Chronic Conditions and Self-Reported Health in a Medicare Advantage Plan Population

Abstract

Self-reported changes in physical and mental health by members are an important dimension by which the quality of a Medicare Advantage (MA) plan is rated by the Centers for Medicare & Medicaid Services. To better target their interventions, MA plans need a better understanding of what observed characteristics—including clinical health conditions—predict self-reported changes in physical and mental health. This study explored how one MA plan's survey of participants' responses regarding changes in physical and mental health is associated with a set of chronic conditions as well as sociodemographic characteristics. Multinomial logistic regressions were used to examine the influence of 9 chronic conditions and age, sex, race, education, dual eligibility status (Medicare/Medicaid eligible), marital and living status, and assistance with survey completion on changes in patient-reported physical and mental health. Six conditions—dementia (P < 0.001), diabetes (P = 0.003), congestive heart failure (P = 0.002), cerebrovascular disease (P = 0.001), coronary artery disease (CAD) (P < 0.001), and rheumatoid arthritis (P < 0.001)—were associated with self-reported worsening of overall physical health. Four conditions—dementia (P < 0.002), diabetes (P = 0.047), CAD (P = 0.001), and decubitus ulcers (P = 0.033)—were associated with self-reported worsening of overall mental health. Females, married respondents, and those needing assistance with survey completion were more likely to report worsening of their mental health. Enrollees older than age 65 actually were less likely to report worsening of overall mental health. Findings provide insight into which members may be more susceptible to reporting that their physical or mental health is worsening.

Introduction

In 2007, the centers for Medicare & Medicaid Services (CMS) introduced the Five-Star Quality Rating System for Medicare Advantage (MA) plans as part of an effort to help educate consumers on the quality of health plans. The Patient Protection and Affordable Care Act of 2010 has since tied federal reimbursement rates for insurance carriers administering MA plans to performance as measured by the CMS Five-Star Quality Rating System. One of the key components of the Star Rating System is self-reported changes in physical and mental health by MA plan members, which CMS collects using surveys.¹ The Consumer Assessment of Healthcare Providers and Systems (CAHPS) and the Health Outcomes Surveys (HOS) are administered to MA plan members annually by CMS and are a primary source of self-reported data for the Star program. Of these, self-reported responses to HOS questions designed to measure perceptions of physical and mental health status are among the triple-weighted outcome measures used by CMS to construct Star Ratings.

The ability to predict which members are most susceptible to feeling that their physical and mental health is worsening over time is useful to MA plans in their effort to design interventions efficiently to manage these members' chronic conditions. Thus, it may be hypothesized that chronic health conditions, defined as “those conditions that last a year or more and require ongoing medical attention and/or limit activities of daily living,”² may play a notable role in member perceptions of their own physical and mental health. This is an especially relevant area of research because the prevalence of chronic conditions is projected to increase over the next decades.³ It has been estimated that people in the United States who report 1 chronic medical condition will increase from 133 million in 2005 to 164 million in 2020, with 81 million reporting 2 or more conditions.^3,4 Among the elderly, hypertension, heart disease, diabetes (DM), arthritis, chronic pulmonary disease, osteoporosis, and cancer are the most common chronic conditions.^3,5

This study used data collected by a large MA plan from an in-house survey that closely mirrors the HOS and CAHPS survey, combined with patient claims data, and analyzed how the following chronic conditions—DM, decubitus ulcers, congestive heart failure (CHF), coronary artery disease (CAD), chronic kidney disease (CKD), cerebrovascular disease (CVD), chronic obstructive pulmonary disease (COPD), arthritis, and dementia—and sociodemographic characteristics (ie, age, sex, race/ethnicity, education level, dual status, living and marital situation, needing assistance with survey completion) predict whether a respondent will report declines, improvements, or no change in his or her overall physical health and mental health.

