Fall risk in elderly nursing home residents: A study of intrinsic, extrinsic and interactive factors

Abstract

An estimated 50% of elderly nursing home (NH) residents fall annually. In studies of fall risk, intrinsic factors are well established, however extrinsic and interactive factors are not. Studies have not developed fall risk models for short-term length of stay (STLOS) and long-term length of stay (LTLOS) residents. The six "triggers" for falls in the Minimum Data Set (MDS) have not been studied as predictors. Primary purposes of this study were to: (1) develop multivariate fall risk models for STLOS and LTLOS NH residents to evaluate "triggers" for falls and establish the intrinsic, extrinsic and interactive factors and (2) determine agreement with which falls are recorded in the MDS versus medical and nursing documentation. A secondary correlational analysis was performed using data from a prospective, longitudinal cohort study of 2,285 NH admissions from 59 randomly selected facilities in the State of Maryland. The current sample included 1,140 residents from 56 facilities meeting inclusion criteria. The sample was split into: (1) STLOS including 429 residents in the NH for ≤90 days and (2) LTLOS including 711 residents in the NH for ≥91 days. Measures and variables were obtained from the following: MDS 1.0, Fall Event Records, Online Survey Certification and Reporting System and Therapeutic Environmental Screening Scale-2+. Negative binomial regression models revealed: (1) four of six for STLOS and three of six for LTLOS of the "triggers" for falls on the MDS were significant; and (2) for both resident groups intrinsic, extrinsic, and interactive factors were significant that are not currently "triggers" for falls. Concordance between falls recorded in the MDS versus medical and nursing documentation occurred for 75% of cases for a 180-day period, with a kappa of.50 (p ≤ .0001), indicating moderate agreement. Findings of the fall risk models for STLOS and LTLOS residents indicate that reevaluation of the current six "triggers" for falls in the MDS process is needed. The study also demonstrated problems exist in the accuracy of reporting of falls in the NMS. These findings have important implications for nursing, health policy and future studies of fall risk in elderly NH residents.