Emmanuel Kiiza Mwesiga

Background Rich clinical information is often stored in handwritten patient charts, but there are limited systems for converting these routine records into usable data for research, care, service improvement, and policy. This process paper describes the development of Hekima, a locally hosted, AI-assisted mental health data platform that converts paper psychiatric records into structured longitudinal datasets for research. Methods We reviewed implementation documents generated during the Digitising Psychosis Project of the Mental Health Data Prize Africa program. We reconstructed the development process, major adaptations, technical and governance decisions, early analytic outputs, capacity-building activities, public communication, and scale-up pathway. Results Hekima evolved from a psychosis chart digitisation project into a broader mental health data platform. Hekima followed a “one pipeline, two outputs” model. Each scanned document was transcribed, reviewed, redacted, and structured into a longitudinal patient record, but also prepared for patient matching with existing electronic medical record systems. The project digitised and archived 203,020 images from 7,164 files. Early OCR-based extraction demonstrated feasibility but performed poorly on handwritten psychiatric notes. This prompted a redesign toward locally hosted vision-language models, redaction of identifiable information, extraction, and mandatory human review. Lived-experience partners contributed to privacy safeguards, engagement interpretation, data-use acceptability, patient-rights protection, public communication, and independent process review. Conclusion Hekima demonstrated that handwritten psychiatric records in a low-resource setting could be transformed into usable mental health data infrastructure but required data governance, lived-experience oversight, capacity building, communication, and engagement with policymakers.

Background: Catatonia is a neuropsychiatric emergency that requires rapid intervention to prevent complications. It is unclear whether African medical practitioners have the requisite knowledge to identify and manage catatonia. Aim: To examine knowledge of catatonia among medical practitioners across partnering institutions of the African Catatonia Registry (ACR), a multi-country resource for research in Africa. Setting: The study was conducted among medical practitioners across the ACR partnering institutions. Methods: The study was a descriptive, cross-sectional online survey comprising 63 questions to examine knowledge of catatonia. Responses were scored with a positive mark if correct and a negative mark if wrong, then standardised and dichotomised to indicate low or sufficient knowledge. Results: Out of 321 participants, nearly half (46.7%) were psychiatrists, and 50.8% had sufficient knowledge of catatonia. Multivariate analysis showed that knowledge varied by sub-domain, country, specialty and prior experience with catatonia. Specialisation in psychiatry was significantly associated with sufficient knowledge of classification (adjusted risk ratio [aRR] 2.15 [95% confidence intervals {CI}: 1.35–3.43]), diagnostic symptoms (aRR = 3.30 [95% CI: 1.62–6.69]), screening (aRR = 4.72 [95% CI: 2.25–9.91]) and rating scales (aRR = 3.34 [95% CI: 1.32–8.46]). Prior experience with catatonia was also associated with sufficient knowledge across these areas (all p ≤ 0.05). Conclusion: While knowledge of catatonia and its diagnostic criteria is sufficient among practitioners, knowledge of its symptoms, treatment and complication management remains low. Contribution: The study highlights the need for the ACR to promote improvements in the identification and management of catatonia to achieve better patient outcomes.

A large number of adolescents with hearing impairment experience socio-emotional and structural challenges that may put them at risk for depression. However, depression is often undiagnosed among them due to lack of appropriate, culturally relevant tools for its assessment. This study set out to determine the prevalence of depression and its associated factors in this vulnerable population. A cross-sectional study was conducted among adolescents aged 10–17 years with hearing impairment, attending three primary schools of the deaf across Kampala City in Uganda. Depression was assessed using the Mini International Neuropsychiatric Interview for children and adolescents (MINI-KID), which underwent linguistic adaptation and translation into Ugandan Sign Language through a collaborative process involving sign language teachers and students, and administered in video format to ensure clarity and inclusion. Data were analyzed with Stata version 17. Modified Poisson regression with robust standard errors was used to examine associations. A total of 116 adolescents participated (mean age = 14.5 ± 1.9 years; 51.7% female). The lifetime prevalence of depression was 60.3%, with point prevalence at 34.5%. Depression was more frequent among females, though these differences were not statistically significant. Living in father-only households was associated with higher likelihood of depression (IRR = 2.11, 95% CI: 1.60–2.77 p < 0.001), while not experiencing or witnessing excessive disciplinary measures at school was associated with a lower likelihood of depression (IRR = 0.65, 95% CI: 0.49–0.87, p = 0.003). Using a linguistically adapted version of the MINI-KID, we found that depression is highly prevalent among adolescents with hearing impairment in Kampala. Strengthening family support, improving access to sign language-trained professionals, and implementing inclusive school-based interventions are key to identifying and managing depression in this population.