Methods

Data used in this study were in-house observational data that were originally collected by an MA health plan in southeast Louisiana as part of the organization's annual patient satisfaction survey process. The survey was administered by a third-party vendor. The survey instrument comprised 52 items that were adapted from the CAHPS survery and HOS. As is required for MA plans, the organization's survey tool was submitted to CMS for approval before use.

This study was reviewed and approved as non–human subject-research by the University of Alabama at Birmingham Institutional Review Board.

Population and sample

The process of survey administration by the third-party vendor was as follows: A prenotification letter was mailed to randomly selected respondents 1 week before mailing the surveys. The survey instrument was mailed with a cover letter explaining the purpose and significance of the survey. A return business reply envelope addressed to an external third party was included with each questionnaire. Neither the organization nor the third-party vendor performed any follow-up calls and each randomly selected respondent received only 1 survey. Data were collected at 3 points in time: September 2011, October 2011, and January 2012. A 99% confidence level was used to determine the sample size at ±5% accuracy. Fifteen thousand surveys were mailed to members in 3 separate 5000-survey increments. A total of 4752 surveys were returned representing a 32% response rate. There were no duplicate responses in the data sample. Data from these 3 satisfaction surveys were combined into 1 data set by a third party.

The satisfaction survey was administered randomly to individual members who were in the MA plan's network of primary care practices. The survey targeted currently enrolled members who were aged 18 years and older at the time of the sample draw and had been continuously enrolled in the health plan for at least 6 months. Member characteristics (age, sex, race, educational level, marital status, length of time enrolled in the health plan, assistance with completing the survey, living situation, total number of chronic conditions, and total number of prescription medications) from the organization's claims and enrollment database were linked to the survey data. This allowed the unique opportunity to study the association between objective data on clinical health status obtained from claims data and self-reported physical and mental health using a survey instrument similar to HOS and CAHPS.

The chronic hierarchical condition categories (HCCs) as defined by CMS were calculated using a 3-year look-back period in the organization's administrative claims database. Two questions asked the respondents to rate their overall physical health and overall mental health.

(1) Compared to 1 year ago, how would you rate your physical health, in general, now?

(2) Compared to 1 year ago, how would you rate your emotional problems (such as feeling anxious, depressed, or irritable), in general, now?

Respondents rate their answers to these questions on a 5-point Likert scale: much worse, slightly worse, about the same, slightly better, and much better.

Data management

Data management and analysis were performed using SPSS statistical software version 21 (IBM Corp., Armonk, NY).⁶ Before analysis of the data, each survey question was evaluated for out-of-range values and eligibility criteria. This resulted in 51 cases that did not meet eligibility criteria and 82 out-of-range responses. The ineligible cases and individual out-of-range responses were eliminated from the analysis. In addition, 4 survey questions (ie, self-reported number of chronic conditions, self-reported number of prescription medications, self-rating of physical health, self-rating of mental health) were evaluated for missing responses; 375 cases with missing responses to these questions were excluded from the data set. Responders were compared to nonresponders to assess response bias.

Empirical approach

Respondents' characteristics and distribution of chronic conditions were summarized using descriptive statistics and are shown in Tables 1 and 2.

Table 1.

Descriptive Statistics for Independent Variables

Variable	N	Mean	SD
Age	4325	76.3	8.0
Plan years	4325	9.2	2.8
Median age	4325	Percentage
<65 years	307	7.1
≥65 years	4018	92.9
Race	4237
White	3119	72.1
Black	966	22.3
All others	152	3.5
Marital status	4280
Married	2043	47.2
Not married	2237	51.8
Sex	4270
Male	1874	43.3
Female	2396	55.4
Living status	4263
Alone	1397	32.3
Not alone	2866	66.3
Assistance with completing survey	4291
Yes	632	14.6
No	3659	84.6
Education	4232
High school and above	3137	72.0
All others	1188	28.0
Plan benefit	4314
Dual eligibility	734	17.0
Not dual eligibility	3580	82.8

SD, standard deviation.

