Dissecting DRGs for Bundled Payment

Hospitals that have applied to participate in the Medicare bundled payment for care improvement initiative (BPCI), and who requested the data that’s available to applicants, will be provided with data for all patient services related to the specific DRGs requested within the selected geographic areas.  The new data provides information about provider types and timeframes that were never previously visible to most hospital and physician providers.  When the data has been linked into inpatient episodes, these organizations will be able to view all services that were billed to Medicare by all providers during the inpatient stay and throughout the selected post-discharge period. 

This paper describes some of the analytics that are possible with this data, and some of the insights into provider services that can be gained through this process.  A more thorough description of the analytic process for bundled payments is described in the “Analytics for Bundled Payments” white paper, which can be obtained from Singletrack Analytics on the “Contact Us” page at www.singletrackanalytics.com. A description of the Medicare bundled payment initiative is available on the “Articles” page of that website.

Under the BPCI, a “contacting organization” agrees to be financially responsible for all Medicare costs for selected DRGs for a timeframe specified in the contract, which must include the inpatient stay and between 30 and 90 days post-discharge.  This contacting organization is generally composed of a hospital and a group of physicians, but may include a physician-only group or other combinations of providers.

This article describes analysis of a medical and a surgical DRG: DRG 470 (Major joint replacement or reattachment of lower extremity w/o MCC), and DRG 392 (Esophagitis, gastrointestinal & miscellaneous digest disorders w/o MCC).  These DRGs provide good examples of the types of analyses that can be undertaken with this data.  Additional variations on these analyses, using the analytical tools that were used to develop these examples, will be necessary to provide sufficient information to prepare a comprehensive bundled payment contracting proposal.

The data used in these analyses was taken from the Medicare 5% sample, which includes claims for 5% of the Medicare population.  Although representative of the population in general, this data may not be representative of the experience for any specific hospital.  These charts show Medicare payments to providers, which represent the “cost” to the Medicare program of these services.  These payments are charged against the bundled payment budget that is contracted with CMS.  Payments that exceed the bundled payment budget are charged to the contacting organization, while payments that are lower than the contracted budget create a savings paid to the contacting organization.

Payments by timeframe

Under the BPCI, contacting organizations are responsible to manage the care of patients after discharge for periods of between 30 and 90 days.  Therefore, it is important to understand the services that are provided during that period, and to assure that the organization can manage the services that are related to the inpatient admission, and can identify and exclude unrelated services from the contract.  In the example below for DRG 470, the majority of services are provided during the inpatient stay.  Relatively few services are provided within the 30 day post-operative period; however a significant amount of cost is incurred during the 30 to 90 day time post-discharge period.

Figure 1 - DRG 470 by Timeframe

In DRG 392 there is considerably more post-discharge care provided.  Also, for the highest-cost patient the post-discharge cost is extremely high.  The reasons for that cost will be explored later in this article.

Figure 2 - DRG 392 by timeframe


Payments by Provider Type

The graphs below show payments to the types of providers who were paid by CMS for those services, which may include home health providers, skilled nursing providers, durable medical equipment suppliers and other similar organizations who are paid by Medicare. This is important because these provider types may not be within the network of the contacting organization.  The “carrier” services shown below are providers that are paid by the Medicare carrier, which are predominantly physicians.

In DRG 470 we can see that most of the cost is attributable to hospital and carrier payments, but that a significant number of patients required SNF services with a few utilizing home health services.

Figure 3 - DRG 470 by provider type

In DRG 392 we see that the carrier payments generally compose a greater proportion of the cost, and that SNF services compose significant portions of cost for some patients.

Figure 4 - DRG 392 by provider type

Payment by physician specialty

Under the Medicare payment rules, hospitals are paid a per-case rate for inpatient services; therefore there is little variation in hospital payment within the same DRG.  Many surgeons are also paid a per-case rate; however medical specialists are paid on a fee for service basis.  Therefore, the costs of physician services by DRG can vary significantly, and an understanding of the services will be critical to managing cost under a bundled payment.

The two graphs below show cost by physicians specialty incurred during the inpatient stay, and in the post-discharge period.  The inpatient period shows predominantly orthopedic services, as well as anesthesia and other hospital based physicians.  The post-discharge graph shows a significant amount of orthopedic services, but also services provided by other provider types, notably clinical laboratory, radiation oncology, and internal medicine.  This indicates a potential presence of diagnoses other than orthopedic conditions which should be examined for potential exclusion from the bundled payment episode.

Figure 5 - DRG 470 - Inpatient Stay

Figure 6- DRG 470 - Post-discharge period

The analysis for DRG 392 shows less consistency in medical specialties, indicating that these types of patients are treated by wider variety of specialists.  Therefore, accepting a bundled payment for this DRG would require inclusion of a larger group of physicians in the contacting organization, or would result in a contract in organization taking risks for physicians outside of the organization.  In addition, the large amount of cost for hematology for the highest-cost patient during the post-discharge indicates a patient with a cancer diagnosis may be unrelated to the esophagitis diagnosis of DRG 392, and therefore may be excludable from the bundled payment for that DRG. 

