JMIR Publications

ABSTRACT

Background:

Falls in hospitalized patients are a serious problem, resulting not only in physical injury but also in secondary complications, impaired activities of daily living, prolonged hospital stays, and increased medical costs. Establishing a fall prediction scoring system to identify the patient population most likely to fall can help in preventing falls among hospitalized patients.

Objective:

The purpose of this study was to identify predictive factors of falls in patients admitted to an acute care hospital, as well as to develop a scoring system using these factors and evaluate its validity.

Methods:

This was a single-center, retrospective cohort study including patients aged 20 years or older at the time of admission to Shizuoka General Hospital between April 2019 and September 2020. Demographic data and candidate predictors at the time of admission, as well as information from fall occurrence reports, were collected from the medical records. The outcome to be predicted was the time from the date of admission to a fall that required the use of medical resources. Two-thirds of all cases were randomly selected as the training set for analysis, and univariate and multivariate Cox regression analysis were used to identify factors affecting patients’ risk of falls. Based on the estimated hazard ratios, the fall risk was scored and we constructed a fall prediction scoring system. The remaining one-third of all cases were used as the test set for analysis to evaluate the predictive performance of the newly developed scoring system.

Results:

A total of 13,725 individuals were included in the analysis. During the study period, 326 (2.4%) of all patients experienced a fall. In the training data set (n=9,150, random sample of approximately two-thirds of all patients), Cox regression analysis identified sex, age, body mass index, independence degree of daily living for the disabled elderly (Bedriddenness Rank), emergency department, and history of falls within 1 year as predictors of falls. Using these factors, scoring was performed based on each hazard ratio, and a new fall prediction scoring system was constructed with patient classification into three fall risk groups (low risk: 0 to 4 points, intermediate risk: 5 to 9 points, and high risk: 10 to 15 points). The c-index, which indicates the predictive performance in the test data set (n=4,575, residual sample), was 0.733 (95% confidence interval, 0.684–0.782).

Conclusions:

We developed a new fall prediction scoring system for patients admitted to acute care hospitals by identifying predictors of falls in Japan. This system may be useful in preventive interventions for patient populations with a high likelihood of falling in many hospitals with acute-care settings.