Table 2.

Frequency of Chronic Conditions (%) (N = 4325)

	DM (%)	COPD (%)	CVD (%)	CHF (%)	CAD (%)	CKD (%)	RA (%)	Decubitus ulcers (%)	Dementia (%)
HCC	42.2	25.7	28.7	19.8	18.7	29.9	7.5	1.6	4.4
No HCC	57.8	74.3	71.3	80.2	81.3	70.1	92.5	98.4	95.6

CAD, coronary artery disease; CHF, congestive heart failure; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVD, cerebrovascular disease; DM, diabetes mellitus; HCC, hierarchical condition category; RA, rheumatoid arthritis.

The main outcomes of interest were respondent-reported changes in their physical health and mental health over the past year. For each, the dependent variable was recoded into 3 categories: worse, which included slightly worse and much worse; better, which included slightly better and much better; and about the same, which was selected as the reference category. Multinomial logistic regressions were used to analyze the relationship of each of the 2 dependent variables and the chronic conditions, with controls for age, race, sex, educational level, marital and living status, needing assistance with survey completion, and dual eligibility status. The chronic conditions of interest were treated as independent variables in the logistic regression model and were coded as a 1 or 0 with a value of 1 when the condition was present and a value of 0 otherwise. All comorbidities were considered simultaneously in the logistic regression model. The predictors X₁–X₁₄ are DM, CHF, CVD, COPD, CKD, CAD, rheumatoid arthritis (RA), decubitus ulcers, dementia, sex, age, race, education, and dual status. All inferential statistical analyses used a level of α = 0.05 as a determinant of statistical significance.

Results

The final sample consisted of 4325 responses. Females represented more than half of the total responses (Table 1). The majority of respondents were white with a mean age of 76.3 years. The mean length of time respondents were enrolled with the health plan was 9.2 years. In all, 41% of respondents reported completing a high school education and 12.3% of respondents reporting completing less than an eighth grade education. In addition, reported statistics for nonresponders revealed that 55.9% were female, 62% were white, with a mean age of 77 years, and the mean number of years with the plan of just over 10 years. Descriptive characteristics are summarized in Table 1. The frequency of the chronic conditions was derived from administrative claims data and is presented in Table 2.

Multinomial logistic regression analysis

This exploratory analysis evaluated the likelihood of an association between the specific chronic conditions as defined by the organization's claims data using HCCs and the self-reported responses to respondents' perception of changes in their physical and mental health compared to 1 year ago.

The results of the multinomial logistic regression analysis (Table 3) revealed that the following chronic conditions were significantly associated with higher odds of respondents reporting a worsening of their physical health compared to remaining about the same: dementia (P < 0.001), DM (P = 0.003), CHF (P = 0.002), CVD (P = 0.001), CAD (P < 0.001), and RA (P < 0.001). Respondents reporting needing assistance with survey completion were more likely to report their physical health as worsening (P < 0.001). Respondents older than age 65 were less likely to report a worsening of their physical health (P = 0.001). Results also revealed that the following chronic conditions were significantly associated with higher odds of respondents reporting improvement in physical health compared to about the same: DM (P = 0.006) and CHF (P = 0.001). With respect to the other sociodemographic predictors, being older than age 65 (P = 0.046) and dual eligibility (Medicare/Medicaid) status (P = 0.030) were associated with higher odds of reporting a better rating of physical health compared to about the same. Nonminority status (P < 0.001), an education level of high school and above (P = 0.001), being female (P = 0.029), and needing assistance with survey completion (P < 0.001) were significantly associated with lower odds of reporting a better rating of physical health compared to about the same. P values, odds ratios (OR), and 95% confidence intervals are presented in Table 3.

Table 3.

