OBJECTIVES: To examine the relationship between anemia and falls in nursing home residents. METHODS: A chart review was conducted in 40 nursing homes across the U.S. Residents were considered eligible if they were: >=18 years old, had >=1 hemoglobin (Hb) level during the period of 1/1/04-2/1/05; had a recorded serum creatinine level; maintained residency in the facility; and did not receive dialysis during the 6-month follow-up period. Demographics, labs, comorbid conditions, medications, functional status, and anemia treatment information were obtained from chart review. All incidents of falls and hospitalizations were identified for each eligible resident. Two logistic regression models were developed to identify factors associated with falling (>=1 fall) and recurrent falls (>=2 falls). Anemia was defined using WHO criteria. RESULTS: Of the 579 residents who met inclusion criteria, 564 met criteria for data completeness. Mean age was 81 years; 70% of residents were female, 56% were identified as anemic, 24% had a documented fall during the 6-month follow-up period, and 12% had recurrent falls. The unadjusted falling rate for anemic residents was 31%, compared to 17% for non-anemic residents (p<0.001). The respective rates for recurrent falls were 15% anemic vs. 7% non-anemic (p=0.003). For the dependent variable of falling, significant associations were seen for anemia (OR=2.22, p<0.001), unable to ambulate (OR=0.59, p=0.025) and use of psychoactive medications (OR=2.18, p=0.001). For the dependent variable of recurrent falls, similar results were observed for anemia and psychoactive medication use. Of the 10 residents studied requiring hospital admission for fracture during the follow-up period, all were anemic. CONCLUSION: Both anemia and the use of psychoactive medications are potentially modifiable factors strongly associated with falling and recurrent falls. Since falls and related fractures are events associated with significant morbidity and mortality, these factors deserve special consideration for potentially reducing the risk of such events.