Risk Factors and Precipitants of Severe Disability Among Community-Living Older Persons

Key Points

Question  What are the relative contributions of traditional risk factors and intervening illnesses or injuries (ie, precipitants) associated with severe disability that develops progressively vs catastrophically?

Findings  In this cohort study of 754 nondisabled community-living persons aged 70 years or older who were followed longitudinally for nearly 19 years, the associations of intervening illnesses or injuries with severe disability were much more pronounced than those of the risk factors, with hazard ratios up to 20-fold greater for progressive disability and 177-fold greater for catastrophic disability.

Meaning  The findings of this study suggest that to reduce the burden of severe disability, more aggressive efforts will be needed to prevent and manage intervening illnesses or injuries and to facilitate recovery after these debilitating events.

Importance  Severe disability greatly diminishes quality of life and often leads to a protracted period of long-term care or death, yet the processes underlying severe disability have not been fully evaluated.

Objective  To evaluate potential risk factors and precipitants associated with severe disability that develops progressively (during ≥2 months) vs catastrophically (from 1 month to the next).

Design, Setting, and Participants  Prospective cohort study conducted in greater New Haven, Connecticut, from March 1998 to December 2016, with 754 nondisabled community-living persons aged 70 years or older. Data analysis was conducted from November 2018 to May 2019.

Main Outcomes and Measures  Candidate risk factors were assessed every 18 months. Functional status and potential precipitants, including illnesses or injuries leading to hospitalization, emergency department visit, or restricted activity, were assessed each month. Severe disability was defined as the need for personal assistance with at least 3 of 4 essential activities of daily living. The analysis was based on person-months within 18-month intervals.

Results  The mean (SD) age for the 754 participants was 78.4 (5.3) years, 487 (64.6%) were women, and 683 (90.5%) were non-Hispanic white participants. The incidence of progressive and catastrophic severe disability was 3.5% and 9.7%, respectively, based on 3550 intervals. In multivariable analysis, 6 risk factors were independently associated with progressive disability (≥85 years: hazard ratio [HR], 1.6; 95% CI, 1.1-2.4; hearing impairment: HR, 1.7; 95% CI, 1.0-2.8; frailty: HR, 2.4; 95% CI, 1.6-3.7; cognitive impairment: HR, 2.0; 95% CI, 1.3-3.1; low functional self-efficacy: HR, 1.8; 95% CI, 1.2-2.8; low peak flow: HR, 1.7; 95% CI, 1.2-2.4), and 4 were independently associated with catastrophic disability (visual impairment: HR, 1.4; 95% CI, 1.1-1.8; hearing impairment: HR, 1.3; 95% CI, 1.0-1.7; poor physical performance: HR, 1.8; 95% CI, 1.3-2.5; low peak flow: HR, 1.3; 95% CI, 1.0-1.7). The associations of the precipitants were much more pronounced than those of the risk factors, with HRs as high as 321.4 (95% CI, 194.5-531.0) for hospitalization and catastrophic disability and 48.3 (95% CI, 31.0%-75.4%) for hospitalization and progressive disability. Elimination of an intervening hospitalization was associated with a decrease in the risk of progressive and catastrophic severe disability of 3.0% (95% CI, 3.0%-3.1%) and 12.3% (95% CI, 12.1%-12.5%), respectively. Risk differences were 0.6% (95% CI, 0.6%-0.6%) and 1.3% (95% CI, 1.3%-1.4%) for emergency department visit and 0.1% (95% CI, 0.1%-0.2%) and 0.4% (95% CI, 0.4%-0.4%) for restricted activity, and ranged from 0.1% (95% CI, 0.1%-0.1%) to 0.3% (95% CI, 0.3%-0.3%) for the independent risk factors, for progressive and catastrophic disability, respectively.

Conclusions and Relevance  The findings of this study suggest that whether it develops progressively or catastrophically, severe disability among older community-living adults arises most commonly in the setting of an intervening illness or injury. To reduce the burden of severe disability, more aggressive efforts will be needed to prevent and manage intervening illnesses or injuries and to facilitate recovery after these debilitating events.