Multinomial Logistic Regression—Rating of Physical Health

			95% Confidence interval for OR
Rating of physical health	OR	P	Lower bound	Upper bound
Worse vs. about the same
Diabetes	1.338	0.003	1.107	1.618
CHF	1.453	0.002	1.149	1.838
CVD	1.427	0.001	1.164	1.750
COPD	1.186	0.110	0.962	1.462
CKD	1.124	0.265	0.915	1.380
CAD	1.581	0.000	1.255	1.993
RA	1.917	0.000	1.408	2.611
Decubitus ulcers	1.440	0.280	0.743	2.788
Dementia	2.094	0.000	1.440	3.044
Female sex	1.188	0.089	0.974	1.499
Age >65	0.573	0.001	0.415	0.791
Nonminority race	0.924	0.489	0.738	1.156
High school and above	0.986	0.900	0.793	1.226
Dual eligibility	1.179	0.195	0.919	1.512
Married	1.083	0.526	0.846	1.388
Living alone	0.964	0.777	0.747	1.244
Assistance with survey completion	1.609	0.000	1.252	2.067
Diabetes	1.220	0.006	1.058	1.405
CHF	1.365	0.001	1.132	1.646
CVD	1.104	0.225	0.941	1.294
COPD	1.064	0.450	0.905	1.252
CKD	0.963	0.640	0.824	1.127
CAD	1.105	0.299	0.915	1.336
Better vs. about the same
RA	1.136	0.359	0.865	1.490
Decubitus ulcers	1.333	0.324	0.753	2.361
Dementia	1.092	0.639	0.755	1.579
Female sex	0.850	0.029	0.735	0.983
Age >65	1.341	0.046	1.006	1.787
Nonminority race	0.448	0.000	0.383	0.524
High school and above	0.771	0.001	0.658	0.905
Dual eligibility	1.232	0.030	1.021	1.487
Assistance with survey completion	0.626	0.000	0.504	0.777
Married	1.027	0.780	0.850	1.241
Living alone	1.090	0.375	0.901	1.318

Reference category is about the same.

The following chronic conditions were significantly associated with higher odds of respondents reporting a worsening of their mental health compared to about the same: dementia (P = 0.002), DM (P = 0.047), CAD (P = 0.001), and decubitus ulcers (P = 0.033) (Table 4). Being female (P = 0.002), married (P = 0.017), and needing assistance with survey completion (P < 0.001) also were associated with higher odds of reporting their mental health as worsening. Being older than age 65 (P < 0.001) was associated with lower odds of categorizing one's mental health as worse versus about the same compared to the reference group. One chronic condition was significantly associated with higher odds of respondents reporting their mental health as better: CVD (OR 1.219, P = 0.013). Being older than age 65 (P < 0.039) and nonminority race (P < 0.001) were associated with higher odds of reporting better mental health, while assistance with survey completion (P < 0.001) and an education level of high school and above (P < 0.001) were associated with lower odds of reporting better mental health. Detailed results are presented in Table 4.

Table 4.

Multinomial Logistic Regression—Rating of Mental Health

			95% Confidence interval for OR
Rating of mental health	OR	P	Lower bound	Upper bound
Worse vs. about the same
Diabetes	1.230	0.047	1.003	1.508
CHF	1.171	0.228	0.906	1.513
CVD	1.200	0.110	0.959	1.501
COPD	1.136	0.272	0.905	1.425
CKD	1.037	0.749	0.829	1.297
CAD	1.535	0.001	1.197	1.968
RA	1.356	0.081	0.963	1.908
Decubitus ulcers	2.076	0.033	1.059	4.069
Dementia	1.856	0.002	1.263	2.726
Female sex	1.403	0.002	1.134	1.737
Age >65	0.326	0.000	0.238	0.447
Nonminority race	0.877	0.280	0.691	1.113
High school and above	0.871	0.241	0.691	1.098
Dual eligibility	0.884	0.382	0.672	1.165
Married	1.393	0.017	1.062	1.826
Living alone	1.173	0.265	0.886	1.552
Assistance with survey completion	1.562	0.001	1.200	2.034
Diabetes	1.129	0.090	0.981	1.300
CHF	1.119	0.232	0.931	1.345
CVD	1.219	0.013	1.042	1.425
COPD	1.024	0.774	0.872	1.202
Better vs. about the same
CKD	0.932	0.373	0.789	1.088
CAD	1.017	0.860	0.844	1.226
RA	1.122	0.424	0.857	1.445
Decubitus ulcers	1.615	0.095	0.921	2.831
Dementia	0.853	0.404	0.588	1.238
Female sex	0.950	0.486	0.822	1.098
Age >65	1.366	0.039	1.015	1.838
Nonminority race	1.081	0.000	0.370	0.505
High school and above	0.735	0.000	0.628	0.861
Dual eligibility	1.081	0.409	0.898	1.300
Assistance with survey completion	0.500	0.000	0.401	0.623
Married	1.042	0.668	0.864	1.256
Living alone	1.133	0.191	0.940	1.367

