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


Review: Multiple risk factor modification reduces falls in elderly persons

ACP J Club. 1998 May-June;128:69. doi:10.7326/ACPJC-1998-128-3-069

Related Content in the Archives
Physical therapy led to modest improvements in mobility in very frail nursing home residents

Source Citation

Gillespie LD, Gillespie WJ, Cumming R, Lamb SE, Rowe BH. Interventions to reduce the incidence of falling in the elderly. In: The Cochrane Database of Systematic Reviews. The Cochrane Library. Oxford: Update Software; 1998, Issue 1. [PubMed ID: 10796348]



Using meta-analysis, what is the effectiveness of interventions to prevent falls in community-dwelling, hospitalized, and institutionalized elderly persons?

Data sources

Studies up to May 1997 were identified from MEDLINE, EMBASE, CINAHL, Current Contents, PsycLIT, Social Science Citation Index, Dissertation Abstracts, Index to UK Theses, and the Cochrane Controlled Trials Register using the Cochrane search strategies combined with subject-specific search terms, including fall, elderly, aged, older, senior, and accidental falls. Bibliographies of relevant articles were reviewed, and workers in the field were contacted.

Study selection

Randomized controlled trials (RCTs) were selected if they examined interventions to minimize or prevent exposure to putative risk factors for falling in elderly persons living in the community or in the hospital or other institutions. Studies that did not report outcomes for falls or that only reported intermediate outcomes (e.g., improved balance) were excluded.

Data extraction

2 independent reviewers extracted data on study settings, participants, interventions, and outcomes and assessed the methodologic quality of each study.

Main results

18 trials and 1 meta-analysis met the inclusion criteria. Meta-analysis of data from 5 studies showed that, compared with usual care, interventions that targeted modification of multiple risk factors on the basis of individual health assessments reduced the number of individuals who fell (Table). Analysis of data from 2 additional studies showed that, compared with usual care, interventions involving the modification of environmental risk factors plus a co-intervention targeting other risk factors reduced the number of individuals who fell but not the number of falls resulting in injury. Untargeted interventions (i.e., those not based on assessment of individual risk) of exercise alone, health education alone, or health education plus exercise did not reduce falls.


Complex interventions that are based on individual health assessments and that target multiple risk factors, and behavioral interventions that include environmental risk factor modification plus a co-intervention, may reduce the number of elderly persons who fall.

Source of funding: Chief Scientist Office, Department of Health, The Scottish Office, UK.

For article reprint: Dr. H. Handoll, Clinical Research Unit, Princess Margaret Rose Hospital, 41-43 Frogston Road West, Edinburgh EH10 7ED, Scotland, UK. FAX 44-131-536-4754.

Table. Health assessment and risk factor intervention vs usual care*

Outcome Assessment and intervention weighted EER Usual care weighted CER RR (95% CI) NNT (CI)
Persons falling 33.4% 39.2% 15% (5 to 23) 18 (11 to 48)

*Abbreviations defined in Glossary; RRR, NNT, and CI calculated from data in article.


During the past 20 years, research about falls has progressed remarkably, from a few studies of risk factors to a meta-analysis of 18 RCTs that examined the effectiveness of interventions to prevent falls. Gillespie and colleagues have identified areas of promise from existing evidence, as well as many remaining questions for further investigation. Their effort, however, was still encumbered by differences in the studies they compared, such as how falls were ascertained. Many promising studies were excluded because they used surrogate outcomes, such as balance. Such differences are one reason why calculating a single estimate of effect in meta-analysis is criticized (1).

Falls usually have multiple causes (2). It is not surprising that meta-analysis shows greater benefit from interventions that target modification of multiple risk factors based on individual assessments. Interventions that do not incorporate individual assessments and have only a single focus, such as exercise, might be expected to be less effective.

The limited number and size of studies using individual assessments and targeted interventions or behavioral interventions preclude assessing the effects of individual intervention components, thereby making it difficult to determine and compare the costs and benefits of establishing such a program. Nevertheless, research about falls has matured to the point where planners and providers can and must move beyond identifying risk factors and start modifying these factors with interventions proven to reduce risk and ultimately falls.

Thomas V. Jones, MD

Thomas V. Jones, MD


1. Bailar JC 3rd. The promise and problems of meta-analysis. N Engl J Med. 1997; 337: 559-61. [PubMed ID: 9262502]

2. Studenski S, Rigler SK. Clinical overview of instability in the elderly. Clin Geriatr Med. 1996;12:679-88. [PubMed ID: 8890110]