HACsDiscussion.3RTIEvaluation

8/8/2019 HACsDiscussion.3RTIEvaluation

1/17

50085Federal Register / Vol. 75, No. 157/ Monday, August 16, 2010/ Rules and Regulations

BILLING CODE 412001C

We refer readers to section II.F.6. ofthe FY 2008 IPPS final rule withcomment period (72 FR 47202 through47218) and to section II.F.7. of the FY2009 IPPS final rule (73 FR 48474through 48486) for detailed analysessupporting the selection of each of theHACs selected through FY 2010.

3. RTI Program Evaluation Summary

a. Background

On September 30, 2009, a contract

was awarded to Research TriangleIncorporated (RTI) to evaluate theimpact of the Hospital-AcquiredCondition-Present on Admission (HACPOA) provisions on the changes in theincidence of selected conditions, effectson Medicare payments, impacts oncoding accuracy, unintendedconsequences, and infection and eventrates. This is an intra-agency projectwith funding and technical supportcoming from CMS, OPHS, AHRQ, andCDC. The evaluation will also examinethe implementation of the program andevaluate additional conditions for future

selection.RTIs evaluation of the HACPOAprovisions is divided into several parts,only some of which were completedprior to the publication date of the FY2011 IPPS/LTCH PPS proposed rule. Inthe FY 2011 IPPS/LTCH PPS proposedrule (75 FR 23883 through 23898), wesummarized the analyses that werecompleted. RTIs analyses of POAindicator reporting, frequencies and netsavings associated with current HACs,and frequencies of previouslyconsidered candidate HACs reflect

MedPAR claims from October 2008through June 2009.

We received a number of publiccomments regarding the evaluationconducted by RTI, despite the fact thatwe did not propose any new policies orpolicy revisions based on theevaluation. Several of these publiccomments are addressed later in anothersection of this preamble, but we believethat it is appropriate to acknowledge thefollowing issues here.

Comment: Several commenters

expressed concern that the RTIevaluation did not include an analysison the costs of complying with theHACPOA provision. According to thecommenters, compliance with ourHACPOA policy results in additionalcosts to providers and individuals, aswell as to the Medicare program bynecessitating additional expensivepreadmission screening tests in order toachieve more accurate admissiondocumentation. The commenters alsostated that the estimated savings toMedicare is not accurate if providers areutilizing additional resources to perform

these expensive tests on their patients.Response: We understand theseriousness of this concern and refer toour original discussion of HACPOAissues in the FY 2009 IPPS final rule (73FR 23547 through 23559) in which weincluded a comprehensive discussion ofwhat we understood to be the fullimpact of this policy. We will continueto evaluate the financial costs ofcompliance with our HACPOAprogram, as well as its impact on ouroverall goal of providing the highestquality of care for Medicare

beneficiaries at the most reasonablecosts.

Comment: Several commenterscommended CMS for making the earlyfindings of the RTI study, as well asHACPOA data, available to the public.The commenters encouraged CMS tocontinue to make additional findingsavailable.

Response: We agree with thecommenters that it continues to beimportant to make HACPOA data andfindings available to the public prior to

proposing any significant updates to theHAC list. As RTI continues its work, wewill share the findings and additionalHACPOA data.

Comment: Several commentersexpressed interest in seeing data on themost common secondary diagnoses onthe CC and MCC list that are reportedalong with an HAC code.

Response: We have asked RTI toinclude a list of the most commonlyreported secondary CC and MCCdiagnoses and display this list alongwith the other HACPOA data on itsWeb site at: http://www.rti.org/reports/

cms.In this final rule, we are updating oursummary of the analyses withadditional data that have becomeavailable since issuance of the proposedrule.

b. RTI Analysis on POA IndicatorReporting Across Medicare Discharges

To better understand the impact ofHACs on the Medicare program, it isnecessary to first examine the incidenceof POA indicator reporting across alleligible Medicare discharges. As

VerDate Mar2010 17:02 Aug 13, 2010 Jkt 220001 PO 00000 Frm 00045 Fmt 4701 Sfmt 4700 E:\FR\FM\16AUR2.SGM 16AUR2
http://www.rti.org/reports/cmshttp://www.rti.org/reports/cmshttp://www.rti.org/reports/cmshttp://www.rti.org/reports/cmshttp://www.rti.org/reports/cms


2/17

50086 Federal Register / Vol. 75, No. 157/ Monday, August 16, 2010/ Rules and Regulations

mentioned previously, only IPPShospitals are required to submit POAindicator data for all diagnosis codes onMedicare claims. Therefore, all non-IPPS hospitals were excluded, as well asproviders in waiver States (Maryland)and territories other than Puerto Rico.