Discussion

This study used in-house survey data similar to those collected in HOS and CAHPS surveys by CMS. The survey is then linked to in-house claims data so as to better understand what factors predicted enrollees' reporting changes in their physical and mental health status. The health burden of chronic conditions and self-perceptions about one's own physical health and mental health have been described as having “multidimensional and reciprocal relationships,”⁷ especially among older adults.^8
–10 In this study, 6 chronic conditions were associated with higher odds of reporting one's physical health as worsening: dementia, DM, CHF, CVD, CAD, and RA. Respondents reporting needing assistance with survey completion were associated with higher odds of reporting their physical health as worsening. Four chronic conditions were associated with higher odds of reporting one's mental health as worsening: DM, CAD, dementia and decubitus ulcers. Respondents who were female, married, or reported needing assistance with survey completion also were more likely to report their mental health as worsening.

Some of the findings appear counterintuitive; for example, results demonstrated that DM and CHF were significantly associated with rating one's physical health as better and worse versus about the same. Furthermore, being older than age 65 was associated with lower odds of reporting worse mental and physical health, whereas higher levels of education were associated with lower odds of reporting better physical and mental health. However, extant studies indicate that there are a number of factors that relate to one's self-perception of health and functional status, including expectations as well as optimism and acceptance. Individuals with chronic conditions can have a long period of time to develop a conceptual belief about their conditions through their own experiences as well as the experiences of others with similar conditions. Idler et al researched the chronic illness models that have implications for self-rated health.¹¹ These include the disease “identity, timeline, causes, consequences, and controllability”—all of which contribute to one's concept of well-being. In the 2009 research article, Self-Comparisons Affect Responses to Health-Related Quality of Life Questionnaires, Robertson concluded 2 key points: (1) some chronic conditions that have symptoms that are “expected” may be perceived to be more manageable and therefore may change the parameters for how an individual responds to questions related to physical health and (2) optimism may be a key factor in evaluating one's health more positively as the elderly accept the limitations resulting from increasing age and chronic conditions.¹² Another potential factor is the unpredictability of symptoms patients can experience with some chronic conditions. A downward progression of physical symptoms that is slower than what the patient expected may be perceived as better physical health, only to be followed by an acute medical crisis, which is reported as a worsening condition. These findings demonstrate the challenges of determining the underlying constructs respondents use when self-reporting on health-related surveys.¹³

With respect to one's overall rating of mental health, it has been estimated that 15% of Americans older than age 65 suffer from significant depressive symptoms.¹⁴ Harman reports that several studies have shown that when depression “co-occurs” with chronic conditions such as DM and CAD, the resulting implications are poorer rates of treatment adherence and poorer health outcomes.¹⁴ Dementia, frequently a long-term chronic condition, may prevent an individual from participating in everyday activities that help maintain cognitive functioning, which also may be a contributing factor in self-reported worsening of mental health.

