Hospitalization, Restricted Activity, and the Development of Disability Among Older Persons FREE

Among older persons living in the community, the inability to perform essential activities of daily living (ADLs), such as bathing, dressing, walking inside the house, and transferring from a chair without the assistance of another person, is common, highly morbid, and costly. Despite recent reductions in the prevalence of disability, the number of chronically disabled individuals aged 65 years or older currently exceeds 7 million in the United States.¹ Disability is associated with increased mortality² and leads to additional adverse outcomes, such as nursing home placement and greater use of formal and informal home services,^3- 6 all of which place a substantial burden on older persons, informal caregivers, and health care resources.^7- 9 In the aggregate, the additional cost of medical and long-term care for newly disabled US elderly individuals is estimated to be $26 billion per year.¹⁰ Given the magnitude of these financial costs and the burden on health care resources, interventions to prevent the onset and progression of disability are greatly needed.

An important impediment to the development of preventive interventions is an incomplete understanding of the mechanisms underlying the disabling process.^11- 12 Previous epidemiological studies have focused almost exclusively on identifying vulnerable older persons at risk for disability.^13- 21 Relatively little is known, in contrast, about the role of intervening events that precipitate disability. While recent evidence suggests that disability may occur insidiously, particularly among older persons who are physically frail, most episodes of disability appear to be preceded by a discernible intervening event.²²

In the current study, we performed monthly assessments of intervening events and disability in ADLs for up to 5 years in community-dwelling older adults to evaluate the relationship between intervening events and the development of disability and to determine whether this relationship is modified by the presence of physical frailty, the factor most often associated with disability.

Participants were members of the Precipitating Events Project, a longitudinal study of 754 community-dwelling persons aged 70 years or older, who were not disabled (ie, required no personal assistance) in 4 essential ADLs—bathing, dressing, walking inside the house, and transferring from a chair.²³ Exclusion criteria included significant cognitive impairment with no available proxy,²⁴ inability to speak English, diagnosis of a terminal illness with a life expectancy of less than 12 months, and a plan to move out of the New Haven, Conn, area during the next 12 months.

The assembly of the cohort, which took place between March 1998 and October 1999, has been described in detail elsewhere.^23,25 In brief, potential participants were identified from a computerized list of 3157 age-eligible members from a large health plan in greater New Haven. Eligibility was determined during a screening telephone interview and was confirmed during an in-home assessment. Persons who were physically frail, as denoted by a timed score of greater than 10 seconds on the rapid gait test (ie, walk back and forth over a 10-ft [3-m] course as quickly as possible), were oversampled to ensure a sufficient number of participants at increased risk for disability in ADLs.^15- 16 In the absence of a criterion standard, operationalizing physical frailty as slow gait speed is justified by its high face validity,²⁶ clinical feasibility,^27- 28 and strong epidemiological link to functional decline and disability.^16,29- 30 Only 4.6% of the 2753 health plan members who could be contacted refused to complete the screening telephone interview; 75.2% of the eligible members agreed to participate in the project. Persons who refused to participate did not differ significantly from those who were enrolled in terms of age or sex (data available on request). The study protocol was approved by the human investigational committee at Yale University and all participants provided verbal informed consent.

Comprehensive home-based assessments were completed at baseline, 18 months, and 36 months, while telephone assessments of intervening events and disability in ADLs were completed monthly for up to 5 years (through March 4, 2003). The research staff who completed the monthly interviews were blinded to the results of the home-based assessments, which were completed by a separate team of research nurses. All research staff were unaware of the study aims and hypotheses. Deaths were ascertained by review of local obituaries and/or from an informant during a subsequent telephone interview. During the 5-year follow-up period, 135 (17.9%) participants died after a median follow-up of 26 months and 30 (4.0%) dropped out of the study after a median follow-up of 19.5 months. Data were otherwise available for 99.2% of the 33 938 monthly telephone interviews.

During the comprehensive assessments, data were collected on demographic characteristics, gait speed using the rapid gait test, cognitive status as assessed by the Mini-Mental State Examination,³¹ depressive symptoms as assessed by the Center for Epidemiologic Studies Depression scale,³² and 13 self-reported, physician-diagnosed chronic conditions: hypertension; myocardial infarction; congestive heart failure; stroke; diabetes mellitus; arthritis; hip fracture; fracture of wrist, arm or spine since age 50 years; amputation of leg; chronic lung disease; cirrhosis or liver disease; cancer; and Parkinson disease. Data on these covariates were 100% complete. Participants were considered to be cognitively impaired if they scored less than 24 on the Mini-Mental State Examination³¹ and to have depressive symptoms if they scored 16 or higher on the Center for Epidemiologic Studies Depression scale.^32- 33 Participants were asked by the trained nurse-researcher to identify their race/ethnicity, which was assessed primarily for descriptive purposes.

