Construct validation of a Frailty Index, an HIV Index and a Protective Index from a clinical HIV database

Standard care for HIV clinical practice has started focusing on age-related problems, but despite this recent change physicians involved in HIV care do not often screen HIV patients for frailty. Our aim was to construct three indexes from an HIV clinical database (i.e. Frailty Index, (FI), HIV Index, (HIVI), and Protective Index (PI)) and to assess levels of frailty, HIV severity and demographic and protective lifestyle factors among HIV patients. Methods and findings We included data from 1612 patients who attended an Italian HIV clinic between September 2016 and December2017 (mean±SD age: 53.1±8 years, 73.9% men).We used 92 routine variables collected by physicians and other health care professionals to construct three indexes: A 72-item FI (biometric, psychiatric, blood test, daily life activities, geriatric syndromes and nutrition data), a 10-item HIVI (immunological, viral and therapeutics) and a 10-item PI (income, education, social engagement, and lifestyle habits data)(the lower the FI and HIVI scores, and the higher the PI scores, the lower the risk for participants).The FI, HIVI and PI scores were 0.19±0.08, 0.48±0.17 and 0.62±0.13, respectively. Men had higher FI (0.19±0.08 vs 0.18±0.08; p = 0.010) and lower HIVI (0.47±0.18 vs 0.50±0.15; p = 0.038) scores than women. FI and HIVI scores both increased 1.9% per year of age (p < 0.001), whereas the PI decreased 0.2% per year (p<0.050). In addition, the FI score increased 1.6% and the PI score decreased 0.5% per year of HIV infection (p < 0.001). Conclusion It is feasible to assess levels of frailty, HIV severity and protective lifestyle factors in HIV patients using data from a clinical database. Frailty levels are high among HIV patients and even higher among older patients and those with a long duration of HIV. Future studies need to examine the ability of the three indices to predict adverse health outcomes such as hospitalization and mortality.