Research project for my MS degree in Statistics from Texas A&M University
The incidence of post-operative nausea and vomiting (PONV) is generally in the range of 20-40%. This condition has negative effects on the health and well-being of patients and is financially costly to healthcare providers. The tradeoff is that preventive therapy has negative side effects and financial costs. So, the challenge is to develop a scoring system that most accurately recommends prophylaxis for the patients at high risk of PONV. In other words, a predictive model that is neither to conservative nor too liberal in determining which patients should be prescribed prophylaxis. This balancing act has been described in the medical literature as the prevent-or-cure dilemma.
This investigation analyzed a data set of 461 patients. The purpose was to develop a predictive model for PONV with performance comparable to or better than models previously published in the medical literature.
The more significant predictors that a patient has, the more likely the patient will have PONV. A patient with all three risk factors in my model has a 72% probability of experiencing PONV, while a patient with none of the three risk factors has a 16% probability of experiencing PONV. This model may have practical application for patient populations similar to the data set of this investigation.
Statistical computing for data analysis was performed using R Markdown.
MS in Statistics