A brief frailty screening tool in Tanzania: external validation and refinement of the B-FIT screen

Background Identifying older people who are most vulnerable to adverse outcomes is important. This is particularly so in low-resource settings, such as those in sub-Saharan Africa (SSA), where access to social and healthcare services is often limited. Aim To validate and further refine a frailty screening tool for SSA. Methods Phase I screening of people aged 60 years and over was conducted using the Brief Frailty Instrument for Tanzania (B-FIT). In phase II, a stratified, frailty-weighed sample was assessed across a range of variables covering cognition, physical function (including continence, mobility, weakness and exhaustion) nutrition, mood, co-morbidity, sensory impairment, polypharmacy, social support and self-rated health. The frailty-weighted sample was also assessed for frailty according to the comprehensive geriatric assessment (CGA), which we used as our ‘gold standard’ diagnosis. Results Of 235 people in the frailty-weighted sample, 91 (38.7%) were frail according to CGA, the median age was 73 years and 136 (57.9%) were female. In multivariable modelling, physical disability (Barthel index), cognitive impairment (IDEA cognitive screen), calf circumference, poor distance vision and problems engaging in social activities were found to be associated with frailty. After developing a scoring system, based on regression coefficients, a modified B-FIT screen (B-FIT 2) had an area under the receiver operating characteristic curve of 0.925, a sensitivity of 86.2% and a specificity of 88.8%. Discussion The inclusion of items assessing nutrition, social support and sensory impairment improved the performance of the B-FIT. Conclusions The B-FIT 2 should be externally validated.