This manual is provided on the World Wide Web courtesy of IRP, Inc. the leading distributor of DRG, APG, Severity-Refined DRG, Coding, and Reimbursement software products for hospital, health professionals, insurance companies, and HMO/managed care providers. We'll be glad to assist you in determining how you can utilize these technologies in your situation. Call us at 1-800-634-0496 ext 161.
Health Care Financing Administration, June 1994
Summary
This paper sets forth a proposed system for incorporating a measure of severity of illness into the Medicare diagnosis-related groups (DRGs). The DRG assignment is one of the main factors in determining the payment made for hospital inpatient services furnished to Medicare beneficiaries. Specifically, the formula used to calculate payment for a single Medicare hospital inpatient case takes the individual hospitals payment rate per case and multiplies it by the relative weights of the DRG to which the case is assigned. Thus, it is easy to see that the DRG relative weights have a large impact on the payment a hospital receives.
In this paper, we describe the Medicare DRG prospective payment system (PPS), evaluate the various classification measurements that are available for assessing severity of illness, describe the analysis we did in formulating our proposal, and present our proposed DRG severity system. We are also including an analysis of the impact that the proposed DRGs would have on hospitals. We are first presenting our proposal on this informal basis because we are extremely interested in consulting with the hospital industry before we make a formal proposal to revise the DRGs as part of a PPS proposed rulemaking document. Because of the short timeframe between the publication of the annual PPS proposed and final rules, we are unable to make major changes in response to public comment. Therefore, we are making our proposal available on a prerulemaking basis so that there will be adequate time to make changes before the revised DRG system is incorporated into an annual PPS proposed rule.
Comments on the proposed modifications for severity are due to HCFA by September 30, 1994. Comments should be mailed to:
Division of Hospital Payment Policy
1-H-1 East Low Rise
6325 Security Boulevard
Baltimore, Maryland 21207
Attn: Nancy Edwards
Similar to our usual practice for proposed DRG changes to PPS, we are making data related to the classification and recalibration of the DRGs for severity available to the public. Special severity versions of the public use data files listed below may be purchased from HCFA. Anyone wishing to purchase data tapes, cartridges, or diskettes should submit a written request along with a company check or money order (payable to HCFA-PUB) to cover the cost, to the following address: Health Care Financing Administration, Public Use Files, Accounting Division, PO Box 7520, Baltimore, Maryland 21207-0520, (410) 597-5151. Please be sure to note that you are requesting the Severity version of the file.
1. Expanded Modified MedPAR-Hospital (National)
The Medicare Provider Analysis and Review (MedPAR) file contains records for 100 percent of Medicare beneficiaries using hospital inpatient services. The records are stripped of most data elements that will permit identification of beneficiaries. Provider numbers and beneficiary claim numbers have been encrypted to protect the privacy of individuals. The hospital is identified by the 6 position Medicare billing number. The file is available to persons qualifying under the terms of the Routine Use Act as outlined in the December 24, 1984 Federal Register (49 FR 49941), and amended by the July 2, 1985 notice (50 FR 27361). The national file consists of approximately 11 million records. Under the requirements of these notices, a data release must be signed by the purchaser before release of these data. For all files requiring a signed data release agreement, please write or call to obtain a blank agreement form before placing your order.
Media: Tape/Cartridge
File cost: $3,415.00
2. HCFA Medicare Case-Mix Index File
This file contains the Medicare case-mix index by provider number based on the proposed severity DRGs. The case-mix index is a measure of the costliness of cases treated by a hospital relative to the cost of the national average of all Medicare hospital cases, using DRG weights as a measure of relative costliness of cases.
Media: Diskette
Price: $145.00
3. Table 5 DRG File
This file contains a listing of current and revised DRG numbers, DRG narrative description, relative weight, geometric and arithmetic mean lengths of stay, and day outlier trim points.
Media: Diskette
Price: $145.00
4. AOR/BOR Tables
This diskette contains data used to develop the DRG relative weights. It contains mean, maximum, minimum, standard deviation and coefficient of variation statistics by DRG for length of stay and standardized charges. The BOR tables are Before Outliers Removed and the AOR is After Outliers Removed. (This refers to statistical outliers, not payment outliers.)
Media: Diskette
File Cost: $145.00
For further information concerning these data tapes, contact Mary R. White at (410) 597-3671
Description of the Medicare DRG-Based Payment System
Over 10 years ago, in response to rising health care costs, a system for paying for hospital inpatient stays under Medicare Part A based on prospectively set rates per discharge was developed. The Social Security Amendments of 1983 (P.L. 98-21) changed the method of payment for inpatient hospital services from a cost-based, retrospective reimbursement system to a diagnosis-specific prospective payment system. This system was primarily intended to provide incentives to hospitals to manage their operations in a more cost effective manner. Section 1886 (d) of the Social Security Act (the Act) established the prospective payment system effective with hospital cost reporting periods beginning on or after October 1, 1983. The Federal regulations governing the hospital inpatient prospective payment system are located at 42 CFR part 412.
Under PPS, inpatient hospital discharges are grouped in DRGs that aggregate cases with similar resource consumption and clinical patterns. Cases are assigned to a DRG based on the principal diagnosis, up to 8 additional (secondary) diagnoses, and up to 6 procedures performed during the stay, as well as age, sex, and discharge status of the patient. DRGs provide an opportunity to define and measure a hospitals case complexity.
In developing the DRGs, a combination of approaches was used. Clinicians evaluation of diagnoses, procedures, patient characteristics, and recommended groupings plus statistical data analysis determined the final DRG categories. Guidelines for defining DRGs included the following criteria: factors used in defining the DRG should be available from the information collected by hospital abstract systems; the number of DRGs should be manageable; and DRGs should contain patients with similar clinical characteristics and similar resource utilization (Averill, 1992b).
Each case is assigned to a DRG, usually within a major diagnostic category (MDC), and may be classified to only one DRG, regardless of the number of conditions treated or services provided. For FY 1991, there were 25 MDCs to which 480 DRGs were assigned; another 9 DRGs are classified outside of MDCs. (The analysis presented in this paper was performed on FY 1991 data.) The MDCs are generally based on a particular organ system of the body; however, some involve multiple organ systems. Within most MDCs, cases are divided into surgical or medical DRGs. These DRGs may be further differentiated based on age or the presence of certain secondary diagnoses that are designated complicating or comorbid (CCs) conditions.
Diagnosis and procedure information is reported by the hospital using codes from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) (US Department of Health and Human Services, 1993). ICD-9-CM is a clinical modification of the World Health Organizations ICD-9 and was developed specifically for use in the United States. The major modification to ICD-9 is the addition of procedure codes and more specificity in the diagnosis coding.
In the original DRG development, each medical and surgical class of patients was evaluated to determine if CCs would consistently affect the consumption of hospital resources. Physician panels classified each diagnosis code based on whether the diagnosis, when present as a secondary condition, would be considered a substantial complication or comorbidity. The same basic list of CCs are used across most DRGs. However, depending on the principal diagnosis of the patient, some diagnoses in the list of CCs may be excluded (the CC exclusions list) if they are closely related to the principal diagnosis (Averill, 1992a).
The basic units of payment under PPS are the standardized amounts and the DRG relative weights. Under this system, Medicare does not pay the costs of individual cases; rather, payment is based on an averaging process, as each DRG contains a range of patient costs and lengths of stay. A DRGs weight represents the average resources necessary to care for cases in that DRG relative to the average resources used to treat cases in all other DRGs. Given a normal distribution, most cases will incur costs close to the average, with some cases costing less and some costing more. It is anticipated that some cases will incur costs in excess of payment and that these cases will be balanced by those in which payment exceeds cost. The incentive to the hospital is to manage its operation more efficiently by evaluating those areas where increased efficiencies can be achieved without adversely affecting the quality of care, and by treating a mix of patients so that payment in excess of cost in one case or DRG will offset cost in excess of payment of another case or DRG.
As noted above, the DRGs are weighted to reflect the resource consumption of each DRG relative to other DRGs based on the average charges within each DRG. Services provided to the patient during the course of treatment are not addressed specifically, but are included in the total charges, which are used as the measure of resource consumption. Each year, the relative weights assigned to the DRGs are recalibrated based on the latest available charge data for Medicare discharges. In general, these data are 2 years old; that is, the Federal fiscal year (FY) 1994 relative weights are based on FY 1992 hospital claims data.
Briefly, the methodology for calculating the relative weights is, first, to establish an average standardized charge (1) per DRG by summing the standardized charges for all cases in the DRG and dividing that amount by the number of cases in the DRG. Then, statistical outliers are eliminated by removing all cases outside of 3.0 standard deviations from the mean of the log distribution of charges per case for each DRG. An average standardized charge for each DRG is recomputed and divided by the national average standardized charge per case to determine the weighing factor. Finally, those weights are adjusted by a factor so the average case weight before recalibration is equal to the average case weight after recalibration.
(1) Charges are standardized to remove the effects of differences in area wage levels, indirect medical education costs,
disproportionate share payments, and, for hospitals in Alaska and Hawaii, the applicable cost-of-living adjustment.
Payment per discharge for operating costs is determined by multiplying the DRG relative weight by the national standardized operating payment amount. Standardized amounts are adjusted for regional differences in wages and, prior to FY 1995, are determined separately for hospitals located in urban and rural areas. Other adjustments are provided for indirect medical education costs, cost of living for hospitals in Alaska and Hawaii, sole community hospitals, Medicare-dependent, small rural hospitals, and hospitals that treat a disproportionate share of low-income patients. Currently, additional payments called outlier payments are made for cases that involve extremely long lengths of stay or extraordinarily high costs to treat when compared to other discharges classified in the same DRG. (Outlier payment for cases with a long length of stay will be phased out over FYs 1995 through 1997.) A separate payment is made for capital-related costs.
