Study design setting
A cross sectional study was conducted among the medical and non-medical undergraduate University students in Dar es Salaam, Tanzania. There are 12 public and private universities/colleges in Dar es Salaam, four having medical training. The population of undergraduates in medical universities/colleges is about 5,000 and about 23,000 in non-medical universities/colleges.
Sample size and sampling
We estimated a sample of about 300 undergraduates for each group, medical and non-medical university/college students. We applied a multi-sampling strategy to get the study population. In the first stage, from the four medical university universities/colleges, we randomly selected three. We also randomly selected three out of 11 non-medical universities/colleges. In the second stage, in each selected university/college, we randomly selected relevant strata based on the year of enrolment. In Tanzania, all undergraduates in medical programmes last for five years. We excluded first and second year medical students because of the least exposure to medical training. Therefore, strata were students in their 3rd, 4th and 5th year of study. For non-medical universities/colleges, most of their programmes last for up to four years. Therefore, strata were the first through fourth year programmes. The required sample size for each group was spread equally to the available strata in each of the selected university/college. In all groups, we excluded undergraduates from abroad because of possible previous training on NCDs. We also excluded students from other health related courses which involved direct provider to patient interaction such as pharmacy and nursing programmes.
Data collection instruments
We used a self-administered the WHO STEPS (STEPwise approach to Surveillance) data collection tool on NCDs risk factors [21]. Although the general tool has three steps, we used STEP-I that entirely focuses on assessing modifiable risk factors through interviews. The tool was slightly modified to enhance better understanding of the study terms. The tool included structured questions organized into two main sections: (a) the background (social and demographic) information and (b) exposure to the five risky behaviours related to NCDs (tobacco use, alcohol consumption, unhealthy diet, salt intake and physical inactivity). The tool was organized and administered in English. Consenting participants were left with the tool to fill-in for about one week, and then we collected filled-in questionnaires for processing and analysis.
Study variables and measurements
Tobacco use:
Tobacco use was defined as those smoking cigarettes, shisha, cigars, nicotine “vaping” instruments or hand rolled cigarettes either. Smokers were either current (within 12 months) or previously (life-time).
Harmful alcohol consumption:
Alcohol users were those reporting consuming alcohol and heavy episodic drinking was classified as consumption of 6 or more alcoholic drinks in one sitting. One standard alcoholic drink was regarded as consuming a 300 ml bottle of regular beer, 30 ml of spirits or 120 ml glass of wine. Any undergraduate drinking alcohol within 30 days before the date of data collection was considered a current alcohol user.
Unhealthy diet:
The proxy marker of unhealthy diet was inadequate vegetables and fruits in a typical week. One serving of vegetables was considered similar to a handful of leafy vegetables or three tablespoons of kidney beans or peas or boiled maize. One serving of fruit was considered similar to one mid-sized banana or orange or mango or a handful of grapes or similar fruits. Total fruit and vegetable consumption was computed as the sum of the average intake of vegetables and fruits per day.
Salt intake:
Salt intake was classified as good (responded affirmatively to more than 3 measures of lowering intake), poor control (1–3 specific salt control measures) and otherwise, no control.
Physical inactivity:
Physical inactivity also here referred to as sedentary life was computed as the sum of the total activity-minutes for work and recreational activities per week. Inactivity was defined as fewer than 75 minutes of vigorous or intensity activity or 150 minutes of moderate intensity physical activity per week.
High risk factors for NCDs
An undergraduate with at least two of the five risky NCDs factors was categorized high risk.
Statistical analysis
Collected data were entered into statistical software (Statistical Package for Social Sciences). They were checked for consistency and coverage before commencing the analysis. We summarized data by running frequencies for categorical variables and calculated the means or median with standard deviations or inter-quartile ranges respectively for quantitative variables. We assessed the association between the outcome variables and selected independent variables using Pearson’s Chi-Square test. Since the proportion of undergraduates with high risk behaviours for NCDs was higher than 10%, use used Poisson regression analysis to estimate the effect of type of university/college when predicting high risk level. We calculated 95% confidence intervals as a measure of strength of the association; using robust adjustment for clustering of the outcome within one institute. The level of significance was set at 5% level.