Figure 7- DRG 392 - Inpatient Stay

Figure 8 - DRG 392 - Post-discharge period

Procedure drilldown

The CMS data provided to contacting organizations contains data at the claim line level, so it is possible to drill down to that level to understand the exact services that created a particular cost.  In the  example below, costs for the highest-cost patients in DRG 470 are displayed, with the General Surgery classification of CPT codes drilled down to show the individual CPT codes billed to Medicare under the that category.  From this drilldown, it is evident that the patient received surgery and care in an admission that occurred during the post-discharge period.  This service should be examined to determine its relationship to the cause for the initial admission, and for the possibility that the services could be excluded from episodes in this DRG.


In addition to procedure code information, the primary diagnosis codes from each claim are also contained in the CMS data.  This data is useful in assessing the medical conditions that caused services to be provided to patients in these DRGs, and is especially useful in examining the services provided during the post-discharge period.

The charts for DRG 470 show, as would be expected, that the predominant primary discharge for these patients was musculoskeletal disorders.  However, in the 90 day post-discharge episode, significant cost was incurred for patients with a diagnosis of neoplasms (cancer), and several other patients have costs for ophthalmological or circulatory disorders.  Absent any other arrangement in a bundled payment contract, these costs would be included in the bundled payment.

Figure 9 - DRG 470 - Primary Diagnoses - All Periods

Figure 10 - DRG 470 - Primary Diagnoses - Post-Discharge Periods

For DRG 392, the highest inpatient cost category was gastrointestinal diagnoses; however significant costs were associated with other diagnoses.  During the post-discharge period, however, musculoskeletal diagnoses accounted for the greatest amount of payment.  This is not unexpected, since many Medicare beneficiaries have multiple health conditions; however these costs must be taken into account when developing the bundled payment proposal.

Figure 11 - DRG 392 Primary Diagnoses - All Periods


Figure 12 - DRG 392 - Primary Diagnoses - Post-Discharge Period


Revenue Code Analysis

The CMS data also allows additional analyses that can be used to identify and compare actual incurred costs among hospital patients.  When adjusted for variations in charge levels, comparisons can also be made among hospitals.  The inpatient data records contain charge detail by hospital revenue code, which can be displayed by DRG to show the hospital services provided to inpatients.  This data can be compared among patients within a hospital to indicate different use of ancillary services, variations in use of medical supplies or devices, or other areas in which different treatment patterns by physicians can lead to variations in hospital costs.  Under the BPCI, if physicians participate in hospital cost reduction efforts, the savings generated by those efforts can be shared with those physicians.  Therefore, an important component of BPCI participation is in identifying these cost reduction opportunities. 

Comparisons within a hospital can be made using the charges reported in the CMS data, since the charge structure within a hospital is consistent.  However, comparisons based on charges between hospitals are inaccurate because of the differences in charge structures among different hospitals.  Therefore, charges need to be adjusted to create cost figures that are comparable.  These costs can be approximated by using the ratios of cost to charges (RCCs) from the Medicare cost reports of each hospital.  After adjusting the charges in each cost center by the respective RCCs for each hospital, cross-hospital comparisons of reasonable accuracy can be performed.  The objective of such comparisons is to identify areas for possible cost reductions (where the contracting organization’s costs are significantly higher than those of other hospitals).

In the example below, RCC-adjusted  average charge per case by cost category for DRG 470 are compared with the adjusted charges for other hospitals. Variations in costs for medical supplies, physical therapy, and pharmacy costs can be seen and can provide a starting point for further research into potential cost-reduction opportunities.

Bundled Payment Analysis at Singletrack Analytics

Singletrack Analytics has developed a comprehensive approach to analytics using the CMS Limited Data Set (LDS), which is the data provided to bundled payment and accountable care organization applicants.  These data sets are quite large, and cannot be processed using common desktop programs.  Using our database tools, we have developed the processes to load the data, prepare the episode bundles, and display the information using Microsoft-based desktop tools.  Controls such as the “DRG Dissection Dashboard” shown below allow quick data selection and filtering of the data presented in the charts and tables in this article.  These tools allow stream-of-consciousness analysis of this data, where the result of one analysis can immediately be used for the next one.  This allows teams of physicians and analysts to rapidly delve into detailed DRG data and extract critical information for the bundled payment proposal.

Figure 13 - DRG Dissection Dashboard

About Singletrack Analytics

Singletrack Analytics is a healthcare consulting firm providing financial and analytical expertise to assist healthcare providers and purchasers achieve to success through better use of data and analytic techniques.  Our 30+ years of experience and unique combination of financial and data-related credentials make us a cost-effective partner for healthcare organizations looking for cutting-edge, data-driven strategies to meet today’s financial and regulatory challenges.

To discuss how Singletrack Analytics can work with your team, please call Jon Pearce at 856-762-0605.  For information about Singletrack Analytics, please visit our website at www.singletrackanalytics.com or email at info@singletrackanalytics.com.