By setting payment rates prospectively, incentives are built into the system to promote efficiency and cost effective provision of inpatient services. By assigning cases to categories that are similar in terms of resource use and clinical characteristics, the intention is to establish a case-mix measure that will account for the variation in resource use among DRGs. In order to ensure equitable payment to hospitals, it is essential that DRG groupings be as homogenous as possible. Therefore, one objective of the DRG classification scheme has been to reduce the variance within case groupings. To the extent that classes of patients differ sufficiently from each other within the same DRG, the equity of payment based on averaging is reduced. That is, it is possible that the averaging process could fail to accommodate legitimate cost differences among hospitals (Queen's University, 1991). Hospitals serving a mix of patients covering a range of both case mix and costs would be predicted to fare well under PPS. However, hospitals that treat a more severely ill population, or specialize in treatment of a select, high cost group of patients, may face greater difficulty.
The PPS was designed to promote efficiency but not at the cost of possibly undercompensating (hospitals with severely ill patients or to promote the avoidance of patients using high level hospital resources (McMahon, 1992). While every effort was made in developing DRGs to relate case mix to hospital resources, it was recognized that, for some cases, payment was more appropriate than for others. The attempt to ensure and maintain equitable payment has led to annual revisions of the DRG classification system (section 1886 (d) (4) (C) of the Act).
To date, there have been 10 revisions to the original DRG classifications. Examples of DRG modifications include adding two new MDCs (MDC 24, Multiple Significant Trauma, and MDC 25, Human Immunodeficiency Virus Infections, effective October 1, 1990); creation of new DRGs (e.g., DRG 475, Respiratory System Diagnosis with Ventilator Support, effective October 1, 1987); redesign of a class of DRGs (e.g., the surgical DRGs in MDC 5, Diseases and Disorders of the Circulatory System, effective (October 1, 1990); and splitting a DRG to increase classification specificity (e.g., DRGs 410 and 492, Chemotherapy with and without Acute Leukemia as Secondary Diagnosis, effective October 1, 1991). The classification of secondary diagnoses as CCs are routinely updated to improve within DRG homogeneity.
A Review of Current Severity Measures
Although many of the previous DRG refinements have resulted in improved variance reduction, we recognize that further modifications could enhance the explanatory power of the classification system. Concerns about the fairness of hospital payment and the ability of the DRG classification to adequately capture differences in levels of patient illness that impact resource consumption have led to increased interest in measures of severity of illness. For several years, we have been analyzing major refinements to the DRG classification system to compensate hospitals more equitably for treating severely ill Medicare patients. As part of this process, we evaluated those measures currently available for assessing severity of illness.
Choice of an appropriate measure depends, in part, on the objective to be met. Severity measurements may be classified according to the following objectives:
Systems designed to measure standards of hospital care, primarily for assessing quality of care and quality assurance, include Medisgroups, Computerized Severity Index (CSI), Severity of Illness Index (SOII), and Patient Management Categories (PMCs). The Acute Physiological and Chronic Health Evaluation (APACHE) and the Medicare Mortality Predictor System (MMPS) were developed as risk management tools to identify the risk of dying. The Yale Refined Diagnosis-Related Groups (RDRGs), the New York All Patient DRGs (AP-DRGs), and the All Patient Refined DRGs (APR-DRGs) exemplify systems developed for payment purposes.
Alternatively, severity measurements may be categorized according to the data source on which the system is based:
- Those systems using the current HCFA classification system.
- Those systems using the same data base as the current HCFA system but not the same classification scheme.
- Those systems using neither the data base nor the classification scheme currently in use by HCFA (U.S. Department of Health and Human Services, 1990).
Refinements to the current system based on HCFA DRGs include Yale RDRGs, the New York AP-DRGs, the APR-DRGs, and the Acuity Index Method (AIM). Replacements of the current system that rely on the Uniform Hospital Discharge Data Set (UHDDS) data with different classifications are Disease Staging, Q Scale, and PMCs.
Examples of measurement methodologies that rely on some additional data available only from the medical record are SOII, CSI, Medical Illness Severity Groups, the APACHE, and the MMPS. Current severity of illness measures that we have reviewed and evaluated are summarized in Table 1.
TABLE 1 - SUMMARY OF SEVERITY MEASUREMENT INSTRUMENTS
Method of
System Data Source Data Collection Objective
AIM Medicare DRGs Electronic Predict Length of Stay
AP-DRGs Medicare DRGs Electronic Predict Resource use for
Non-Medicare Population
APACHE II Medical Record Medical Chart Review Predict Risk of
Death
APACHE-L Medical Abstract Electronic Predict Mortality
CSI Medical Record Medical Chart Clinical Outcomes
and Electronic
Disease Medical Abstract Electronic Within DRG Comparison
Medisgroups Medical Record Medical Chart Review Clinical Outcomes
MMPS Medical Record Medical Chart Review Predict Death
PMCs Medical Abstract Electronic Normative Standards
SOII Medical Record Medical Chart Treatment Characteristics
RDRGs Medicare DRGs Electronic Predict Resource use
for Medical Population
APR-DRGs AP-DRGs Electronic Predict Resource use
for All Patients
Distinctions between severity of illness, case-mix complexity, intensity of service, and resource intensity have been noted in several studies (Averill, 1993) (Houchen, 1989). Our objectives in measuring severity have been to determine the extent to which patient resource use within DRGs varies due to severity of illness, to evaluate the effect of a severity adjustment on total hospital payment, and to develop a cost effective method for adding severity dimension to patient classification for purposes of Medicare payment for hospital inpatient services. Our evaluation focused on whether an alternative methodology improved the predictive capabilities of the current system in terms of resource utilization (using length of stay and total charges as proxies for resource consumption), the stability of the system across classification categories as well as over time, the vulnerability of the system to manipulation, the costs associated with adopting the method, and the number of final classification groups.
We found that all the severity systems we evaluated explained more variation in resource use than the current Medicare DRGs alone. However, explanatory power across DRGs has been demonstrated to exhibit considerable variation across differential severity measures. Iezzoni found, for example, that Medisgroups showed only modest improvement over current DRGs for select DRGs, with an increase in explanatory power greatest among medical DRGs (Iezzoni, 1991b). These results parallelled those found using similar measurement systems that rely on computerized UHDDS data. DRG refinement systems that require special abstraction of data would impose significant administrative burdens, involving substantial data collection, verification, and processing. Pennsylvania, with a mandate to collect the medical chart data required for Medisgroups for all cases in the State, incurred a cost of $10 per case (Iezzoni, 1991c). For Medicare, with over 10 million discharges per year, this translates to a significant financial burden for both HCFA and hospitals. Thus, although we recognize that a refinement that relies on information already provided and readily available rather than additional information obtained from the patient's medical record may show only modest improvement in explaining variation in hospital costs and lengths of stay, we included a criterion of administration ease and limited increased financial burden in our decision making.
In developing a methodology appropriate for Medicare purposes, the following criteria were used:
- Within-group variation in resource use must be reduced, resulting in improved homogeneity within DRG.
- The final number of classification groups must be manageable and administratively feasible.
- Necessary data must be easily obtainable and consistent across hospital.
- Administrative costs must be reasonable.
In addition, we sought a system that would be seen as fair, nonpunitive, and easy to understand by hospitals, physicians, and beneficiaries. To ensure adaptability to existing hospital data and claims payment systems, the number of expanded DRGs was limited to no more than 999. While increasing the number of patient classes generally improves accuracy in predicting resource consumption, it also increases the opportunity for manipulation of the system by shifting patients into classes with higher payments. Also, increasing the number of categories increases the potential for incorrect DRG assignment based on coding errors.
The extent to which the severity measures under consideration met the HCFA criteria is summarized in Table 2.
TABLE 2 - ABILITY OF SEVERITY MEASUREMENT SYSTEM TO MEET HCFA CRITERIA
READILY
MANAGEABLE OBTAINABLE REASONABLE REDUCE NUMBER OF & CONSISTENT ADMINISTRATION SYSTEM VARIANCE GROUPS DATA COST
AIM YES NA YES NA
AP-DRGS YES YES YES YES
APACHE II YES NA NO NO
APACHE-L YES NA NO NO
CSI YES NA NO NO
DISEASE
STAGING YES NA NO NO
MEDIS-
GROUPS VARIES NA NO NO
MMPS LIMITED NA YES YES
PMCs YES NA NO NO
SOII YES NA NO NO
RDRGs YES NO YES YES
APR-DRGs YES NO YES YES
As we stated above, all the measures reduced variation within the DRGs, with varying degrees of success, and improved the ability to predict resource consumption. However, the number of resulting groups often was not included in the description of the system or depended upon whether the system was used to overlay existing DRGs or applied independently to individual case records. Measures such as the MMPS limited observations to a small group of DRGs. Those systems that relied on additional medical record information were eliminated from consideration as being too costly to administer. Based on these criteria, the most promising DRG refinements are the Yale RDRGs, the New York AP-DRGs, and the APR-DRGs.
The Yale RDRGs were developed by the Health Systems Management Group at Yale University under a cooperative agreement with HCFA. The RDRGs are closely related to HCFA DRGs, assigning cases to a DRG based on principal diagnosis, secondary diagnoses, and surgical procedures. In developing the RDRGs, HCFA DRGs are first collapsed to combine paired groupings (DRGs with and without CCs). For example, DRG 272 (Major Skin Disorder with CC) is combined with DRG 273 (Major Skin Disorders without CC) to form the new adjacent DRG (ADRG) 272 (Major Skin Disorders).
ADRGs are then divided into subclasses. Each medical DRG is structured to contain 3 subclasses; surgical DRGs are divided into 4 subclasses. These subclasses are known as refined DRGs (RDRGs). These class levels correspond to the level of resource intensity, as measured by the effect secondary diagnoses have on resource requirements. The medical classes are minor or no effect, moderate effect, and major effect. The surgical RDRGs include the same categories plus one for catastrophic effect. Comparable CCs are assigned to the major effect RDRG for medical cases and to the catastrophic effect RDRG for surgical cases. Although, for the most part, there are standard sets of diagnosis codes that define the classes for all medical and surgical cases, there are some exceptions for specific ADRGs.(2) Not unexpectedly, the catastrophic RDRGs are more costly, have longer lengths of stay, and are also more disparate in terms of resource use than the noncatastrophic RDRGs (U.S. Department of Health and Human Services, 1990).
