ABSTRACT OBJECTIVE To investigate the use of health services among adults living in Manaus, Amazonas. METHODS This was a panel of two cross-sectional studies conducted in Manaus in 2015 and 2019. Individuals aged ≥ 18 years were selected by probabilistic sampling and interviewed at home. The study outcomes were doctor visits and hospitalizations in the previous 12 months, and unmet surgical needs. Variations between 2015 and 2019 were tested using chi-squared goodness-of-fit test. Poisson regression with robust variance was employed to calculate the prevalence ratios (PR) of the outcomes with 95% confidence intervals (95%CI). RESULTS The surveys included 5,800 participants in total. Visits to the doctor decreased from 2015 (78.7%) to 2019 (76.3%; p < 0.001), hospital admissions increased from 2015 (7.9%) to 2019 (11.5%; p < 0.001), and unmet surgical needs decreased in the period (15.9% to 12.1%; p < 0.001). These variations were particularly observed in vulnerable individuals - sicker; poorer; non-whites; and those belonging to lower social classes, with less access to education, formal jobs, and health insurance (p < 0.05). Doctor visits were higher in people with fair health status (PR = 1.09; 95%CI 1.06-1.12), health insurance (PR = 1.13; 95%CI 1.09-1.17), and chronic diseases (p < 0.001) but lower in men (PR = 0.87; 95%CI 0.84-0.90) and informal workers (PR = 0.89; 95%CI 0.84-0.94). Hospitalizations were higher in people with worse health statuses (p < 0.001), without partners (PR = 1.27; 95%CI 1.05-1.53), and with multimorbidity (PR = 1.68; 95%CI 1.33-2.12) but lower in men (PR = 0.55; 95%CI 0.44-0.68), older adults (p < 0.001), informal workers (PR = 0.67; 95%CI 0.51-0.89), and unemployed (PR = 0.72; 95%CI 0.53-0.97). Unmet surgical needs were higher in older adults (p < 0.001), middle-class people (PR = 1.24; 95%CI 1.01-1.55), worse health statuses (p < 0.001), and chronic diseases (p < 0.001) but lower in men (PR = 0.76; 95%CI 0.65-0.86). CONCLUSIONS From 2015 to 2019, less people visited the doctor, more were admitted to hospitals, and less were in need of surgery or aware of that need, potentially indicating poorer access to health services.