Independent Predictors of Diabetes Outcomes: A Binary Logistic Regression Analysis of Sociodemographic and Clinical Factors at Specialist Hospital Yola, Adamawa State, Nigeria

1*Oboh, S. O., 2Bororo Agatha, 1Gambo I. Bature, 3Mohammed, Fadimatu B.

1Department of Statistics, Federal University Wukari, Nigeria

2Department of Mathematics, College of Education Zing, Taraba State, Nigeria

3Department of Statistics, Adamawa State Polytechnic, Yola, Nigeria

*Corresponding author’s Email: obohsamuel@fuwukari.edu.ng, doi.org/10.55639/607.020100123


ABSTRACT

Diabetes mellitus is a growing public health crisis in Nigeria; however, facility-level epidemiological data from the northeastern region are critically lacking. This retrospective cross-sectional study investigated the sociodemographic and clinical factors associated with diabetes among 5,288 patients at the Specialist Hospital Yola, Adamawa State, Nigeria. Data extracted from clinical records were analyzed using chi-square tests and binary logistic regression. The prevalence of diabetes in the study population was 6.5% (342/5,288). Bivariate analysis revealed significant associations between diabetes and age group, blood glucose level, body mass index, hypertensive history, and stroke history (p < 0.05). The logistic regression model was statistically significant (χ² = 563.186, p < 0.001) with adequate fit (Hosmer–Lemeshow p = 0.386). Blood glucose level emerged as the strongest independent predictor; compared to patients with glucose >6.9 mmol/L, those with levels 5.6–6.9 mmol/L had 57.3% lower odds (OR = 0.427, p < 0.001), and those with < 0.001) of diabetes. Diastolic blood pressure was also a significant positive predictor (OR = 1.031 per mmHg; p < 0.001). Patients aged >50 years had significantly higher odds than those aged < 0.001), consistent with established comorbidity. The study concluded that elevated fasting plasma glucose and raised diastolic blood pressure are key indicators for diabetes risk stratification at the Specialist Hospital Yola. Routine glucose screening for all outpatients and integrated hypertension–diabetes management protocols are recommended.

KEYWORDS
Fasting plasma glucose, Diastolic hypertension, risk stratification, Cross sectional