The Yale RDRGs recognize two special classes: medical cases involving early death (within 2 days of admission) and cases requiring tracheostomy procedures. DRGs for these two conditions are present within each MDC except MDC 3 (Ear, Nose, Mouth, and Throat) and MDC 15 (Newborns and Other Neonates), and cases are assigned to them prior to other DRG determinations. In MDC 3, the initial tracheostomy group contains only medical cases since the procedure is often part of the normal course of treatment for surgical patients in this MDC. The Yale revisions expanded the number of patents classes to 1,263 DRGs. Health Care Investment Analysts (HCIA) currently maintains and updates the RDRG system to remain consistent with the Medicare DRGs. The RDRG system is currently used by the Ohio Department of Health (Leary, 1992) and is under consideration for use by the Washington State Health Care Authority (Health Care Investment Analysts, Inc., 1992).
In 1987, New York State enacted legislative that required a prospective payment system for all non-Medicare patients. The State Department of Health was required to assess the appropriateness of HCFA DRGs for a non-Medicare population, including a specific evaluation of neonates and patients with Human Immunodeficiency Virus (HIV). When first implemented on January 1, 1988, the New York All-Patient DRGs (AP-DRGs) expanded the Medicare DRG classification system to include newborn and neonate categories based on birth weight, which were modified versions of the neonate
(2) For example, pleural effusion is a member of the major class for medical cases except when it occurs with pulmonary embolism, respiratory neoplasms, major chest trauma, heart failure and shock, and other circulatory system diagnoses, where it is a member of the moderate class. This is because pleural effusion commonly occurs as part of the disease process for these principal diagnoses.
categories of the pediatric modified DRG (PM-DRG) system developed by the National Association of Childrens Hospitals and Related Institutions (NACHRI).
In 1990, New York refined its DRG system by the addition of a severity measure. New York developed a list of secondary diagnoses that were considered to have a major effect on resource use when present in a case. This list was based on the Yale secondary diagnoses that are designated catastrophic for surgical cases and major for medical cases, with modifications to eliminate those diagnoses that New York did not believe to be consistently catastrophic or major or that are susceptible to coding manipulation and to add other diagnoses that are not considered catastrophic or major in the RDRGs.
New York's analysis showed that, within any MDC, the surgical patients with major CCs were similar in terms of resource use and the medical patients with major CCs were also similar. That is, the presence within an MDC of the major CC was a better indicator of the resources used than was the type of surgery performed or the principal diagnosis. Therefore, New York created major DRG categories by MDC.
Some MDCs have only two major CC AP-DRGs, one for surgical cases and one for medical cases (for example MDC 2, Diseases and Disorder of the Eye). In other MDCs, further distinctions are made in either the surgical or medical partitions. For example, MDC 11 (Diseases and Disorders of the Urinary System) has two medical major CC AP-DRGs and MDC 6 (Diseases and Disorders of the Digestive System) has three surgical major CC AP-DRGs. These revisions resulted in 54 new groups for major CC cases. As of January 1, 1994, there are a total of 632 AP-DRGs.
Most recently, the APR-DRGs were developed as part of a joint research effort between 3M/Health Information Systems (HIS) and NACHRI to address the following limitations in the Yale RDRGs:
- The base DRGs are the Medicare DRGs, which do not address the non- Medicare population.
- There is no recognition of the impact of multiple CCs.
- The secondary diagnoses that would assign a patient to a CC subclass are limited to the Medicare list of CCs.
- The structure of four surgical subclasses and three medical subclasses is inconsistent and confusing.
The AP-DRGs were used as the base DRGs in the formation of the APR-DRGs, the APR-DRGs are developed from the information contained in the medical record abstract or UB-92 billing form, including principal diagnosis, secondary diagnoses, operating room procedures, age, sex, and discharge disposition, as well as birth weight and days on mechanical ventilator for neonates.
The APR-DRGs consist of consolidated DRGs together with four complexity subclasses: minor or no CC, moderate, major, and extreme. The assignment of a patient to a subclass is a three phase process. In the first phase, the complexity level of each secondary diagnosis is determined. The second phase determines a base complexity subclass for the patient based on the patients secondary diagnoses. In the third phase, the final complexity subclass for the patient is determined by incorporating the impact of principal diagnosis, age, nonoperating room (non-OR) procedures, and combinations of categories of secondary diagnoses.
The secondary diagnoses considered extreme are primarily serious acute conditions that are often life-threatening and require extensive amounts of resources. Major CCs are primarily significant acute diseases or chronic diseases for which an acute exacerbation would present a significant problem for the patient and would require a substantial amount of additional resources. A moderate CC includes acute and chronic diseases that have only a modest impact on resource use. Minor secondary diagnoses have little or no impact on total resource use. The complexity level for some secondary diagnoses differs depending on whether the APR-DRG is medical or surgical. Certain neuromuscular diseases (e.g., myopathy) have a greater impact on the amount of resources used for medical patients than for surgical patients. Conversely, certain complications (e.g., malfunctioning of graft or device) can be more resource intensive for surgical patients.
The 348 basic APR-DRGs, each with four complexity subclasses, combined with the 45 neonatal APR-DRGs results in a total of 1,437 APR-DRGs. The neonatal APR-DRGs do not follow the same subclass divisions, but, rather, are based on subclasses for multiple major neonatal problems, major neonatal problems, significant neonatal problems, and other neonatal problems. The APR-DRGs were first available on January 1, 1993 and will be updated annually by physician specialists for both adult and pediatric patients and research analysts from 3M/HIS and NACHRI (Averill, 1993). The updates will be made available each spring, with the first update scheduled for release in spring 1994.
Severity measurement systems that refine the current DRGs have many advantages over methods that replace this system. The current DRG system has been in use for over 10 years and its rationale and methodology are relatively well understood. Hospital systems are designed to process claims based on this system, as are those of the Medicare intermediaries (the contractors HCFA uses to process and pay claims). The mechanisms to retrieve and process patient data necessary to the present system are already in place. Implementation time and costs associated with a revision are less than those associated with replacement for both hospitals and HCFA.
Since the Yale RDRGs, New York AP-DRGs, and the 3M/HIS APR-DRGs are all systems originally based on the Medicare DRGs and use the same data source and elements, they could easily be adapted and used for Medicare payment. With this in mind, we evaluated these three classification systems as possible severity systems for the Medicare population.
The Yale system represents a significant increase in the number of DRGs, giving rise to the following issues:
- The number of low-volume RDRGs.
- The stability of the relative weights over time.
- The ability of the RDRGs to capture the difference in the amount of resources needed to treat cases with increasing degrees of complication.
We found that using the Yale RDRGs resulted in a sizable increase in the number of low-volume DRGs (those with fewer than 10 cases) and an even more significant increase in the number of DRGs with 30 or fewer cases. Thus, compared to the current Medicare DRGs, the Yale RDRGs would result in a relatively large increase in the number of low-volume classes.
Examining the stability of relative weights between 2 years of data, we found that the relative weights of 48 percent of all Yale RDRGs changed by 5 percent or more. For the same 2 years, only 24 percent of the Medicare DRGs changed by 5 percent or more. This indicates that the relative weights under the Yale system are less stable over time than under the Medicare DRGs.
Finally, we analyzed the differences in relative weights between adjacent severity classes. For medical DRGs, the relative weights of the Moderate class DRGs are, on average, almost 40 percent higher than the minor or no CC class and the major DRGs have on average, a relative weight 65 percent higher than the moderate class. For surgical DRGs, the average relative weight is 23 percent higher for moderate DRGs compared to minor or no CC DRGs, the major DRGs have average relative weights 34 percent higher than moderate DRGs, and the difference between the average relative weights of the catastrophic DRGs and the major DRGs is 66 percent.
Thus, it appears that the Yale severity classification consistently capture the differences in the amounts of resources needed to treat more severe cases. Moreover, in both the medical and surgical DRGs, the catastrophic classifications capture the cases that are significantly more costly to treat than cases in other classifications with relative weights that average 65 percent higher than the immediately preceding severity category.
Based on our evaluation of the Yale RDRGs, we believe that a system that incorporates catastrophic CCs results in a significant improvement in explanation of severity of illness. The New York AP-DRGs incorporate this aspect of the Yale system, with the addition of only 54 DRGs. In comparing low-volume DRGs, we found that New York compares favorably with the Medicare DRGs and would not result in much, if any, difference in number of low-volume DRGs.
Regarding the stability of the relative weights from year to year, New York is superior to Yale and improves on the HCFA DRGs with only 23 percent of the AP-DRGs experiencing a greater than 5 percent change in weight from one year to the next. One problem we did find, however, is that by consolidating major CC cases at the MDC level, the New York system often combines groups of severely ill patients who require on average, substantially different resources. In addition, some of those severely ill groups did not vary sufficiently in terms of resource use from other clinically similar groups in the MDC to justify placing them in a major CC DRG.
Although we did not perform the same analysis with the APR-DRGs as we did for the Yale and New York systems, because of their similarity to those systems, we believe any analysis would produce similar results. Like Yale, the significant increase in the number of DRG classes and the resulting probability of increased low-volume DRG and instability in relative weights from one year to the next would offset the systems significant improvement in case-level homogeneity and ability to explain resource use for severely ill patients. In addition, the relatively complicated algorithm that is used to determine the complexity subclass of a case is not easily explained or understood. We believe this would make it more difficult for a typical hospital to have enough experience to allow meaningful comparative analysis to be performed.
In light of these findings, we believe the best approach would be to develop our own system, which would incorporate aspects of both the Yale RDRGs and the New York AP-DRGs. Therefore, similar to both Yale and New York, we are proposing to modify the DRG classification system by developing a list of secondary diagnoses that have major effect on the resources used by hospitals in treating patients across DRGs. However, unlike Yale, we would not automatically create a major CC class for every DRG. Conversely, unlike New York, we would not create major CC DRGs on an MDC level. Rather, we would evaluate the need to create DRGs for groups of patients with major CCs or CCs on a DRG-by-DRG basis.
Therefore, we are proposing to modify the DRG classification system by including a severity measurement based on secondary diagnoses that are classified as major CCs. Our purpose is to increase DRG homogeneity, improve payment equity, and to recognize the impact of varying severity levels on resource consumption.
