Kruger DJ, Cupal S, Zimmerman MA. Effect of financial stress and positive financial behaviors on cost-related nonadherence to health regimens among adults in a community-based setting. Prev Chronic Dis. 2016;13:E46. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Sawyer SM, Aroni RA. Sticky issue of adherence. J Paediatr Child Health. 2003;39(1):2–5. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Butterworth SW. Influencing patient adherence to treatment guidelines. J Manag Care Pharm. 2008;14(6 Suppl B):21–24. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Kreps GL, Villagran MM, Zhao X, McHorney CA, Ledford C, Weathers M, et al. Development and validation of motivational messages to improve prescription medication adherence for patients with chronic health problems. Patient Educ Couns. 2011;83(3):375–381. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Peterson AM, Takiya L, Finley R. Meta-analysis of trials of interventions to improve medication adherence. Am J Health Syst Pharm. 2003;60(7):657–665. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Kahana S, Drotar D, Frazier T. Meta-analysis of psychological interventions to promote adherence to treatment in pediatric chronic health conditions. J Pediatr Psychol. 2008;33(6):590–611. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- McDonald HP, Garg AX, Haynes RB. Interventions to enhance patient adherence to medication prescriptions: scientific review. JAMA. 2002;288(22):2868–2879. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Mihalko SL, Brenes GA, Farmer DF, Katula JA, Balkrishnan R, Bowen DJ. Challenges and innovations in enhancing adherence. Control Clin Trials. 2004;25(5):447–457. [Crossref], [PubMed], [Google Scholar]
- Russell CL, Ruppar TM, Matteson M. Improving medication adherence: moving from intention and motivation to a personal systems approach. Nurs Clin North Am. 2011;46(3):271–281, v. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Almadana Pacheco V, Pavon Masa M, Gomez-Bastero Fernandez AP, Muniz Rodriguez AM, Tallon Moreno R, Montemayor Rubio T. Patient profile of drop-outs from a pulmonary rehabilitation program. Arch Bronconeumol. 2017;53(5):257–262. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Cassidy S, Turnbull S, Gardani M, Kirkwood K. Attendance at pulmonary rehabilitation classes: an exploration of demographic, physiological and psychological factors that predict completion of treatment. Chron Respir Dis. 2014;11(2):95–102. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Gershon AS, Dolmage TE, Stephenson A, Jackson B. Chronic obstructive pulmonary disease and socioeconomic status: a systematic review. COPD. 2012;9(3):216–226. [Taylor & Francis Online], [Google Scholar]
- Oates GR, Jackson BE, Partridge EE, Singh KP, Fouad MN, Bae S. Sociodemographic patterns of chronic disease: how the mid-south region compares to the rest of the country. Am J Prev Med. 2017;52(1S1):S31–S39. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Fan VS, Giardino ND, Blough DK, Kaplan RM, Ramsey SD, Nett Research G. Costs of pulmonary rehabilitation and predictors of adherence in the national emphysema treatment trial. COPD. 2008;5(2):105–116. [Taylor & Francis Online], [Google Scholar]
- Cleland JA, Lee AJ, Hall S. Associations of depression and anxiety with gender, age, health-related quality of life and symptoms in primary care COPD patients. Fam Pract. 2007;24(3):217–223. [Crossref], [PubMed], [Web of Science ®], [Google Scholar]
- Panagioti M, Scott C, Blakemore A, Coventry PA. Overview of the prevalence, impact, and management of depression and anxiety in chronic obstructive pulmonary disease. Int J Chron Obstruct Pulmon Dis. 2014;9:1289–1306. [PubMed], [Web of Science ®], [Google Scholar]