The BPCI Advanced data provided to applicants by CMS these past several weeks consists of target price workbooks, an episode-level summary data file, and Medicare claims data files by service type (hospital inpatient, hospital outpatient, HHA, SNF, hospice, Part B, and DME). The episode summary file contains information for each potential episode initiated by the applicant during the January 1, 2014 through December 31, 2016 period based on the discharge date of the “anchor” hospital stay. The target price workbooks contain preliminary target prices for each clinical episode group for each hospital or hospital/physician group practice (PGP) combination.
Combining the episode-level summary file with the target price workbooks will allow applicants to compare historical spending (included on the episode-level file that has already been inflated to the first performance period and winsorized at the national 1st and 99th percentile) with the target price. This comparison offers a first approximation of how historical episode spending performance compares with the target prices for each clinical episode group. Hospitals and PGPs will need this information to decide which clinical episodes to move forward with.
However, it should be noted that the patient case mix adjustment (PCMA) factors used to calculate the target prices are preliminary and are subject to change each year in the baseline period, as well as when facilities move forward to the performance years of the program. Therefore, CMS also included the coefficients used to calculate the patient case-mix adjusted spending (stage 1 compound log-normal model) as part of the target price workbooks. The patient specific characteristics in the episode-level data were also included. When combined with the regression coefficients, these patient characteristics allow for calculating the patient case mix adjusted spending for each episode. Performing this step will allow participants to estimate a PCMA factor for any specific period to account for changes in patient case mix for that period relative to the 4-year baseline used by CMS, which can be used to refine target prices.
The detailed claims data will provide additional power to the database when combined with the episode-level file and target price workbooks to obtain more information about spending during the episodes. These data allow for performing discharge destination analyses (e.g., what percent of cases are discharged to institutional PAC settings versus home with or without home health?). The detailed claims data also allow for examining episode spending for other providers and services incurred during the specific episode and can be used to determine certain outcomes (i.e., readmissions or mortality rates).
Combining these datasets will produce a robust set of data for determining the best opportunities under BPCI Advanced. In addition, by using Medicare public use claims data to build peer-group benchmarks on specific key performance metrics (i.e., episode spending by service type, discharge destination statistics, and readmission rates) participants can further identify opportunities to reduce episode spending and improve patient care, achieving the primary goal of the BPCI Advanced program.
For more information, email Randy Haught at firstname.lastname@example.org