Current issues of ACP Journal Club are published in Annals of Internal Medicine


7 risk factors predicted repeated hospital admissions in elderly patients

ACP J Club. 1994 Mar-April;120:49. doi:10.7326/ACPJC-1994-120-2-049

Source Citation

Boult C, Dowd B, McCaffrey D, et al. Screening elders for risk of hospital admission. J Am Geriatr Soc. 1993 Aug;41:811-7.



To define and assess the validity of screening criteria for identifying elderly persons at risk for repeated hospitalization.


Cohort study of elderly persons followed for at least 4 years (Longitudinal Study of Aging). Half the participants were used to identify risk factors for hospitalization and the other half to validate the risk factors.


United States.


A random sample of noninstitutionalized, civilian participants > 70 years of age at baseline (1984) who had complete Medicare data (n = 5876, 91% white, 57% women).

Assessment of risk factors

Participants were interviewed to assess 1 enabling variable (income < $10 000/y), 6 predisposing variables (age, sex, race, years of education, locus of control, and living alone), and 21 need-based variables (general health categories; disease categories; health resource use categories; functional ability; informal caregiver availability; falls; incontinence; and visual, cognitive, or hearing impairment).

Main outcome measures

Repeated hospitalization was defined as ≥ 2 hospitalizations over 4 years.

Main results

During follow-up, 1670 participants (28%) were hospitalized more than once and 646 (11%) died. Predictors of hospitalization were self-rated poor general health (odds ratio [OR] 2.2, 95% CI 1.3 to 3.6), fair general health (OR 1.7, CI 1.2 to 2.6), or good general health (OR 1.4, CI, 1.0 to 2.0), coronary artery disease (OR 1.5, CI 1.1 to 2.0), diabetes in the past year (OR 1.4, CI 1.0 to 1.9), hospitalization in the previous year (OR 1.7, CI 1.3 to 2.3), > 6 physician visits in the previous year (OR 1.4, CI 1.1 to 1.8), informal caregiver availability (OR 2.0, CI 1.2 to 3.3), age 75 to 79 years (OR 1.3, CI 1.0 to 1.6), age 80 to 84 years (OR 1.4, CI 1.0 to 1.9), age ≥ 85 years (OR 1.7, CI 1.2 to 2.7), and male sex (OR 1.3, CI 1.0 to 1.6) using data from the first half of the participants. Based on these factors, the rest of the participants were classified as high risk (7%) or low risk (93%) for repeated hospitalization. Compared with participants at low risk for hospitalization, participants at high risk had a higher cumulative rate of repeated hospitalizations (42% vs 26%, P < 0.001) and a higher cumulative rate of death (44% vs 19%, P < 0.001).


Risk factors predicting repeated hospitalizations over 4 years in a cohort of elderly participants were coronary artery disease; diabetes in the past year; hospital admission; poor, fair, or good self-rated general health; > 6 physician visits in the previous year; informal caregiver availability; male sex; and age ≥ 75 years.

Sources of funding: National Institute on Aging; Minnesota Medical Foundation; University of Minnesota Center for Urban and Regional Affairs; Alfred P. Sloan Foundation.

For article reprint: Dr. C. Boult, University of Minnesota Health Center, Box 381, 516 Delaware Street Southeast, Minneapolis, MN 55455, USA. FAX 612-627-4314.


Health care expenditures are not distributed equally among elderly persons. This applies especially to funds allocated for repetitive hospitalization near the end of life. For example, 5% of Medicare enrollees account for 62% of all hospital expenses; these are often incurred in the last year of life (1).

The study by Boult and colleagues asks the important question: Are there reliable clinical indicators to predict this subset of patients so that targeted interventions could be developed before the need for hospitalization? This study found 7 risk factors from a brief questionnaire, which, when applied to elderly, noninstitutionalized Medicare enrollees, identified 7% as "high risk." This group had about twice the cumulative incidence of repeated admission and death over the 4 years of observation. In addition, the authors provided ORs and a predictive formula to estimate the relative probability of repeated hospitalizations for individual patients.

The authors make the interesting observation that "need-based" risk factors that reflect coexisting illness, including self-assessment of health, frequency of physician visits, recent hospital admissions, and coronary artery disease, made a greater contribution to cumulative risk than "predisposing" factors, most notably chronologic age itself. This relative neutrality of age as a prognostic factor has also been observed in studies of outcomes of hospitalization, especially technologically intensive care (2). It remains to be shown, however, whether the rate of hospitalization can be altered by interventions emerging from the identification of a "high-risk" subset of elderly persons.

William J. Hall, MD
University of Rochester School of Medicine & DentistryRochester, New York, USA


1. Riley G, Lubitz J, Prihoda R, Stevenson MA. Changes in distribution of Medicare expenditures among aged enrollees, 1969-82. Health Care Financ Rev. 1986;7:53-63.

2. Swinburne AJ, Fedullo AJ, Bixby K, Lee DK, Wahl GW. Respiratory failure in the elderly. Analysis of outcome after treatment with mechanical ventilation. Arch Intern Med. 1993;153:1657-62.

Addendum to the 1994 commentary

The lead author, Dr. C. Boult, recommends the following newer articles as supplementary evidence:

1. Coleman EA, Wagner EH, Grothaus LC, et al. Predicting hospitalitzation and functional decline in older health plan enrollees: are administrative data as accurate as self-report? J Am Geriatr Soc. 1998;46:419-25.

2. Boult L, Boult C, Pirie P, Pacala JT. Test-retest reliability of a questionnaire that identifies elders at risk for hospital admission. J Am Geriatr Soc. 1994;42:707-11.

3. Pacala JT, Boult C, Boult L. Predictive validity of a questionnaire that identifies older persons at risk for hospital admission. J Am Geriatr Soc. 1995;43:374-7.

4. Pacala JT, Boult C, Reed RL, Aliberti E. Predictive validity of the Pra instrument among older recipients of managed care. J Am Geriatr Soc. 1997;45:614-7.