The onset of severe disability, defined as the need for personal assistance with 3 or more essential activities of daily living, greatly diminishes quality of life and often leads to a protracted period of long-term care or death.^1-3 These episodes of severe disability are expensive,^4-6 and they are highly stressful to both patients and their caregivers.^7,8

Previous research, based on annual assessments of functional status, has attempted to distinguish between severe disability that develops progressively through an intermediary stage of mild disability and severe disability that develops abruptly or catastrophically.^9-12 This work has provided important insights into potential risk factors for severe disability,¹³ including its 2 major subtypes,^11,12 and has suggested that severe disability, particularly the catastrophic subtype, is commonly precipitated by an intervening illness or injury leading to hospitalization.¹⁰ However, the applicability of these findings to clinical practice may be limited for at least 3 reasons. First, annual assessments cannot fully account for the dynamic nature of disability, which is often characterized by frequent transitions in functional status during much shorter periods.^14,15 Second, temporal precedence between a potential precipitant and functional transition is difficult to establish when assessment intervals are as long as a year. Third, increasing evidence suggests that disability may be precipitated or worsened by less serious illnesses and injuries that do not lead to hospitalization.^16-19

To address these limitations, we used high-quality data from a unique longitudinal study of community-living older persons that includes monthly assessments of functional status and potential precipitants, including emergency department visits, episodes of restricted activity, hospitalizations, and a large array of potential risk factors that were assessed every 18 months for nearly 19 years. Our objective was to evaluate potential risk factors and precipitants associated with severe disability that develops progressively vs catastrophically. The results of this study may inform the development of evidence-based interventions to prevent the onset of severe disability and, hence, to reduce the need for long-term care and associated increases in the cost of care.

Participants were members of the Precipitating Events Project, an ongoing longitudinal study of 754 community-living persons aged 70 years or older who were initially nondisabled in their essential activities of daily living.^20,21 Potential participants were members of a large health plan and were excluded for significant cognitive impairment with no available proxy,²² life expectancy less than 12 months, plans to move out of the area, or inability to speak English. Only 126 of 2735 individuals (4.6%) refused screening, and 754 of 1002 (75.2%) who were eligible agreed to participate and were enrolled from March 1998 to October 1999. The study was approved by the Yale Human Investigation Committee, and all participants provided verbal informed consent. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Data on the candidate risk factors were collected during comprehensive home-based assessments, whereas those on disability and potential precipitants were obtained from monthly telephone interviews and Medicare claims. The comprehensive assessments were completed by trained nurse researchers at baseline and every 18 months (completion rate, approximately 95%), whereas the telephone interviews were completed by a separate team of researchers through December 2016 (completion rate, approximately 99%). For participants who had significant cognitive impairment or were otherwise unavailable, we interviewed a proxy informant using a rigorous protocol.²² Participants were asked to identify their race/ethnicity primarily for descriptive purposes. Deaths were ascertained by review of obituaries, from an informant during a subsequent interview, or both.

In addition to demographic factors, we considered risk factors from 4 domains found to be associated with disability in previous studies.²³ The health-related factors included 9 self-reported physician-diagnosed chronic conditions,¹⁵ vision,²⁴ hearing,²⁵ and frailty.²⁶ The cognitive-psychosocial factors included cognitive status,²⁷ depressive symptoms,²⁸ functional self-efficacy,²⁹ and social support.³⁰ The behavioral factors included smoking and body mass index.³¹

The physical capacity factors included a modified version of the Short Physical Performance Battery.^32,33 The other physical capacity factors included nondominant limb, upper and lower extremity muscle strength,³³ manual dexterity,³⁴ gross motor coordination,³⁴ and peak expiratory flow.³⁵ To enhance clinical interpretability and facilitate calculation of risk differences, the risk factors were dichotomized with accepted cut points. Additional operational details are provided in Table 1.^{26,32,33,36,37}

Potential precipitants included intervening illnesses or injuries leading to hospitalization, emergency department (ED) visit, or restricted activity. The primary source of information on hospitalizations and ED visits was linked Medicare claims data, which were available for nearly all hospitalizations and for ED visits among fee-for-service participants.¹⁵ For periods when participants had managed Medicare, hospitalizations were ascertained with Medicare Provider Analysis and Review files, whereas information on ED visits and some hospitalizations was obtained during the monthly interviews. Participants were asked whether they had visited an ED or stayed overnight in a hospital since the last interview. The accuracy of this information was high.^16,26 Hospitalizations were subsequently classified as critical illness, major surgery, or neither, as previously described.^38,39 For descriptive purposes, the reasons for ED visits and hospitalizations were subsequently grouped into distinct diagnostic categories according to revised versions of a previously described protocol.^10,16,17

In accordance with a standardized protocol with high reliability, participants were asked 2 questions related to restricted activity to ascertain less potent precipitants.²⁰ Participants who answered yes to 1 or both questions were asked to identify the reason(s) for their restricted activity. We have previously shown that older persons usually attribute their restricted activity to several concurrent health-related problems.²⁰

The potential precipitants were organized into 3 mutually exclusive hierarchic categories: hospitalization, ED visit but no hospitalization, and restricted activity but no hospitalization or ED visit.

