Differential associations of the timing and amount of social media use with sleep quality and academic performance among undergraduate medical and dental students in Pakistan: a cross-sectional study

Abstract

Poor sleep and reduced academic performance among medical and dental students have been linked to social media use; however, whether the amount, purpose, and timing of use are independently associated with these outcomes remains unclear. In this cross-sectional study, 364 students enrolled in MBBS or BDS programs at ten institutions in Khyber Pakhtunkhwa, Pakistan, were recruited through convenience sampling and completed a paper-based questionnaire between October 2023 and April 2024. Sleep quality was measured using the Pittsburgh Sleep Quality Index (PSQI), and academic performance was assessed using self-reported percentage marks from the most recent professional examination. The models were adjusted for background characteristics and other social media exposures. Poor sleep quality (PSQI > 5) was present in 312 students (85.7%, 95% CI 81.7–88.9). Duration of pre-sleep social media use was the only exposure independently associated with the PSQI global score (B = 1.594, 95% CI 1.081 to 2.106, p  < 0.001), while total daily use was not. Conversely, each one-category increase in total daily use was associated with examination scores that were 1.40 percentage points lower (B = −1.404, 95% CI −2.053 to −0.756, p  < 0.001). The classification of platform-derived purpose was not associated with either outcome; however, a secondary measure of purpose, the proportion of use devoted to academic work, was associated with a higher examination percentage (B = 2.148, 95% CI 1.513 to 2.782, p  < 0.001). Thus, conclusions about purpose depended on how it was measured. Nocturnal checking was not associated with either outcome. Fitting both outcomes in a single model supported outcome specificity, most clearly for pre-sleep use (difference in standardized β = −0.271, p  = 0.006; total daily use, 0.235, p  = 0.023). Sleep quality was primarily associated with when social media was used, while examination performance was related to how much social media was used. These exploratory associations from a convenience sample cannot establish causal direction, apply to the participating institutions rather than to Pakistani students generally, and require replication with prospective designs and objective measures in future studies.