AI Diagnostics in Resource-Constrained Healthcare: A Comparative Exploration in Malawi's Urban Settings
Bibliographic record
Abstract
AI diagnostics have shown promise in resource-constrained healthcare settings, particularly in rural areas where access to trained professionals is limited. In Malawi, urban healthcare systems face similar challenges with a high prevalence of infectious diseases and limited medical resources. The study employed a mixed-methods approach involving data collection from electronic medical records at three urban health clinics in Malawi: one public and two private. Data were analysed using machine learning algorithms to predict disease diagnoses with an accuracy rate of up to 85% (95% CI). AI models demonstrated a higher diagnostic accuracy in the public clinic, particularly for respiratory infections, suggesting that public clinics may benefit more from AI diagnostics due to their larger patient volume and varied case mix. The findings indicate that AI can complement traditional healthcare practices but require further refinement and validation before widespread adoption. The study highlights the need for tailored AI solutions addressing local healthcare needs and resource constraints. Public health authorities should prioritise training of AI diagnostic models on local data to enhance accuracy and relevance. Additionally, ongoing research into AI ethics and privacy is essential for ethical deployment in urban settings. AI diagnostics, Malawi, urban healthcare, machine learning, clinic performance Model estimation used $\hat{\theta}=argmin_{\theta}\sum_i\ell(y_i,f_\theta(x_i))+\lambda\lVert\theta\rVert_2^2$, with performance evaluated using out-of-sample error.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.007 | 0.024 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.003 |
| Science and technology studies | 0.002 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.004 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".