PHILADELPHIA REFLECTIONS
Musings of a Philadelphia Physician who has served the community for six decades

Return to Home

Related Topics

Reflections on Impending Obamacare
Reform was surely needed to remove distortions imposed on medical care by its financing. The next big questions are what the Affordable Care Act really reforms; and, whether the result will be affordable for the whole nation. Here are some proposals, just in case.

Final Rule: Outpatient Reimbursement Regulations

CY 2013 OPPS/ASC Final Rule with Comment Period

Medicare 2013 OPPS Final Rule Claims Accounting

Calculating OPPS payment rates consists of calculating relative resource cost and calculating budget neutrality adjustments, which are applied to estimates of resource cost and the conversion factor to create a budget neutral prospective payment system. The purpose of the following discussion is to provide a detailed overview of CMS manipulation of the CY 2011 claims data to produce the final prospective CY 2013 OPPS payment rates. The following information supports an already detailed discussion of data manipulation in the CY 2013 OPPS/ASC final rule. This discussion is divided into two parts: the traditional accounting of claims behind the cost calculations and an accounting of claims behind the budget neutrality, outlier, and impact calculations.

PART 1 - COST CALCULATIONS

CMS used information from 120 million single procedure, generated single procedure, and generated single “session” composite claim records to set the final APC rates to be paid under Medicare OPPS for CY 2013. This number of “pseudo” and “natural” single bills is comparable to the 110 million single bills that we used in the CY 2012 OPPS/ASC final rule. The increase in claims data used is largely attributable to an increase in the actual number of claims, although other “pseudo” single claims creation processes could potentially contribute to that increase. We now estimate that we used some portion of the data from 95 percent of the CY 2011 claims containing services payable under the OPPS for CY 2013 rate setting.

For the CY 2013 OPPS, we retain all HCPCS codes on the CY 2012 bypass list and include HCPCS codes that are not on the CY 2012 bypass list that, using either CY 2012 OPPS final rule or February 2012 HOP Panel data, met the established empirical criteria for inclusion on the bypass list. We note that, under our policy of basing the CY 2013 OPPS relative weights on geometric mean costs, we also revised the median cost of packaging criterion to reflect the geometric mean cost of packaging instead. We typically include codes that violate our empirical criteria in response to public comment recommending certain codes and when our clinicians believe that the services would rarely have packaging and that any packaging associated with the services would be very limited due to the clinical nature of the service.

Attached is a narrative description of the accounting of claims used in the setting of payment rates for Medicare’s 2013 Outpatient Prospective Payment System (OPPS). We note that while we have historically developed the OPPS relative payment weights based on median costs, we finalized a policy to develop CY 2013 OPPS relative payment weights based on geometric mean costs, which would continue to include all services (i.e. HCPCS codes) in an APC. For a detailed discussion of the CY 2013 OPPS policy of basing the final CY 2013 OPPS relative weights on geometric mean costs, please see section II.A.2.f. of the CY 2013 OPPS/ASC final rule.

CY 2013 OPPS/ASC Final Rule with Comment Period

As described in detail in the material that follows, geometric mean costs were calculated from claims for services paid under the Medicare OPPS and cost report data for the hospitals whose claims were used. The geometric mean costs were converted to payment weights by dividing the geometric mean for each APC (a group of HCPCS codes) by the geometric mean cost for APC 606, the mid-level outpatient visits APC in CY 2013. As discussed in Part 2 below, the resulting unscaled weights were scaled for budget neutrality to ensure that the effect of the final recalibration of APC weights for CY 2013 was removed. The scaled weights were multiplied by the final CY 2013 conversion factor to determine the final national unadjusted payment rate for the APCs for CY 2013. Calculation of payment rates for drugs and biological are an exception, as their payment rates are a percentage of average sales price and are not scaled.

The purpose of this part of the claims accounting is to help the public understand the order in which CMS processed claims to produce the final CY 2013 OPPS geometric mean costs and the reasons that not all claims could be used.

General Information :

In order to calculate the APC costs that form the basis of OPPS payment rates, CMS must isolate the specific resources associated with a single unique payable procedure (which has a HCPCS code) in each APC. Much of the following description, Pre-stage 1 through Stage 3, covers the activity by which CMS 1) extracts the direct charge (i.e. a charge on a line with a separately paid HCPCS code) and the supporting charge(s) (i.e. a charge on a line with a packaged HCPCS or packaged revenue code) for a single, major payable procedure for one unit of the procedure and 2) packages the supporting charges with the charges for the single unit of the major procedure to acquire a full charge for the single unit of the major procedure. In order to calculate the costs for composite APCs, CMS must isolate the specific resources associated with a single “session” of the composite service. Although these single session claims have more than one payable service, the direct charge for these services would be combined with supporting packaged charges to identify a full charge for the composite session.