Complete details regarding the assessment of disability, including reliability and accuracy, are provided elsewhere.^21,22 Each month, participants were asked, “At the present time, do you need help from another person to [complete the task]?” for the 4 following essential activities: bathing, dressing, walking, and transferring. Disability was operationalized as the need for personal assistance, and the severity of disability was denoted by the number of disabled activities (0-4) in a specific month. Disability in 1 to 2 activities was considered mild, whereas disability in 3 to 4 activities was considered severe.^14,21 For participants with major cognitive impairment, the monthly interviews were completed with a designated proxy.

The analysis was based on person-months within 18-month intervals. Participants could contribute more than one 18-month interval (ie, observation) to the analysis, but they had to be nondisabled and living in the community at the start of each interval. Of the 4702 person-intervals through December 2016, 3550 (75.5%) were eligible. For each observation, we evaluated time to first occurrence of severe disability from the start of the interval. Severe disability was classified as catastrophic if it developed from 1 month to the next and progressive if it developed during 2 or more months through an intermediary stage of mild disability.⁹ In accordance with previous work,⁴⁰ we evaluated the association between the potential precipitants and onset of severe disability (progressive and catastrophic) monthly.

The characteristics of the candidate risk factors, assessed at the start of each 18-month interval, were summarized according to type of severe disability per interval (ie, progressive, catastrophic, or neither). For each of these 3 groups, exposure to the potential precipitants was calculated per 100 person-months with an intercept-only Poisson model with generalized estimating equations and autoregressive correlation.

We used pooled logistic regression with a complementary log-log link and generalized estimating equations with robust standard errors based on an autoregressive correlation structure to evaluate the bivariate and multivariable relationships between candidate risk factors and potential precipitants, respectively, and discrete time (in months from the start of the interval) to first occurrence of either of the 2 severe disability outcomes, referred to also as progressive and catastrophic disability. This analytic strategy permits the calculation of hazard ratios.^41,42 For the multivariable analyses, we used backward selection with P < .20. The 18-month interval was included as a count variable to account for calendar time. As previously described,^43,44 a series of sensitivity analyses was performed to assess for potential bias caused by the competing risk of death. To enhance clinical interpretability, we calculated risk differences with bootstrapped CIs for each of the statistically significant risk factors and precipitants from the final multivariable models.⁴⁵ The risk difference denotes how much the population-level risk of an outcome would be reduced if the relevant factor were removed. We also ran an alternative form of the final multivariable model that subcategorized hospitalization as critical illness,³⁸ major surgery³⁹ (exclusive of critical illness), or other.

The amount of missing data for the candidate risk factors ranged from 0.3% for smoking to 8.1% for peak flow, with the exception of upper and lower extremity muscle strength, for which 10.2% and 9.7% of the observations were missing, respectively. To address these missing data, we used sequential Markov chain Monte Carlo imputation for multivariate normal data.⁴⁶ Missing data were not imputed for the potential precipitants (0%, 0.4%, and 0% of the observations for hospitalization, ED visits, and restricted activity, respectively).

All analyses were conducted with SAS version 9.4 (SAS Institute). Statistical significance was defined as a 2-tailed P < .05.

The mean (SD) age for the 754 participants was 78.4 (5.3) years, 487 (64.6%) were women, and 683 (90.5%) were non-Hispanic white participants. Severe disability developed in 469 (13.2%) of the 3550 person-intervals. The incidence of catastrophic disability (9.7%) was considerably greater than that of progressive disability (3.5%), whereas the mean (SD) time to develop progressive disability (ie, 10.9 [4.9] months) was modestly longer than that of catastrophic disability (9.0 [5.3] months). The characteristics of the 3 disability groups are provided in Table 1. With few exceptions, the most favorable characteristics were observed for the group without severe disability, whereas the least favorable were observed for the progressive severe disability group.