Background: We previously examined the burden and predictors of sustained mental health care engagement in Ugandan first episode psychosis patients by retrospective chart review methods. However, the extensive requirements of chart reviews meant that we could only extract data from a random 10% sample of 1,677 newly enrolled Ugandan first episode psychosis patients at Butabika National Referral Mental Hospital in 2018. The Hekima Platform has been designed to transform handwritten files into datasets for analysis. Objectives: This study aims to: (1) utilize the Hekima Platform to transform paper-based clinical charts of all 1,677 Ugandan psychosis patients enrolled at Butabika Hospital for the first time in 2018 into a standardized, anonymized longitudinal database; and (2) re-examine predictors of sustained MHC engagement in this cohort. Methods: We will digitize and archive all patient charts. We will then use the Hekima Platform to extract handwritten clinical data into machine-readable text using user-trained machine learning and deep learning models and natural language processing (NLP) techniques to generate a structured, anonymized database. A minimum 10% random sample of extracted data will be manually validated using Cohen’s kappa. For the analytical aim, descriptive statistics, bivariate analysis, and multivariable logistic regression will model predictors of sustained engagement, with exploratory machine learning approaches used as a complementary analytical strategy. Ethical approval has been obtained from the Uganda National Council for Science and Technology and Butabika Hospital’s Research Ethics Committee. Expected outcomes: Patient clinical charts are a rich data source but there are extensive requirements to be able to use them for research. This study will generate the first AI-assisted, standardized longitudinal database from handwritten psychiatric records in Uganda, enabling well-powered analyses of predictors of MHC engagement. Findings will inform targeted interventions to improve retention in care and will offer a scalable model for mental health research in low- and middle-income countries.

Introduction: Engagement with mental health services (MHCS) during the first episode of psychosis (FEP) is critical for symptom control, quality of life, and relapse prevention. However, disengagement rates remain high in Uganda with severe consequences for patients and caregivers. This study protocol describes a mixed-methods investigation which aims to examine the relationship between patients and caregivers lived experiences and mental health service engagement during first-episode psychosis. Methods and Analysis: The mixed-methods study will recruit 82 patients with first-episode psychosis and their primary caregivers from Butabika National Referral Mental Hospital in Kampala, Uganda. Inclusion criteria are ages 18-60, less than 12 weeks on antipsychotic medications, living in the greater Kampala Metropolitan Area, with a consenting caregiver. Caregivers must be an adult (> 18years) providing full-time care for at least 6 months prior. Patients with substance use disorders will be excluded. Qualitative data on the lived experiences of patients and caregivers will be collected using the draw-write-and-tell method, while quantitative data on service engagement and associated factors will be collected using semi-structured questionnaires. The data will be analysed using Stata version 18, and participants will be reimbursed for their time. Ethics and Dissemination: Ethical clearance has been obtained from the School of Medicine Research and Ethics Committee (SOMREC) Ref: Mak-SOMREC-2024-1002 and institutional approval from Butabiika National Referral Mental Hospital. All participants will provide informed consent prior to participation. Data will be de-identified and securely stored, with results disseminated through peer-reviewed academic publications, conferences and community stakeholder workshops. Strength and limitations of the study: ▯ This longitudinal mixed-methods study integrates patient and caregiver perspectives, enabling a holistic understanding of mental health service engagement during the first episode of psychosis. Particularly the use of an innovative draw-write-and-tell method for qualitative research in this population is novel. ▯ Inclusion of lived-experience narratives alongside service engagement trajectories strengthens the relevance of findings for mental health service design and early intervention strategies in low-resource settings. ▯ Conducting the study at a national referral mental health hospital allows in-depth examination of care pathways within a specialist service context; however, findings may not be fully generalisable to community-based or non-specialist settings. ▯ Exclusion of individuals with comorbid substance use disorders may limit the applicability of findings, given the high prevalence of substance use among people experiencing first-episode psychosis.

