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.

Facilitating high quality acute care in resource-constrained environments: Perspectives of patients recovering from sepsis, their caregivers and healthcare workers in Uganda and Malawi.

Sepsis is a major global health problem, especially in sub-Saharan Africa. Improving patient care requires that healthcare providers understand patients' priorities and provide quality care within the confines of the context they work. We report the perspectives of patients, caregivers and healthcare workers regarding care quality for patients admitted for sepsis to public hospitals in Uganda and Malawi. This qualitative descriptive study in two hospitals included face-to face semi-structured interviews with purposively selected patients recovering from sepsis, their caregivers and healthcare workers. In both Malawi and Uganda, sepsis care often occurred in resource-constrained environments which undermined healthcare workers' capacity to deliver safe, consistent and accessible care. Constraints included limited space, strained; water, sanitation and hygiene (WASH) amenities and practices, inadequate human and material resources and inadequate provision for basic needs including nutrition. Heavy workloads for healthcare workers strained relationships, led to poor communication and reduced engagement with patients and caregivers. These consequences were exacerbated by understaffing which affected handover and continuity of care. All groups (healthcare workers, patients and caregivers) reported delays in care due to long queues and lack of compliance with procedures for triage, treatment, stabilization and monitoring due to a lack of expertise, supervision and context-specific sepsis management guidelines. Quality sepsis care relies on effective severity-based triaging, rapid treatment of emergencies and individualised testing to confirm diagnosis and monitoring. Hospitals in resource-constrained systems contend with limitations in key resources, including for space, staff, expertise, equipment and medicines, in turn contributing to gaps in areas such as WASH and effective care delivery, as well as communication and other relational aspects of care. These limitations are the predominant challenges to achieving high quality care.