Medicare: A Look at the Largest Single Payer System in the United States

Matthew Farris
12/14/2016

Medicare Background

Medicare is the only federally mandated insurance carrier, providing primary health insurance for those 65 and older. The Center for Medicare and Medicaid Services (CMS) is the governing entity that provides oversight specifically in the following:

Billing and Coding
Requlations for Managed Medicare Insurance Groups
Inpatient/Outpatient Prospective Payment System

Proposal

In the initial proposal of this project, the focus was Medicare Payments and the following points were the focus:

Medicare Payments vs. Socioeconmic indicator (poverty rates)
Medicare Payments and Readmission Rates
Inpatient vs Outpatient payments

Data Collections

All our data was collected primarily through data.cms.gov's website in conjunction with CMS. They have all the indicator data along with medicare payments and charges.

CMS

Medicare Payments vs. Poverty Rate in 2014

plot of chunk unnamed-chunk-2

Readmission Rates

Medicare has recently re-evaluated its payment system to incorporate payment reductions for hospitals with higher readmission rates. In 2012, CMS started collecting data, and with the IPPS regulations, finally implemented the rule to recoup payment from hospitals with higher readmissions then the national average.Furthermore, Hospitals can longer bill Medicare for readmissions within 30 days, the must combine the bills, and take a loss for the second admission.

Linear Regression Analysis-COPD

plot of chunk unnamed-chunk-3


Call:
lm(formula = ny_copd$READM ~ ny_copd$Average.Medicare.Payments)

Residuals:
    Min      1Q  Median      3Q     Max 
-3.7518 -0.8424 -0.1687  0.8910  3.9066 

Coefficients:
                                   Estimate Std. Error t value Pr(>|t|)
(Intercept)                       2.033e+01  2.987e-01  68.068   <2e-16
ny_copd$Average.Medicare.Payments 7.716e-05  3.506e-05   2.201   0.0294

(Intercept)                       ***
ny_copd$Average.Medicare.Payments *  
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 1.423 on 135 degrees of freedom
Multiple R-squared:  0.03465,   Adjusted R-squared:  0.0275 
F-statistic: 4.845 on 1 and 135 DF,  p-value: 0.02942

Linear Regression Analysis-Heart Failure

plot of chunk unnamed-chunk-5


Call:
lm(formula = ny_hf$READM ~ ny_hf$Average.Medicare.Payments)

Residuals:
    Min      1Q  Median      3Q     Max 
-3.1224 -1.2482 -0.0151  0.9316  4.7398 

Coefficients:
                                 Estimate Std. Error t value Pr(>|t|)    
(Intercept)                     2.180e+01  3.405e-01  64.022  < 2e-16 ***
ny_hf$Average.Medicare.Payments 1.591e-04  3.497e-05   4.551 1.16e-05 ***
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 1.584 on 139 degrees of freedom
Multiple R-squared:  0.1297,    Adjusted R-squared:  0.1234 
F-statistic: 20.71 on 1 and 139 DF,  p-value: 1.155e-05

Observation Status and Modeling

As we know, Hospitals are businesses, and in the long run it is important for them to try and make money. However, with CMS now auditing and recouping payments for these readmissions.

What is Observation?

Short-Stay Hospital visit (1 to 2 days)
Care Provide: Inpatient = Observation
Payment: Inpatient \( \neq \) Observation

Hypothetical Inpatient/Observation Model

Plan and Results

Now that we have a hypothetical model, the goal is to test this with real world data. With a few more tweaks to the hypothetical, we want to see on average if it is better to admit patients, or keep them under observation status. Of course, this is at the discretion of the Doctor treating the patient, but as of yet, a lot of hospitals do not have dedicated observation units. And this could be used to assess whether or not they are financial viable.