The proposed modification enhances the ability of the Medicare DRGs to predict resource consumption. It must be noted, however, that multiple factors in addition to severity of illness affect health care outcomes and the cost of hospitalization. Length of stay and charges may vary based on the extent to which the diagnosis is known at the time of admission, eliminating costly tests to determine cause for admission, whether the treatment is aggressive or palliative, or whether or not there are multiple stays for the same patient (Jencks, 1987). The burden of underlying chronic disease, patient functional impairment, and the psycho-social status of the patient (Iezzoni, 1990) may have an impact on the hospital stay and resource utilization. None of these factors are captured in the Medicare data.
In developing the proposed methodology, we performed a variety of statistical analyses of length of stay and standardized charge data, in addition to consulting with physicians on issues that required clinical judgement. We used FY 1991 charge and length of stay data and FY 1992 GROUPER DRG classifications for our analysis.
PROPOSED DRG CLASSIFICATION SYSTEM FOR SEVERITY
The following is a brief description of the methodology we used to create our severity DRGs:
- We collapsed the current paired DRG groupings (DRGs with and without CCs).
- We combined 24 sets of DRGs based on current Medicare utilization and charge data as well as clinical similarity.
- Using an iterative process, we evaluated each individual diagnosis code to determine if its presence as a secondary condition resulted in increased resource use for hospitals across all DRGs and designated each codes as either a major CC, a CC, or a non-CC.
- Finally, we evaluated each collapsed DRG to determine if it should be split on the basis of the presence of a major CC, a CC, both, or neither.
Collapsing DRGs
Prior to identifying potential DRG severity subclasses, the first analysis we undertook was to review the current classification groupings to determine if any consolidation was possible. In order to do this, we combined existing DRG pairs (i.e., those DRGs that currently split based on CCs) so that we could compare overall DRG charge and length of stay data. When the combined DRG pairs also contained an age split, the age distinction was maintained. Next, based on clinical judgement and statistical analysis, we identified DRGs that should be combined because they contained patients with similar clinical patterns and resource consumption. As a result of this analysis and clinical review, we combined 24 select groups of DRGs (Table 3).
TABLE 3 - CONSOLIDATED DRGS
Current Consolidated
DRG Current Title DRG Revised Title
1 Craniotomy Age >17 Except 1 Craniotomy Age >17 for Trauma
2 Craniotomy for Trauma Age >17
42 Intraocular Procedures 42 Intraocular Procedures Except Retina, Iris and Lens Except Iris and Lens
43 Hyphema 46 47 Other Disorder of 46 47 Other Disorder of the Eye * the Eye
50 Sialoadenectomy 51 Salivary Gland 51 Salivary Gland Procedures Procedures Except Sialoadenectomy
55 Miscellaneous Ear, Nose, 56 Miscellaneous Ear, Mouth, and Throat Procedures Nose, Mouth, and 56 Rhinoplasty Throat Procedures
72 Nasal Trauma and Deformity 73 74 Other Ear, Nose 73 74 Other Ear, Nose, Mouth, and Mouth and Throat Throat Diagnoses Diagnoses
89 90 91 Simple Pneumonia 89 90 91 Simple Pneumonia, and Pleurisy Pleurisy and 92 93 Interstitial Lung Disease Interstitial Lung Disease
85 86 Pleural Effusion 94 Pneumothorax and Pleural Effusion 94 95 Pneumothorax
146 147 Rectal Resection 148 149 Major Small and 148 149 Major Small and Large Large Bowel Bowel Procedures Procedures
185 186 Dental and Oral Disorders 185 186 Dental and Oral Except Extractions and Disorders
Restorations
187 Dental Extractions and
Restorations
199 Hepatobiliary Diagnostic 199 Hepatobiliary Procedure for Malignancy Diagnostic 200 Hepatobiliary Diagnostic Procedures Procedure for Nonmalignancy
202 Cirrhosis and Alcoholic 205 206 Disorders of Liver Hepatitis Except Malignancy 205 206 Disorders of Liver Except
Malignancy, Cirrhosis, and
Alcoholic Hepatitis
223 Major Shoulder and Elbow 223 224 Shoulder, Elbow and Procedures, or Upper Forearm Procedures Extremity Procedures w CC
224 Shoulder, Elbow or Forearm
Procedures Except Major Joint
Procedures w/o CC
228 Major Thumb or Joint 228 229 Hand and Wrist Procedures, or Other Hand Procedures or Wrist Procedures w CC
229 Hand or Wrist Procedures,
Except Major Joint Procedures,
w/o CC
244 245 Bone Diseases and Specific 246 Bone Diseases and Arthropathies Arthropathies 246 Nonspecific Arthropathies
250 251 252 Fracture, Sprain, 253 254 256 Fracture, Sprain, Strain, and Dislocation Strain, and of Forearm, Hand, and Foot Dislocation of Lower Leg of Upper Extremity 253 254 255 Fracture, Sprain, Strain,
and Dislocation of Upper Arm
and Lower Leg Except Foot
257 258 Total Mastectomy for 259 260 Mastectomy for Malignancy Malignancy 259 260 Subtotal Mastectomy for
Malignancy
268 Skin, Subcutaneous Tissue, 269 Skin, Subcutaneous and Breast Plastic Procedures Tissue, and Breast 269 Other Skin, Subcutaneous Procedures Tissue, and Breast Procedures
271 Skin Ulcers 272 273 Major Skin Disorders 272 273 Major Skin Disorders
294 Diabetes Age >35 295 Diabetes
295 Diabetes Age 0-35
296 297 298 Nutritional and 296 297 298 Nutritional and Miscellaneous Metabolic Metabolic Disorders Disorders
299 Inborn Errors of Metabolism
338 Testes Procedures for 339 Testes Procedures Malignancy
339 340 Testes Procedures,
Nonmalignancy
411 History of Malignancy 412 History of w/o Endoscopy Malignancy 412 History of Malignancy
w Endoscopy
465 Aftercare w History of 465 Aftercare
Malignancy as Secondary
Diagnosis
466 Aftercare w/o History of
Malignancy as Secondary
Diagnosis
* A single title combined with two DRG numbers is used to signify a pair. Generally, the first DRG is for cases with CC and the second is for cases without CCF. If a third number is included, it represents cases of patients who are ge 0-17. Occasionally, a pair of DRGs is split age >17 and age 0-17.
Low Volume DRGs
DRGs with too few cases for valid or reliable statistical results have always been problematic for Medicare. Although the majority of Medicare beneficiaries are over age 65, approximately 10 percent are under 65 and qualify for health care coverage on the basis of disability. The claims data for some of the DRGs to which this younger population are classified have not been of sufficient volume for routine analysis. The original DRG classification system was developed from analysis of claims data representative of the total inpatient population. When we calculated the original Medicare weights for the DRGs to which newborn, children, and maternity patients are classified, we used non-MedPAR discharge records from Maryland and Michigan hospitals because there were either non MedPAR cases or too few cases classified to these DRGs to provide a reasonably precise estimate of the average cost of care. (See September 1, 1983 prospective payment final rule with comment period (48 FR 39768).)
Since that time, because of the lack of Medicare claims data, these low-volume DRGs have not been analyzed and refined, and the relative weights assigned to them may no longer be entirely reflective of the resources needed to treat the patients. In addition to the severity changes, we also intend to improve the classification and relative weights of the pediatric, newborn, and maternity DRGs. Our current plans is to use the DRG classifications that have been adopted by New York (In the AP-DRGs) and CHAMPUS as a basis for our changes. We again intend to rely on data bases outside the MedPAR file to supplement our data.
Among the low volume DRGs are those that split on the bases of age (ie.e, age 0-17). When collapsing DRGs into clinically homogenous groups, we combined some DRGs that split based on age with DRGs that did not. In creating the new DRG, we retained the age distinction. Thus, the newly created DRG still groups cases into age 0-17 or age >17 subgroups but these now contain cases that were not previously split on that basis. For example, when combining DRG 43 (Hyphema) with DRG 46 (Other Disorders of the Eye Age >17), the new DRG 46 (Other Disorders of the Eye Age >17), includes all cases over age 17 from both DRGs, and hyphema cases under age 17 are assigned to DRG 47 (Other Disorders of the Eye Age 0-17).
The 42 DRGs that split on the basis of age 0-17 were not included in the severity analysis and remain unchanged. The 22 DRGs in MDCs 14 (Pregnancy, Childbirth and the Puerperium) and 15 (Newborns and Other Neonates with Conditions Originating in the Perinatal Period) were also excluded from consideration when forming severity subclasses since we will be modifying these DRGs in a separate analysis as discussed above. After combining the CC pairs for DRG 370 and 371 (Cesarean Section), no further modifications were made to the MDC 14 or 15 DRGs.
Analysis of Secondary Diagnoses
After collapsing CC pairs and combining other clinically similar DRGs, we evaluated each diagnosis, as identified by the ICD-9-CM diagnosis coded, in order to determine its appropriate CC status, that is, whether it should be designated and treated as a non-CC, a CC, or major CC (MCC) when it is present as a secondary diagnosis. Several factors were considered in evaluating the diagnoses including charges, clinical aspects, and the coding of the diagnosis. Surgical cases were compared to medical cases, and cases in which the patient was discharged alive were compared to those in which the patient died. We also evaluated diagnoses that, when present as a secondary diagnosis, are used to determine assignment to certain DRGs. (These select analysis are described below.)
Methodology
In order to evaluate the effect of each diagnosis on the total resource use for cases in which the diagnosis is a secondary condition, these cases were grouped into three subsets based on the presence and CC status of the other secondary diagnoses of the case. In our first analysis, we defined secondary diagnoses as CCs if they are currently CCs for Medicare. A diagnosis was classified as an MCC if it was defined that way for the New York AP-DRGs. We also used the current CC and the MCC exclusions.
We devised a system that would allow us to measure the value of each secondary diagnosis when it is present in a patients case under three different scenarios:
- The patient has no other secondary diagnosis or only non-CC secondary diagnoses.
- The patient has at least one other secondary diagnosis that is a CC but none that is an MCC.
- The patient has at least one other secondary diagnosis that is an MCC.
The numerical values we assigned to each diagnosis are as follows:
Computational Values for Secondary Diagnosis
Value Meaning
0 Significantly below expected value for the non-CC subclass 1 Approximately equal to expected value for the non-CC subclass.
