E-Frailty Replaces Gait Speed to Identify Frailty in Geriatric Incontinence

Objectives: The geriatric incontinence syndrome (GIS) is defined by the presence of urinary incontinence (UI) concomitantly with physical or cognitive impairments and is clinically important because it leads to frailty. UI severity of ≥2 UI episodes/day is a hallmark feature of GIS. Frailty is associated with increased post-procedure morbidity; thus, its identification in clinical practice is critical. The presence of frailty may be determined by gait speed, a clinical assessment, or an electronic medical record (EMR) based frailty index (eFI). We hypothesize that eFI may be an efficient proxy for gait speed in the clinical setting to determine frailty risk among women with GIS.

Methods: We performed a retrospective cohort study of 84 women, 70 years and older seeking care for moderate-to-severe UI symptoms enrolled in prospective cohort studies between 2016 and 2023. All subjects underwent clinical UI and geriatric assessment at baseline. The presence and severity of UI was determined by 3-day bladder diary data. The Short Physical Performance Battery (SPPB) was applied to measure physical function; this instrument assesses balance, strength, and gait speed. The eFI is calculated by abstracting read codes from the EMR and counting the number of deficits divided by 36. A cut of score of >0.12 to defined frailty. The primary outcome was the strength of the associations between the gait speed measurement and eFI score based on UI severity. Tests of differences were examined using t-tests for continuous factors and chi-square tests for categorical factors. Univariate relationships between functional measures and eFI based on UI severity were examined using Spearman’s rank correlation. Logistic regression models were fitted to examine eFI, chair-stand pace, and gait speed as predictors of the dichotomized UI outcome (severe ≥ 2 UI episodes/day vs moderate >2 UI episodes/day).

Results: 84 subjects were included in this analysis. The mean age was 77.5±5.6 years. There were no demographic differences based on UI severity. The mean eFI was 0.22±0.1; based on frailty categories, 15% were normal and 85% had at mild or more severe frailty. Gait speed was significantly slower and chair-stand pace longer in subjects with severe UI symptoms. UI severity did not correlate with eFI, but had positive correlation with chair-stand pace and negative correlation with gait speed (Figure 1). The eFI was negatively correlated with gait speed [r] -0.28, p=0.02. The eFI was not a significant predictor of UI severity for either unadjusted or adjusted models, OR 1.15 (0.71-1.9) adjusted (Table 1)

Conclusions: The eFI may be an efficient proxy for gait speed measurements in clinical practice evaluating frailty in women with GIS. However, due to its poor predictive ability and correlation with UI severity, prospective studies are needed prior to its use in replacement of clinical assessments of frailty in GIS.