Abstract Purpose This study aimed to investigate changes in physical work demands in association with self-rated health and musculoskeletal symptoms. Methods Data from five waves over the period 2019–2021 of the Netherlands Working Conditions Survey COVID-19 were available for 7191 participants aged 19–64 years who worked (partly) on-site during at least two consecutive waves. Logistic generalized estimated equations (GEE) were used to estimate the odds ratios (OR) with 95% confidence interval (CI) for changes (increase or decrease compared to no change) in physical work demands between two waves and poor self-rated health and musculoskeletal symptoms in the following wave, adjusted for the health outcome at the first wave, age, educational level, working hours and hours worked from home. Results In females, a statistically significant association was found between an increase in physical work demands compared to no change and musculoskeletal symptoms (OR 1.39, 95% CI 1.17–1.65). A decrease in physical work demands in females was not statistically significantly associated with musculoskeletal symptoms (OR 0.93, 95% CI 0.80–1.08). Similar trends were found for poor self-rated health, although non-statistically significant. For males, comparable but attenuated associations were found. Conclusion While our study showed that increasing physical work demands are associated with adverse health (self-reported and musculoskeletal), it did not appear to benefit worker’s health to reduce work demands. Future research with multiple measurements in a shorter period and additionally using devices to measure physical work demands will be needed to confirm our study results.