The intervening events included illnesses and injuries leading to either hospitalization or restricted activity. During the monthly telephone interviews, participants were asked whether they had stayed at least overnight in a hospital since the last interview (ie, during the past month). The accuracy of these reports, based on an independent review of hospital records among a subgroup of 94 participants, was high (κ=0.94).²² Participants who were hospitalized were also asked to provide the primary reason for their admission. These reasons were subsequently grouped into distinct diagnostic categories using a revised version of the protocol described by Ferrucci et al.³⁴ Agreement relative to an independent review of hospital records among a subgroup of 172 admissions was 82%.

To ascertain less potent intervening events, participants were asked 2 questions related to restricted activity using a standardized protocol with high reliability (κ=0.90) as determined during pilot testing (mean interval between assessments was 4.1 days among 20 persons)²³: (1) “Since we last talked on (date of last interview), have you cut down on your usual activities due to an illness, injury or other problem?” and (2) “Since we last talked on (date of last interview), have you stayed in bed for at least half a day due to an illness, injury or other problem?” Participants who answered “yes” to one or both of these questions were considered to have restricted activity during a specific month.²³ These participants were subsequently asked to identify the reason(s) for their restricted activity using a standardized protocol that included 24 prespecified problems and an open-ended response.²³ We have previously demonstrated that older persons usually attribute their restricted activity to several concurrent health-related problems,²³ and that the occurrence of restricted activity (assessed monthly) is strongly associated with decline in ADLs (assessed twice during an 18-month period).³⁵

Complete details regarding the assessment of disability in ADLs, including formal tests of reliability and accuracy, are provided elsewhere.^24- 25 During the monthly telephone interviews, participants were assessed for disability in ADLs (hereafter referred to simply as disability) using standard questions that were identical to those used during the screening telephone interview.²⁴ For each of the 4 essential ADLs, we asked, “At the present time, do you need help from another person to (complete the task)?” Participants who needed help with any of the tasks were considered to be disabled. Participants were not asked about eating, using the toilet, or grooming because the incidence of disability in these 3 ADLs is low among nondisabled, community-dwelling older persons.^15- 16 Furthermore, it is highly uncommon for disability to develop in these ADLs without concurrent disability in bathing, dressing, walking inside the house, or transferring from a chair.^15- 16,36 Among a subgroup of 91 participants who were interviewed twice within a 2-day period by different interviewers, our disability assessment had substantial reliability³⁷ (κ=0.75 for disability in ≥1 of the 4 ADLs). For the 18 paired interviews that were completed independently by different interviewers on the same day, κ was 1.0. For participants with significant cognitive impairment, the monthly telephone interviews were completed with a designated proxy. The accuracy of these proxy reports for disability, as determined during a substudy in which 20 participants who were cognitively intact and their designated proxies were interviewed separately via the telephone each month for 6 months, was excellent (κ=1.0).²⁴

The primary outcome was time to the first occurrence of any disability during the 5-year follow-up period. Because disability for 1 month could be due to measurement error rather than a real change in functional status, we also considered persistent disability defined as a new disability that was present for at least 2 consecutive months²⁴ as a secondary outcome. To enhance the clinical relevance of our study, we also evaluated disability with admission to a nursing home during the same month as a secondary outcome. Information on nursing home admissions was obtained from participants during the monthly telephone interviews.

We evaluated the time to the first occurrence of any disability, persistent disability, and disability with nursing home admission, respectively, according to physical frailty at baseline using the Kaplan-Meier method and the log-rank test for statistical comparisons. Participants who had not developed the relevant disability outcome were censored at the time of death or the last completed interview prior to March 5, 2003.

To distinguish the effects of new intervening events from those of prior events, we defined the exposure period for new events as the month prior to the assessment of disability (t) and the exposure period for prior events as the time from the baseline assessment to 2 months prior to the onset of disability (t – 1) or to a censoring event for participants who did not develop the relevant disability outcome. We calculated the median number of months (per 100 months) of exposure to intervening events (new and prior combined) and compared these values between participants who did and did not develop each of the 3 disability outcomes using the Wilcoxon rank test.