Background: Early Intervention for Psychosis Services (EIPS) enhance outcomes for individuals experiencing their first episode of psychosis (FEP). However, in low-resource settings, there is limited knowledge about i) the pathways patients take to access EIPS, ii) the proportion and factors associated with acceptance of referral to EIPS, and iii) if different pathways to EIPS services affect clinical outcomes. Uganda’s first EIPS, the Specialised Treatment Early in Psychosis Service at Makerere University Hospital (STEP_MaKH), presents a unique opportunity to explore these important questions. Aims: We aimed to examine the pathways to EIPS, the factors associated with referral to specialised psychosis care and the impact of initial treatment-seeking behaviour on long-term symptom remission and quality of life. Methods: We conducted a multiple-method study. Pathways to care were assessed retrospectively using the WHO Encounter Form among adults with FEP eligible for referral to STEP_MaKH. Among those who completed referral and enrolled in STEP_MaKH. Symptom severity and quality of life were followed prospectively for 12 months. Modified Poisson regression identified predictors of referral completion. Kaplan–Meier methods and Cox proportional hazards models examined time to symptom remission and time to achieving a good quality of life. Results: Of the 187 adults with first-episode psychosis eligible for referral to STEP_MaKH, Native/religious healers (n = 86) were the predominant first point of contact. Only 56 (29.9%) accepted referral to STEP_MaKH. Participants referred from Mulago National Referral Hospital more likely to enrol than those referred from Butabika (RR = 4.7; 95% CI: 2.90–7.87). Longer delays from first treatment contact were associated with reduced likelihood of reaching STEP_MaKH (RR = 0.99 per month; p = 0.041). After enrolment, symptoms improved rapidly with 60% achieving PANSS remission by Month 1, and fewer than 10% remained non-remitted by Months 2–3. In adjusted Cox models, participants initially seen by mental health workers achieved remission more quickly than those initially seen by non-medical personnel (HR = 1.48; 95% CI: 1.05–2.10). Older age was associated with slower remission (HR = 0.94; p = 0.023). Quality of life improved over the follow-up period, with earlier attainment of good quality of life among those initially managed by mental health workers. Conclusions: Pathways to care for FEP in Uganda are complex and culturally mediated, with substantial attrition before specialised early psychosis care is reached. Referral completion is strongly shaped by referral site and by delays in the care pathway. Once in specialised care, clinical outcomes improve rapidly, and initial contact with mental health workers is associated with faster symptom remission and earlier gains in quality of life. Strengthening referral systems, reducing pathway delays, and developing collaborative detection-and-referral links with community and frontline providers are key priorities for optimising early psychosis outcomes in low-resource settings.

Comorbid depression and anxiety in patients with diabetes mellitus (DM) is associated with poor disease outcomes, yet its burden and associated factors are not well-characterized in low-resource settings like Uganda. This study aimed to determine the prevalence and factors associated with comorbid depression and anxiety among patients with DM in Uganda. A cross-sectional study was conducted at the diabetes clinic of Mulago National Specialised Hospital. The study consecutively enrolled 223 adult patients (≥18 years) with DM. The Mini-International Neuropsychiatric Interview (MINI) was used to diagnose depression and anxiety disorders. Data on socio-demographics, clinical characteristics, alcohol use (AUDIT-C), social support (MSPSS), and quality of life (WHOQOL-BREF) were collected. Logistic regression was used to identify factors associated with comorbid depression and anxiety. The mean age of participants was 54.6 years (SD = 13.1), and 72.2% were female. The prevalence of comorbid depression and anxiety was 14.3% (95% CI: 10.3–19.6). In the multivariable analysis, factors significantly associated with increased odds of comorbidity included having a higher number of children (Adjusted Odds Ratio, aOR=1.15, 95% CI: 1.02–1.32), longer duration since DM diagnosis (aOR=1.07, 95% CI: 1.01–1.12), high blood pressure (aOR=2.19, 95% CI: 1.94–5.08), and moderate/high alcohol use (aOR=1.46, 95% CI: 1.09–5.14). Conversely, diagnosis of Type II DM (aOR=0.40, 95% CI: 0.16 - 0.91), older age (aOR=0.97, 95% CI: 0.94–0.99) and higher scores across all WHOQOL-BREF domains (physical, psychological, social, environmental) were associated with significantly reduced odds of comorbidity. Nearly one in seven patients with DM in this Ugandan cohort had comorbid depression and anxiety. The findings underscore the need for integrated mental health screening and intervention within diabetic care services, particularly targeting younger patients, those with Type 1 DM, longer disease duration, hypertension, and hazardous alcohol use.

