MHA FPX 5017 Assessment 3 Predicting an Outcome Using Regression Models Regression Modeling for Predictive Analysis

 

In predicting the reimbursement amount, a regression model incorporating age, risk, and satisfaction datasets reveals an explanatory variance of 11% (Gaalan et al., 2019). It’s important to note that not all independent variables contribute equally to this variance; rather, each variable’s percentile contribution must be considered to understand the model’s fitness accurately. The multiple regression model demonstrates statistical significance, with F(3,181) = 7.69, P < .001, and R2 = .11.

Statistical Results and Decision Making

Utilizing data from the provided dataset, multiple regression equations can support healthcare decisions regarding predicted reimbursement costs for individual patients. The reimbursement cost for each patient can be calculated using the equation: y = 6652.176 + 107.036(age) + 153.557(risk) – 9.195*(satisfaction). Examples of predicted reimbursement costs for specific patients from rows 13, 20, and 44 are presented below.

Conclusion

To optimize healthcare reimbursement costs, it may be prudent to exclude the satisfaction variable from predictive models, as it appears incongruent with other predictor variables. However, employing various regression models remains essential for making informed decisions and aligning with long-term organizational goals. Despite potential regulatory adjustments, healthcare organizations can leverage regression analysis to navigate uncertainties and plan for future reimbursement costs effectively.

Reference

Casson, R. J., & Farmer, L. D. M. (2014). Understanding and checking the assumptions of linear regression: A primer for medical researchers. Clinical & Experimental Ophthalmology, 42(6), 590–596.

Gaalan, K., Kunaviktikul, W., Akkadechanunt, T., Wichaikhum, O. A., & Turale, S. (2019). Factors predicting quality of nursing care among nurses in tertiary care hospitals in Mongolia. International Nursing Review, 72(5), 53-68.

IntroToIS BYU. (2016). Creating a multiple linear regression pred