We used the time-dependent Cox proportional hazards method³⁸ to evaluate the bivariate and multivariable relationships between the following independent variables and the development of any disability: age, sex, race/ethnicity, living alone, years of education, chronic conditions having an overall prevalence at baseline of at least 4%, cognitive impairment, depressive symptoms, physical frailty, and new and prior intervening events, which included hospitalization and restricted activity without hospitalization, referred to hereafter as restricted activity only. The proportional hazards assumption was assessed by survival and Schoenfeld residual plots and by inclusion of an interaction term with time and the natural logarithm of time in the model.³⁸ The exact method was used to handle tied outcome times. For the new events, the calculated hazard ratios (HRs) refer to the risk of developing disability at month t + 1 based on exposure to hospitalization or restricted activity only, respectively, during the preceding month (t). The reference group for these analyses included participants without hospitalization or restricted activity during the preceding month. Exposure for the prior events was defined as the number of months with hospitalization and the number of months with restricted activity only, respectively, during the exposure period as previously defined. The results did not differ substantively when exposure for the prior events was defined instead as the proportion of months with hospitalization and restricted activity only. For both the bivariate and multivariable analyses, the time-dependent covariates were updated using data from the 18- and 36-month follow-up assessments. The multivariable analyses were repeated for persistent disability and disability with nursing home admission, respectively, and were subsequently stratified according to the presence of physical frailty at baseline. Formal tests for statistical interactions between baseline physical frailty and the new intervening events were performed.

To help interpret the magnitude of the HRs, we calculated the incidence rates and 95% confidence intervals (CIs) of the 3 disability outcomes among the respective comparison groups for participants without a new intervening event by dividing the number of persons with the outcome by the number of person-months of exposure using a Poisson distribution.³⁹ We also calculated the population-attributable fractions of the 3 disability outcomes for each of the 2 new intervening events from the fully adjusted models according to the following formula⁴⁰: [prevalence of new event × (HR of new event – 1)]/[prevalence of hospitalization × (HR of hospitalization – 1) + prevalence of restricted activity only × (HR of restricted activity only – 1) + 1]. We estimated the prevalence as the number of person-months of exposure to the new event divided by the total number of person-months of observation.⁴¹ The population-attributable fractions represent the proportions of the disability outcomes that would not have occurred in the absence of exposure to each of the 2 new events.⁴²

All statistical tests were 2-tailed and P<.05 indicated statistical significance. All analyses were performed using SAS statistical software (version 8.2, SAS Institute Inc, Cary, NC).

The baseline characteristics of the study participants are shown in Table 1. Compared with participants who were not physically frail, those who were physically frail were older, were more likely to be female, to be living alone, to be cognitively impaired, and to have depressive symptoms. These individuals also were less likely to be non-Hispanic white, had less education, and had a higher prevalence of most chronic conditions. A total of 135 individuals died, of whom 24 (17.8%), 56 (41.5%), and 68 (50.4%) died without having developed any disability, persistent disability, and disability with nursing home admission, respectively.

During the 5-year follow-up period, any disability developed among 417 (55.3%) of the 754 participants, persistent disability in 278 (36.9%), and disability with nursing home admission in 199 (26.4%). Kaplan-Meier curves for the 3 disability outcomes by physical frailty at baseline are shown in the Figure. Participants who were physically frail were more likely than those who were not physically frail to develop each of the 3 disability outcomes.

The median duration of follow-up was 35.5 months for any disability, 45 months for persistent disability, and 47 months for disability with nursing home admission. Table 2 provides information on exposure to intervening events (per 100 months) according to disability status. With only 1 exception, participants who developed disability were significantly more likely to have been hospitalized or to have had restricted activity than those who did not develop disability. For example, participants who developed any disability had a median of 14 months of restricted activity compared with 7.8 months for those who did not develop disability (P<.001). The incidence rates for any disability, persistent disability, and disability with nursing home admission per 100 person-months of exposure to hospitalization were 31.9 (95% CI, 27.4-36.5), 17.2 (95% CI, 14.2-20.2), and 18.8 (95% CI, 16.0-21.7), respectively. The corresponding results for exposure to restricted activity were 3.4 (95% CI, 2.8-4.0), 1.5 (95% CI, 1.1-1.9), and 0.31 (95% CI, 0.15-0.48).

