Appendix E
Harvard University Price Adjustment Memorandums
HARVARD INPUT-PRICE ADJUSTMENT MEMORANDUM (11.04.11)
Background
The input price adjustment is meant to reflect variation in input prices required to provide care. Harvard uses input price indices developed by the Centers for Medicare & Medicaid Services (CMS) to adjust raw spending.
The input price adjustment is performed separately for each of two claim types: (1) inpatient facility diagnosis-related groups (DRGs) and (2) inpatient professional current professional terminologies (CPTs) and outpatient CPTs. The specifications for each are described below, but for each the unit of analysis is the “claim-day.” A “claim-day” is an aggregation of all spending for a given person for a given procedure code for a given claim type (DRG, inpatient CPT, outpatient CPT). For example, a person may have multiple claims for a given code on a given day. We add up the spending for all of the claims for that person for the same code and claim type on the same date (in the case of inpatient stay, we use the admission date) to calculate average spending for the code. For around 15 percent of medical (nondrug) spending, there is an outpatient claim line that is missing a CPT code. For these claims we distribute the spending over claims that do have a CPT on the same day and for the same person according to the proportion of total daily spending each CPT represents. We believe that the outpatient claims that are missing a CPT code are facility payments that have been carved out of the professional payments, and that this procedure
reapportions spending. For the purpose of input price adjustment, we treat claims that are missing a CPT code and that do not have any CPT codes on the same day as if they were an inpatient facility claim.
Inpatient Facility (DRGs)
Each claim day with a DRG code is adjusted by the Area Wage Index Values that include the occupational mix adjustment.1 The wage index data is merged into MarketScan data by county. 69.7% of the claim amount is adjusted by the wage index value, a proportion based on the labor-related share outlined in the Hospital Inpatient Prospective Payment System Final rule.2 This is the same procedure we perform on any outpatient claims that cannot be apportioned as described above.
(0.697 × raw claim day amount / wage index) + 0.303 × raw claim day amount
Inpatient Professional and Outpatient Claims (CPTs)
Each claim day with a CPT code is adjusted by the fully-implemented relative value unit (RVU)3 share multiplied by the relevant GPCI.4 Inpatient professional and outpatient claims incorporate the practice expense facility component of the RVU.
Raw claim day amount / (%Work × Work GPCI + %PE (facility) * PE GPCI + %MP × MP GPCI)
All other claim types (claims without a valid CPT that cannot be allocated to other claims on the same day, durable medical equipment, prescription drugs) are left unadjusted.
_________________
1The wage index file for 2007 can be found at http://www.cms.gov/AcuteInpatientPPS/WIFN/itemdetail.asp?filterType=none&filterByDID=0&sortByDID=3&sortOrder=ascending&itemID=CMS1187403&intNumPerPage=10.
2The final rule for fiscal year 2007 can be found at http://www.cms.gov/AcuteInpatientPPS/IPPS/itemdetail.asp?filterType=none&filterByDID=-99&sortByDID=4&sortOrder=ascending&itemID=CMS1229138&intNumPerPage=10.
3The RVU lists are based on the October release for each year. 2007 can be found at http://www.cms.gov/PhysicianFeeSched/PFSRVF/itemdetail.asp?filterType=none&filterByDID=-99&sortByDID=1&sortOrder=descending&itemID=CMS1203203&intNumPerPage=1.
4The GPCI lists are based on the October release for each year. 2007 can be found at http://www.cms.gov/PhysicianFeeSched/PFSRVF/itemdetail.asp?filterType=none&filterByDID=-99&sortByDID=1&sortOrder=descending&itemID=CMS1203203&intNumPerPage=10.
Capitated Claims
We outline our overall approach to capitated claims in a separate memo. We will input-price adjust the imputed values for capitated claims.
Aggregation
After the spending is adjusted as described above, the adjusted spending (including spending on claims that are left unadjusted) is summed across a person-year. This results in one input-price adjusted value per person per year.