Findings from this study may have strategic implications for MA plans as they continue to design care coordination models to improve members' quality of life and reduce health care expenditures. Organizations need to understand the factors that influence members' perceptions of their physical and mental health to improve outcomes and remain competitive. As organizations move to a population health focus, the ability to stratify the enrolled population not just by predicted expense but also by social determinants of health and the ability to perform activities of daily living (ADLs)—which in this study were captured by the proxy variable of “requiring assistance to complete daily tasks”—becomes imperative in allocating limited resources to improve outcomes. Annual evaluation of an enrollee's health status utilizing an instrument such as a health risk assessment tool can provide valuable information on ADLs and instrumental ADLs.¹⁵ At the same time, this study's results suggest that, when it comes to self-reported health, some of the observable indicators of social determinants of health (eg, education, marital status, age) may not influence self-perceptions of physical and mental health in the anticipated direction. The adoption of predictive modeling tools can assist an organization to identify those enrollees in low-, moderate-, and high-risk categories of reporting worsening health based on sociodemographic characteristics, chronic conditions, and other observable factors—and allow targeted care coordination interventions to improve or, at a minimum, maintain enrollees' health status.

In addition to stratifying the population, assessing and understanding enrollees' level of engagement in their health care is an important factor in improving outcomes. Utilization of the Patient Activation Measure (PAM-13) tool can give providers and care coordinators insight on how to more effectively enhance an individual's level of self-activation and engagement in the care process. Respondents' self-reported assessment of their health includes self-knowledge, knowledge of conditions, and the interaction between chronic conditions. Greater patient-level self-activation—defined as “understanding one's own role in the care process and having the knowledge, skills, and confidence to take on that role”¹⁶—may give patients a better sense of control, realistic expectations, and a more positive outlook regarding their own health.

Another factor that may moderate the relationship between multiple chronic conditions and self-perceptions of health may be better care coordination among various providers because patients with multiple chronic conditions often experience suboptimal health care.¹⁷ This requires a shift from a patient receiving care for a specific chronic condition from a specialist in isolation to coordination between complex treatment regimens.¹⁸

Limitations

The study was limited to survey respondents who were enrolled in an MA plan in southeastern Louisiana; therefore, generalizability beyond the specific geographic region should be considered carefully. The survey used self-reported assessments of respondents' overall physical health and overall mental health but lacked data on the social support systems of respondents and the quality of their interactions with health care providers. The possibility of response bias may serve as a limitation to the accuracy of self-reports. Using claims data presents a limitation in that the data are only as complete as the health care providers' accuracy of submitted claims to the health plan, and there is potential for underreporting. In addition, this study lacked information about the severity of chronic conditions as well as the length of time since initial diagnosis, which also may be important factors in self-perceptions of health.

Conclusion

The significant growth of the aging population will require the health care industry and the organizations that care for the elderly to further their understanding of the physical and psychological ways in which people adapt to aging, to changes in their sociodemographic situation, and to chronic conditions. MA plans should more broadly consider collecting survey data from members using survey instruments similar to what is used by CMS, and linking it to claims data and clinical history to receive insights into what factors are linked to members being more susceptible to perceptions of worsening physical or mental health. This helps MA plans to develop targeted interventions that have the potential advantage of improving members' quality of life, which in turn can help lead to improvements of the MA plans' Star Ratings.

Areas for future research include developing a more complete understanding of what range of factors contribute other attributes respondents consider when reporting their health status and determining if those considerations change over time and with age. With an aging population who are living longer with more chronic conditions, developing a more comprehensive understanding of the physical and emotional impacts of chronic conditions occurring over time will be important considerations for MA plans as they make strategic business decisions related to the components of their delivery models. In addition, one policy suggestion is for CMS to consider developing a risk adjustment model that examines the potential effects of severity of illness for survey respondents, as the burden of chronic conditions continues to increase in the population.