The bivariate associations of the candidate risk factors with the 2 disability outcomes are also provided in Table 1. For each outcome, the strongest associations were observed for frailty (progressive severe disability: hazard ratio [HR], 6.2; 95% CI, 4.3-8.9; catastrophic severe disability: HR, 3.1; 95% CI, 2.5-3.9), low Short Physical Performance Battery score (progressive severe disability: HR, 5.5; 95% CI, 3.5-8.7; catastrophic severe disability: HR, 3.7; 95% CI, 2.8-4.7), and low functional self-efficacy (progressive severe disability: HR, 4.6; 95% CI, 3.2-6.8; catastrophic severe disability: HR, 2.8; 95% CI, 2.2-3.4). No other factor achieved an unadjusted hazard ratio of 3 or greater for catastrophic disability, and only 3 achieved this level for progressive disability: upper extremity muscle weakness (HR, 3.4; 95% CI, 2.2-5.2), lower extremity muscle weakness (HR, 3.1; 95% CI, 1.9-4.9), and low manual dexterity (HR, 3.2; 95% CI, 2.3-4.6).

As shown in Table 2, the rates of hospitalizations and ED visits were substantially lower in the group without severe disability than in the groups with progressive severe disability and with catastrophic severe disability (hospitalization mean [SD] exposure rate: 2.2 [0.1] vs 10.6 [0.9] vs 10.8 [0.6]; ED visit mean [SD] exposure rate: 1.9 [0.1] vs 5.6 [0.8] vs 3.4 [0.3]), whereas the rate of restricted activity was only modestly lower (mean [SD] exposure rate, 11.5 [0.4] vs 12.3 [1.2] vs 14.3 [0.9]). The bivariate associations between the potential precipitants and 2 disability outcomes were very large, especially for hospitalization (progressive severe disability: HR, 62.9; 95% CI, 40.9-96.7; catastrophic severe disability: HR, 375.3; 95% CI, 228.4-616.5). For each precipitant, the associations were much stronger for catastrophic than progressive severe disability (eg, ED visits: HR, 38.3 [95% CI, 19.2-76.2] vs 12.4 [95% CI, 5.9-26.0]). The reasons for hospitalizations, ED visits, and restricted activity are provided in eTables 1, 2, and 3 in the Supplement. For each disability group, the most common reasons for hospitalization were cardiac (without severe disability: 342 of 1400 [24.5%]; progressive severe disability: 27 of 172 [15.6%]; catastrophic severe disability: 60 of 395 [15.2%]) and infection (without severe disability: 175 [12.5%]; progressive severe disability: 35 [20.2%]; catastrophic severe disability: 67 [17.0%]); for ED visits, musculoskeletal (without severe disability: 279 of 1005 [27.8%]; progressive severe disability: 28 of 74 [37.8%]; catastrophic severe disability: 24 of 89 [27.0%]); and for restricted activity, fatigue (mean [SD] episodes per 100 person-months without severe disability: 61.3 [0.5]; progressive severe disability: 72.8 [0.4]; catastrophic severe disability: 71.0 [0.5]) and pain or stiffness in joints (without severe disability: 38.8 [0.5]; progressive severe disability: 46.2 [0.5]; catastrophic severe disability: 49.0 [0.5]). Differences across the 3 groups were most pronounced for hospitalizations for fall-related injury (without severe disability: 54 [3.9%]; progressive severe disability: 12 [7.0%]; catastrophic severe disability: 52 [13.2%]); for ED visits, cardiac (without severe disability: 105 [10.5%]; progressive severe disability: 10 [13.5%]; catastrophic severe disability: 2 [2.3%]), falls or mobility problems (without severe disability: 48 [4.8%]; progressive severe disability: 11 [14.9%]; catastrophic severe disability: 11 [12.4%]); for restricted activity, dizziness or unsteadiness on feet (mean [SD] episodes per 100 person-months without severe disability: 27.0 [0.4]; progressive severe disability: 45.1 [0.5]; catastrophic severe disability: 34.9 [0.5]).

