Influenza Staffing Project
A medical staffing agency provides temporary workers to clinics and hospitals on an as-needed basis. This analysis will examine past influenza trends to proactively anticipate staffing needs for clinics across the country.
-
Generate a staffing plan that helps inform the timing and spatial distribution of medical personnel during flu season.
Determine the seasonality of influenza deaths.
Identify states with higher populations of vulnerable individuals (defined here as 65 years and older).
-
Translating business requirements
Data cleaning, integration, and transformation
Statistical hypothesis testing
Visual analysis
Forecasting
Storytellling in Tableau
Presenting results to stakeholders
-
Influenza deaths by geography (Source: CDC)
Population data by geography, time, age, and gender (Source: US Census Bureau)
NOTE: These numbers are estimates.
-
Excel
Tableau
Question #1 - Who is our key demographic?
The first task is to determine which populations are the most vulnerable so we can allocate staff to states with higher populations of vulnerable individuals. Due to the data available, age was one way to define a vulnerable population, and when looking at the number of influenza deaths between the over 65 years and under 65 years age groups (shown right), it’s abundantly clear that individuals who are older than 65 are more likely to die from influenza than someone younger than 65.
Question #2 - Where are our most vulnerable populations?
This question was a bit more difficult to determine. There are several metrics which could be used, two of which are seen here. First, the top graph shows the percentage of the population that died from the flu. Good to know, but those numbers are dependent on the population, and how do you compare (for example) New York and Hawaii?
Another option is to send staff to states with higher numbers of individuals 65 years and older. This gets closer to a solution, but looking purely at population ignores the influenza aspect, so again, while this is helpful information to have, it doesn’t adequately answer the question of where we should send supplemental staffing.
To hopefully identify some states that met both criteria, I created a map that represents both the average mortality rate for each state and the 65+ population for each state. California, Florida, Texas, New York, and Pennsylvania all have high populations of individuals 65 years and older, as well as a high influenza mortality rate. Therefore, those states should be considered the most at risk.
Question #3 - When should staffing be deployed?
I set the “start” and “end” of the year as July and June, respectively, because influenza spikes in the winter, and changing how the year is grouped allows for a better look at those spikes. As seen above, influenza starts to ramp up in December with a peak in January, along with an occasional resurgence in March.
Recommendations to Staffing Agency
Recommendation #1:
Plan to deploy staff in December, and they should be prepared to stay at their assigned clinic until March. That period represents the peak of flu season, and that is when clinics will need supplemental staffing.
Recommendation #2:
Because there is such overlap between flu deaths and vulnerable populations, staffing should be distributed as a percentage of total flu deaths by state. For example, California was responsible for 13% of the total flu deaths in the United States, so it should receive 13% of the available staffing to distribute amongst its clinics.
Tableau Storyboard
