Differences in Risk Factors for Falls Among Psychiatry and Medicine/Surgery Patients

Background: Patient falls are common adverse events in hospitals.  Reported fall rates are higher among psychiatry patients than among other patient populations, but it is unclear whether risk factors associated with falls in psychiatry patients differ from risk factors in non psychiatry patients.

Objective: To evaluate specific patient characteristics and medications as risk factors for falls among psychiatry patients and determine whether these risk factors differ from risk factors for falls among medicine and surgery patients.

 

Methods: 76 adult psychiatry inpatients who fell and 228 randomly selected control psychiatry patients who did not fall (1:3) were prospectively identified between 7/3/07 and 8/13/08 at Barnes-Jewish Hospital.  Patient demographics, health status, and medication orders were analyzed.  Multivariable logistic regression models were used to predict patient and medication-related risk factors for falling.  The results of this analysis were compared to results from our case control study of 230 medicine/surgery patients who fell and 690 controls to identify differences in fall risk factors.

Results: The mean age for psychiatry patients who fell was 56 years and mean length of stay was 22 days.  On univariate analysis, patients who fell were more likely than controls to have a history of falls (22% vs. 8%, p=0.001), hearing aid (23% vs. 7%, p<0.001), dizziness (13% vs. 3%, p=0.001), need assistance for ambulation (43% vs. 15%, p<0.001), overestimate their ability (34% vs. 7%, p<0.001) and be unable to estimate the time (28% vs. 13%, p<0.01).  Medications associated with falls in univariate analysis included calcium channel blockers (22% vs. 11%, p<0.05), analgesics (93% vs. 83%, p<.05), cholinesterase inhibitors (11% vs. 3%, p<0.01), proton pump inhibitors (32% vs. 19%, p<0.05) and adrenergic bronchodilators (18% vs. 8%, p<0.01).  Anticholinergic antiparkinson agents were associated with reduced risk (9% vs. 29%, p=0.001).  A comparison of multivariable risk factor models for psychiatry and medicine/surgery patients revealed notable differences (see attached table).  Risk factors for falls among psychiatry patients included overestimating ability, dizziness, adrenergic bronchodilators, narcotic analgesic combinations, salicylates and acetomenophen/tramadol.  Risk factors among medicine/surgery patients, in contrast, included fall history, need for assistance to ambulate, underweight (BMI≤18), inability to estimate the time, incontinence, hydantoin and benzodiazepine anticonvulsants, haloperidol, tricyclic antidepressants, and insulin. 
Conclusions: Risk factors for falls among psychiatry patients differ from risk factors for falls among medicine and surgery patients.  It may be necessary to develop separate risk prediction models to identify psychiatry patients at risk for falls and to better target fall prevention strategies among this patient population.