CMS estimates resource costs from the billed charges by applying a cost-to-charge ratio (CCR) to adjust the charges to cost. CMS uses the most recent CCRs in the CMS Hospital Cost Report Information System (HCRIS) file in the calculation of the payment weights. Wherever possible, department CCRs rather than each hospital’s overall CCR are applied to charges with related revenue codes (e.g. pharmacy CCR applied to charges with a pharmacy revenue code). The order of matching department CCRs to revenue codes is laid out in the OPPS revenue code-to-cost center crosswalk (http://www.cms.gov/Medicare/Medicare-Fee-for-ServicePayment/HospitalOutpatientPPS/crosswalk.html). In general, CMS carries the following data elements from the claim through the weight setting process: revenue code, date of service, HCPCS code, charges (for all lines with a HCPCS code or if there is no HCPCS code, with an allowed revenue code), and units. Some specific cost modeling calculations may require more data elements.

CY 2013 OPPS/ASC Final Rule with Comment Period

Definitions of terms used:

“Excluded” means the claims were eliminated from further use.

“Removed to another file” means that we removed them from the general process but put them on another file to be used in a different process; they did not remain in the main run but were not eliminated because the claims were used to model specific costs.

“Copied to another file” means that we copied information off the claims but did not eliminate any of the copied information.

“STAGE” means a set of activities that are done in the same run or a series of related runs; the STAGE numbers follow the stages identified in a spreadsheet that accounts for the claims.

“*” Indicates a component of the limited data set (LDS) available for purchase from CMS.

Pre-STAGE 1: Identified gross outpatient claim population used for OPPS payment and applied the hospital CCRs.

Pulled claims for calendar year 2011 from the national claims history, n=156,135,150 records with a total claim count of 153,206,242. This is not the population of claims paid under OPPS, but all outpatient claims processed by fiscal intermediaries.

Excluded claims with condition code 04, 20, 21, 77 (n=398,934). These are claims that providers submitted to Medicare knowing that no payment will be made. For example, providers submit claims with a condition code 21 to elicit an official denial notice from Medicare and document that a service is not covered.

Excluded claims with more than 300 lines (n=2,219).

Excluded claims for services furnished in Maryland, Guam, US Virgin Islands, American Samoa and the Northern Marianas (n=2,149,165).

Balance = 152,787,224

Divided claims into three groups:

1) Claims that were not bill type 12X, 13X (hospital outpatient bill types), 14X (laboratory specimen bill types), or 76X (CMHC bill types). Other outpatient bill types are not paid under OPPS and, therefore, their claims were not used to set OPPS payment (n=31,713,169).

CY 2013 OPPS/ASC Final Rule with Comment Period

2) Bill types 12X, 13X, or 14X. 12X and 13X claims are hospital outpatient claims. Claims with bill type 14X are laboratory specimen bill types, of which we use a Sub set for the limited number of services in these claims that are paid under the OPPS (n=120,916,045).

3) Bill type 76X (CMHC) These claims are used to set the per diem partial hospitalization rate for CMHCs (n=158,010).

Balance for Bill Types 12X, 13X, and 14X=120,916,045

Incorporated all new Category I and III CPT codes and new Level II HCPCS codes that were effective as of April 1, 2012, July 1, 2012, and October 1, 2012 or will be effective January 1, 2013.

Applied hospital specific and where possible departmental specific CCRs to claims and flagged hospitals with CCRs that will be excluded in Stage 1 below. We used the most recent CCRs that was available in the CMS HCRIS system.

STAGE 1: Further refined the population of claims to those with a valid CCR and removed claims for those procedures with unique packaging and cost calculation processes to separate files.

Began with the set of claims with bill types 12X, 13X, and 14X, without Maryland, Guam or USVI, and with flags for invalid CCR set (n=120,916,045).

Excluded claims with CCRs that were flagged as invalid in Pre-Stage 1. These included claims for hospitals without a CCR, for hospitals paid an all-inclusive rate, for critical access hospitals, for hospitals with obviously erroneous CCRs (greater than 90 or less than .0001), and for hospitals with CCRs that were identified as outliers (3 standard deviations from the geometric mean after removing error CCRs) (n=5,049,768).

*Identified claims with condition code 41 and removed to another file (n=93,857). These claims were used to calculate the partial hospitalization per diem rate for hospital-based partial hospitalization programs.

Excluded claims without a HCPCS code (n=16,256).

Removed to another file claims that contain nothing but flu and PPV vaccine (n=196,860).

