Exploring Unknown Subgroups in Sleep Disturbance: A Mixture Regression Approach for College Students Pre- and During COVID-19 in Hawaii

Study Objectives College students have a high propensity for disordered sleep. Combined with disruptions due to the COVID-19 pandemic, there may be an increased tendency for disordered sleep patterns. The current study focuses on both identification of subgroups of disordered sleep college students before and after the COVID-19 pandemic as well as determining demographic differences between the groups. Methods This study used mixture regression model analysis (MRM) to evaluate the effects of age, gender, Body Mass Index (BMI), Grade Point Average(GPA) and lesbian, gay, bisexual, transgender and queer or questioning (LGBTQ) identification on sleep disturbance scores. Two independent samples from a university in Hawaii, collected before (n = 216) and during the COVID-19 pandemic (n = 177), were analysed to explore the heterogenous patterns of sleep quality and related factors, addressing inconsistent findings on the pandemic’s impact. Results The MRM revealed two latent groups within Pittsburgh Sleep Quality Index (PSQI) scores across datasets from 2020 and 2021: “typical college sleepers” and “disordered college sleepers”. Before the pandemic, typical sleepers had an average PSQI score of 4.89, while disordered sleepers averaged 8.36. During the pandemic, heterogeneity persisted: typical sleepers scored 6.81, while disordered sleepers averaged 14.38. Demographic and academic factors influenced the two groups differently. Conclusions The MRM effectively identified latent groups with different sleep disturbance scores, offering insights into previously undetected differential relationships between students’ characteristics, risk factors, and sleep outcomes. Furthermore, our study demonstrated that MRMs are a suitable tool for addressing heterogeneity in sleep quality research.