Objective First responders’ mandatory reports of mental health episodes requiring emergency hospital care contain rich information about patients and their needs. In Queensland (Australia) much of the information contained in Emergency Examination Authorities (EEAs) remains unused. We propose and demonstrate a methodology to extract and translate vital information embedded in reports like EEAs and to use it to investigate the extreme propensity of incidence of serious mental health episodes.Methods The proposed method integrates clinical, demographic, spatial and free text information into a single data collection. The data is subjected to exploratory analysis for spatial pattern recognition leading to an observational epidemiology model for the association of maximum spatial recurrence of EEA episodes.Results Sentiment analysis revealed that among EEA presentations hospital and health service (HHS) region #4 had the lowest proportion of positive sentiments (18 %) compared to 33 % for HHS region #1 pointing to spatial differentiation of sentiments immanent in mandated free text which required more detailed analysis.At the postcode geographical level, we found that variation in maximum spatial recurrence of EEAs was significantly positively associated with spatial range of sentiments (0.29, p < 0.001) and the postcode-referenced sex ratio (0.45, p = 0.01). The volatility of sentiments significantly correlated with extremes of recurrence of EEA episodes. The predicted (probabilistic) incidence rate when mapped reflected this correlation.Conclusions The paper demonstrates the efficacy of integrating, machine extracted, human sentiments (as potential surrogates) with conventional exposure variables for evidence-based methods for mental health spatial epidemiology.Such insights from informatics-driven epidemiological observations may inform the strategic allocation of health system resources to address the highest levels of need and to improve the standard of care for mental patients while also enhancing their safe and humane treatment and management.