We assessed each line on the claim to determine whether the charge was reported under a revenue code that we allow, for purposes of OPPS rate setting, on the OPPS revenue code-to cost center crosswalk. If the revenue code is allowed, we applied the most specific available hospital specific CCR to the charge on the line. See the OPPS revenue code-to-cost center

CY 2013 OPPS/ASC Final Rule with Comment Period

crosswalk for the hierarchy of cost centers for each revenue code; where none of the revenue code specific cost centers applied, we used the hospital specific overall ancillary OPPS CCR to reduce the charges on the line to costs. If the revenue code under which a charge is reported is not allowed for OPPS rate setting, that charge is not reduced to cost nor used in calculation of the statistics that determine the OPPS weight. Typically, the OPPS does not allow revenue codes for OPPS rate setting that also are not allowed for payment by the Integrated Outpatient Code Editor (IOCE).

Balance = 115,559,304

Copied line items for drugs, radiopharmaceuticals, blood, and brachytherapy sources (the lines stay on the claim but are copied off onto another file) to a separate file (n=341,620,889). No claims were deleted. We use these line items to calculate per unit per day cost information for drugs (including therapeutic radiopharmaceuticals) and blood. We trimmed units at +/- 3 standard deviations from the geometric mean unit, and then +/- 3 standard deviations from the geometric mean unit cost, before calculating costs per unit and per day. For drugs and biological, we used the July 2012 ASP plus 6 percent and multiplied that amount by the average number of units per day for each drug or biological to arrive at its per day cost. For items that did not have an ASP, we used CY 2011 hospital claims data to determine the per day cost. We use per day cost to determine whether a drug or biological is packaged.

In the past several years, we have developed payment policy for nonpass-through separately paid drugs and biological based on a redistribution methodology that accounts for pharmacy overhead by allocating cost from packaged drugs to separately paid drugs. This typically would have required us to reduce the cost associated with packaged coded and uncoded drugs in order to allocate that cost. However, for CY 2013, we are proposing to pay for separately payable drugs and biological under the OPPS at ASP plus 6 percent, based upon the statutory default described in section 1833(t)(14)(A)(iii)(II) of the Act. Therefore, under this proposal, we would not redistribute the packaged cost. We refer readers to section V.B.3. of the final rule for a complete discussion of our final policy to pay for separately paid drugs and biological in CY 2013.

The payment rates for blood and blood products were based on simulated geometric mean costs under a different methodology that is explained in the CY 2013 OPPS/ASC final rule.

STAGE 2: Excluded claims with codes not payable under OPPS, conducted initial split of claims into single and multiple bills, and prepared claims for generating pseudo single claims.

Removed lines from claims that had payable status indicators both in the year the claim was billed and in the prospective payment year, which received no payment. This line item based trim, described in section II.A.2.c. of the CY 2013 OPPS/ASC final rule, was implemented to ensure that we are using valid claims that represent the cost of payable services to set payment rates for the prospective year. Having logic that requires both the status indicator on the claim and the prospective status indicator to be payable, preserves charges for services that would not have been paid in the claim year but for which some estimate of cost is needed for the prospective year (n=2,350,684). For the CY 2013 OPPS final rule, we are excluding line item data for pass-through drugs and biologicals (status indicator G for CY 2011 claims data) and non-pass through drugs and biological (status indicator K for CY 2011 claims data) that do not receive payment (n=205,262).

Prior to splitting the claims, we identified which status indicator Q2 codes (Tpackaged) would be paid when appearing with an S, V, or X service. If a Q2 code appeared with a separately paid procedure with a status indicator of T on the same date of service, we identified the code as packaged. If the Q2 code appeared with a separately paid procedure(s) with a status indicator of S,V, or X and no other Q2 codes were on the same date of service we forced the units to 1 and changed the major-minor designation to major, identifying the Q2 code as separately paid. Finally, if more than one Q2 code appeared on a claim with a separately paid procedure(s) with a status indicator of S, V, or X we would rank the Q2 codes using their final rule 2012 APC designations and associated scaled weight. We would change the major minor designation of the Q2 code with the highest weight to major status and force the units to 1. We designated the other Q2s on the claim packaged, status indicator of N, and left their status as minor.

Divided claims into 5 groups using the indicators (major, minor, or bypass) that are assigned to each HCPCS code. Major procedures are defined as procedure codes with status indicator S, T, V, or X. Minor procedures are defined as procedures that have status indicator F, G, H, K, L, N, R, or U.

1)*Single Major File: Claims with a single unit of one separately payable procedure (SI= S, T, V or X, which are called “major” procedures, including codes with status indicator Q3), all of which will be used in rate setting; claims with only one unit of a status indicator Q1 code that was an STVX-packaged code where there was no other code on the claim with status indicator S, T, V, or X, on the same claim on the same date; or claims where there was a status indicator Q2 (T-packaged) code with one unit where there was no code with a status indicator S, T, V, or X on the same claim on the same date (n=39,451,982).

