Diabetes testing, care cascade and the burden of undiagnosed diabetes in Kenya. Evidence from the 2015 Kenya STEPwise survey.

Background: Diabetes is a growing public health challenge globally, with a disproportionate burden occurring in low- and middle-income countries (LMICs). In Kenya, limited evidence exists on type 2 diabetes mellitus (T2DM) screening coverage, undiagnosed diabetes, and progression through the diabetes care cascade. This study assessed the prevalence and determinants of T2DM testing, quantified undiagnosed T2DM, and examined the T2DM care cascade in Kenya. Methods: We analysed data from the 2015 Kenya WHO STEPwise survey, a nationally representative cross-sectional survey of adults aged 18-69 years. The analysis included 4,167 participants after excluding individuals with missing fasting plasma glucose (FPG) measurements and those below the age of 18 and above 69 years. Diabetes testing was defined as self-reported prior blood glucose measurement by a health worker among individuals without known diabetes. Diabetes was defined as FPG [≥]7 mmol/L, previous diagnosis, or current diabetes treatment. Weighted prevalence estimates with 95% confidence intervals (CI) were calculated. Univariable and multivariable logistic regression analyses were conducted to identify factors associated with diabetes screening while accounting for the complex survey design. Results: Overall, 11.9% (95% CI: 9.7-14.5) of participants reported ever being tested for diabetes. The weighted prevalence of diabetes was 2.66% (95% CI: 2.06-3.42), while 53.4% (95% CI: 39.0-67.4) of diabetes cases were previously undiagnosed. Testing prevalence was highest among older adults, individuals with tertiary education, obese participants, and those with hypertension, but lowest among younger adults, individuals with no formal education, and residents of the North Eastern region. Among individuals with diabetes, 46.9% were aware of their diagnosis, 40.8% of those aware were on treatment, and only 26.4% of those on treatment had achieved glycaemic control. In multivariable analysis, older age (60-69 years: aOR=6.29, 95% CI: 3.66-10.79), tertiary education (aOR=2.65, 95% CI: 1.33-5.27), poorest wealth quintile (aOR=8.42, 95% CI: 4.11-17.24), obesity (aOR=2.13, 95% CI: 1.41-3.21), and hypertension (aOR=1.64, 95% CI: 1.21-2.23) were significantly associated with increased odds of diabetes testing. Conclusion: Diabetes testing coverage in Kenya remains low, with more than half of diabetes cases remaining undiagnosed and substantial losses occurring across the diabetes care cascade. These findings highlight critical gaps in early diagnosis, linkage to care, treatment uptake, and glycaemic control. Strengthening integrated diabetes screening and care services, particularly for underserved and high-risk populations, is essential to improve early detection and continuity of care in Kenya.