An Analytical Approach to Estimate Prevalence of Depression from a Non-Probability Sample of College Students in Karachi Using Beck Depression Inventory

Abstract

Objective: To estimate and compare the prevalence of depression among college students in Karachi. The analytical approach was aimed to minimize biases produced by profession, gender, and age of samples on the prevalence rate.
Methods: A cross-sectional study design was employed. Two hundred and sixty students were chosen from January 2018 to March 2019 through a purposive non-probability sampling technique from different professional colleges in Karachi. Beck Depression Inventory, a self-report questionnaire comprising 21 items was used as a study tool. After taking informed consent, each sample was inquired about the profession, age, gender, and personal and family history of psychiatric illness. Diagnosed cases of any psychiatric illness were excluded. Samples were stratified on the basis of profession, gender, and age. A minimum sample size of 70 students was extracted from the total sample through quasi randomization. Statistical software SPSS version 16 was used for data entry, sample randomization and analyses.
Results: Overall prevalence rate was 26.15%. Profession based stratification of samples showed prevalence rates of 21.81% and 29.33% in business and medical college students, respectively. Substrata of medical colleges i.e. public and private medical college students showed prevalence rates of 38.66% and 20%, respectively. Gender-based stratification showed a prevalence rate of 33.33% in female and 17.24% in male students. Age-based stratification showed a prevalence rate of 15.49% in teens (17-19 years) and 21.69% in the post-teen group (20-25 years). Independent sample t-tests showed a statistically significant difference between the strata of the profession, gender, and age and substrata of public and private medical colleges. A prevalence rate of 25.71% was estimated in quasi-randomized samples.
Conclusion: Randomization, stratification and large sample sizes are employed to minimize biases in prevalence studies. The present study showed the importance of sample stratification based on age, gender, and profession in estimating the prevalence of depression. Results also showed the validity of the minimum sample size is calculated appropriately.