2)*Multiple Major File: Claims with more than one separately payable procedure and/or multiple units of “major” procedures, including codes with status indicator “Q3”; claims with code that has a status indicator “Q2” that has been designated as major and separately paid (no procedure with a status indicator “T” on the same date of service and no higher weighted Q2 code on the same date of service); claims that contain conditional and independent bilateral codes when the bilateral modifier is attached to the code (n=30,151,312). Multiple major claims are examined carefully in stage 3 for dates of service and content to see if they can be divided into simulated or “pseudo” single claims.

3)*Single Minor File: Claims with a single unit of a single HCPCS to which we assigned the status indicator of N (packaged item or service), F, G, H, K, L, R, or U (n=267,862). We retain this file as insurance against last minute changes in packaging decisions.

4)*Multiple Minor File: Claims with multiple HCPCS codes, multiple services on the same date of service, and/or that have multiple units of one or more procedure codes with status indicator of F, G, H, K, L, N, R, or U; claims containing status indicator Q1 (STVX-packaged) or status indicator Q2 (T-packaged) codes with more than one unit of the code or more than one line of these codes on the same date of service and no other separately paid procedures (n =205,217).

5) Non-OPPS claims: These claims have no services payable under OPPS on the claim and are excluded (n=45,482,931). These claims have codes paid under other fee schedules such as the DMEPOS fee schedule, clinical laboratory fee schedule, and physician fee schedule. These claims have no major or minor procedures on them. The only procedure codes on these claims have a status indicator other than S, T, X, V, N, F, G, H, K, L, R, or U.

STAGE 3: Generated additional single claims or “pseudo singles” from multiple claims files

From the 30,151,312 multiple major claims, we were able to use some portion of 27,205,882 claims to create 80,202,374 pseudo single claims. As noted above, the multiple major claims already contained the final payment disposition of codes with status indicator Q2 (T-packaged codes) when they appeared with a S, T, V, or X services, making these services part of the pseudo single process. We also created 1,040,934 single “session” imaging composite claims through this process. In this preliminary rule data set, pseudo single bills were created in several different ways.

We begin by removing all line items for separately payable procedures that are thought to contain limited packaging (bypass codes) from the multiple major claims as pseudo single claims. Because bypass codes are thought to have limited packaging, we also used the line item for the bypass code as a pseudo single by estimating a unit cost and weighting any descriptive statistics.

Because some of the services on the bypass list also are included in the multiple imaging composites, we suppressed these “overlap bypass codes,” in order to retain all pertinent imaging HCPCS codes to identify a single session composite claim. Overlap codes are HCPCS codes that are both on the bypass list and are members of the multiple imaging composite APCs. The specific “overlap bypass codes” are in the Addendum N promulgated with this CY 2013 OPPS/ASC final rule.

We then broke claims by dates of service and reassessed each new claim for its eligibility as a single major claim, or in the case of the multiple imaging composite APCs, a single session claim. To improve the quality of the extended assessment and management composite APCs, we include logic to ensure that any pseudo singles for visits with G0378, observation services, billed on the same date of service, did not have a T status procedure on the day before that would preclude eligibility for the composite payment.

We created one set of pseudo singles by taking dates of service that now had only one separately paid service.

We created another set of pseudo single bills taking line-items within dates of service that contain multiple major procedures with unit=1 and no additional packaging on the date of service.

We created single session claims for estimating the multiple imaging composite APCs by identifying dates of service that contain more than one unit of a code in the same imaging family and no other separately payable codes. We later classified the dates of service for CT and CTA family and MRI and MRA family into those with and without contrast to create single session claims for the APC cost calculation.

Having identified all pseudo singles and single session claims, we reassessed the claims without suppression of the “overlap bypass codes” under our longstanding “pseudo” single process to determine whether we could convert additional claims to “pseudo” single claims.

Prior to CY 2012, we did not reassess any remaining claims or portion of claims after execution of this pseudo single procedure logic to ascertain whether packaged Q1 and Q2 codes could be elevated to major status and used as a single procedure claim for rate setting. For the CY 2013 OPPS, we will continue our CY 2012 OPPS policy of including an additional step to create pseudo single claims by treating conditionally packaged codes (identified by status indicators Q1 and Q2) that do not meet the criteria for packaging as if they were separately payable major codes. We then apply the pseudo single process to these claims to create single procedure claims from them if they meet the criteria for single procedure claims.

We were not able to use 3,135,140 claims because these claims continued to contain multiple separately payable procedures with significant packaging and could not be split (n=2,945,463) or because the claims contained services with SI=N and no separately payable procedures on the claim (n=189,677). We also were not able to use claims with the following characteristics: major procedure with a zero cost (n=36,888), major procedure with charges less than $1.01 (n=30,349); or packaging flag of 3 (n=325,177), suggesting token charges.