2 Approximately equal to expected value for the CC subclass. 3 Approximately equal to expected value for the MCC subclass.
4 Significantly above the expected value for the MCC subclass
Each diagnosis for which Medicare data were available was evaluated to determine its effect on resource use and to determine which subclass (non-CC, CC, or MCC) the cases in which it is a secondary diagnosis most closely approximate based on resource use. In order to make this determination, the average charge, or the subclass value, for each subset of cases was compared to the expected charge, or the expected value, for cases in that subset. The following format was used to evaluate the diagnoses:
Code Diagnosis Name Cnt1 C1 Cnt2 C2 Cnt3 C3
where Cnt is the number of patients in each subclass and C1, C2, and C3 refer to the three different subsets of patient described above, respectively.
The following are examples from our analysis of each type of secondary diagnosis (non-CC, CC and MCC) and their computational values:
CC
Code Diagnosis Status Cnt1 C1 Cnt2 C2 Cnt3 C3
553.3 Diaphragmatic Non-CC 8120 1.1 6103 1.7 488 2.1 hernia
568.81 Hemoperitoneum CC 18 1.9 73 2.8 43 3.5 112.5 Disseminated MCC 13 3.4 70 3.8 122 4.0
The first column (C1) shows the value of the diagnosis when it is the only secondary diagnosis or the only other secondary diagnoses are non-CCs. Code 553.3 is approximately equal to the expected value for the non-CC subclass (1), code 568.81 is approximately equal to the expected value for the CC class (2.0, and the value for code 112.5 (3.4) is higher than the expected value for the MCC class (3.0).
Both the consistency of the trend of the subclass values across the three subclasses and the volume of cases were considered in the weight given to each subclass value. Determinations involving diagnoses or subclasses with fewer than 10 cases were based both on the charges of those cases and on the charges associated with diagnosis codes for related conditions. In addition to the statistical analysis, each diagnosis was also evaluated from a clinical perspective in order to determine which CC designation would be most consistent with the conditions that it represents. In general, we changed the current CC designation of a diagnosis if the average charges or values of its subclasses more closely approximate the expected value of another CC subclass, or if we believe the diagnosis is clinically significant.
The following are examples of diagnoses whose original CC classification was revised based on the values of the computational analysis. Diagnosis code 569.81 (Intestinal fistula) is currently a non-CC in the Medicare system. However, the analysis of this code produced the following subclass values:
Cnt1 C1 Cnt2 C2 Cnt3 C3
48 2.1 168 3.0 73 3.7
Because the 48 cases in which code 569.81 is the only secondary diagnosis have a value that is equal to the value for the CC subclass, and the values consistently increase across the subclasses, we redesignated this code as a CC.
Conversely, diagnosis code 042.0 (HIV with specified infection) was designated an MCC by New York. Our analysis produced the following values:
Cnt1 C1 Cnt2 C2 Cnt3 C3
28 2.3 73 1.9 57 3.3
We redesignated 042.0 as a CC rather than an MCC based on its value of 2.3 (close to the expected value of the CC class) for the 28 cases in which it is the only secondary diagnosis as well as the fact that it decreases in value when present with another secondary diagnosis that is designated a CC. One reason we believe that 042.0 does not produce the values expected of an MCC is because of the way we currently pay for HIV cases. When HIV is a secondary diagnosis and the principal diagnosis is an HIV-related condition, the case groups to MDC 25 (HIV Infections) along with cases that have a principal diagnoses of HIV. Thus, when the HIV code is a secondary diagnosis and it not linked to a significant condition (that is, it does not group to MDC 25), it appears to require only a moderate amount of additional resources. Therefore, it produces computational values similar to a CC rather than an MCC.
We reestimated the computational values of the secondary diagnoses as changes to the list of CCs and MCCs were made. We did this because as the designation of a diagnosis was revised, the cases in which it appears as a secondary must be reevaluated using the new designation. After several iterations, the computational values and the designation of the diagnoses stabilized. When we completed a revised calculation of computational values for which we made no additional designation changes in CC status, we assigned the final designations to the diagnoses and proceeded to our next step. As a part of our overall analysis, we also conducted several focused analyses.
o Select Analysis: Medical/Surgical
We initially evaluated the resource use for surgical cases separately from medical cases. Although there were instances in which the subclass values for a secondary diagnosis present in surgical cases ranked higher that when present in medical cases, in a comparable number of circumstances, the reverse was true. The subclass values for these two types of cases indicated that, overall, for a given diagnosis, the resource use for a particular secondary diagnosis present in surgical cases did not vary significantly or consistently from that of medical cases. Therefore, distinctions between medical/surgical cases were not sufficiently definitive to warrant separate categorization. Thus, a secondary diagnosis is uniformly categorized as a non-CC, a CC, or and MCC across all medical and surgical cases.
o Select Analysis: Alive/Dead
We evaluated the secondary diagnosis data to determine whether or not there is a difference in resource use in cases in which the patient is discharged alive or died during the hospital stay. We found that, for most secondary diagnoses, the charges for these two groups were similar. However, there were a few diagnoses fore which the difference in charges and clinical considerations supported the assignment of a different CC designation for patients who died before discharge. For these diagnoses, the patients who were discharged alive required significantly more hospital resources than the patients who died., That is, the resource use (as indicated by the subclass values) for the live patients approximated the expected value for the MCC subclass. The subclass values for the cases in which the patient died approximated the expected value for the CC subclass. Therefore, each of these diagnoses is designated as an MCC in cases in which the patient is discharged alive and as a CC in cases in which the patient died. The affected diagnoses are listed below by diagnosis code.
Code Description
427.41 Ventricular fibrillation
427.5 Cardiac arrest
785.51 Cardiogenic shock
785.59 Other shock without mention of trauma
799.1 Respiratory arrest
998.0 Postoperative shock
o Select Analysis: Secondary Diagnoses Currently Required for DRG Assignment
Special attention was paid to specific secondary diagnoses that are necessary in order to determine DRG assignment. For example, specific secondary diagnoses are required in order for a case to be assigned to the DRGs for acute myocardial infarction (DRGs 121 through 123), multiple trauma cases (DRGs 484 through 487), and human immunodeficiency virus (HIV) cases (DRGs 488 through 490). For each of these specified secondary diagnoses, we compared the cases that are assigned to these DRGs to cases with the same secondary diagnosis assigned to other DRGs. Our analysis indicated that the pattern of resource utilization incurred by these secondary diagnoses was similar to all other secondary diagnoses.
As these secondary diagnoses are required to determine the DRG assignment, we believe that they should not be used to determine assignment to either a CC or an MCC subclass for these same DRGs. Our rationale is similar to that of the CC exclusions, which precludes a secondary diagnosis from being treated as a CC in those cases in which it is closely related to the principal diagnosis. For example, a secondary diagnosis of congestive heart failure (diagnosis code 428.0) was excluded from subclass determination for DRGs 121 (Circulatory Disorders with Acute Myocardial Infarction and Cardiovascular Complications Discharged Alive) and 124 (Circulator Disorders Except Acute Myocardial Infarction with Cardia Catheterization and Complex Diagnosis).
o Select Analysis: Quality Issues
We also evaluated the CC subclass of those conditions that indicate adverse results of treatment during hospitalization. We did not designate these conditions as MCCs even though the charges of the cases in which they occur are similar to those of other cases in an MCC subclass. Our rationale for excluding these conditions from MCC status was to avoid rewarding the hospital for substandard care or inadvertently providing incentives for care that does not meet quality standards. For example, diagnosis code 995.4 (Shock due to anesthesia) is currently an MCC under the New York AP-DRGs and its subclass value is like other MCCs. However, for the reasons stated above, we designated this condition as a CC.
In other instances, we examined diagnoses that we believe indicate a lack of quality of care. Diagnosis code 998.4 (Foreign body accidently left during a procedure) is currently a CC for Medicare cases, and based on our analyses, its subclass value is similar to other CCs. Although it is true that this is an outcome of surgery that may increase a patient's resource requirements, we decided not to designate it as a CC in order to avoid rewarding poor medical treatment.
o Select Analysis: Coding Issues
The ease of assigning a diagnosis code to a condition was also considered in determining the CC status for certain diagnoses. For example, many diagnosis code categories include codes for conditions that are not elsewhere classified (NEC) and for conditions that are not otherwise specified (NOS). The code for the term including NEC is to be used only when the coder lacks the information necessary to code the term to a more specific category (U.S. Department of Health and Human Services, 1993). The NOS abbreviation is the equivalent of unspecified; that is, the medical record does not include enough information to code the case to a more specific category. We are cognizant of the impact that potential assignment to a higher-weighted DRG has on coding practices; our objective is for improved, accurate coding practices as one outcome of these proposed severity modifications. Thus, whenever feasible and appropriated, we want to encourage coding to the highest level of specificity. To this end, if possible, we attempted to maintain families (related diagnoses within the same 3-digit coding structure) at the same subclass designation.
In evaluating these codes, we considered each code both individually and with the other diagnosis codes in its category of similar conditions.; In some code categories, all the diagnosis codes were assigned the same CC status, such as diagnosis code 681 (Cellulitis and abscess of finger and toe) where all 4th and 5th level designations are non-CC as well as the NOS categories (681.00, Cellulitis, finger NOS; 681.10, Cellulitis, toe, NOS; and 681.9 Cellulitis of digit, NOS).
However, in other code categories, we made other decisions. For example, code category 263 consists of five codes, as follows:
263.0 Malnutrition of moderate degree
263.1 Malnutrition of mild degree
263.2 Arrested development following protein-calorie malnutrition
263.8 Other protein-calorie malnutrition (NEC) 263.9 Unspecified protein-calorie malnutrition (NOS)
These codes are all currently designated as CCs. In evaluating their subclass values, however, we found that 263.8 and 263.9 had higher values than the other codes. In fact, their values were more like the MCC subclass.
We were extremely reluctant to designate an NOS code category as an MCC because of the potential for miscoding. Given the choice between specific codes that are designated CCs and a nonspecific code designated as an MCC, it is easy to imagine the incentive to code the NOS diagnosis. This should not be as true with regard to the NEC classification, which designates an actual specified condition in the category that is not covered by a more specific code. Therefore, the final CC designation of this code category is CC for every code except 263.8, which is designated as an MCC.
