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

Quality Improvement

Hospital discharge planning for high-risk patients decreased readmissions and increased number of patients living at home

ACP J Club. 1993 Sept-Oct;119:59. doi:10.7326/ACPJC-1993-119-2-059

Source Citation

Evans RL, Hendricks RD. Evaluating hospital discharge planning: a randomized clinical trial. Med Care. 1993 Apr;31:358-70.



To determine if early discharge planning given to high-risk patients reduces length of hospital stay or risk for readmission or increases the likelihood of a successful return to home.


Randomized controlled trial with 9-month follow-up.


A Veterans Affairs medical center.


Patients who were hospitalized for > 2 days and were considered at high risk for increased health resource use. High risk was defined as having 3 or more of the following: 2 or more chronic conditions, poor mental status, psychiatric comorbidity, previous hospital admission, age ≥ 70 years, living alone or in a nursing home, dependent ambulation, or being unmarried. Of 6859 screened patients, 923 were eligible and 835 (mean age 67 y, 794 men) were randomized.


The experimental group (417 patients) received a social work assessment on day 3 to initiate discharge planning. The control group (418 patients) was given usual care, including social worker intervention only after medical referral.

Main outcome measures

Hospital length of stay, discharge location, and hospital readmission were determined by chart audit and patient contact.

Main results

All experimental group patients received discharge planning beginning 3 days after admission. 124 control patients also received discharge planning, starting a mean of 9 days after admission. The groups did not differ for mortality, initial length of stay (11.9 d for experimental vs 12.5 d for control patients), or 9-month readmission rates (55% vs 61%, P = 0.08). The experimental group, compared with controls, had fewer 30-day readmissions (24% vs 35% {95% CI for difference 4% to 17%}*, P < 0.001), more home discharges (79% vs 73% {CI for difference 0.3% to 12%}*, P < 0.05), and more patients remaining at home at 9 months (62% vs 54% {CI for difference 2% to 15%}*, P < 0.05). The experimental group received more home care, help with finances and housing, counseling, and other social services (97% vs 30% {CI for difference 63% to 72%}*, P < 0.001).


Early discharge planning for high-risk patients decreased 30-day readmissions and increased the number of patients living at home at 9 months, but did not change length of hospital stay.

Source of funding: Department of Veterans Affairs Health Services Research and Development Program.

For article reprint: Mr. R.L. Evans, Veterans Affairs Medical Center (122), 1660 South Columbian Way, Seattle, WA 98108, USA. FAX 206-764-2514.

*Calculated from data in article.


Repeated hospitalizations account for half of the admissions and 60% of total charges (1). This study by Evans and Hendricks provides strong evidence that anticipating social needs through early discharge planning can improve the likelihood of successfully remaining at home. I view the lack of effect on length-of-stay as a positive finding: Addressing social needs does not prolong hospitalization.

Although unstated, I assume that the wards studied were largely run by housestaff. Physicians in training often have little experience in multidisciplinary care and underestimate the importance of psychosocial issues (this is suggested by the 9-day delay in requesting social work services). Further, the nature of residency training may create perverse incentives for ignoring post-hospital needs (2). Finally, it is unlikely that the hospital physicians had contact with the patients either before or after discharge. These factors highlight the importance of early social work involvement in academic settings but may limit the generalizability of these results.

Who should receive early discharge planning is a major unresolved issue. This study targeted a specific high-risk group that represented a fifth of all admissions. The authors' targeting criteria, however, had only a 60% sensitivity for identifying adverse outcomes. Although intensive interventions (e.g., geriatric evaluation and management units) do require proper targeting to be cost effective, a social worker's assessment is relatively inexpensive and may be cost-effective even when offered to all patients (3).

Joseph Francis, MD, MPH
University of TennesseeMemphis, Tennessee, USA


1. Zook CJ, Moore FD. High cost users of medical care. N Engl J Med. 1980;302: 996-1002.

2. Mizrahi T. Getting rid of patients: contradictions in the socialization of physicians. New Brunswick: Rutgers University Press; 1986.

3. Safran C, Phillips RS. Interventions to prevent readmission: the constraints of cost and efficacy. Med Care. 1989;27:204-11.