Predictors of inpatient mortality following colorectal cancer surgery in New York City 2000--2004

This study characterized the relationship between demographic variables, surgical procedure volume, co-morbidities, adverse events, and admitting diagnosis (colorectal cancer or other) on inpatient mortality for patients undergoing colorectal cancer surgery at New York City Hospitals in 2000-2004.;The New York Statewide Planning and Research Cooperative System database was queried for patients with principle diagnosis of colorectal cancer and admission diagnosis, demographic factors, an economic profile, determined by the median household income by zip code; and the surgical volume of treating hospital. The 66 hospitals included in the study were grouped by colorectal cancer surgical volume into 4 groups: low, medium, high and highest. Distribution of factors within hospital groups was analyzed first by bivariate and then step-wise multiple logistic regression for variables that independently impacted or enhanced model prediction.;The univariate analysis of data revealed that socioeconomic factors, hospital surgical procedure volume, co-morbidities, patient safety indicators (PSI) and admission diagnosis individually have an impact on inpatient mortality. A regression model including only socioeconomic factors indicated that patients living in the low, middle and high median household income (MHI) areas were 1-2 times more likely to die from these surgical procedures than those in the highest MHI areas. Additionally, patients with Medicare and Medicaid insurance coverage were 2-3 time more likely to die than those with commercial insurance.;Multivariate regression analysis for all variables uncovered that older patients and low and medium hospital procedure volume were significant predictors of increased risk for death after surgery. Having one or more co-morbidities including heart disease, diabetes, heart disease with obesity were also significant predictors of increased risk for death following colorectal surgery. Patients having one or more PSI had increased risk for death following colorectal cancer surgery for patients in this data set. Finally the admission diagnosis did predict a higher risk for death for patients in this study if they were admitted for another diagnosis other than colorectal cancer.;More research is needed to better understand the effect of other factors such the potential difference in the nursing process between low- and high-volume hospitals has not been addressed in the literature. Additionally, the issue of PSI, as utilized in this study, has never been investigated or utilized to improve care. Finally, for the highest volume hospitals in this study two thirds of their patients came from outside of NYC. The potential impact of geographic influences on volume studies has never been reported and research is need in this area. Nursing research is needed to determine what nursing systems and processes lead to improved care and see that these are disseminated in research reports and quality initiatives.