Similarly, we evaluated the CC status of symptom codes that we believe are relatively easy to code. For example, diagnosis code 998.5 (Postoperative infection) has subclass values similar to the MCC subclass. However, given an MCC subclass, it would be extremely tempting for hospitals to code minor postoperative effects, such as slightly elevated temperatures, as postoperative infection. Therefore, we did not move code 998.5 from its current CC status to MCC status.
List of Proposed Classification of Secondary Diagnoses
In our analysis of each diagnosis code to determine its appropriate CC designation, we considered all of the preceding. As discussed above, we performed several iterations of the analysis, since changing the CC designation of diagnosis codes affected the subclass values of other diagnoses. When the data for each diagnosis appeared to be stable, we set the final CC designations, which are listed in Appendix A. The E-codes, which are diagnosis codes used to classify external causes of injury and poisoning, are not included in this list. All E-codes are designated as non-CCs under the current DRG system and our evaluation supports this non-CC designation as appropriate.
Redefining DRGs on the Basis of CCs
In designing an improved DRG classification system, two of our major goals were to create DRGs that would more accurately reflect the severity of the cases assigned to them and to create groups that would have enough cases so that they would be meaningful and stable. As noted above, we excluded the DRGs in MDCs 14 and 15 as well as the DRGs for patients age 0-17 years from consideration because these are generally low-volume DRGs.
In designating a DRG as one that will be split on the basis of an MCC or CC, we developed a set of criteria to facilitate our decision-making process. In order to warrant creating of an MCC or CC subgroup within a DRG, the subgroup had to meet the following g five criteria:
- A reduction in variance in charges of at lest 32 percent - At least 5 percent of the patients in the DRG fall within the MCC or CC subgroup.
- At least 50 cases must fall into the MCC or CC subgroup.(3) - There must be at lease a 20 percent difference in average charges between the subgroups.
- There must be a $2,000 difference in average charge between subgroups.
Our objective in developing these criteria was to create homogeneous subgroups that are significantly different from one another, that have enough volume to be meaningful, and that improve our ability to explain variance.
We also evaluated the number of subgroups we would create using reductions in variance of 10 and 5 percent as one of the criteria. This proved to be overly stringent and did not recognize many of the CC groups that we currently use and which have no reason to believe are inappropriate. We finally decided to use a 3 percent criterion for both the MCC and CC subgroups.
To begin our analysis, we split each of the DRG groups into three subgroups: no CC, CC, and MCC. Each subgroup was then analyzed by itself and in relation to the other two using the volume, charges, and reduction in variance criteria. Using this method, it was possible to determine the strength of the effect the secondary diagnoses in the DRG group had on each subgroup. For any given DRG group, our evaluation may show that a split between a non-CC and CC/MCC group is all that is warranted; that is, there is not a great enough difference between subgroups to justify a split between CC and MCC. Similarly, even though an MCC subgroup is justified, there may not be enough differences between the remaining cases to justify a non-CC/CC split.
(3) Since our analysis used a 10 percent sample of the entire MedPAR file, this represents approximately 500 Medicare cases in a year.
Based on this methodology, DRGs may be split according to the following three subgroups, rather than the current with CC and without CC.
- DRG with MCC
DRG with CC
DRG without CC
- DRG with CC or MCC
DRG without CC
- DRG with MCC
DRG without MCC
The most straightforward type of DRG is that for which there are no subgroups (120 DRGs). An example of this is current DRG 6 (Carpal Tunnel Release), for which the data do not indicate any division into CC subgroups is warranted. Therefore, in the revised DRGs, as in the current ones, this DRG will have no MCC or CC differentiation.
The DRGs with an MCC designation are new and are made up of cases that have secondary diagnoses that have been designated as major, and for which a split was warranted based on the criteria listed above. For example, current DRGs 1 and 2 (Craniotomy) were combined, as explained earlier, and the, based on the secondary diagnoses, they were split into 3 DRGs; Craniotomy with MCC, Craniotomy with CCs that are not major, and Craniotomy without CC.
The last two types of paired DRG are those in which some division is warranted but for which application of the criteria did not indicate a split based only on CCs or MCCs. Rather, a separation based on a combination of CC types was indicated. Thus, we created the subgroups with CC or MCC and without MCC. These types of DRG will contain cases with a mixture of CC type. A DRG with CC or MCC will consist of cases that have at least one secondary diagnosis that is classified as either a CC or an MCC, but for which a split for either of these CC types alone was unwarranted. An example of this type of split is current DRG 13 (Multiple Sclerosis and Cerebellar Ataxia). In the revised system, DRG 13 will split to become Multiple Sclerosis and Cerebellar Ataxia with CC or MCC and Multiple Sclerosis and Cerebellar Ataxia with CC.
A DRG without MCC will contain cases that may have a secondary diagnosis that is classified as a CC, but will not have any that are MCCs, and for which a CC/non-CC split was not warranted. An example of this is current DRG 87 (Pulmonary Edema and Respiratory Failure). In the revised system, DRG 87 will split to become Pulmonary Edema and Respiratory Failure with MCC and Pulmonary Edema and Respiratory Failure without MCC.
As noted above, there are 12 DRGs to which cases are assigned based on both their principal and secondary diagnosis,; for example, the DRGs in MDC 24 (Multiple Significant Trauma). Because a secondary diagnosis is responsible for the DRG assignment, we did not believe it would be reasonable to consider these secondary diagnoses when splitting these DRGs into MCC or CC subgroups based on secondary diagnoses. That is, a secondary diagnosis used for the DRG classification should not also be used to move the case into ann MCC or CC classification. As stated in our discussion of diagnosis level analysis, we believe that this is similar to our CC exclusions policy in which a secondary diagnosis that is clinically similar to the principal should not be used to move the case to a higher-weighted DRG with CC. The following is a complete list of the DRGs for which secondary diagnoses count in assignment:
Current DRG Title
121 Circulatory Disorders with AMI and Cardiovascular Complications, Discharged Alive
122 Circulatory Disorders with AMI without Cardiovascular Complication, Discharged Alive
123 Circulatory Disorders with AMI, Expired
124 Circulator Disorders except AMI, with Cardiac Catheterization and Complex Diagnosis
259 Subtotal Mastectomy for Malignancy with CC
484 Craniotomy for Multiple Significant Trauma
485 Limb Reattachment, Hip and Femur Procedure
for Multiple Significant Trauma
486 Other O.R. Procedures for Multiple Significant Trauma 487 Other Multiple Significant Trauma
489 HIV with Major Related Condition
490 HIV with or without Other Related Condition
492 Chemotherapy with Acute Leukemia as Secondary Diagnosis
Current DRGs 468 (Extensive O.R. Procedure Unrelated to Principal Diagnosis), 476 (Prostatic O.R. Procedure Unrelated to Principal Diagnosis), and 477 (Nonextensive O.R. Procedure Unrelated to Principal Diagnosis) are reserved for those cases in which none of the OR procedures performed during a hospital stay is related to the principal diagnosis. These DRGs are intended to capture atypical cases; that is, those cases not occurring with sufficient frequency to represent a distinct, recognizable clinical group.
Since each of these DRGs consists of cases that do not share the characteristics that are used to group all other DRGs, (principal diagnosis, clinical-relatedness, and similar resource consumption), we are uncertain about what is the best method for handling them in our revised system. Currently, we do not split these DRGs on the basis of CC. However, because application of our criteria indicates that each of these DRGs should split into MCC, CC, and non-CC groups, we have incorporated that split into the proposed DRGs.
Because the cases in these DRGs lack clinical similarity and cases are only to be assigned here when there is no better placement, we do not want to provide any incentive to have cases group here. We believe these DRGs are necessary to handle special cases, but we routinely review them to identify those procedures occurring in conjunction with certain principal diagnoses with sufficient frequency to justify adding them to one of the surgical DRGs for the MDC in which the diagnosis falls.
If these DRGs have MCC subgroups, there may be an incentive for assignment to these DRGs that is not appropriate. Since this is a departure from our current methodology, we are particularly interested in comment on this issue.
The following table presents a breakdown of the number of the different proposed DRG subgroups.
Number of Number of
Subgroups Collapsed DRGs Refined DRGs
No subgroups 126 126
With CC or MCC, Without CC 72 144
With MCC, Without CC 52 104
With MCC, With CC, Without CC 85 255
Subtotal 335 629
MDCs 14 and 15 21 21
Total 356 650
Thus, after adding the DRGs for Principal Diagnosis Invalid as Discharge Diagnosis and Ungroupable (current DRGs 469 and 470), there are a total of 652 DRGs.
The severity-refined DRGs achieve a variance reduction of 42 percent, compared to 38 percent under the current system. This is only 1 percent less than the 43 percent that is achieved by splitting all DRGs three ways and thus creating 987 groups. Although this is fewer than 999 DRGs (our systems limit), it would not leave much room for any possible future DRG expansion.
The resulting 652 DRGs as well as their relative weights are listed in Appendix B as Table 5. We have also included a Table 3C that lists each hospitals case-mix index value using the refined DRGs (Appendix C). (We have maintained the Table number designations used each year in the PPS proposed and final rules.) The relative weights and the case-mix index values are based on FY 1992 MedPAR data, so they are equivalent to the current FY 1994 DRG relative weights. Also included in the table are the current DRG numbers and the geometric and arithmetic means. For simplicity, we have numbered these DRGs according to the order in which cases are assigned; that is, pre-MDC, MDC1, MDC2, etc. From experience, however, we know that, over time, consecutive numbering becomes disrupted with the revisions we make to the DRGs. We have considered two strategies in order to minimize future numbering disruption. One would be to leave several unused DRG numbers in each MDC, which will be available for future use as needed. We also considered using an alpha-numeric system that would designate both MDC and DRG. For example, DRG 7, the first DRG in MDC 1 would be A1 and DRG 90, the 20th DRG in MDC3 would be C20. Both of these alternates are somewhat problematic to incorporate into our current MedPAR system. In addition, an alpha-numeric system would limit us to 26 MDCs, only one more than we currently have.
