Dynamics and solidification of multidimensional physical fitness in college students: a joint Markov chain and latent transition analysis

Background Existing research has focused mainly on cross-sectional fitness levels or on average changes in total scores, and relatively little is known about how the multidimensional structure of physical fitness evolves and whether it stabilizes over the college years. This study aimed to deconstruct the dynamic evolution, longitudinal transitions, and stability of multidimensional physical fitness among college students over a 4-year period. Methods We analyzed longitudinal physical fitness data (2020–2023) from 3,012 undergraduates in Chengdu. Based on six core indicators (forced vital capacity, 50 m dash, sit-and-reach, standing long jump, endurance running, and muscle strength), a first-order homogeneous Markov chain model and latent transition analysis (LTA) were used to identify latent fitness classes and to quantify transition probabilities, with gender and baseline body mass index (BMI) included as covariates. Results The Markov chain model suggested strong persistence at the Pass grade in the evolution of total scores, with the limiting steady-state distribution indicating that approximately 80.9% of students would remain at the Pass grade in the long term. Latent transition analysis identified three latent classes that showed cross-time measurement invariance, namely a low-fitness group, a basic-pass group, and a high-fitness group. The longitudinal transition matrix suggested marked within-state persistence in fitness trajectories. The annual probabilities of remaining in the three groups were 93.6%, 93.9%, and 88.8%, respectively, whereas the probability of moving from the low-fitness to the high-fitness group was only 2.6%. In addition, male gender was associated with a lower likelihood of being assigned to the high-fitness group at baseline (odds ratio = 0.33), a pattern possibly related to the difficulty of the pull-up strength test for males. By contrast, an abnormal baseline BMI was associated with higher odds of being assigned to the basic-pass group (OR = 3.12) and the high-fitness group (OR = 1.42) relative to the low-fitness group. Conclusion College students' physical fitness appears to exhibit pronounced path dependence and stability across the college years. These findings suggest that establishing early screening and interventions delivered in the freshman year may help to improve the fitness trajectories of low-fitness groups.