Footnotes

Author Disclosure Statement

Drs. Guerard, Omachonu, Hernandez, and Sen declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Dr. Guerard is employed at Peoples Health; Drs. Omachonu and Sen are independent contractors for Peoples Health. Dr. Hernandez declared no conflicts of interest. The authors received the following financial support: This study was partially supported by Peoples Health, a Medicare Advantage plan in southeast Louisiana. The data used in this work are the property of the Peoples Health organization. The organization had no role in the design or conduct of the study or approval of the article. The content is solely the responsibility of the authors and does not represent the official views of the organization.

References

Jones

III , Jones

, Miller

. The Medicare health outcomes survey program: overview, context, and near-term prospects. Health Qual Life Outcomes, 2004; 2:33–43.

Hwang

, Weller

, Ireys

, Anderson

. Out-of-pocket medical spending for care of chronic conditions. Health Aff (Millwood), 2001; 20:267–278.

Anderson

, Horvath

. The growing burden of chronic disease in America. Public Health Rep, 2004; 119:263–270.

Bodenheimer

, Wagner

, Grumbach

. Improving primary care for patients with chronic illness: the chronic care model, Part 2. JAMA, 2002; 288:1909–1914.

Vogeli

, Shields

, Lee

, et al. Multiple chronic conditions: prevalence, health consequences, and implications for quality, care management, and costs. J Gen Intern Med, 2007; 22(Suppl 3):391–395.

IBM Corp.. IBM SPSS Statistics for Windows. Armonk, NY: IBM Corp.; 2011.

Kelley-Moore

, Ferraro

. A 3-D model of health decline: disease, disability, and depression among black and white older adults. J Health Soc Behav, 2005; 46:376–391.

Kempen

, Ormel

, Brilman

, Relyveld

. Adaptive responses among Dutch elderly: the impact of eight chronic medical conditions on health-related quality of life. Am J Public Health, 1997; 87:38–44.

Ormel

, Rijsdijk

, Sullivan

, Van Sonderen

, Kempen

GIJM

. Temporal and reciprocal relationship between IADL/ADL disability and depressive symptoms in late life. J Gerontol B Psychol Sci Soc Sci, 2002; 57:338–347.

10.

Aneshensel

, Frerichs

, Huba

. Depression and physical illness: a multiwave, nonrecursive causal model. J Health Soc Behav, 1984; 25:350–371.

11.

Idler

, Leventhal

, McLaughlin

, Leventhal

. In sickness but not in health: self-ratings, identity, and mortality. J Health Soc Behav, 2004; 45:336–356.

12.

Robertson

, Langston

, Stapley

, et al. Meaning behind measurement: self-comparisons affect responses to health-related quality of life questionnaires. Qual Life Res, 2009; 18:221–230.

13.

Bosworth

, Steinhauser

, Orr

, Lindquist

, Grambow

, Oddone

. Congestive heart failure patients' perceptions of quality of life: the integration of physical and psychosocial factors. Aging Ment Health, 2004; 8:83–91.

14.

Harman

, Edlund

, Fortney

, Kallas

. The influence of comorbid chronic medical conditions on the adequacy of depression care for older Americans. J Am Geriatr Soc, 2005; 53:2178–2183.

15.

Benyamini

, Idler

. Community studies reporting association between self-rated health and mortality. Res Aging, 1999; 21:392–401.

16.

Hibbard

, Stockard

, Mahoney

, Tusler

. Development of the Patient Activation Measure (PAM): conceptualizing and measuring activation in patients and consumers. Health Serv Res, 2004; 39:1005–1026.

17.

Parekh

, Goodman

, Gordon

, Koh

, HHS Interagency Workgroup on Multiple Chronic Conditions. Managing multiple chronic conditions: a strategic framework for improving health outcomes and quality of life. Public Health Rep, 2011; 126:460–471.

18.

Grembowski

, Schaefer

, Johnson

, et al. A conceptual model of the role of complexity in the care of patients with multiple chronic conditions. Med Care, 2014; 52(Suppl 3):S7–S14.