FARGO HEALTH GROUP: MANAGING THE DEMAND FOR MEDICAL EXAMINATIONS USING PREDICTIVE ANALYTICS Case Solution And Analysis, HBR Case Study Solution & Analysis of Harvard Case Studies

FARGO HEALTH GROUP: MANAGING THE DEMAND FOR MEDICAL EXAMINATIONS USING PREDICTIVE ANALYTICS Case Solution

Business Problem & Necessity for Data-Analytic Solution

The management, owner and the director of the Fargo Health Group, Jay Rubin,have always emphasized upon the importance of using data driven planning in order to plan the examining physicians at the HCs of the organization. According to Mr. Rubin, the organizational problem of planning could be best met by using the predictive analytics and this is what the organization needs the most at this time in order to reduce the fees paid to the Regional Office of Health Oversight (ROHO) and the costs of Outpatient Clinics (OCs).

The main business problem currently being faced by the management is that the organization does not have enough number of the examining physicians and this is limiting the capacity of the HCs to meet the mandated demand within a 30 day time frame. As a result of this, the examining physicians always get late and they report to the Local Offices (LOs) by exceeding their 30 day deadlines. Most of the time the HCs also reject the requests back to the LOs and explaining them this due to understaffing of the HCs.

Furthermore, the rerouting of the examination requests among the different HCs of the organization is also proving to be another source of the reputational and financial burden for the organization. Due to these major business problems, it has become necessary for the organization to seek for the external efforts for developing the predictive analytic product which has always been advocated by Mr. Rubin over the years. This model could then be used to predict the volume of the incoming medical requests and thus, the management can better schedule and plan the examining physicians.

Case Analysis

The detailed analytical models have been applied in order to formulate the best model possible for the company to predict the number of the incoming medical requests. However, before the formulation of different time series models, the data provided in the excel spreadsheet for the historical medical examinations has been trimmed and cleaned by using a range of different data cleaning strategies as described below.

Nature & Structure of Data

Since we are required to focus on just the Abbeville HC for the Fargo Health Group and formulate the models for the cardiovascular or heart related examinations only, therefore, the data provided in the sheet Abbeville, LA sheet is for this HC only. If we analyze the structure of the data that it consists of a range of errors, missing values, outliers, inconsistencies and duplications. The nature of the data is simple. We are provided with scale data for the incoming number of the examinations which have been received in a particular month for a specific year. The data has been provided for the period of January 2006 to December 2013. The original data received is also mixed up and not in any proper format. Therefore, before the generation of different forecasting models, the data has been cleaned by using different data cleaning strategies in the excel spreadsheet. These are described in detail in the next section.

Data Cleaning Strategies

The original data set provided in the excel sheet is all mixed up therefore, the complete data set has been sorted in an ascending order since this is a time series data. The data has been first sorted in ascending order based on the years and then for each of the years the data has been sorted in ascending order based on the month. Once the complete data set has been sorted, then all the outliers in the data set which seemed unusual and also certain inconsistencies such as 99999999 values had been deleted from the complete data set.......................