The bivariate and multivariable HRs for the development of any disability are provided in Table 3. In the multivariable analysis, factors associated with any disability included age, living alone, myocardial infarction, stroke, congestive heart failure, depressive symptoms, physical frailty, and new hospitalization and restricted activity only. The risks associated with new hospitalization and restricted activity only were pronounced, with multivariable HRs of 59.8 for hospitalization and 5.11 for restricted activity only. Hospitalization or restricted activity only occurring more than 2 months earlier were not associated with the development of any disability in the multivariable analysis. For persistent disability, the multivariable HRs for new hospitalization and restricted activity only were 43.0 (95% CI, 32.2-57.5) and 3.27 (95% CI, 2.28-4.69), respectively. The corresponding HRs for disability with nursing home admission were 223 (95% CI, 138-362) and 3.51 (95% CI, 1.72-7.19).

In the fully adjusted models, new hospitalization and restricted activity only were significantly associated with each of the 3 disability outcomes regardless of level of physical frailty at baseline with only 1 exception (Table 4). Exposure to hospitalization or restricted activity only more than 2 months earlier did not increase the likelihood of developing either persistent disability or disability with nursing home admission in the multivariable analyses (results available on request).

Among participants without a new intervening event, the incidence rates (per 100 person-months) of any disability, persistent disability, and disability with nursing home admission were 0.61 (95% CI, 0.50-0.72), 0.40 (95% CI, 0.31-0.48), and 0.07 (95% CI, 0.04-0.11), respectively. The population-attributable fractions associated with new exposure to hospitalization and restricted activity only, respectively, were 0.48 and 0.19 for any disability, 0.46 and 0.13 for persistent disability, and 0.82 and 0.05 for disability with nursing home admission. Among participants who were physically frail, the corresponding values were 0.40 and 0.20 for any disability, 0.41 and 0.17 for persistent disability, and 0.81 and 0.07 for disability with nursing home admission. Among those who were not physically frail, the values were 0.61 and 0.16 for any disability, 0.59 and 0.08 for persistent disability, and 0.86 and 0.03 for disability with nursing home admission.

Table 5 provides information on the primary reasons for hospitalization. While cardiac (coronary heart disease, congestive heart failure, arrhythmia, etc) was the most common diagnostic category, fall-related injury conferred the highest risk of disability for each of the outcomes, with 79.4% of admissions for a fall-related injury leading to any disability, 45.2% to persistent disability, and 58.8% to disability with nursing home admission. Table 6 provides comparable information for episodes of restricted activity. The most common reasons for restricted activity leading to disability were fatigue, pain or stiffness in joints, pain or stiffness in back, and dizziness or unsteadiness on feet. However, of all the reasons, a fall or injury conferred the highest risk of disability with 11.0% of the fall-related episodes leading to any disability, 5.3% to persistent disability, and 1.7% to disability with nursing home admission.

In this prospective cohort study of community-dwelling older persons, we found that intervening events, including illnesses and injuries leading to either hospitalization or restricted activity, were strongly associated with the development of disability in essential ADLs. These associations were limited to events occurring within a month of disability onset, were observed for 3 distinct disability outcomes, persisted despite adjustment for several potential confounders, and were present among persons who were and were not physically frail. Because several of the most common intervening events are either preventable^43- 45 or amenable to aggressive in-hospital management^46- 47 or restorative interventions after hospitalization,^48- 51 they provide an attractive target for the prevention of disability among community-dwelling older persons.

While the deleterious effects of illnesses and injuries leading to hospitalization have been previously suggested,^34,52- 53 the magnitude of these effects has not been well defined. In the current study, the HRs associated with hospitalization ranged from 43.0 for persistent disability to 223 for disability with nursing home admission. These large relative increases in risk reflect not only the disabling effects of serious illness coupled with the potential hazards of hospitalization,⁵⁴ but also the low incidence of disability in the comparison groups of persons without an acute hospital admission or restricted activity, which was most striking for disability with nursing home admission. In absolute terms, illnesses and injuries leading to hospitalization accounted for about 50% to 80% of the disability outcomes. Another 5% to 19% of the disability outcomes were attributable to illnesses and injuries leading to restricted activity but not to hospitalization. Depending on the specific disability outcome, the risk of disability was elevated more than 5-fold in the setting of restricted activity. Because older persons usually attribute their restricted activity to several concurrent health-related problems,²³ clinicians may want to avoid focusing on any single problem, especially because the likelihood of disability does not differ greatly based on the underlying reasons for restricted activity. However, because falls and fall-related injuries resulting in hospitalization or restricted activity conferred the highest risk of disability, fall prevention has promise as a primary strategy for reducing the burden of disability among community-dwelling older persons.⁴³

