Diagnostic performance of the Finnish diabetes risk score for undiagnosed dysglycaemia in Western Kenya

Abstract

Background: In Kenya, the screening of dysglycaemic states, such as prediabetes and diabetes, is conventionally done through blood glucose testing, which is often impractical to implement. The Finnish Diabetes Risk Score (FINDRISC), an easy-to-use, valid, and freely available non-invasive pre-screening tool has the potential to improve Kenya’s dysglycaemia screening strategy as part of the multistage screening strategy recommended by the WHO for resource-constrained settings. FINDRISC is a one-page questionnaire containing eight questions of non-invasively measured risk factors for dysglycaemia that was derived from a ten-year prospective study for identification of individuals with a high risk for developing diabetes among the Finnish population and has been validated in many populations. The diagnostic performance of FINDRISC in a pragmatic setting in Kenya remains unknown, hence the need for this study. Objectives: To evaluate the diagnostic performance of FINSRISC and determine its optimal cut-off scores for detecting adults with undiagnosed dysglycaemia in a rural population of Western Kenya. Methods: This was a cross sectional study conducted among 382 participants within Trans-Nzoia county of Western Kenya between November 2020 and February 2021. Participants were enrolled via simple random sampling and stratified according to age group. Data was collected using an adopted FINDRISC questionnaire and subsequently participants were tested using OGTT, which was the gold standard test to determine glycaemic status. Continuous variables were presented as mean (SD or 95% CI), while categorical variables were presented as proportions. Comparisons between normally distributed continuous variables were performed using Student’s t-test, while associations between categorical variables were done using Fisher’s exact test. Measures of diagnostic accuracy (sensitivity, specificity, positive predictive value [PPV] and negative predictive value [NPV]) were calculated for various FINDRISC cut-off points for both prediabetes and diabetes using OGTT results as the gold standard test for dysglycaemia. Discrimination was determined by calculating the area under the receiver operating characteristics curve (AUROC). Results: The study population was predominantly (92.9%) rural. The mean age was 45.5 years, and majority of the participants (68%) were female. The overall prevalence of undiagnosed diabetes and prediabetes was 3.9% (95% CI 1.97-5.88) and 7.9% (95% CI 5.14-10.56) respectively. Using OGTT as the gold standard test for dysglycaemia, FINDRISC detected undiagnosed diabetes with 67% (95% CI 38-88) sensitivity and 93% (95% CI 90-96) specificity at a cut-off score of ≥14, AUROC 0.80 (95% CI 0.75-0.84). It detected undiagnosed prediabetes with 57% (95% CI 37-75) sensitivity and 75% (95% CI 70-80) specificity at a cut-off score of ≥10, AUROC 0.64 (95% CI 0.59-0.69). FINDRISC detected both prediabetes and diabetes with 60% (95% CI 44-74) sensitivity and 77% (95% CI 72-81) specificity at an optimal cut-off point of ≥10, AUROC 0.69 (95% CI 0.64-0.74). It demonstrated a low PPV (26% [95% CI 20-32]) but a high NPV (94% [95% CI 90-95]) for dysglycaemia. Conclusion: FINDRISC had a high NPV for dysglycaemia at an optimal cut-off score of ≥10. Recommendation: FINDRISC should be used as a pre-screening tool, and diagnostic laboratory testing for dysglycaemia limited to individuals with a score of ≥10.