We also created additional single bills from the multiple minor file. We broke status indicator Q1 (STVX-packaged) and status indicator Q2 (T-packaged) codes by date, packaged all packaged costs, including other Q1 and Q2 costs, into the code with the highest CY 2012 payment weight based on CY 2012 APC assignment, forced the units to one to match our policy of paying only one unit of a code with SI=Q1 or Q2, and treated these claims as pseudo single claims. We created 32,440 pseudo singles from the multiple minor claims. We were not able to use 189,677 multiple minor claims because these claims contained minor codes that could not be elevated to major status when billed alone: largely drugs or packaged HCPCS coded procedures.

We were not able to use any of the 267,862 single minor claims because minor claims, by definition, contain only minor codes: drugs or packaged HCPCS coded procedures. Claims with a single Q1 or Q2 code with a single unit would have been classified as a single major in the initial split logic.

Balance = 120,727,730 (the sum of single majors =39,451,982, and pseudo singles from multiple majors, multiple minors, and the single “session” composite claims =81,275,748).

STAGE 4: Packaged costs into the payable HCPCS codes

We package the costs 1) on lines with packaged HCPCS codes and allowed revenue codes as shown in the revenue code-to-cost center crosswalk and 2) on lines without HCPCS but with revenue codes on the packaged revenue code file in Table 3 of the CY 2013 OPPS/ASC final rule. This included the reduced cost for coded packaged drugs and biological with an ASP and reduced cost for other packaged drugs and biological, especially estimated costs associated with uncoded pharmacy revenue codes.

We began with 120,727,730 single procedure claim records that still had costs at the line item level. We summed the costs on the claim to complete packaging and we standardized the total cost using 60 percent of each hospital’s IPPS pre-reclassification wage index. Specifically, standardized cost for the single bill or single session bill = sum of estimated line costs for the single bill or single session bill/((.6 * pre-reclassification wage index) + .4).

We left stage 4 with 120,727,730 single procedure claim records containing summarized costs for the payable HCPCS and all packaged codes and revenue centers on the claim.

Balance= 120,727,730

STAGE 5: Calculated HCPCS and APC costs

Began with 120,727,730 single procedure claim records with summarized costs.

We excluded 1,393 claim records that had zero costs after summing all costs on the claim in Stage 4.

We excluded no claim records because CMS lacked an appropriate wage index.

We excluded 1,040,372 claim records that were outside +/- 3 standard deviations from the geometric mean cost for each HCPCS code.

We excluded 1,557 claims records that contained more than 50 units of the code on the claim.

Balance = 119,684,408

We used the balance of 119,684,408 single procedure claims records to calculate HCPCS code geometric mean costs for the “2 times” examination and APC payment weight development. Section 1833(t)(2) of the Act provides that, subject to certain exceptions, the items and services within an APC group cannot be considered comparable with respect to the use of resources if the highest median (or mean cost, if elected by the Secretary) for an item or service in the group is more than 2 times greater than the lowest median cost for an item or service within the same group (referred to as the “2 times rule”). Under our final policy of basing the CY 2013 OPPS relative payment weights using geometric mean costs, we are also applying the “2 times rule” using geometric mean costs.

We added additional geometric mean costs calculated outside this process. We added a geometric mean per diem cost for APC 0173 (Level II Partial Hospitalization (4 or more services) for CMHCs) and APC 0172 (Level I Partial Hospitalization (3 services) for CMHCs), calculated from the bill type 76x claims from Pre-Stage 1. We also added a geometric mean per diem cost for APC 0176 (Level II Partial Hospitalization (4 or more services) for Hospital-based PHPs) and APC 0175 (Level I Partial Hospitalization (3 services) for Hospital-based PHPs), calculated from the bill type 13X claims with condition code 41 written off in Stage 1.

We added blood geometric mean costs that were calculated with the use of a simulated departmental CCR for blood for hospitals that do not have cost centers for blood. We added APC geometric mean costs for composite APCs, as well as other customized or “offline” geometric mean costs discussed in the final rule, such as the extended assessment and management composites and device dependent APCs. The unique assumptions behind each composite or alternative geometric mean calculation methodology are discussed in greater detail in the CY 2013 OPPS/ASC final rule.

PART 2 –BUDGET NEUTRALITY, OUTLIER THRESHOLD, AND IMPACT CALCULATIONS

After converting geometric mean costs into unscaled weights by dividing the geometric mean cost for each APC by the geometric mean cost for APC 606, the mid-level outpatient visit APC in CY 2013, we began the process of calculating budget neutrality adjustments and the outlier threshold to determine final payment rates. The result of all prospective payment policies are presented in the impact table in Section XXII Regulatory Impact Analysis of the CY 2013 OPPS/ASC final rule. The following discussion provides greater detail about our manipulation of the claims to calculate budget neutrality adjustments, to estimate outlier thresholds, and to create the impact table and overall beneficiary copayment percentage. The discussion below supplements discussion already provided in the final rule about calculation of the weight scaler, the conversion factor, the hospital and CMHC outlier thresholds, and the impact table columns.

