To investigate the relationship between patient mortality and the educational preparation (graduateness) of the nurses who cared for them.There have been 18 studies over the last two decades examining the effect of nurses' educational qualifications on mortality. All but three have used mortality data aggregated at the hospital level that has been combined with surveys of nurses to estimate the level of graduateness in the population. Data collection and extraction generally has been done at different points in time.A retrospective, cross-sectional study.Routine administrative patient data were extracted (May-August 2015). The primary outcome was all-cause patient mortality at discharge. We were able to identify the individual nurses who provided care during patients inpatient stay using an identification number. We were then able to calculate the 'graduateness' of the nursing care patients received by dividing the number of recorded episodes of care provided by baccalaureate prepared nurses with the total number of care episodes.After adjusting for confounding, we observed a significant association between patient mortality and nurse graduateness. Our observations suggest an optimum level of baccalaureate prepared nurses of approximately 70%. Above this level, there appears to be no additional decrease in mortality rates.This study represents an important methodological step forward over previous approaches. Our observations are generally consistent with existing literature and confirm the importance of baccalaureate nurse education.