Caffeine metabolic ratios (CMR) following monitored caffeine intake are the gold standard to probe cytochrome P450 (CYP) 1A2 activity, which metabolizes antipsychotics like clozapine and olanzapine. Given caffeine's ubiquity, we tested whether random CMR from dietary caffeine were associated with (1) clinical, genetic, and epigenetic factors linked to CYP1A2 activity; (2) plasma concentrations of clozapine and olanzapine; and (3) psychotropic treatment response. First, we analyzed two population-based studies (CoLaus|PsyCoLaus, N = 4898; SKIPOGH, N = 2054) to investigate random CMR associations with clinical, genome-wide, and epigenome-wide factors associated with CYP1A2 activity. Second, in psychiatric cohorts, we tested CMR associations with dose-normalized plasma concentrations (C/D) of clozapine (N = 164) and olanzapine (N = 222) and with psychotropic treatment response, including hospital admission risk (N = 1019) and prolonged stays (N = 1349). CMR were positively associated with age, CYP1A2 inducers including smoking, and negatively with female sex. CMR were negatively associated with clozapine C/D, explaining up to 14.9% of the variance; over six-fold the variance explained by genetic factors. A one-unit increase in CMR was associated with a 26% increased likelihood of hospital admission (p = 0.002) and reduced short-stay chance by 11% (p < 10^-3 ). Random CMR provides a useful method to probe CYP1A2 activity, contributing, alongside other variables, to personalizing clozapine doses and identifying psychiatric patients at risk of hospital admission and lengthy stays. Incorporating routine measurement of random CMR before introduction of clozapine could be considered to allow early assessment of CYP1A2 activity, a key determinant of personalized clozapine dose titration.