Background: A large number of adolescents with hearing impairment experience socio-emotional and structural challenges that may put them at risk for depression. However, depression is often undiagnosed among them due to lack of appropriate, culturally relevant tools for its assessment. This study set out to determine the prevalence of depression and its associated factors in this vulnerable population. Methods: A cross-sectional study was conducted among adolescents aged 10 to 17 years with hearing impairment, attending three primary schools of the deaf across Kampala City in Uganda. Depression was assessed using the Mini International Neuropsychiatric Interview for children and adolescents (MINI-KID), which underwent linguistic adaptation and translation into Ugandan Sign Language through a collaborative process involving sign language teachers and students, and administered in video format to ensure clarity and inclusion. Data were analyzed with Stata version 17. Modified Poisson regression with robust standard errors was used to examine associations. Results: A total of 116 adolescents participated (mean age = 14.5 ± 1.9 years; 51.7% female). The lifetime prevalence of depression was 60.3%, with point prevalence at 34.5%. Depression was more frequent among females, though these differences were not statistically significant. Living in father-only households (IRR = 2.11, 95% CI: 1.60–2.77 p < 0.001), and witnessing excessive disciplinary practices (IRR = 0.65, 95% CI: 0.49–0.87, p = 0.003), were significantly associated with depression. Conclusion: Using a linguistically adapted version of the MINI-KID, we found that depression is highly prevalent among adolescents with hearing impairment in Kampala. Strengthening family support, improving access to sign language-trained professionals, and implementing inclusive school-based interventions are key to identifying and managing depression in this population.

Background: Social protection programs, such as Uganda’s Senior Citizens Grant (SCG), are designed to provide financial support to improve the welfare of older adults. Neurocognitive impairment affects financial competence and may impair beneficiaries’ ability to manage financial resources effectively. However, evidence from low-income settings remains limited. This study therefore examined the association between neurocognitive impairment and financial competence in elderly Ugandans receiving social assistance grants. Methods: A cross-sectional study was conducted among 588 elderly SCG beneficiaries sampled from ten Ugandan districts. Data were collected between 4th October and 19th November 2024 during the scheduled SCG disbursement period. Neurocognitive impairment was assessed using the Mini-Mental State Examination. Financial competence was assessed across six domains: involvement in financial decision-making, saving, borrowing, investment knowledge, insurance knowledge, and retirement planning. Each domain was treated as a separate binary outcome. Binary Logistic regression was used to estimate adjusted odds ratios (aORs) for each outcome while controlling for sociodemographic, health, and functional variables. Results: The prevalence of neurocognitive impairment was 60.0% (95% CI: 56.0–63.9%). Engagement across different financial domains was generally low. Only about 27.0% were involved in financial decisions or reported saving (24.8%). Only 10.2% had taken a loan, 8.9% had investment knowledge, 4.1% had insurance knowledge, and 13.2% reported having any retirement plans. After adjustment, neurocognitive impairment was strongly associated with reduced odds of involvement in financial decisions in several domains, including financial decision-making (aOR 0.49, p<0.001), saving (aOR 0.62, p=0.022), investment knowledge (aOR 0.30, p<0.001), and insurance knowledge (aOR 0.13, p=0.002). Conclusions: Neurocognitive impairment is widespread among elderly Ugandans receiving SCG. The impairment is associated with multiple domains of financial competence, which may undermine individuals' ability to fully benefit from social protection interventions. Routine screening for neurocognitive impairment and tailored financial management support may enhance program effectiveness and safeguard older persons’ welfare.

The link between weather elements and mental disorders is often described in high-income countries, with hardly any data from low-and middle-income countries where the resources to cope with the negative mental health impacts of climate change are extremely constrained. In this paper, we examined the association between weather elements and the incident presentation with a mental disorder at Butabika National Referral Mental Hospital. We used secondary data from two datasets: i) a mental health data set collated from all patients presenting at Butabika National Referral Mental Hospital in 2019; and ii) a climatic dataset for the geographic location of Butabika National Referral Mental Hospital for the same year (2019). The mental health data set included socio-demographic variables and mental disorder diagnoses, while the climatic data set included data on, atmospheric pressure (hPa), rainfall (mm), sunshine (hours/month), humidity (%), temperature (°C) and wind speed (m/s). We performed descriptive statistical analyses to summarize the frequency of mental disorder diagnoses and the monthly averages of weather variables. We then undertook correlation and multiple logistic regression analyses to investigate the associations between specific weather elements and the incident presentation of different mental disorders. In the mental health data, we had 2,827 participants, males were 56.1%(n=1,584), and the median age was 29 years (IQR 23-38). Psychotic disorders were the most common diagnosis at 43.8% (n=1,239). Overall, various weather elements correlated at different strengths with incident presentation of various mental disorders, particularly humidity and rainfall. On controlling for age and sex in the multiple regression models, the strongest associations were between heat elements and incident presentation for psychotic disorders [AOR1.12, 95%CI (1.04; 1.27) p<0.001]. No associations were demonstrated between weather elements and incident presentation for neurocognitive and neurodevelopmental disorders. These preliminary findings point to a possible relationship between incident presentation of mental disorders to a tertiary psychiatric hospital in Uganda and various weather elements. There is need for longitudinal studies to confirm these associations and to explore underlying social and biological mechanisms.