Implementing focused echocardiography and AI-supported analysis in a population-based survey in Lesotho: implications for community-based cardiovascular disease care models

In settings where access to expert echocardiography is limited, focused echocardiography, combined with artificial intelligence (AI)-supported analysis, may improve diagnosis and monitoring of left ventricular hypertrophy (LVH). Sixteen nurses/nurse-assistants without prior experience in echocardiography underwent a 2-day hands-on intensive training to learn how to assess parasternal long axis views (PLAX) using an inexpensive hand-held ultrasound device in Lesotho, Southern Africa. Loops were stored on a cloud-drive, analyzed using deep learning algorithms at the University Hospital Basel, and afterwards confirmed by a board-certified cardiologist. The nurses/nurse-assistants obtained 756 echocardiograms. Of the 754 uploaded image files, 628 (83.3%) were evaluable by deep learning algorithms. Of those, results of 514/628 (81.9%) were confirmed by a cardiologist. Of the 126 not evaluable by the AI algorithm, 46 (36.5%) were manually evaluable. Overall, 660 (87.5%) uploaded files were evaluable and confirmed. Following short-term training of nursing cadres, a high proportion of obtained PLAX was evaluable using AI-supported analysis. This could be a basis for AI- and telemedical support in hard-to-reach areas with minimal resources.

High Rates of Undiagnosed Target Organ Damage Among Adults with Elevated Blood Pressure or Diabetes Mellitus in a Community-Based Survey in Lesotho

INTRODUCTION

Prevalence of elevated blood pressure (BP) and diabetes mellitus (DM) is increasing in sub-Saharan Africa. Data on target organ damage such as retinopathy, left ventricular hypertrophy (LVH), renal impairment and peripheral neuropathy (PN) among persons with elevated BP and/or DM in sub-Saharan Africa remain scarce.

AIM

To determine at community-level the prevalence of retinopathy, LVH, renal impairment, and PN among adults with elevated BP and/or DM, and assess the association of elevated BP and/or DM with target organ damage in Lesotho.

METHODS

During a household-based survey, a sub-sample of adults with elevated BP (≥ 140/90 mmHg) and/or DM (glycosylated hemoglobin ≥ 6.5%), as well as comparators (BP < 140/90 mmHg, HbA1c < 6.5%) were screened for retinopathy, LVH, renal impairment, and PN. We used multivariable logistic regression for inferential analysis.

RESULTS

Out of 6108 participants screened during the survey, 420 with elevated BP only, 80 with DM only, 61 with elevated BP and DM, and 360 comparators were assessed for target organ damage. Among those with elevated BP, and among those with DM with or without elevated BP, prevalence of retinopathy was 34.6% (89/257) and 14.4% (15/104); renal impairment was 45.0% (156/347) and 42.4% (56/132), respectively. Among those with elevated BP, 2.3% (7/300) and 65.7% (224/341) had LVH and left ventricular concentric remodeling, respectively. PN, only assessed among those with DM, was present in 32.6% (42/129). Elevated BP was associated with increased odds of retinopathy (aOR, 19.13; 95% CI, 8.52-42.94; P < 0.001) and renal impairment (aOR, 1.80; 95% CI, 1.27-2.55; P = 0.001). Presence of both elevated BP and DM was associated with an increased odds of retinopathy (aOR, 16.30; 95%CI, 5.69-46.68; P < 0.001), renal impairment (aOR, 2.55; 95% CI, 1.35-4.81; P = 0.004), and PN (aOR, 2.13; 95% CI, 1.04-4.38; P = 0.040).

CONCLUSION

We found a high prevalence of undiagnosed target organ damage among adults with elevated BP and/or DM during community-based screening. These findings emphasize the importance of regular prevention and screening activities in this setting.

Community-based type 2 diabetes care by lay village health workers in rural Lesotho: protocol for a cluster-randomized trial within the ComBaCaL cohort study (ComBaCaL T2D TwiC)

BACKGROUND

Type 2 diabetes (T2D) poses a growing public health burden, especially in low- and middle-income countries (LMICs). Task-shifting to lay village health workers (VHWs) and the use of digital clinical decision support systems (CDSS) are promising approaches to tackle the current T2D care gap in LMICs. However, evidence on the effectiveness of lay worker-led T2D care models, in which VHWs initiate and monitor drug treatment in addition to community-based screening and referral services, is lacking.

