High blood pressure and high blood sugar diseases increase the risk of heart disease, stroke and kidney disease. These are the leading causes of death worldwide collectively accounting for more than 17 million deaths worldwide in 2018.  Whilst diseases that can spread from person to person also known as communicable diseases are responsible for more deaths in poor countries, societal shifts and urbanization over the last decade have been driving an increase in death and disease-related complications from heart disease. In these regions, high blood pressure and high blood sugar diseases are frequently not discovered by healthcare professionals and therefore untreated in up to two thirds of adults according to the International Diabetes Federation (IDF) and the World Health Organization. High blood sugar disease remains significantly associated with death from heart disease and the association increases as the level of blood sugar control becomes poorer. The healthcare professionals use a measurement which can tell them the level of blood sugar control for a period of the past 3 months in a person. This measurement is called HbA1c and refers to how much sugar has coated the blood during this period.

 In Kenya, heart disease is the second leading cause of death only rivalled by diseases that spread from person to person also called communicable diseases. Advances in computer-based healthcare also called digital health may provide opportunities to improve access to healthcare and enhance how high blood sugar and high blood pressure can be discovered by health care workers in settings with limited healthcare infrastructure.

This study aims to use one such computer-based healthcare system called HealthyRoute to predict blood pressure, level of sugar coating in the blood and kidney function using images or photographs of the eye.

Various research groups have trained computers to diagnose and grade complications of the eye arising from high blood sugar using computerized images or photographs of the eye and also, make predictions of blood pressure and level of sugar coating in the blood. Further development of this technology could deliver on-the-spot detection of high blood pressure, high blood sugar and kidney disease and could be deployed in virtually any setting. Improved diagnosis could facilitate earlier discovery of disease and earlier treatment and subsequently lead to useful improvements in health and mortality in low- income countries where the ratio of healthcare workers per patient are not ideal.

The study will enrol approximately 300 participants for the entire study. There are 2 sites in Kenya and the CRDR KEMRI Nairobi site will enrol 150 participants and the second site, the Aga Khan University Hospital will also enrol 150 participants. CRDR KEMRI will have the potential of increasing enrolment in case the second site has difficulties to reach their enrolment goal. This is the case because the former is a more experienced site with clinical research.

These participants will be enrolled in two phases; roll-in and actual study. The first phase would be a roll-in, this means a few participants are enrolled for taking of the retinal images solely for the purposes of ensuring high quality of images and ensuring the site staff becomes acquainted or comfortable with the retinal image technology. The site will then hold enrollment while the Study Sponsor assesses retinal image quality and this would take a few days ideally less than a week. If the images are deemed to be of high quality, the site will be given a go ahead to commence enrolment in to the second phase which is the actual study. The two phases have separate consent forms so that if one wants to exit after roll in phase that would be allowed. Roll in phase participants can be rescreened to enter the actual study.

Rationale:

Establish scalable methodology for collection of retinal images, blood pressure (BP) and laboratory-based assessments

Compare the results of a machine-learning algorithm in predicting BP, glycated haemoglobin (HbA1c) and estimated glomerular filtration rate (eGFR) from digital retinal images with clinical and laboratory-based measures

Determine the required sample size needed to support a future study to fully validate the machine-learning algorithm

Objectives and Endpoints

Primary Objective

• To assess successful use of methodology for retinal image acquisition, blood pressure, HbA1c and eGFR measurement

Endpoint.

• Proportion of participants with completed study procedures and interpretable results including both retinal images, systolic and diastolic BP, HbA1c (laboratory based or point of care test) and eGFR at Visit 1.

Overall Design

Disclosure Statement: This is a non-interventional pilot study

Number of Participants: Approximately 300 participants will be enrolled.

Note: "Enrolled" means a participant's, or their legally acceptable representative’s, agreement to participate in a clinical study following completion of the informed consent process. Potential participants who are screened for the purpose of determining eligibility for the study, but are not assigned in the study, are considered “screen failures”, unless otherwise specified by the protocol.

Intervention Groups and Duration: Non-interventional; total study duration up to 66 days.

Data Monitoring Committee: Not applicable

Sample size:

Formal sample size calculation is not required for this pilot study; recruitment of approximately 300 participants is empirically determined and deemed adequate to assess and optimise methodology and estimate sample size required for a future larger study. The results of this study will be used to inform powering and sample size calculation for a future larger validation study.

Statistical analyses:

There will be no formal hypothesis testing. All data collected during the study will be presented as summary statistics. No formal statistical models will be required in this study.