Pitfalls of Single Measurement Screening for Diabetes and Hypertension in Community-Based Settings

Abstract

Background: Cross-sectional screening programs are used to detect and refer individuals with non-communicable diseases to healthcare services. We evaluated the positive predictive value of cross-sectional measurements for Diabetes Mellitus (DM) and hypertension (HTN) as part of a community-based disease screening study, 'Vukuzazi' in rural South Africa. Methods: We conducted community-based screening for HTN and DM using the World Health Organization STEPS protocol and glycated haemoglobin A1c (HbA1c) testing, respectively. Nurses conducted follow-up home visits for confirmatory diagnostic testing among individuals with a screening BP above 140/90 mmHg and/or HbA1c above 6.5% at the initial screen, and without a prior diagnosis. We assessed the positive predictive value of the initial screening, compared to the follow up measure. We also sought to identify a screening threshold for HTN and DM with greater than 90% positive predictive value. results: Of 18,027 participants enrolled, 10.2% (1,831) had a screening BP over 140/90 mmHg. Of those without a prior diagnosis, 871 (47.6%) received follow-up measurements. Only 51.2% (451) of those with completed follow-up measurements had a repeat BP>140/90 mmHg at the home visit and were referred to care. To achieve a 90% correct referral rate, a systolic BP threshold of 192 was needed at first screening. For DM screening, 1,615 (9.0%) individuals had an HbA1c > 6.5%, and of those without a prior diagnosis, 1,151 (71.2%) received a follow-up blood glucose. Of these, only 34.1% (395) met criteria for referral for DM. To ensure a 90% positive predictive value i.e. a screening HbA1c of >16.6% was needed. conclusions: A second home-based screening visit to confirm a diagnosis of DM and HTN reduced health system referrals by 48% and 66%, respectively. Two-day screening programmes for DM and HTN screening might save individual and healthcare resources and should be evaluated carefully in future cost effectiveness evaluations.

Quantifying met and unmet health needs for HIV, hypertension and diabetes in rural KwaZulu-Natal, South Africa

“…Secondly, Vukuzazi only enrolled half of the eligible population which may have biased description of health needs and their associations in directions hard to anticipate based on known differences between the sampled and unsampled population (12). We acknowledge that people who screened positive for DM and HPTN required confirmatory testing prior to confirmation of diagnosis, and that this testing could rule out disease requiring immediate treatment; thus, we may have overestimated the burden of undiagnosed disease (30). Lastly, we acknowledge that it is an oversimplification to ascribe "no health needs" to people who screen negative for disease because it neglects the need for interventions targeting disease prevention, which may be critical for optimal community health.…”

Convergence of HIV and non-communicable disease epidemics: geospatial mapping of the unmet health needs in an HIV hyperendemic community in South Africa

“…Therefore, we acknowledge that people who screened positive for diabetes and hypertension required confirmatory testing prior to confirmation of diagnosis, and that this testing could rule out disease requiring immediate treatment. Thus, we may have overestimated the burden of undiagnosed disease 45. Furthermore, only three chronic disease conditions were considered in this study, and it must be taken into consideration that there are several other additional NCDs which have not been included in the Vukuzazi study including cancer, chronic respiratory diseases, mental health and other NCDs, limiting our ability to draw comprehensive conclusions about population health and multimorbidity.…”

Prevalence of chronic conditions and multimorbidity among healthcare workers in Zimbabwe: Results from a screening intervention

“…The pragmatic, cross-sectional design of the health-checks means that our estimates were based on testing at a single time point using point-of-care tools and there were considerable missing data for HIV and HbA1c. Previous studies have suggested that blood pressure measurement on a single day may overestimate hypertension prevalence [ 32 ]. HbA1c has high specificity but low sensitivity (58%) for diabetes in Africa; HbA1c-based screening may therefore underestimate diabetes prevalence [ 33 ].…”