Postoperative visual outcome prediction after cataract surgery at DCT Mvumi hospital, Chamwino-Dodoma

This is the authors' abstract. We don't add key points for this paper.

This research aimed to apply logistic regression to model the associations between demographic, clinical factors and preoperative visual outcomes with postoperative visual outcomes among patients undergoing cataract surgery at DCT Mvumi Hospital, Dodoma, Tanzania. A retrospective cohort study design was employed, with analysis of secondary data from 1155 ophthalmology patients seen between March and September 2020. Descriptive statistics, chi-square tests, and logistic regression were used to examine associations and construct a predictive model. We assessed model performance using receiver operating characteristic (ROC) analysis and calibration tests. Among the patients studied, 61.6% underwent cataract surgery. Logistic regression showed that age (OR=8.6), preoperative visual acuity (OR=8.7), comorbidities (OR=2.9), follow-up duration (OR=1.84), and socioeconomic status (OR=1.80) were statistically significant, strong predictors of postoperative visual outcomes ( \(p<0.001\) ). The model had excellent discriminatory ability (AUC=0.952) and adequate calibration, making it reliable for predicting individual patient outcomes. The analysis demonstrates the usefulness of adding patient age and preoperative vision status to predictive models of cataract surgical outcome prediction. The model outlined is a helpful instrument for clinicians for stratifying high-risk patients, enhancing preoperative counseling, and promoting the appropriate allocation of limited eye care resources. The findings are of significant value to the development of cataract care in Tanzania and other comparable low-resource settings.