Developing Data Products - Forecast of Heathcare vs GDP

Forecasting escalating Healthcare costs vs GDP

  1. BLS Medical Expenditure Panel Survey (MEPS) account data from 2000-2010
  2. GDP and topline healthcare spending fit to a time series
  3. The app forecasts into the future both the GDP and Healthcare costs
  4. Forecasts are plotted with the 95% confidence interval

Results

  1. Healthcare costs are forecast to expand at the current rate
  2. Surprisinging GDP did not increase, instead the 95% interval widened
  3. Subsequent slides show the R calculations
  4. Check the Shiny app to follow along

50 year forecast server code run output

##   Year     GDP   Health
## 1 2011 14756.7 2072.350
## 2 2012 14756.7 2121.792
## 3 2013 14756.7 2172.413
## 4 2014 14756.7 2224.242
## 5 2015 14756.7 2277.308
## 6 2016 14756.7 2331.640

Forecasting Code snippet


  tsHlth <-  ts(dat$Health, start = 2000)
  tsGdp <-  ts(dat$GDP, start = 2000)

  fcastHlth <- data.frame(forecast(tsHlth,  h = numYears, level=95, allow.multiplicative.trend=TRUE))
  fcastGdp <- data.frame(forecast(tsGdp,  h = numYears, level=95, allow.multiplicative.trend=TRUE))

Conclusion

Healthcare spending is unsustainable !!