This study assessed occupational health and safety practices on building construction sites in Uganda and examined which practices were associated with lower injury rates and fewer serious accidents. A quantitative case-study design was adopted, and a coded dataset of 120 building sites was analysed using descriptive statistics, Pearson correlations, multiple linear regression of the injury rate and binary logistic regression of serious accidents, supported by diagnostic tests. The dataset was simulated for study purposes and was designed as a template to be replaced by real site records. The study results indicated that injury rates averaged 3.4 per 100 workers per year, and that the employment of a safety officer (B = -1.19, p = 0.032) and safety training hours (B = -0.134, p = 0.004) were significantly associated with lower injury rates. Safety officers (odds ratio = 0.37, p = 0.021) and regular inspections (odds ratio = 0.67, p = 0.027) reduced the odds of a serious accident. Heteroscedasticity and non-normal residuals were detected and addressed through robust standard errors, which left the conclusions unchanged. The study concluded that organisational safety arrangements mattered more than site size or contractor type, and it recommended mandatory safety officers, structured training, more frequent inspections and improved accident reporting.