Identifying predictors of nursing dropout and attrition before and after Bachelor's Graduation based on the IPOD model: A machine learning approach

Abstract

Aim

To identify predictors of academic dropout before graduation and professional attrition after graduation among nursing students, using the theoretical IPOD model and a machine learning approach.

Background

Academic dropout and professional attrition are global challenges.

Design

A retrospective study design.

Method

This study included 878 undergraduate nursing students enrolled between 2007 and 2018. Data were collected from Education Department records and follow-up interviews conducted via phone, email, or social media platforms. To predict academic dropout before graduation, four machine learning models were used: Extreme Gradient Boosting (XGBoost), Support Vector Machine (SVM), Decision Tree (DT), and Multinomial Logistic Regression (mLR). To predict professional attrition after graduation, Random Forest (RF), SVM, DT, and Binary Logistic Regression (BLR) models were applied.

Outcomes

The academic dropout rate was 2.2 %, while the professional attrition rate was 28.3 %. The XGBoost model, with 91 % accuracy, identified dropout predictors including a higher ratio of failed semesters to total semesters, lower GPA in the first and second semesters, younger age at admission, abnormal early academic status, tuition payment, and male gender. The Random Forest model, with 90 % accuracy, linked professional attrition to higher clinical competency, higher overall GPA, longer waiting time before employment, job burnout, and longer work experience.

Conclusions

Academic performance indicators, particularly during the early semesters, were associated with nursing student dropout, while professional factors such as job burnout and employment delays were linked to post-graduation attrition. These findings may inform targeted interventions to improve retention across both academic and professional stages.