WFJ Health has been in a rapid growth phase for the last five years. On average, the health system hires 20 new providers each year across its largest internal medicine clinics throughout the northern Arizona area. It takes 6 to 12 months to recruit, hire, and credential a new physician at these locations. However, it has been difficult for the system to identify when a new provider is needed until it seems too late. The clinic is forced to handle visit volume that exceeds provider capacity during that 6- to 12-month hiring time frame. This has resulted in unsatisfactory wait times for patients and reduced staff satisfaction because of unreasonable workload.

The health system has a great relationship with the local medical school and often hires graduating residents. However, the graduates are all looking for placement at the same time, either in January or June of each year. WFJ Health would like to develop an analytical model that would help predict when and where a new physician will be needed to successfully hire a larger proportion of the graduating residents each cycle.

The internal medicine leadership team and WJH Health’s predictive analytics team plan to meet this week.

Questions 

1. What are some of the major key variables that will need to be extracted for a predictive hiring model?

2. How can the team measure whether the model is successful post-implementation?

3. What are some of the associated risks with relying on this type of model?

4. Design a sample confusion matrix for projected visit volumes comparing “new” and “returning” patient demand. Which of the four quadrants of the matrix are of the largest potential concern and why?

5. Write a simple IF statement that could serve as the foundation of the predictive model. What additional information do you need to test whether your conditional statement can provide valuable output?