Identification and management of frail patients in English primary care: An analysis of the General Medical Services 2018/2019 contract dataset

Objectives The aim of this study was to explore the extent of implementation of the General Medical Services 2018/2019 € frailty identification and management' contract in general practitioner (GP) practices in England, and link implementation outcomes to a range of practice and Clinical Commissioning Group (CCG) factors. Design A cross-sectional study design using publicly available datasets relating to the year 2018 for all GP practices in England. Settings English general practices. Data The analysis was conducted across 6632 practices in 193 CCGs with 9 995 558 patients aged 65 years or older. Outcomes Frailty assessment rates, frailty coding rates and frailty prevalence rates, plus rates of medication reviews, falls assessments and enriched Summary Care Records (SCRs). Analysis Summary statistics were calculated and multilevel negative binomial regression analysis was used to investigate relationships of the six outcomes with explanatory factors. Results 14.3% of people aged 65 years or older were assessed for frailty, with 35.4% of these - totalling 5% of the eligible population - coded moderately or severely frail. 59.2% received a medications review, but rates of falls assessments (3.7%) and enriched SCRs (21%) were low. However, percentages varied widely across practices and CCGs. Practice differences in contract implementation were most strongly accounted for by their grouping within CCGs, with weaker but still important associations with some practice and CCG factors, particularly healthcare demand-related factors of chronic caseload and (negatively) % of patients aged 65 years or older. Conclusion CCG appears the strongest determinant of practice engagement with the frailty contract, and fuller implementation may depend on greater engagement of CCGs themselves, particularly in commissioning suitable interventions. Practices understandably targeted frailty assessments at patients more likely to be found severely frail, resulting in probable underidentification of moderately frail individuals who might benefit most from early interventions. Frailty prevalence estimates based on the contract data may not reflect actual rates.