We are particularly interested in comment from the public on a DRG numbering system. We note that the DRG designation should not exceed three characters so that hospitals, intermediaries, and HCFA will not be required to change current computer software. Thus, we would appreciate suggestions on how we could assign numbers to the revised set of DRGs in such a way as to create as clear and orderly a system as possible and prevent confusion in the future when new DRGs are created or changed.
All comments received on our proposed refinements to the DRG system by September 30, 1994 will be considered in the formal proposed refinement to the DRGs for severity.
Impact Analysis
In the analysis that follows, we examine the effects on hospital payment of implementing the severity-adjusted DRGs holding other payment variables constant. The data used in developing the quantitative analysis presented below are taken from the FY 1992 MedPAR file and the hospital-specific data that are used for payment purposes. These are the same MedPAR data used to set the severity-refined DRG relative weights. Although the analysis does not incorporate any actual cost data, the most recently available hospital cost report data were used to create some of the variables by which hospitals are categorized.
The PPS for hospital inpatient services encompasses nearly all general, short-terms acute care hospitals that participate in the Medicare program. We have included in our analysis only those hospitals paid under PPS. There are 55 short-term, acute care hospitals that remain excluded from the prospective payment system under section 1814(b)(3) of the Act (in Maryland) or a demonstration project (in the Finger Lakes region of New York State).l Thus, as of August 1993, approximately 5,300 hospitals were receiving prospectively based payment for furnishing inpatient services. This represents about 83 percent of all Medicare-participating hospitals. This impact analysis focuses on this set of hospitals.
Our analysis has several qualifications. First, we could not make adjustments for behavioral changes that hospitals may adopt in response to this policy change., Second, due to the interdependent nature of the PPS, it is very difficult to precisely isolate and quantify the impact associated with a given change. Third, the results of our analysis are, of course, dependent on the quality of the data employed. We have attempted to construct each variable using the best available source, and we are confident that our simulation accurately projects the likely impact on various hospital groups. For individual hospitals, however, data biases may occur.
The simulation estimates total payments under the PPS for inpatient operating costs using both the current FY 1994 and the severity-adjusted DRG relative weights. We use an estimated FY 1995 single standard rate for rural and other urban hospitals. Large urban hospitals receive a higher standard rate. In addition, we have adjusted outlier thresholds for each set of DRGs to meet required outlier payment levels.
As noted above, short-terms acute care hospitals not paid under the prospective payment system (hospital sin the New York Finger Lakes demonstration project, and hospitals in Maryland) are excluded from the simulation. Payments under the capital PPS, or payments for other than inpatient operating costs, are not estimated. For purposes of determining which method payment to apply to sole community hospitals (SCHs or Medicare-dependent, small rural hospitals (MDHs) (the Federal payment rate or the applicable hospital-specific payment rate as prescribed by section 1886 (d)(5)(D)(i) of the Act), we assume that all such hospitals have a cost reporting period that coincides with the Federal fiscal year.
The following impact table categorizes hospitals by various geographic and special payment groups to illustrate the varying impacts on different types of hospitals. The top row of the table shows the overall impact on the 5,301 hospitals included in the analysis. The next three rows of Table I contain hospitals categorized according to their geographic location (large urban, other urban or rural) based on the Metropolitan Statistical Area (MSA) definitions. There are 1,636 hospitals located in large urban areas (populations of 1 million or fewer), and 2,321 hospitals in rural areas.
The next two groupings are by hospital bed size and urban or rural MSA designations. The final groupings under geographic location are by census divisions, determined on the basis of geographic location in either an urban or rural county under the new MSA definitions.
The next two grouping are by hospital bed size and urban or rural MSA designations. The final groupings under geographic location are by census divisions, determined on the basis of geographic location in either an urban or rural county under the new MSWA definitions.
The next three groupings examine the impacts of the proposed changes on hospitals grouped by whether they have residency programs (teaching hospitals), whether they receive disproportionate share (DSH) payments, and whether they receive some combination of these two adjustment. Major teaching hospitals in this analysis are those having 100 or more residents.
Disproportionate share hospitals are grouped according to their payment status during FY 1994. That is, hospitals located in rural counties that have been reclassified as urban by the Medicare Geographic Classification Review Board (MGCRB) for purposes of assigning the standardized amount are categorized here as urban, since they are considered urban in determining the amount of their DSH adjustment. The rural DSH hospitals, therefore, including those in the rural referral center (RRC) and SCH categories, represent hospitals that were not reclassified for the standardized amount. The next category groups urban hospitals in terms of whether they receive the indirect medical education adjustment or the DSH adjustment, or both.
The next six rows categorize rural hospitals by special payment groups (SCHs, RRCs, and MDHs). Rural hospitals reclassified for purposes of the standardized amount for FY 1994 are not included here. The MDH and RRC rows include all hospitals that we have identified as MDHs or RRCs that were not reclassified for purposes of the standardized amount or both the wage index and the standardized amount. Because the Omnibus Budget Reconciliation Act of 1993 (Public Law 103-66) permits MDHs and RRCs that failed to qualify as a result of a reclassification decision by the MGCRB for either FY 1993 or FY 1994 to decline their reclassification, it was necessary to project which hospitals would exercise this option. Those projected to do so are included here.
The RRCs (197), SCHs (619), and RRC/SCHs (49) shown here were not reclassified for purposes of the standardized amount.
The next two groupings are based on type of ownership and the hospitals Medicare utilization expressed as a percent of total patient days. Data needed to calculate Medicare utilization percentages were unavailable for 66 hospitals.
Column 3, labeled Effect on Payment, shows the results of comparing projected FY 1995 payments using the severity DRGs to payments using the FY 1994 DRGs.
Because the refined DRGs better account for severity, payments should shift from hospitals treating less severe cases to those treating more severe cases. As expected, rows 2, 3, and 4 show that payments to large urban hospitals would increase, while those to other urban and rural hospitals would fall. Also, payment s increase to both large teaching hospitals and urban hospitals serving a disproportionate share of low-income patients. These are the hospitals that have traditionally been thought to be treating the most severely ill Medicare patients. Although virtually all categories of rural hospitals would experience a decrease in payments (hospitals in New England and nonspecial status hospitals would be unaffected), the degree of the change varies from a decrease of 0.2 for DSH hospitals to 2.2 for hospitals in Puerto Rico.
IMPACT ANALYSIS OF REFINED DRGs
Effect on Number of Number of Payment Hospitals Cases (Percent) (1) (2) (3) (By Geographic Location)
All Hospitals 5,301 10,394,116 0.0
Large Urban Areas 1,636 4,731,994 0.2 (Population over 1 Million)
Other Urban Areas 1,344 3,635,186 -0.3
Rural Areas 2,321 2,026,936 -0.6
Urban Hospitals 2,980 8,367,180 0.0
0- 99 Beds 750 479,424 -0.2
100-199 Beds 899 1,653,923 0.0
200-299 Beds 611 2,121,934 0.0
300-499 Beds 529 2,643,821 0.1
500 or more Beds 191 1,468,078 0.1
Rural Hospitals 2,321 2,026,936 -0.6
0- 49 Beds 1,179 373,404 -0.6
50- 99 Beds 708 628,119 0.6
100-149 Beds 222 378,969 -0.6
150-199 Beds 106 257,329 -0.8
200 or more Beds 106 389,115 -0.5
Urban by Region
New England 172 534,996 -0.4
Middle Atlantic 447 1,620,445 1.0
South Atlantic 453 1,380,352 -0.4
East North Central 498 1,488,283 0.1
East South Central 170 530,314 -0.5
West North Central 187 531,983 0.0
West South Central 380 817,664 -0.4
Mountain 121 325,555 -0.2
Pacific 502 1,042,605 -0.1
Puerto Rico 50 94,982 -1.4
Rural by Region
New England 53 57,961 0.0
Middle Atlantic 85 150,477 -0.3
South Atlantic 302 391,596 -0.5
East North Central 313 329,870 -1.0
East South Central 289 324,813 -0.5
West North Central 542 297,184 -0.2
West South Central 361 259,284 -0.9
Mountain 226 112,040 -0.6
Pacific 145 98,385 -1.1
Puerto Rico 5 5,326 -2.2
By Payment Classification
All Hospitals 5,301 10,394,116 0.0
Teaching Status
Non-Teaching 4,260 6,073,560 -0.2
Fewer than 100 821 3,088,760 -0.1 Residents
