Impact of a national diabetes risk assessment and screening programme in England: a quasi-experimental study

Abstract

BackgroundThe global prevalence of type 2 diabetes is increasing but the effectiveness of large-scale screening programmes to detect high-risk individuals (eg, those with non-diabetic hyperglycaemia) is debated. We assessed the impact of a national diabetes risk assessment and screening programme, NHS Health Checks, on detection of incident cases of non-diabetic hyperglycaemia and type 2 diabetes and on management of cardiovascular risk factors in newly diagnosed cases. MethodsIn a quasi-experimental propensity score adjusted study, we extracted data on 370 454 individuals aged 40–74 years registered with 455 general practices in England between 2009 and 2016 without pre-existing type 2 diabetes. Individuals were characterised into three groups on the basis of their general practice's programme coverage (low, medium, high). We assessed differences in diagnosis of non-diabetic hyperglycaemia and type 2 diabetes and changes in cardiovascular risk (QRISK2 algorithm) using multilevel Cox and linear regression models. Propensity score, based on the probability of individuals being registered with a low, medium, or high coverage practice, was calculated. Analyses were by intention to screen. FindingsThere were 138 364, 131 679, and 100 411 individuals in the low, medium, and high coverage groups, respectively. The incidence rate of detected non-diabetic hyperglycaemia was 19% higher in the high coverage group than in the low coverage group (2781 vs 2479 cases; hazard ratio 1·19, 95% CI 1·01 to 1·41), and rates of type 2 diabetes diagnosis (4058, 4657, and 3827 cases in low, medium, and high coverage groups) were 10% higher in the medium coverage group (1·10, 1·03 to 1·18) and 11% higher in the high coverage group (1·11, 1·03 to 1·19). Individuals with detected non-diabetic hyperglycaemia in the high coverage group had a 1·1% larger reduction in cardiovascular risk than did those in the low coverage group (β=–1·12, 95% CI −1·61 to −0·63; mean follow-up 43·9 months), and those with detected type 2 diabetes a 0·4% larger reduction (−0·42, −0·78 to −0·06; 49·8 months). InterpretationAlthough only routinely collected information was used in this study, a robust methodology with a representative sample of the English population allows causal inference. General practices actively participating in a national diabetes screening programme had higher detection of non-diabetic hyperglycaemia and type 2 diabetes and, when newly detected, reduced cardiovascular risk. FundingPart funded by the National Institute for Health Research (NIHR) under the Collaboration for Leadership in Applied Health Research and Care programme Northwest London.