Predictive validity of a two-step tool to map frailty in primary care

Abstract

Background: EASY-Care Two step Older people Screening (EASY-Care TOS) is a stepped approach to identify frail older people at risk for negative health outcomes in primary care, and makes use of General Practitioners' (GPs) readily-available information. We aimed to determine the predictive value of EASY-Care TOS for negative health outcomes within the year from assessment. Methods: A total of 587 patients of four GP practices in and around Nijmegen (The Netherlands) consented to participate in a longitudinal primary care registry based cohort study. Participants' frailty was judged by their GP following the EASY-Care TOS procedure and by a Comprehensive Geriatric Assessment (CGA) at baseline. After one year health outcomes of the participants were measured by reassessment with the EASY-Care TOS procedure. Results: Follow up information was available for 520 of 587 participants. In the non-frail group 9% showed any negative health outcomes (death, ADL decline, institutionalisation, too ill to undergo assessment), against 30% in the frail group (95% confidence interval of the difference (CI): 14%-28%). Area under the receiver operating curve (AUC) of the EASY-Care TOS frailty judgement for a composite of negative health outcomes mentioned was 0.67 (95% CI: 0.62-0.73). Compared with discrimination on the basis of age, sex and GP practice (AUC 0.70), adding EASY-Care TOS frailty judgement increased the AUC to 0.75 (+0.05, p=0.02). The AUC on the basis of a full CGA is almost comparable to the AUC of the model with age, sex, and frailty judgement with EASY-Care TOS: 0.76 (+0.07, p=0.005). Conclusions: GPs applying the EASY-Care TOS procedure, where they only perform additional assessment when they judge this as necessary, can predict negative health outcomes in their older populations efficiently and almost as accurately as a complete specialist CGA.