The frequency of our assessments increases the likelihood that the intervening events preceded the disability outcomes, thereby strengthening temporal precedence and supporting a causal association. In previous studies,^{34- 35,52- 53} temporal precedence was not well established because functional status was assessed only twice during the course of 12 to 18 months. While the frequency of our assessments allowed us to demonstrate that the deleterious effects of intervening events are limited to a relatively brief exposure period of 1 month, our data do not allow us to determine how often the intervening events resulted immediately in disability, as may occur with a sudden acute process such as a stroke or hip fracture.

In addition, we were not able to collect information on the length of the intervening events. We did not ask participants to estimate the number of days they had restricted their activities. To our knowledge, the reliability of these estimates has not yet been determined. Furthermore, other investigators have documented high rates of missing data and a highly skewed distribution for days of restricted activity.⁵⁵ For similar reasons, we did not ask participants about the length of their hospital stay. Because participants have been admitted to more than 50 different hospitals (to date), including more than 40 that are out of state (ie, during a vacation or after a move), obtaining length of stay from review of hospital records was not feasible.

In contrast to an earlier study,⁵² which included an assessment interval of 12 months, we found that the likelihood of developing disability in the setting of an illness or injury leading to hospitalization was greater among persons who were not physically frail than those who were physically frail. Because functional status in the current study was reassessed within 1 month of hospital admission rather than several weeks to months after hospitalization,⁵² participants had less opportunity to recover independent function. This difference in study designs is important because rates of recovery are substantially higher among persons who are not physically frail.^25,56 While the lower “base” rate of disability (ie, in the absence of an intervening event) among persons who are not physically frail²² will lead to HRs that are higher than those among persons who are physically frail, the absolute risk of disability in the setting on an intervening event will be higher among persons who are physically frail.

The validity of our results is strengthened by the nearly complete ascertainment of intervening events and disability, the high reliability and accuracy of these assessments, the low rate of attrition, adjustment for several relevant covariates at 18-month intervals with few missing data, and the consistency of the associations across 3 different disability outcomes. While our participants were members of a single health plan in a small urban area, our participation rate was greater than 75%, enhancing generalizability. Moreover, our study population reflects the demographic characteristics of persons aged 65 years or older in New Haven County, Conn, which are comparable with the United States as a whole, with the exception of race (New Haven County has a larger proportion of non-Hispanic whites in this age group than the United States, 91% vs 84%).⁵⁷ However, our analyses were limited to the first episodes of any disability, persistent disability, and disability with nursing home admission, and given the dynamic nature of disability among older persons,^25,58 additional research is needed to elucidate the role of intervening events on the subsequent course of disability.

Our results provide empirical evidence to support several related models of disability.^59- 61 Brocklehurst⁵⁹ postulated that many frail older persons manage to remain in the community by balancing assets (which help maintain independence) with deficits (which threaten independence); and that “breakdown” occurs either by the addition of an acute or subacute medical problem to the deficit side of the balance or by the loss of social support from the asset side of the balance. Rockwood et al⁶⁰ extended this model to describe a more dynamic process in which perturbations in assets and deficits lead to changes in functional status. Finally, Campbell and Buchner⁶¹ proposed that disability may arise from a single catastrophic event, such as a stroke or traumatic amputation, in an otherwise robust individual, or from a less potent event, such as an attack of bronchitis, in a frail older person. While frailty is thought to confer high risk for an array of adverse outcomes, including functional decline and disability,^61- 65 it does not presuppose the presence of concurrent disability or significant comorbidity. Indeed, among participants in the Cardiovascular Health Study, 46% of those who were classified as frail had comorbid disease, 6% had disability, 22% had both comorbid disease and disability, and 27% had neither disability nor comorbidity.⁶²

The results of the current study highlight the importance of intervening events as a potential target for the prevention of disability, regardless of the presence of physical frailty. Interventions that boost reserve capacity and augment compensatory strategies have been shown to prevent the progression of disability among physically frail older persons.^28,66- 67 In the setting of an acute illness or injury leading to hospitalization, functional outcomes are improved by management of older persons on specialized inpatient services^68- 69 and, posthospitalization, by highly coordinated gerocentric care provided in the home.⁵⁰ Based on our results, comparable interventions may be warranted in the setting of an illness or injury that leads to restricted activity, but does not require hospitalization.