fagan2888/optimization-of-healthcare-network-facility-staff

This repository contain a python notebook in which we create solution of allocating healthcare staff.

Optimizing a Healthcare Network for Improved Service Delivery

Table of Content:

PROBLEM

1. INTRODUCTION

1.1 SAMPLE PROBLEM
1.2 DESIRED SOLUTION
1.3 DATA SOURCES

2. DELIVERABLE

CODED SOLUTION OF GIVEN PROBLEM

1. Data Gathering
2. Data Understanding
3. Data Cleaning
4. Data Manipulation
5. Optimization
6. Visualizing Distances Using Folium Map

Conclusion

PROBLEM

1. INTRODUCTION

The Washington State Health Ministry would like to optimally upgrade staff or resources in existing facilities across Washington so that they can allocate resources to where they are most needed, based on demand for services in different geographic regions/areas. To address this problem, the health ministry needs to:

• Make services more accessible by allocating resources to where they are most needed; increasing capacity in some facilities and decreasing capacity in others.
  
• Minimize people’s travel time so that they only travel to their nearby facility.
  

Table of Content

1.1 SAMPLE PROBLEM

Majority of the population living in Area A need access to health services, but would need to travel on average 2 hours per return trip to get their needs met in Facility B, instead of travelling 0.5 hours to their nearby Facility A. This is because Facility A has only 21 staff/resources as opposed to 52 staff in Facility B. People living in Area A cannot travel this far, so they need to get their needs met by Facility A, otherwise they go without healthcare.

Table of Content

1.2 DESIRED SOLUTION

To model how many staff members should be shifted from Facility B to Facility A, so that the population living in Area A can travel to their nearby Facility A instead of Facility B, improving travel cost/access to healthcare. Depending on the level of care, a rule of thumb is that the healthcare worker to patient ratio should be approximately 1:2808 in order to not overwork staff members. So 1 healthcare worker can reasonably serve 2808 standard patient needs a year. Not the entire population living nearby a facility would need healthcare services, but we assume at least more than half would seek services for a variety of problems.

Table of Content

1.3 DATA SOURCES

Facility ID	Facility Area-Zipcode	Facility Staff Count
Facility A	98007	21
Facility B	98290	52
Facility C	98065	43
Facility D	98801	9
Facility E	98104	64

1.3.2 Maps

Example sources: Google Maps, Bing Maps, HERE, etc. These sources can be used for estimating average travel time for people traveling to a facility in their area or nearby their area, and the average travel time to all other facilities.

1.3.3 Population

Example sources: US Census Bureau, World Population Review, etc.

These sources can be used to grab population estimates in facility zip areas and nearby areas.

Table of Content

2. DELIVERABLE

Write a short technical document addressing the following:

• Introduction: What is your understanding of the problem? Can you write the introduction and the problem statement in formal conference paper-like format?
• Assumptions: What assumptions would you make to simplify the problem?
• Data gathering, handling, cleaning, processing: How would you acquire the data? What steps would you perform to process or clean the data? How would you extract useful data inputs for this task?
• Proposed Solution: Share your optimization technique and any necessary details.
• References: Cite any libraries, APIs, publications. We believe in standing on the shoulders of giants. Please reuse any existing research papers, source code, libraries but make sure to cite them.
• Source code: Share your R or Python source code.

Table of Content

CODED SOLUTION OF GIVEN PROBLEM

1. Data Gathering:

First of all, we collect data from "Washington Demographics" and make its data frame.

Table of Content

2. Data Understanding:

Then we start understanding data by applying basic pandas statistical methods on the above data frame.

Table of Content

3. Data Cleaning:

Now we prepare data for analysis by removing unnecessary columns that exist in the above table(data frame) and changing dtype of columns from object to int32.

Table of Content

4. Data Manipulation:

After that, we 'll manipulate or prepare data in the format that is suitable for finding the solution to the problem.

Table of Content

5- Optimization:

After data preparation or manipulation we'll code the solution of the given problem.

Table of Content

6- Visualizing Distances Using Folium Map:

After finding the distances from each area to other areas, we'll visualize them on the folium map.

Area A (Bellevue)

Area A (Bellevue) with marker

Distance from Area A to B (Bellevue to Snohomish Country)

Distance from Area A to C (Bellevue to Montalbano Elicona)

Distance from Area A to D (Bellevue to Wenatchee)

Distance from Area A to E (Bellevue to ძველი თბილისის რაიონი, თბილისი)

Table of Content

Conclusion:

So in this project, we modeled the number of staff members needed in each area depending on the level of care & rule of thumb. Furthermore, we found the distances from each area to other areas and then we created leaflet maps of those distances with folium library.