The Figure provides the multivariable results for the risk factors and precipitants associated with time to onset of progressive and catastrophic severe disability, respectively. In the final models, 6 risk factors were independently associated with progressive disability (≥85 years: HR, 1.6; 95% CI, 1.1-2.4; hearing impairment: HR, 1.7; 95% CI, 1.0-2.8; frailty: HR, 2.4; 95% CI, 1.6-3.7; cognitive impairment: HR, 2.0; 95% CI, 1.3-3.1; low functional self-efficacy: HR, 1.8; 95% CI, 1.2-2.8; low peak flow: HR, 1.7; 95% CI, 1.2-2.4), and 4 were independently associated with catastrophic disability (visual impairment: HR, 1.4; 95% CI, 1.1-1.8; hearing impairment: HR, 1.3; 95% CI, 1.0-1.7; low Short Physical Performance Battery score: HR, 1.8; 95% CI, 1.3-2.5; low peak flow: HR, 1.3; 95% CI, 1.0-1.7). The associations for progressive disability were strongest for frailty and cognitive impairment, whereas those for catastrophic disability were strongest for low Short Physical Performance Battery score and visual impairment. For both types of severe disability, the associations of the precipitants were much more pronounced than those of the risk factors, with the most potent associations observed for hospitalization, followed by ED visit and restricted activity. For example, the adjusted HR was as high as 321.4 (95% CI, 194.5-531.0) for hospitalization and catastrophic disability and 48.3 (95% CI, 31.0%-75.4%) for hospitalization and progressive disability. The results did not change when age was analyzed as a continuous rather than dichotomous variable. In the sensitivity analyses, the associations between the precipitants and 2 severe disability outcomes remained robust to the competing risk of death (eTable 4 and eTable 5 in the Supplement). Severe disability was preceded by a precipitant for 68.0% (85 of 125) of the progressive outcomes and 91.0% (313 of 344) of the catastrophic outcomes.

The corresponding results for risk differences are provided in Table 3. Elimination of an intervening hospitalization was associated with a reduction in the risk of progressive and catastrophic severe disability of 3.0% (95% CI, 3.0%-3.1%) and 12.3% (95% CI, 12.1%-12.5%), respectively. The corresponding values were 0.6% (95% CI, 0.6%-0.6%) and 1.3% (95% CI, 1.3%-1.4%) for ED visit and 0.1% (95% CI, 0.1%-0.2%) and 0.4% (95% CI, 0.4%-0.4%) for restricted activity, respectively. The values for the independent risk factors were much lower than those for hospitalization and ED visit and were comparable to those for restricted activity, ranging from 0.1% (95% CI, 0.1%-0.1%) to 0.3% (95% CI, 0.3%-0.3%).

In the alternative form of the final multivariable model, with no precipitant as the reference group, the adjusted HRs for critical illness, major surgery, and other hospitalizations were, respectively, 133.5 (95% CI, 64.0-278.4), 99.3 (95% CI, 50.6-194.7), and 38.4 (95% CI, 23.9-61.6), respectively, for progressive disability and 1004.6 (95% CI, 563.5-1791.2), 676.8 (95% CI, 394.3-1161.8), and 203.4 (95% CI, 121.9-339.3), respectively, for catastrophic disability.

In this prospective longitudinal study of community-living older persons, we evaluated potential risk factors and precipitants associated with severe disability that develops progressively vs catastrophically. Seven major findings warrant comment. First, the incidence of catastrophic disability was nearly 3 times higher than that of progressive disability. Second, frailty and cognitive impairment were the risk factors most strongly associated with progressive disability, whereas low Short Physical Performance Battery score, an indicator of physical capacity, and visual impairment were the risk factors most strongly associated with catastrophic disability. Third, only 2 risk factors were independently associated with both disability outcomes: hearing impairment and low peak flow. Fourth, associations were considerably stronger for each of the 3 precipitants than any of the risk factors for both outcomes. Fifth, hospitalization was the most potent precipitant, followed by ED visit and restricted activity. Sixth, hospitalizations involving critical illness or major surgery were particularly deleterious. Seventh, risk differences were much lower for the independent risk factors than for hospitalization and ED visit and were comparable to those for restricted activity. These findings suggest that attention to intervening illnesses or injuries will be necessary to prevent the onset of severe disability among older persons and to reduce the resulting need for long-term care.

