Using artificial intelligence on dermatology conditions in Uganda: a case for diversity in training data sets for machine learning

Background

In pursuit of applying universal non-biased Artificial Intelligence (AI) in healthcare, it is essential that data from different geographies are represented.

Objective

To assess the diagnostic performance of an AI-powered dermatological algorithm called Skin Image Search on Fitzpatrick 6 skin type (dark skin) dermatological conditions.

Methods

123 dermatological images selected from a total of 173 images were retrospectively extracted from the electronic database of a Ugandan telehealth company, The Medical Concierge Group (TMCG) after getting their consent. Details of age, gender, and dermatological clinical diagnosis were analysed using R on R studio software to assess the diagnostic accuracy of the AI app along with disease diagnosis and body part. Predictability levels of the AI app were graded on a scale of 0 to 5, where 0- no prediction was made and 1-5 demonstrated a reduction incorrect diagnosis prediction rate of the AI.

Results

76 (62%) of the dermatological images were from females and 47 (38%) from males. Overall diagnostic accuracy of the AI app on black dermatological conditions was low at 17% (21 out of 123 predictable images) compared to 69.9% performance on Caucasian skin type as reported from the training results. There were varying predictability levels correctness i.e., 1-8.9%, 2-2.4%, 3-2.4%, 4-1.6%, 5-1.6% with performance along individual diagnosis highest with dermatitis (80%).

Conclusion

There is need for diversity of image datasets used to train dermatology algorithms for AI applications to increase accuracy across skin types and geographies.

Qualitative study of mental health attribution, perceptions and care-seeking in Kampala, Uganda

Background

Mental health problems contribute to a substantial proportion of the global burden of disease. In Uganda, the World Health Organization estimates that 2.2 million people are affected by mental health disorders. Further research is needed to highlight people’s views about mental health in order to ensure that services are appropriate, accessible and effective.

Aim

This qualitative study aimed to explore perceptions, experiences and care-seeking preferences to inform stakeholders looking to provide contextually appropriate mental health programmes.

Setting

A diverse neighbourhood in central Kampala, Uganda.

Methods

The authors conducted 56 in-depth semi-structured interviews with people over the age of 37 years from November 2018 to May 2019.

Results

Participants discussed interpersonal and systemic issues that affect mental health in their community and the existing coping mechanisms that people employ. Social factors were often associated with mental health problems, with 36% of participants attributing them to economic stressors in particular. Mental health services were often perceived to be unavailable, costly or stigmatised, which can mean that care-seeking is delayed until problems become severe. Some people said they prefer to turn to prayer (25%) or counselling within their family or community (12.5%).

Conclusion

Mental health problems are often attributed to socioeconomic factors, which can also hinder access to services. An understanding of perceptions about mental health can help to align programmes for appropriateness and effectiveness. Our study suggests that beneficial additional services for people living in low-income urban settings in Uganda could include those which are free, community-based or offering financial support.

HiSTEP: A Single-Arm Pilot Study of a Technology-Assisted HIV Self-testing Intervention in Kampala, Uganda

We developed and pilot tested a 3-month HIV self-testing intervention called HiSTEP ("HIV Self-testing Engagement Project") among 95 adult (18+ years) at-risk (condomless sex < 3 months) adults in Kampala, Uganda. HiSTEP leverages theoretically-grounded (in the Information-Motivation-Behavioral Skills model) text messages, a telehealth centre with live support, and "last-mile" HIV self-testing kit delivery to a location chosen by the participant. Nearly 94% of participants were retained at month 3. HIV self-testing was highly acceptable across age and gender groups (94% very satisfied), although older women had slightly lower acceptability ratings (92% very satisfied). Only 13% of participants used HIV self-testing prior to enrollment. Over the 3-month study period, 86% of participants ordered a total of 169 HIV self-testing kits (69% for participant use; 31% for use by others). Findings show that the intervention approach taken in HiSTEP may be particularly valuable for engaging at-risk Ugandan adults in HIV self-testing using a novel technology-assisted promotion and delivery method.

Feasibility and acceptability of a digitally supported channel for behavioral messaging on modern contraception: A pilot cross-section study in Uganda (Preprint)

Background

With the region’s highest population growth rate (30%), Uganda is on the brink of a population explosion, yet access to and utilization of public health control measures like modern contraception is a challenge. This is due to remotely located health facilities, noncustomized health content, and poor or nonfunctional post-facility follow-up.

Objective

The aim of our study was to evaluate the feasibility and acceptability of a telehealth engagement platform primarily targeting men; the platform provided behavioral and informational messaging on modern contraception (ie, family planning) and its impact on shaping sexual and reproductive health and knowledge and uptake of family planning services.

Methods

A longitudinal cohort of men aged 18 years and older gave consent to receive mobile phone messages on family planning; follow-up was performed at months 1, 4, and 6 to assess key study-related outcomes on knowledge transfer and acquisition on modern contraception, partner communication, and spousal uptake of family planning. Qualitative interviews with the study participants’ spouses were also performed.

Results

The study included 551 study participants, 450 of whom were men, the primary study participants, who received the family planning mobile messages and 101 of whom were their spouses. Of the 450 primary participants, 426 (95%) successfully received the messages and only 24 (5%) reported not receiving them. The average response (ie, participation) rate in weekly quizzes was 23%. There was a noted 18.1% increase in couple communication attributed to the intervention; couples opened up more to each other on matters concerning family planning.

Conclusions

Using digital channels to address the concerns and inquiries of participants in real time or as fast as possible helped to increase the likelihood that couples adopted family planning.

Point-of-care testing for HIV and sexually transmitted infections reveals risky behavior among men at gambling centers in Uganda"

In sub-Saharan Africa (SSA), men are generally difficult to engage in healthcare programs. However, sports gambling centers in SSA can be used as avenues for male engagement in health programs. We offered point-of-care HIV and syphilis testing for men located at five gambling centers in Uganda and assessed HIV risky sexual behavior. Among 507 men, 0.8% were HIV-positive and 3.8% had syphilis. Risky sexual behavior included condomless sex with partner(s) of unknown HIV status (64.9%), having multiple sexual partners (47.8%), engaging in transactional sex (15.5%), and using illicit drugs (9.3%). The majority at 64.5% were nonalcohol consumers, 22.9% were moderate users, and 12.6% had hazardous consumption patterns. In 12 months of follow-up, the incidence rate of syphilis was 0.95 (95% CI: 0.82-1.06) among 178 men. Thus, men in SSA have a high prevalence of syphilis and risky sexual behavior which should be more effectively addressed to reduce the risk of HIV acquisition.

Continuity of health service delivery during the COVID-19 pandemic: the role of digital health technologies in Uganda

In response to coronavirus disease-2019 pandemic (COVID-19), the government of Uganda instituted movement restrictions to curb disease spread. However, this affected accessibility to medical services in a setting where the healthcare system is not equipped to handle most healthcare needs of the populace outside hospital premises. This gap led to the prominence and unprecedented rise in the use of digital health technologies to deliver health information and services at a distance (telehealth) during the COVID-19 outbreak. The use of telehealth modalities including tele-consultation, tele-psychiatry, call centers and mobile phone health information dissemination increased. The COVID-19 pandemic augmented the rising role of digital health technologies as a much needed aspect of medical service delivery in our times. However, the efficacy and impact on clinical outcomes across various healthcare thematic areas need to be explored further and more evidence generated.