Identification of Prognostic Factors Associated with RT-PCR–Confirmed COVID-19 Using Logistic Regression in a Controlled Setting

Ravi  Vajala; Gurprit  Grover; Chandra  Bhan Yadav

Ravi Vajala PhD, Principal of Sri Venkateswara College, University of Delhi, Delhi, India
Gurprit Grover PhD, Department of statistics, Faculty of Mathematical Science, University of Delhi, Delhi, India
Chandra Bhan Yadav Assistant Professor, Department of statistics, Hindu College, University of Delhi, India

Keywords: COVID-19, Cross-Sectional Study, Logistic Regression, Risk Factors, Vaccination, Diabetes, Asymptomatic Infection

Abstract

Managing asymptomatic patients with communicable diseases poses significant diagnostic and public health challenges. Understanding factors associated with infection status is essential for risk stratification
and targeted interventions in closed populations. This cross-sectional study analysed 472 individuals from an Indian army hospital who underwent RT-PCR testing for COVID-19 during a single time frame.
A logistic regression framework assessed associations between demographic/clinical factors (sex, age, diabetes, vaccination, obesity, and asymptomatic status) and COVID-19 positivity. Significant factors associated with COVID-19 positivity included diabetes (OR =2.21, 95% CI: 1.10–4.43, p = 0.025), vaccination (OR = 0.43, 95% CI:
0.28–0.66, p < 0.001, 57% reduced odds), and asymptomatic status (OR = 0.113, 95% CI: 0.05–0.25, p < 0.0001, 89% lower odds). Age showed modest association (OR = 1.17 per year, 95% CI: 1.05–1.31,p = 0.004), while sex and obesity were not statistically significant. The interaction model did not significantly improve fit over the main effects model (LRT p = 0.412). This cross-sectional analysis identified diabetes, vaccination, and asymptomatic status as key determinants of COVID-19 positivity. Findings support targeted diabetic screening,
continued vaccination efforts, and routine testing in closed populations to identify asymptomatic cases. Although single-centre design limits generalisability, the logistic regression framework offers a flexible tool
for rapid risk factor assessment during future outbreaks. Multi-site validation and longitudinal studies are warranted.

DOI: https://doi.org/10.24321/0019.5138.202650

How to cite this article:
Vajala R, Grover G, Yadav C B. Identification of Prognostic Factors Associated with RT PCR–Confirmed COVID-19 Using Logistic Regression in a Controlled Setting. J Commun Dis. 2026;58(3):18-26.

References

McKendrick AG. Applications of Mathematics to Medical Problems. Edimburgh Mathematical Society [Internet].

[Google Scholar]

Bartlett MS. Some evolutionary stochastic processes. Journal of the Royal Statistical Society. Series B (Methodological). 1949 Jan 1;11(2):211-29. [Google Scholar]

Weiss GH. On the spread of epidemics by carriers. Biometrics. 1965 Jun 1;21(2):481-90. [Google Scholar]

Downton F. A note on the ultimate size of a general stochastic epidemic. Biometrika. 1967 Jun 1;54(1-

:314-6. [Google Scholar]

Severo NC. The probabilities of some epidemic models. Biometrika. 1969 Mar 1;56(1):197-201. [Google Scholar]

Daniels HE. An exact relation in the theory of carrierborne epidemics. Biometrika. 1972 Apr 1;59(1):211-3.

[Google Scholar]

Becker N. Carrier-borne epidemics in a community consisting of different groups. Journal of Applied

Probability. 1973 Sep;10(3):491-501. [Google Scholar]

Jerwood D. The cost of a carrier-borne epidemic. Journal of Applied Probability. 1974 Dec;11(4):642-51. [Google

Scholar]

Routleff C. A variation of Weiss’s carrier-borne epidemic model. Journal of Applied Probability. 1982

Jun;19(2):403-7. [Google Scholar]