CURRENT SITUATION
A medical staffing agency requires assistance in planning for the influenza season when there is a high demand for temporary workers in clinics and hospitals.
GOAL
Perform an analysis that will examine influenza trends and provide insights to proactively plan for staffing needs across the country. The results will help the agency to be better prepared for the season.
DATA PREPARATION
During this stage, using Excel, I carried out the following:
-
Data Integrity & Quality Measures: created a data profile for each of the data sets, which includes information on data types, data integrity issues (accuracy and consistency), any cleaning conducted, as well as summary statistics in each profile. Besides that, checked the data for completeness, uniqueness, and timeliness.
-
Data Transformation & Integration - Integrated data from two sources into one cohesive data set using the VLOOKUP function.
Some processes you can see below.
ANALYSIS
To provide recommendations for medical staff demand during the flu season, I carried out using Excel:
-
Conducted statistical analyses by calculating the variance and standard deviation for key variables.
-
Formulated hypothesis that states that “if a patient's age is either 65+ or lower than 5, they will be more susceptible to death from influenza”.
-
Performed statistical hypothesis testing (one-tail t-test) which confirmed the project hypothesis with a 95% level of confidence.
INSIGHTS & VISUALIZATION
-
Based on the provided analysis, there seems to be a weak correlation coefficient of -0.07 between the Influenza Mortality rate and the population considered to be Vulnerable across states. Therefore, I would suggest prioritizing the distribution of medical staff based on the Influenza Mortality Rate as the primary factor, followed by the Vulnerable population as the secondary factor.
-
Influenza season occurs at the same time once per year in all US regions. It starts in December and finishes in March. The peak of Influenza season is in January.
Some interactive temporal, statistical, and spatial visualizations are presented below.
SOME OF RECOMMENDATIONS
Spatial Medical Staff Distribution
Considering Influenza mortality rate and vulnerable population spatial distribution, thought US states should be:
-
High-need-based: California, New York, Pennsylvania, and Florida.
-
Medium-need-based: Illinois, Texas, Ohio, North California, Michigan, Tennessee, Massachusetts, and Missouri.
-
Low-need-based: The rest of the US states.
Forecasting
Considering that the Influenza season occurs at the same time once per year in all US regions, gradually increasing quantity of temporary medical staff starting from November till January, after February – gradually decreases.
RETROSPECTIVE
Project Challenges
The analysis of vulnerable populations was limited to adults over 65 years old and children under 5 years old due to data limitations. Pregnant women, individuals with chronic medical conditions, and healthcare workers were not included in the analysis.
To protect privacy, the CDC suppressed 82% of data points due to fewer than 10 deaths in a demographic, which could cause potential bias in the result of the analysis.
Future Improvements
I would analyze vaccination data to determine its influence on mortality and effectiveness as a preventive measure.
Besides that, I would analyze medical staffing data to identify understaffing or overstaffing in US states and provide guidance on the number of additional resources required by each state.