STAGE 6: Created Summary Service Utilization Files for Current and Prospective OPPS Year by Provider

We began the budget neutrality calculations by making the services, utilization, and APC assignment on the CY 2011 claims look like they would if they were paid under the current OPPS, CY 2012, and the prospective OPPS year, CY 2013. We create a summary utilization file of services for each provider in the CY 2011 claims database that would be paid under the prospective payment system and a summary utilization file of services that would be paid under the current system for the same set of providers. In essence, this step runs the claims with payable OPPS services through a mock Integrated Outpatient Code Editor (IOCE) and Pricer for the current and prospective year and then summarizes utilization by provider, APC, HCPCS, and status indicator. Updated October 2012 IOCE specifications (v13.3) are available at: http://www.cms.gov/Medicare/Coding/OutpatientCodeEdit/Downloads/IntegOCEspecsV 133.pdf.

We constructed a summary utilization file for the CY 2013 OPPS final rule using single and multiple bills from STAGE 2 of this document (n=70,076,373), the partial hospitalization claims (n=93,857) from STAGE 1, and those from CMHCs (n=158,010) from Pre-STAGE 1. In this summary process, we identified line-items that were not payable under OPPS, including units on drugs and biological greater than the upper trim level identified in the units trim discussed in STAGE 1, units greater than 100 for procedure codes, a status indicator that is not payable under OPPS (SI=A, B, E, C, D, F, L, M), and 0 units on a claim line without an associated charge. We specifically included the 32,440 pseudo singles for claims with a separately paid Q2 or Q1 code created from the multiple minor claims in Stage 3 of the claims process. We trimmed 11,709 claims with no line items relevant to OPPS after edits. After changes in utilization and the addition of final CY 2013 payment policies, we summarized these files to a single CY 2013 summary file of 2,991,400 observations from 3,975 hospitals (including cancer and children’s hospitals) and 159CMHCs, which only provide one service, partial hospitalization. We used this summary file as the basis for modeling the final rule CY 2013 weight in the weight scaler calculation and estimated payment in CY 2013 in the impact table.

We also constructed a baseline summary utilization file to reflect the existing CY 2012 OPPS. For the CY 2012 OPPS baseline file, we began with the single and multiple bills from STAGE 2, the pseudo single claims for codes with status indicator Q1 and Q2 created from the multiple minor claims, and the same partial hospitalization and CMHC claims listed above. We removed 11,710 claims with no line items relevant to the 2012 OPPS. We summarized this second set of files to a single file of 2,995,528 services by hospitals and CMHCs. We used this summary file as the basis for modeling the current CY 2012 weight in the weight scaler calculation and estimated payment in CY 2012 of the impact table.

Utilization in both of these files includes changes for “discounting,” which is any change in payment, applied to the line-item units for a specific service on a claim, resulting from application of the multiple procedure discounting to services with status indicator T or the presence of a modifier indicating that the procedure was terminated. For 2013, we used unscaled weights, the APC geometric mean cost divided by the geometric mean cost for APC 606, to order services on each claim for application of multiple procedure discounting because scaled weights are not yet available. For 2012, we relied on CY 2012 final rule scaled weights, after modification from the subsequent correction notices.

We took a few additional steps to prepare both files for budget neutrality calculations. We adjusted units to accommodate changes in HCPCS descriptions and new HCPCS between 2011 and 2013. The final summary utilization file for the prospective CY 2013 OPPS contains 3,167,920 (including CMHCs) observations for 4,134 providers, and the final summary utilization file for the current 2012 OPPS contains 3,070,808 (including CMHCs) observations for 4,134 providers.

Each observation in these summary files includes one provider OSCAR, one HCPCS code, the SI for the HCPCS code, the APC to which the HCPCS is assigned and the sum of discounted units of that HCPCS code furnished by that hospital.

Balance prospective CY 2013=3,167,920 HCPCS, by SI, by APC, by Provider

Balance baseline CY 2012=3,070,808 HCPCS, by SI, by APC, by Provider

STAGE 7: Calculated the Weight Scaler

The weight scaler is the budget neutrality adjustment for annual APC recalibration and its calculation is discussed in section II.A. of the CY 2013 OPPS/ASC final rule. The weight scaler compares total scaled weight under the current OPPS for 4,127 providers to total unscaled weight under the prospective OPPS for the same providers, holding wage adjustment and rural adjustment constant to the current year’s adjustments. We estimated wage adjusted weight for each provider using the formula provided in section II.H. of the CY 2013 OPPS/ASC final rule without multiplying by the conversion factor, which is held constant. For example, for a procedure with SI=S provided by an urban hospital, the total weight for a service would be calculated:(UNSCALED_2013_WEIGHT*.4+UNSCALED_2013_WEIGHT*.6 *CY2012_WAGE_INDEX)*TOTAL_DISCOUNTED_UNITS