METHODS

We are conducting a cluster-randomized trial nested within the Community-Based Chronic Disease Care Lesotho (ComBaCaL) cohort study (NCT05596773) using the trial within cohort (TwiC) design to assess the effectiveness of a VHW-led, CDSS-assisted T2D care model in rural Lesotho. Participants are non-pregnant members of the ComBaCaL cohort study with T2D. The ComBaCaL cohort study is conducted in approximately 100 villages in two rural districts in Lesotho and is managed by trained and supervised VHWs. In intervention villages, VHWs offer a community-based T2D care package including lifestyle counselling, first-line oral antidiabetic, lipid-lowering, and antiplatelet treatment guided by a tablet-based CDSS to participants who are clinically eligible, as well as treatment support to participants who prefer or clinically require facility-based T2D care. In control clusters, all participants will be referred to a health facility for T2D management. The primary endpoint is the mean glycosylated haemoglobin (HbA1c) 12 months after enrolment. Secondary endpoints include the 10-year risk for cardiovascular events estimated using the World Health Organization risk prediction tool.

DISCUSSION

The trial was launched on May 13, 2023, and has enrolled 226 participants at the date of submission (October 6, 2023). To our knowledge, the trial is the first to assess task-shifting of T2D care to VHWs at the community level, including the prescription of first-line antidiabetic, lipid-lowering, and antiplatelet medication in sub-Saharan Africa, and will thus provide the missing evidence on the effectiveness of such a T2D care model in this setting. The study is operating within the established Lesotho VHW programme. Similar community health worker programmes which exist across sub-Saharan Africa may benefit from the findings.

TRIAL REGISTRATION

ClinicalTrials.gov NCT05743387. Registered on February 24 2023.

Head-to-head comparison of the WHO STEPwise approach with immediate unattended and delayed unattended automated blood pressure measurements during household-based screening: a diagnostic accuracy study in Lesotho

Background

WHO introduced the STEPwise approach to surveillance (STEPS) to monitor trends in non-communicable diseases. For arterial hypertension, the STEPS protocol takes the average of the last two out of three standard blood pressure measurements (SBPM). This study assesses the diagnostic accuracy of SBPM, same-day and next-day unattended automated measurement (uABP), with 24 h ambulatory measurement (24 h-ABPM) as reference.

Methods

This diagnostic accuracy study was done within a population-based household survey on cardiovascular risk factors in two districts in Northern Lesotho. Adults (aged ≥ 18 years) with elevated SBPM (defined as ≥140/90 mmHg), and 2:1 age- and sex-matched participants with normal SBPM during the survey were recruited. Following SBPM, first uABP readings were obtained on survey day. Afterwards, participants received a 24 h-ABPM device. Second uABP readings were taken 24 h later, after retrieval of the 24 h-ABPM. The main outcome was overall diagnostic accuracy of all screening measurements (SBPM, first uABP, and second uABP), determined using area under the receiver operating characteristic curve (AUROC), with 24 h-ABPM as a reference.

Findings

Between November 2, 2021 and August 31, 2022, 275 participants (mean age 58 years (SD: 16 years), 163 (59%) female) were enrolled, 183 of whom had elevated and 92 had normal SBPM. Mean difference between systolic daytime 24 h-ABPM and screening measurements was highest for SBPM (mean difference: -13 mmHg; 95% CI: -14 to -11). Mean difference between diastolic daytime 24 h-ABPM and diastolic SBPM was -2 mmHg (95% CI: -4 to -1), whereas no difference was found for mean diastolic first uABP (mean difference: -1 mmHg; 95% CI: -2.0 to 0.3); and mean diastolic second uABP (mean difference: 1.0 mmHg; 95% CI: -0.4 to 2.3). White coat hypertension was highest with SBPM (55 [20%]), followed by first uABP (27 [9.8%]), and second uABP (18 [6.5%]). Using systolic daytime 24 h-ABPM as a reference, the uABPs had higher AUROC (first uABP: 87% [95% CI: 83-91]; second uABP: 88% [95% CI: 84-92]); SBPM: (79% [95% CI: 74-85]). This difference was significant between first uABP and SBPM (P = 0.0024), and between second uABP and SBPM (P = 0.0017).

Interpretation

uABP had better diagnostic performance than SBPM. Integration of uABP into STEPS protocol should be considered.

Funding

Swiss Agency for Development and Cooperation under the ComBaCaL project, and the World Diabetes Foundation.