HealthCare Analytics project providing insights to policymakers

Healthcare-Quality-and-Expenses-Analysis - Overview

• Built statistical models such as fixed, random effect to identify the impact of variable effecting the healthcare quality and expenses
• collected the data from OCED and WHO websites of 39 countries over 2010-16 years
• identified that healthcare resources and insurance type plays a vital role in country's health quality and expenses

Code and Resources Used:

R packages : stargazer, plm, pheatmap

Data Source:

OECD - https://stats.oecd.org/Index.aspx

WHO - https://www.who.int/gho/database/en/

YouTube Project Walk-Through

Data Cleaning:

• imputed the null values of variables using the KNN techniques

Exploratory Data Analysis :

Distribution of the Health_Expenditure and Life Expectancy

Correlation Plot

Health Expenditure vs LifeExpectancy

Model Building

As this is a multi-level data with lower level as time(years), we built the Panel regression models such as fixed, Random and also pooling model as a baseline model using the plm packages in R

Insights and recommendation

• A quantitative measure for making this strategical decision is to build 10 hospitals per 1 million population, to increase Life expectancy by 6.7 years.
• To manage the health expenses effectively, the government should strive to increases the % of total population under public insurance.
• Quantitative measure for making this strategical decision would be to look at decrease in expenses by 1% per person with increasing 1% population into public insurance.
• Also, we observed that the peculiar case of USA with High health expenses is mostly due to less percentage of people under Public Insurance i.e., 30% if increased will decrease the health expenses and the recent presidential election 2020 campaign is all around “Medicare for All”. Hope this model explains the argument of Campaigners.