Staff Schedule
Chicago,
USA
Location
September 2020 -
December 2020
Date
Andrew Patronick
Andy Balcazer
Role
Team
Optimization
Skills
Algorithm Developer
Time Series: Patient Volume
The goal is to develop a 4-week staffing schedule of full-time and part-time nurses in the Hospital Medicine Department at NMH, using a dataset containing the hourly patient counts of each day from the beginning of 2009 to mid-2012.
To eliminate the process inefficiency's in healthcare, hospitals rely on advanced forecasting technology to create an effective nurse staffing schedule. The scheduling system gets complicated without accurately predicting the demand (hospitalized patients). Healthcare providers must adjust the staffing schedule by balancing full-time and part-time nurses.
Research
Data Analysis
Analyzed Data from the hospital regarding the number of incoming patients
Prediction Modeling
Analyzed Data from the hospital regarding the number of incoming patients
Optimizing Staff
Optimized the number of nurses needed in AMPL to balance the cost of full time, part-time and shortfall workers
Data Analysis
We began by averaging the number of patients by the days over a 24 hour period.
How to determine the number of shifts?
Using AMPL the team determined:
-
the total number of shifts needed across the week at each time t.
-
the number of full-time and part-time nurses needed to be hired.
* Full-time nurses work 12-Hour in each shift 3 shifts/week
* Part-time nurses work 8-Hour in each shift 5 shifts/week

Mathematical Formulation to determine the number of nurses & shifts in AMPL
Creating a Nurse Schedule using Number of Shifts
Using AMPL we created a schedule that took into account the nurse demands for each shift start time and shift type (12-HR and 8-HR).
The model assigned Full-time and Part-time nurses to specific shits on specific days while minimizing costs and fitting the constraints.

Mathematical Formulation to assign nurses to shifts in AMPL
Nurse to Patient Ratio
To ensure every patient will be taken care of in a timely manner, we ran the model with three values of patient-to-staff ratio.
-
5-to-1 ratio is usually applicable to Medical/Surgical units
-
3-to-1 ratio is suitable for step down units
-
4-to-1 ratio is preferred for most units including Emergency Departments
The increase in cost from 5-to-1 ratio to 4-to-1 is outweighed by safety measures and a 4-to-1 ratio with a total cost of $197,400
Takeaways
-
Data Analysis helps analyze the number of incoming patients
-
Optimization Algorithms determine the number of nurses needed per shift and minimize the cost of staffing by assessing the best combination of full-time and part-time nurses.
My Ventures
Stocks
Analyzing banks S&P 500 companies recommendations. Surprisingly, there were no trackers following the performance of analyst picks over the long term and I decided to build one.
Logistics
Stroke Code
Redesigning the CODE stroke activation process to reduce the Door-in-door out time for stroke patients
PROJECT WILL BE AVAILABLE AUGUST 2021