100 or more 220 1,231,796 0.5 Residents
Disproportionate Share Hospitals (DSH)
Non-DSH 3,480 5,677,792 -0.3
URBAN DSH
100 Beds or More 1,302 4,242,179 0.2
Fewer than 100 Beds 140 63,083 0.3
RURAL DSH
Sole Community (SCH) 112 74,267 -0.2
Referral Centers (RRC) 48 151,839 -0.2
OTHER RURAL DSH HOSPITALS
100 Beds or More 60 94,570 -0.3
Fewer than 100 Beds 159 90,386 -0.9
URBAN TEACHING AND DSH
Both Teaching and DSH 605 2,431,196 0.3
Teaching and no DSH 386 1,757,462 -0.2
No Teaching and DSH 837 1,874,065 0.0
No Teaching and No DSH 1,409 2,610,182 -0.2
Rural Hospital Types
Nonspecial Status 3,974 9,080,061 0.0 Hospitals
RRC 197 560,748 -0.6
SCH 619 411,341 -0.3
Medicare-Dependent Hospitals (MDH)
SCH and RRC 49 132,307 -0.5
SCH and MDH 1,065 66,670 -0.5
Type of Ownership
Voluntary 3,058 7,690,660 -0.1
Proprietary 776 1,228,348 0.1
Government 1,467 1,475,109 0.0
Medicare Utilization as a Percent of Inpatient Days
0-25 308 292,296 0.4
25-50 1,613 3,746,846 -0.1
50-65 2,301 5,040,031 -0.1
Over 65 1,012 1,257,668 -0.1
Unknown 66 57,275 1.3
BIBLIOGRAPHY
1. Adams, E.K.; Houchens, R.; Wright, G.; Robbin, J.; Predicting Hospital Choice for Rural Medicare Beneficiaries: The Role of Severity of Illness, Health Services Research 26 (5): 583-612, Dec 1991
2. Alemi, F.; Rice, J.; Hawkins, R.; Predicting In Hospital Survival of Myocardial Infarction: A Comparative Study of Various Measures, Medical Care 28(9): 762-775, Sept 1990
3. Appleby, C.; New Jersey to Study Tools for Reimbursement Overhaul, Healthweek, p.5, Aug 14, 1989
4. Averill, R.; McGuire, T.; Manning, B.; Fowler, D.; Horn, S.; Dickson, P.; Coye, M.; Knowhon, D.; Bender, J.; A Study of the Relationship Between Severity of Illness and Hospital Costs in New Jersey Hospitals, Health Services Research 27(5):587-606, Dec 1992a
5. Averill, R.; Mullin, R.; Steinbeck, B.; Elia, E.; DRGs: Diagnosis Related Groups Definitions Manual, Version 10.0, 3M/Health Information Systems, HCFA contract 500-89-0015, 1992b
6. Averill, R.; Goldfield, N.; Steinbeck, B.; Muldoon, J.; Beaudry, P.; All Patient Refined Diagnosis Related Groups, 3M/Health Information Systems, 1993
7. Burns, L.; Wholey, D.; Effects of Patient, Hospital and Physician Characteristics on Length of Stay and Mortality, Medical Care 29(3):251-271, Mar 1991
8. Coulan, R.F.; Gaumer, G.L.; Medicare's Prospective Payment System: A Critical Appraisal, Health Care Financing Review, Supplement 45-77, 1991
9. Damiano, A.; Berger, M.; Draper, E.; Knous, W.; Wagner, D.; Reliability of a Measure of Severity of Illness: Acute Physiology of Chronic Health Evaluation II, Journal of Clinical Epidemiology 45(2):93-101, 1992
10. Farley, D.; Hogan, C.; Case Mix Specialization in the Market for Hospital Services, Health Services Research 25(5):757-783, Dec 1990
11. Feldman, R., Commentary: Are We Splitting Hairs over Splitting DRGs?, Health Services Research 27(5):613-617, Dec 1992
12. Feldman, R., Averill, R. et al., Research Dialogue, Health Services Research 27(5):607-612, Dec 1992
13. Fetter, R.; Thompson, J.; Inventors of DRGs look at PPS now, Hospitals, p. 136, July 5, 1992
14. Gianfrancesco, F.D.; The Fairness of Prospective Payment Reimbursement Methodology, Health Services Research 25(1): 1-23, Apr 1990
15. Goldfarb, M.; Coffey, R.; Changes in the Medicare Case Mix Index in the 1980s and the Effect of the Prospective Payment System, Health Services Research 27(3):385-415, Aug 1992
16. Green, J.; Wintfeld, N.; Sharkey, P.; Passman, L.; The Importance of Severity of Illness in Assessing Mortality, Journal of the American Medical Association 263: 241-246, Jan 12, 1990
17. Green, J.; Passman, L.S.; Analyzing Hospital Mortality: The Consequences of Diversity in Patient Mix, Journal of American Medical Association 265(4): 1849-1857, April 10, 1991
18. Health Care Investment Analysts, Inc., Evaluation of Alternative Patient Classification Systems, Apr 1992
19. Horn, S.; Sharkey, P.D.; Buckle, J.M.; Backofer, J..E.; Averill, RF, Horn, RA, The Relationship Between Severity of Illness and Hospital Length of Stay and Mortality, Medical Care 29(4): 305-317, Apr 1991
20. Houchen, R.L.; Conkline, J.E.; Briscoe, W.N.; Measures of Complexity of Illness within DRGs, Final Report, Prospective Payment Contract T47540316, Task Order 9, Washington DC, ProPAC, Mar 1989
21. Houchens, R.; Briscoe, W.; Within DRG Complexity Change in Fiscal Year 1990, Prospective Payment Commission Extramural Report E-92-04, Apr 1992
22. Iezzoni, L.I.; Ash, A.S.; Cobb, J.; Moskowitz, M.A.; Admission Medisgroup Score and the Cost of Hospitalizations, Medical Care 26(11): 1068-1080, Nov 1988
23. Iezzoni, L.I.; Ash, A.S., Coffman, G.A.; Moskowitz, M.A.; Admission and Midstay Medisgroup Scores as Predictors of Hospital Charges, Medical Care 29(3): 210-220, Mar 1991a
24. Iezzoni, L.I.; Ash, A.S., Coffman, G.A.; Moskowitz, M.A.; Admission and Midstay Medisgroup Scores as Predictors of Death within 30 Days of Hospital Admission, American Journal of Public Health, 81(1): 74- 78, Jan 1991b
25. Iezzoni, L.I.; Restuccia, J.; Schwartz M.; Schaumberg, D.; Coffman, G.; Kreger, B.; Butterly, J.; Selker, H.; The Utility of Severity of Illness Information in Assessing the Quality of Hospital Care, Medical Care 30(5): 428-444, May 1992
26. Iezzoni, L.I.; Schwartz, M.; Restuccia, J.; The Role of Severity Information in Health Policy Debates: A Survey of State and Regional Concerns, Inquiry 28(2): 117-128, Summer 1991c
27. Iezzoni, L.I.: Severity of Illness Measures, Medical Care 28(9): 757- 761, Sept 1990
28. Jencks, S.; Daley, J.; Draper, D.; Thomas, N.; Lenhart, G.; Walker, J.; Interpreting Hospital Mortality Data: The Role of Clinical Risk Adjustment, Journal American Medical Association 260(24): 3611-3628, Dec 23/30, 1988
29. Jencks, S.; Dobson, H.; Refining Case Mix Adjustment: The Research Evidence, New England Journal of Medicine 317(11): 679-686, Sept 10, 1987
30. Kahn, K.; Rogers, N.H.; Rubenstein, L.V.; Sherwood, M.J.; Reinisch, E.J.; Keeler, E.B.; Draper, D.; Kosecoff, J.; Brook, R.H.; Measuring Quality of Care with Explicit Process Criteria Before and After Implementation of DRG Based PPS. Journal of American Medical Association 264 (15): 1984-1994, Oct 17, 1990
31. Keeler, E.; Kahn, K.; Rubenstein, L.V.; Sherwoood, M.J.; Reinisch, E.J.; Draper, D.; Kosecoff, J.; Brook, R.H.,.; Changes in Sickness at Admission Following Introduction of PPS, Journal of American Medical Association 264(15): 1962-1979, Oct 17, 1990
32. Leary, Renee; Leary, R.; Dove, H.; RDRGs: A New Perspective on Patient Classification, pre-publication copy, Topics in Health Information Management, May 1993
33. Leary, Robert; Leary, R.; Dove, H.; Refined DRG Version 9.0: Upgrade for the Ohio Department of Health, 1992
34. MacKenzie, E.J.; Steinwachs, D.; Ramzy, AS.I.; Ashworth, J.W.; Shankar, B.; Trauma Case Mix and Hospital Payment: The Potential for Refining DRGs Health Services Research 26(1): 10-26, Apr 1991
35. MacMahon, Jr., L.F.; Hayward, R.A.; Bernard, A.M.; Rosevear, J.S.; Weissfeld, L.A.; APACHE-L: A New Severity of Illness Adjuster for Inpatient Medical Care, Medical Care 30(5): 445-452, May 1992
36. McGuire, Thomas, Recent DRG Refinements, Unpublished Manuscript, Mar 1990
37. McNeil, B.; Kominski, G.; Williams-Ashmen, A.; Modified DRGs as Evidence for Variability in Patient Severity, Medical Care 26(1): 53- 61, Jan 1988
38. Mullin, Robert, DRG Statistics FY 1985 to FY 1993, Hospital of Saint Raphael, New Haven, Conn., 1992
39. Park, R.E.; Brook, R.A.; Kosecoff, J.; Keesey, L.; Rubenstein, L.; Keeler, C.; Kahn, K.L.; Rogers, W.H.; Chassin, M.R.; Explaining Variations in Hospital Death Rates, Journal of American Medical Association 264(4): 484-490, July 25, 1990
40. Pompei, P.; Charlsom, M.E.; Alex, K.; MacKenzie, C.R.; Norton, M.; Relating Patient Characteristics at the Time of Admission to Outcomes of Hospitalization, Journal of Clinical Epidemiology 44(10): 1063- 1069, 1991
41. Queen's University, Patient Classification Systems: An Evaluation of the State of the Art, Volume I, Case Mix Research, Queen's University, Kingston, Ontario, Feb 1991
42. Relman, Arnold; Assessment and Accountability, New England Journal of Medicine 319(18): 1220-1222 Nov 3, 1988
43. Rogers, W.H.; Draper, D.; Kahn, K.L.; Keeler, E.B.; Rubenstein;, L.V.; Kosecoff, J.; Brook, R.H.; Quality of Care Before and After Implementation of the DRG-Based Prospective Payment System, Journal of American Medical Association 264(15): 1989-1998, Oct 17, 1990
44. Schoeder, S.A.; Kabcenell, A.I.; Do Bad Outcomes Mean Substandard Care, Journal of American Medical Association 265(15):1995, Apr 17, 1991
45. Stearns, S.C.; Hospital Discharge Decisions, Health Outcomes, and the Use of Unobserved Information on Case Mix Severity, Health Services Research 26(1):27-51, Apr 1991
46. Thomas, J.W.; Ashcraft, M.; Measuring Severity of Illness: A Comparison of Interrater Reliability Among Severity Methodologies, Inquiry 26(4):483-492, Winter 1989
47. Thompson, J.D.; DRG Prepayment: Purpose and Performance, Bulletin New York Academy of Medicine 64(1):28-52, Feb 1988
48. US Department of Health and Human Services, Health Care Financing Administration, International Classification of Diseases, 9th Revision, Clinical Modification, Volumes 1,2,3 Med-Index, 4th printing, 1993
49. US Department of Health and Human Services, Report to Congress: Chapter Three: The Case Mix Adjustment, Unpublished Document, 1990
50. Zuckerman, S.; Feder, J.; Hadley, J.; Hospital Responses to Medicare's Prospective Payment System, Bulletin New York Academy of Medicine 64(1): 52-62, Jan-Feb 1988
 |