Second CMS Data Release Confirms Unexplained Variability in Sticker Prices and Steady Cost Increases

September 15, 2014 3:38 pm
Posted By: Stephen Manti


by Xue Guo, Sean Barnes, Margrét Bjarnadóttir 

Center for Health Information & Decision Systems, Robert H. Smith School of Business

 

In the past two years, the Centers for Medicare and Medicaid Services (CMS) has released data containing provider charges (i.e., sticker prices) and payments for the 100 most common inpatient services and 30 common outpatients services for fiscal years 2011[1] and 2012[2]. The first data release motivated numerous media reports, resulting in headlines like “New hospital billing data is enough to make you sick“[3]. The tone of the headlines was inspired by the variability observed in the data. The prices charged by providers vary in some cases by an order of magnitude and cannot be explained entirely by the provider location, type of service, and population demographics[4]. Figure 1 includes examples of the distribution average provider charges for three different DRG codes that demonstrate this variability. The three different DRG codes[1] represent an inexpensive, a moderately expensive, and an expensive procedure.

 

fig1

Figure 1. Box-plots of providers’ average charges for three DRG codes.

DRG 897 (Alcohol/drug abuse or dependence w/o rehabilitation therapy w/o  

                  MCC)—Inexpensive procedure,

DRG 252 (Other vascular procedures w MCC)—Moderately expensive procedure, and

DRG 870 (Septicemia w MV 96+ hours)—Expensive procedure.


[1] DRG code: The code identifying the MS-DRG. MS-DRGs are a classification system that groups similar clinical conditions (diagnoses) and procedures furnished by the hospital during the stay.

We created a regression model to explain average sticker prices in 2012, and consistent with researchers’ observations from the 2011 data, commonly used provider, procedure and population characteristics only explain 67.45% of the variability in the average charges.

In addition, we compared the average sticker prices in 2012 with those from 2011 for the inpatient services, in order to investigate any differences in provider charging behavior. National health spending is expected to grow 5.8 percent[5] not adjusted for inflation) per year from 2010 through 2020. For the period 2011-13, national health spending increased around 4.9 percent [5] annually. We expect to see, on average, the prices increase at the rate of health care inflation. This is confirmed by the study as described below.

We conclude that the 2012 charge data confirms the variability observed with the initial information release. In addition the data is consistent with the observed increases in health care costs. We hypothesize that the initial information release may have had an effect on charging behavior, an effect that can possibly be observed when 2013 charge data will be released next year.

 

Analysis of Cost Increases

We calculated the ratio of average provider charges in 2012 to 2011, for each provider-DRG pair. We excluded records with fewer than 25 discharges billed by the provider  (either year) so that our results were not overly sensitive to outliers. We summarize the distribution of the ratios in Figure 2.

Fig2

Figure 2. Distribution of ratios of average provider chargers in 2012 to 2011

The ratios range from 0.29 to 3.45 and the mean is 1.052. We found that approximately 95% of our records’ ratios fall into the interval (0.695, 1.394). When we exclude the observations that fall outside this interval, the mean ratio for the new data is 1.042. Both means of ratio are close to the projected increase previously mentioned.

In addition, we divided the diagnosis-related group (DRG) codes into three price-based categories: codes with major complications or comorbidities (W_MCC), codes with complications or comorbidities (W_CC), and the remaining procedure codes (Else). Figure 3 summarizes the distribution of the ratios for the three categories. The ranges and means of the ratio for the three groups are: (0.3381, 3.4460), (0.5542, 2.2250), (0.2976, 3.1460) and 1.0430, 1.0450, 1.0600 respectively. In general, the mean ratio for each category is within range of the expected inflation rate for health spending and do not show a substantial changes from the previous year. The W_MCC category has the lowest mean and the largest range. The W_CC category has the least number of outliers.

We did not observe a significant change in the distribution of provider charges from the first year to the next.

 

Fig3

Figure 3. Distribution of ratios of average provider chargers in 2012 to 2011 of group W_MCC, W_CC and Else respectively.

 

 

 

References:

1. Centers for Medicare & Medicaid Services. Inpatient Charge Data FY 2011 [Internet, cited 2014 Jun 3]. Available from:

https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Medicare-Provider-Charge-Data/Inpatient2011.html

2. Centers for Medicare & Medicaid Services. Inpatient Charge Data FY 2012 [Internet, cited 2014 Jun 3]. Available from:

https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Medicare-Provider-Charge-Data/Inpatient2012.html

3. Nancy Kaffer. New hospital billing data is enough to make you sick. [Internet, cited 2014 July 2]. Available from:

http://www.freep.com/article/20130508/BLOG25/305080081/medicare-billing-hospital-rates

4. Untied States Census Bureau. The 2012 Statistical Abstract—Population. [Internet, cited 2014 June 25]. Available from:

http://www.census.gov/compendia/statab/cats/population.html

5. Sean P. Keehan, Andrea M. Sisko, Christopher J. Truffer, John A. Poisal, Gigi A   Cuckler, Andrew J. Madison, Joseph M. Lizonitz and Sheila D. Smith National Health Spending Projections Through 2020:Economic Recovery And Reform Drive Faster Spending Growth Health Affairs, no. (2011) Available from: http://content.healthaffairs.org/content/early/2011/07/27/hlthaff.2011.0662.full.pdf+html

 

 

 

 

 

 

 


1 Comment

  1. chenweihaobang@gmail.com says:
    Posted May 2, 2017 at 1:23 pm | Permalink

    Thanks for sharing!