For a procedure with SI=S provided by a rural sole community hospital, the total weight for a service would be calculated: (UNSCALED_2013_WEIGHT*.4+UNSCALED_2013_WEIGHT*.6 *CY2012_WAGE_INDEX)*TOTAL_DISCOUNTED_UNITS *1.071

For a specified covered outpatient drug with SI=K provided by any hospital, the total weight for a service would be calculated: UNSCALED_2013_WEIGHT*TOTAL_DISCOUNTED_UNITS

Scaling does not apply to OPPS services that have a predetermined payment amount, especially separately paid drugs and biologicals and new technology APCs. Items with a predetermined payment amount were included in the budget neutrality comparison of total weight across years by using a weight equal to the payment rate divided by the CY 2012 final rule conversion factor. However, scaling of the relative payment weights only applies to those items that do not have a predetermined payment amount. Specifically, we remove the total amount of weight for items with predetermined payment amount in the prospective year from both the prospective and current year and calculate the weight scaler from the remaining difference. In doing this, those services without a predetermined payment amount would be scaled by the proportional amount not applied to the services with a predetermined payment amount. We do not make any behavioral predictions about changes in utilization, case mix, or beneficiary enrollment when calculating the weight scaler. After application of separately paid weights, there are seven providers that we removed from additional calculations based on not having cost report data, billing services that had no modeled payment, or missing provider information: three short term care hospitals, two long term care hospitals, and two rehabilitation hospitals.

Balance prospective CY 2013= 4,127 providers

Balance baseline CY 2012= 4,127 providers

CY 2013 weight scaler = 1.3596

STAGE 8: Calculated the Wage and Provider Adjustments

We used the same providers to estimate the budget neutrality adjustment for adopting the final IPPS FY 2013 post reclassification wage index for the CY 2013 OPPS, discussed in section II.C. of the CY 2013 OPPS/ASC final rule. Using the same wage-adjusted weight formulas presented above, the wage adjustment compares differences in total scaled, prospective CY 2013 weight providers varying only the wage index, CY 2012 and CY 2013, and using the 2012 rural adjustment. The budget neutrality adjustment for changes in the wage index is 0.9998. We did not make changes to our rural adjustment policy this year. Therefore, the final budget neutrality adjustment for the rural adjustment is 1.0000.

We used the same providers to estimate the budget neutrality adjustment for the final dedicated cancer hospital adjustment for the CY 2013 OPPS, discussed in section II.F. of the CY 2013 OPPS/ASC final rule. We calculated a CY 2013 budget neutrality adjustment factor by comparing the estimated total CY 2013 payments under section 1833(t) of the Act, including the CY 2013 cancer hospital adjustment relative to the CY 2012 cancer hospital adjustment under section 1833(t)(18)(B) and 1833(t)(2)(E) of the Act, to hospitals described in section 1886(d)(1)(B)(v) of the Act, excluding the TOPs adjustment. This difference in the estimated payments due to the cancer hospital adjustment in CY 2013 relative to the estimated payments due to the cancer hospital adjustment in CY 2012 does not have a significant impact on the budget neutrality calculation. Therefore the final budget neutrality adjustment for the final CY 2013 cancer hospital adjustment is 1.0000.

We used the same providers to estimate the budget neutrality adjustment for the final dedicated cancer hospital adjustment for the CY 2013 OPPS, discussed in section II.F. of the CY 2013 OPPS/ASC final rule. We calculated a CY 2013 budget neutrality adjustment factor by comparing the estimated total CY 2013 payments under section 1833(t) of the Act, including the CY 2013 cancer hospital adjustment relative to the CY 2012 cancer hospital adjustment under section 1833(t)(18)(B) and 1833(t)(2)(E) of the Act, to hospitals described in section 1886(d)(1)(B)(v) of the Act, excluding the TOPs adjustment. This difference in the estimated payments due to the cancer hospital adjustment in CY 2013 relative to the estimated payments due to the cancer hospital adjustment in CY 2012 does not have a significant impact on the budget neutrality calculation. Therefore the final budget neutrality adjustment for the final CY 2013 cancer hospital adjustment is 1.0000.