HARVARD CLAIM-DAY METHOD MEMORANDUM (11.04.11)
Background
There are frequent occurrences of multiple claims for a single procedure coded on a single day for a single person. This is due to the way claims are coded by providers and processed by insurers. We believe the vast majority of these claims are not separate events, but a single event coded on multiple claim lines. In addition, about 14 percent of total medical spending (non-drug) has no associated CPT code. The vast majority (>95 percent) of this spending is represented in the outpatient facility file. After exploring these claims with experts at Thomson Reuters, we believe that the majority are facility fees carved out of outpatient claims.
Harvard’s Approach
The general approach is to collapse all claims for a specific service (e.g., CPT code), performed on the same day for the same enrollee to 1 claim-day observation. This is the unit of quantity that will be used throughout much of the analysis, such as creating price lists for output price adjustment. Harvard recognizes that this approach will mean sometimes counting 2 distinct procedures performed on the same day as 1 claim day. As a result those days will be codes as having a higher price as opposed to greater quantity.
More detailed methods are described here:
• For inpatient facility claims, all claims coded as the same DRG on the same date of admission for the same person will be collapsed to 1 claim-day. There are no inpatient facility claims that are missing an associated DRG code.
• CPTs are used for inpatient professional, outpatient professional, and outpatient facility claims. There are three components paid
for a CPT claim a work component, a practice expense and a malpractice component. If the service is provided in a facility (which is always the case for inpatient services and may be the case for outpatient services) the practice expense component is reduced and the facility gets paid separately (by DRG in the case of inpatient or CPT in the case of outpatient). We will collapse inpatient professional claims with the same enrollee, day, and procedure code into 1 claim-day observation.
• For outpatient claims we cannot identify if and how the professional and technical components are broken out, and therefore we will collapse outpatient claims with the same procedure code, enrollee, and day into 1 “claim-day” observation, thus rolling together the professional and facility components.
• Around 45,000 outpatient claims (14 percent of total medical spending in 2007) do not have an associated procedure code. For these claims, we apportion spending across all CPT claims that occurred on the same day for each enrollee based on the proportion of spending that each CPT represents on that day. For instance, if $100 lacked a procedure code, and on the same day we observe two CPTs: CPT “A” for $200 and CPT “B” $400, we apportion the $100 based on a 1/3 and 2/3 ratio, respectively. The new value for CPT “A” is $225, and CPT “B” is $475. We run this apportioning step before performing input or output price adjustments, and before imputing capitated claims.
HARVARD OUTPUT-PRICE ADJUSTMENT MEMORANDUM (10.05.12)
Background
The purpose of the output price adjustment is to eliminate the effect of different prices between locations. Harvard’s approach is to create a nationalized standard price for each procedure code and type of claim (inpatient DRG, inpatient CPT, and outpatient CPT) based on the national mean payment for each procedure per day it was rendered. This standard price is then applied back to each claim day of the same type. For example, we compute the mean spending per person per day for a specific code for computed tomography (CT) scan (conditional on being greater than zero for that person on that day). We then apply that price to each person day for the same CT code regardless of where they live or what was actually paid.
Trimming Before Calculating the National Mean Price
We do not trim claims other than those with spending less than or equal to zero. We experimented with trimming and found it made the decomposition into price and quantity effects difficult without substantially altering conclusions about aggregate variation.
Types of Claims Included
The output price adjustment is performed for all inpatient and outpatient claims with a procedure code. Claims that do not include a procedure code and that cannot be apportioned to other claims on the same day are left unadjusted. Capitated claim days are omitted when calculating means/ medians. Drug claims are also left unadjusted
Summing Procedure
After the spending is adjusted as described above, the adjusted spending amounts and residual unadjusted spending (drug claims and claims that are missing procedure code after the apportioning step) are summed across a person-year. This results in one output price adjusted value per person per year.
Zero-Dollar and Capitated Claims
We outline our overall approach to capitated claim-days in a separate memo. We will apply an output price adjustment value to all capitated claim days (even those with observed spending of $0). We will not apply an output price adjustment to non-capitated claim days that have $0 spending. This is because we believe capitated claim days with zero dollar spending generally represent procedures that occurred, while non-capitated claim days with zero dollar spending are likely corrections and do not represent actual services.