Prevention of severe disability is important for several reasons. First, the likelihood of recovering independent function is considerably lower for a severe disability than a mild one.^14,21 Second, the onset of severe disability usually necessitates receipt of long-term care services that are expensive and emotionally taxing.^3-6,8 Third, most deaths in older persons are preceded by severe disability.⁴⁷

With few exceptions,¹⁰ previous studies of severe disability have focused on identifying predisposing risk factors.¹³ Although these factors may help to identify older persons who are more susceptible to developing severe disability, their contribution to this outcome is small relative to that of intervening illnesses or injuries, especially those leading to hospitalization. Our results suggest that the most effective strategies for reducing severe disability in older persons will likely include minimizing preventable illnesses and injuries leading to hospitalization,^48-51 decreasing the adverse functional consequences of hospitalization,^52-56 bolstering restorative therapies after hospitalization,^57,58 and substituting hospital at home for traditional inpatient care.⁵⁹ Because fall-related injuries were the third leading cause of hospitalizations, resulting in both progressive and catastrophic disability, fall prevention is a particularly attractive strategy for reducing severe disability,⁴⁸ especially if the results of 2 large pragmatic trials are supportive.^60,61

A unique feature of our study is the availability of data from monthly interviews, which allowed us to more precisely determine the occurrence and severity of disability, more accurately distinguish between progressive and catastrophic disability, and more completely ascertain exposure to potential precipitants (or intervening illnesses and injuries). The frequency of our assessments increases the likelihood that the precipitants preceded the disability outcomes, thereby strengthening temporal precedence and supporting a causal association. We found that severe disability was preceded by a precipitant for nearly all catastrophic outcomes and approximately three-quarters of the progressive outcomes. These values are much higher than those from previous studies that were based on annual assessments of disability and evaluated only hospitalization as a precipitant.¹⁰ As discussed elsewhere,¹⁷ the large HRs reflect not only the deleterious effects of the intervening illnesses or injuries but also the low incidence of disability in the absence of such an event. In addition to ascertaining the reason(s) for each precipitant, we also distinguished hospitalizations that included critical illness and major surgery from other admissions, allowing us to demonstrate a dose-response association across these 3 types of hospital-based precipitants.

This study has limitations. We did not ascertain intervening illnesses or injuries that did not lead to hospitalization, ED visit, or restricted activity. A previous study found no relationship between a series of new or worsening conditions, when ascertained without such thresholds, and decline in physical or cognitive performance.⁶² Because this was an observational study, the reported associations, whether assessed as HRs or risk differences, cannot be construed as causal. Even if the associations were causal, the proportion of severe disability that could be prevented through currently available interventions is uncertain. Because our participants were members of a single health plan in a small urban area and because our analyses were restricted to participants who were nondisabled and living in the community, our findings may not be generalizable to older persons in other settings or those with mild disability. However, generalizability depends not only on the choice of the study sample but also on the stability of the sample over time.⁶³ A strength of our study is the low attrition rate and nearly complete ascertainment of disability and intervening illnesses or injuries. The generalizability of our findings is also enhanced by our high participation rate, which was greater than 75%.

The findings of this study suggest that whether it develops progressively or (especially) catastrophically, severe disability among community-living older persons arises most commonly in the setting of an intervening illness or injury. To reduce the burden of severe disability in an aging society, more aggressive efforts will be needed to prevent and manage intervening illnesses or injuries and to facilitate recovery after these debilitating events.

Accepted for Publication: March 23, 2020.

Published: June 2, 2020. doi:10.1001/jamanetworkopen.2020.6021

Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Gill TM et al. JAMA Network Open.

Corresponding Author: Thomas M. Gill, MD, Department of Internal Medicine, Yale School of Medicine, 874 Howard Ave, New Haven, CT 06519 (thomas.gill@yale.edu).

Author Contributions: Drs Gill and Murphy had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Gill, Murphy.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Gill, Murphy.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Han, Murphy.

Obtained funding: Gill.

Administrative, technical, or material support: Gill, Leo-Summers.

Supervision: Gill.

Conflict of Interest Disclosures: None reported.

Funding/Support: The work for this report was funded by grant R01AG17560 from the National Institute on Aging. The study was conducted at the Yale Claude D. Pepper Older Americans Independence Center (grant P30AG21342 from the National Institute on Aging). Dr Gill is the recipient of Academic Leadership Award K07AG043587s from the National Institute on Aging.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: We thank Denise Shepard, BSN, MBA, Andrea Benjamin, BSN, Barbara Foster, and Amy Shelton, MPH, for assistance with data collection; Wanda Carr and Geraldine Hawthorne, BS, for assistance with data entry and management; Peter Charpentier, MPH, for design and development of the study database and participant tracking system; and Joanne McGloin, MDiv, MBA, for leadership and advice as the project director. Each of these persons was a paid employee of Yale School of Medicine during the conduct of this study.