Balance CY 2013 providers = 4,127

Total wage index adjustment to the conversion factor = 0.9998

Total rural adjustment to the conversion factor = 1.0000

Total cancer hospital adjustment to the conversion factor = 1.0000

Total budget neutrality adjustment to the conversion factor = 0.9998

STAGE 9: Calculated Hospital Outlier Threshold

We started with aggregated claims from the single and multiple bills, pseudo singles from the multiple minor file, and partial hospitalization files to model the hospital fixed dollar hospital outlier threshold. After removing 11,709 claims with no line items relevant to OPPS, we used 70,619,115 claims to estimate the outlier threshold as well as anticipated outlier payment by provider. We created a CCR for every hospital in our hospital base file of 4,127 hospitals using the July 2012 update to the Outpatient Provider Specific File, which contains the actual overall CCRs the fiscal intermediaries or MACs are using to make outlier payments in CY 2012. We used internally calculated CCRs to substitute for any missing CCRs on the July OPSF update, and we substituted the statewide CCR for providers with CCRs greater than the 1.4 upper limit. We did not estimate the CMHC threshold this year, continuing our policy of 3.4 times payment for APC 0173 (Level II Partial Hospitalization (4 or more services))regardless of the level of partial hospitalization provided. We are proposing to continue applying the standard OPPS outlier policy for all other hospitals to the hospital-based PHP APCs.

As discussed in section II.G. of the CY 2013 OPPS/ASC final rule, we simulated CY 2013 costs by applying a charge inflation factor of 1.0866 to charges on the CY 2011 claims and by applying the CCR adjustment of 0.9880 to the July 2012 OPSF CCRs. We compared estimated cost to wage adjusted payment for each separately paid service on each claim. Holding the multiple threshold constant at 1.75 times the APCpayment amount, we iterated total outlier payment calculations, changing the size of the fixed dollar threshold each time, until total outlier payments matched our estimate of 1.0 percent of total payment on all included claims. Using the resulting $2,025 fixed dollar threshold, we estimated outlier payments for 3,066 hospitals for column 8 of the impact table.

We repeated this exercise for the current year CY 2012 OPPS. After removing claims with no line items relevant to OPPS, we used 70,619,114 claims to estimate the percentage of total payment attributable to outlier payments in 2012. We inflated charges on the CY 2011 claims by an inflation factor for one year, 1.0424, and using the CCRs from the July 2012 update to the Outpatient Provider Specific File, we estimated CY 2012 costs and compared them to wage-adjusted CY 2012 payment for each service. Ultimately, we estimated outlier payments for 3,047 hospitals for column 8 of the impact table. We also estimated total outlier payments to be 0.9% of total CY 2012 payments.

Balance CY 2013= 4,127 hospitals

Balance baseline CY 2012=4,127 hospitals

STAGE 10: Created the Impact Table and Calculated the Beneficiary Impact Percentage

The impact table in section XXII Regulatory Impact Analysis of the CY 2013 OPPS/ASC final rule compares OPPS payment for 4,127 providers in the baseline CY 2012 file to the final CY 2013 OPPS payment for the same set of hospitals, in aggregate and across classes of hospitals. We began with the summary utilization files created in Stage 6 and recreated each of the above total weight calculations (weight scaler, wage adjustment, rural adjustment) as payments by adding in the conversion factor. We compared the difference in payments between those under the CY 2013 final rule to the baseline CY 2012 payment and we show this result in column 2. The detailed calculations behind the table columns are discussed in section XXII of the CY 2013 OPPS/ASC final rule. Final rule payment presented in Column 8 of the impact table compares total estimated payment, including outlier payments, but excludes pass-through payment for the current and prospective years.

In order to group types of hospitals, we constructed a file of descriptive information from the cost report and IPPS provider files identifying different classes of hospitals. This file contains the variables we use to model adjustments including the wage index, geographic location, and provider type, as well as other descriptive information, such as bed size. We have complete information for the 4,127 hospitals with any claim used to model the prospective OPPS. We do not have complete descriptive information for 159CMHC’s because their cost report is not included in HCRIS and because they are not hospitals paid under IPPS. We make available an impact file that contains all descriptive information for the providers that we used in our calculations, as well as estimated CY 2013 payments, including outlier payments, by provider for the subset of 3,905 hospitals excluding children’s and cancer hospitals, which are permanently held harmless, and 159 CMHCs for which we present detailed information in the impact table that accompanies the CY 2013 OPPS/ASC final rule.

Finally, we estimated the overall beneficiary copayment percentage for the current and prospective OPPS years. We applied the calculated, adjusted (wage, rural, and cancer) copayment to all separately paid HCPCS, and we capped copayment at the inpatient deductible for 2012, since the CY 2013 inpatient deductible was not yet available. We summed total copayments for each year and divided by respective total payment. We estimate that total beneficiary liability for copayments would decrease as an overall percentage of total payments, from 22.0 percent in CY 2012 to 21.5 percent in CY 2013.

(2511)

Please Let Us Know What You Think


(HTML tags provide better formatting)

Because of robot spam we ask you to confirm your comment: we will send you an email containing a link to click. We apologize for this inconvenience but this ensures the quality of the comments. (Your email will not be displayed.)
Thank you.