In the FY 2011 IPPS/LTCH PPSproposed rule (75 FR 23880 through

23898), we provided a preliminary

analysis on claims data from October2008 through June 2009. Sincepublication of that proposed rule, anadditional 3 months of data for FY 2009that include claims from July 2009through September 2009 have becomeavailable. Below we present thecumulative results of RTIs findings forFY 2009.

Using MedPAR claims data fromOctober 2008 through September 2009,RTI found a total of approximately 65.22million secondary diagnoses acrossapproximately 9.3 million discharges.As shown in Chart A below, themajority of all secondary diagnoses(83.69 percent) were reported with aPOA indicator ofY, meaning the

condition was POA.

c. RTI Analysis on POA IndicatorReporting of Current HACs

Following the initial analysis of POAindicator reporting for all secondarydiagnoses, RTI then evaluated POAindicator reporting for specific HAC-associated secondary diagnoses. The

term HAC-associated secondarydiagnosis refers to those diagnoses thatare on the selected HAC list and werereported as a secondary diagnosis. ChartB below shows a summary of the HACcategories with the frequency in whicheach HAC was reported as a secondarydiagnosis and the corresponding POAindicators assigned on the claims. It isimportant to note that, because morethan one HAC-associated diagnosis codecan be reported per discharge (that is,on a single claim), the frequency ofHAC-associated diagnosis codes may bemore than the actual number of

discharges that have a HAC-associateddiagnosis code reported as a secondarydiagnosis. Below we discuss thefrequency of each HAC-associateddiagnosis code and the POA indicatorsassigned to those claims.

RTI analyzed the frequency of eachreported HAC-associated secondary

diagnosis (across all 9.3 milliondischarges) and the POA indicatorassigned to the claim. Chart B belowshows that the most frequently reportedconditions were in the Falls and TraumaHAC category, with a total of 153,284HAC-associated diagnosis codes being

reported for that HAC category. Of these153,284 diagnoses, 5,684 reported aPOA indicator ofN for not POA and147,257 diagnoses reported a POAindicator ofY for POA. The lowestfrequency appears in the Surgical SiteInfection (SSI) Following BariatricSurgery for Obesity HAC category withonly 17 HAC-associated secondarydiagnosis codes (and procedure codes)reported. It is important to note that thenumber of secondary diagnosis codesclassified as POA is likely overstateddue to coding practices, and, therefore,the number of HACs not POA are

expected to be greater than indicated inCharts B and C. As a result, these datalikely underestimate the number ofcomplications some would consideracquired in the hospital or other healthcare setting. For example, the HACslisted as present on admission (POA =Y) include those instances where the

HAC condition was present onadmission from the emergency room orother outpatient settings within theadmitting institution. The POAindicator ofY is also used to identifycases where a patient was dischargedand then readmitted one calendar day or

more after the date of discharge due tocomplications from a HAC. In addition,the POA indicator ofY may alsoinclude patient transfers to the acutecare hospital from other health carefacilities, like nursing homes, or from ahome health setting, where thesecondary diagnosis considered to be aHAC was initially acquired. Usingcurrent coding guidelines, all of theabove scenarios can be correctly andappropriately classified as POA (wherePOA = Y) on an inpatient claim, andCMS does not have data from which todetermine where the condition

described in the secondary diagnisoswas acquired. Therefore, while afraction of the HACs reported as POAwere acquired outside the hospital priorto admission, some conditions couldalso have been acquired at the hospitalin an outpatient setting or through aprior admission.



3/17



In the FY 2011 IPPS/LTCH PPSproposed rule (75 FR 23885), wewelcomed public comments on the datapresented that could provide insightinto the accuracy of those data, the useof comparative data sets or analysis, andhow aspects of the coding system mightinfluence these data.

Comment: One commenter expressedits past and continuing support of the

HACPOA program. This commenterapplauded CMS efforts to evaluate thepayment and clinical impacts of theHACPOA policy and for making thepreliminary data available for publiccomment. However, the commenterreported that it found the preliminarypublished POA data for certainconditions interesting. Specifically, thecommenter noted that the POA data forthe catheter-associated urinary tract

infection (CAUTI) condition wasunexpected in that 85 percent of thecases reporting that condition as asecondary diagnosis were assigned aPOA indicator ofY (meaning that thecondition was present on admission).The commenter further noted that therewere other conditions whose POA dataanalysis results were equallyunexpected. This commenter stated itlooked forward to reviewing further



4/17


analyses and understanding how thePOA indicator is being documented andthe accuracy of the documentation.

Response: We appreciate andacknowledge the commenters supportof the HACPOA provision. As statedearlier, one aspect of the HACPOAprogram evaluation is to examine theaccuracy of coding, which includes a

review of the POA indicator data. RTIwill continue to study these data and,when they become available, we plan topublish the results.

Comment: Commenters expressedconcern about the accuracy of POAindicator reporting for the HACs relatedto intracranial injury with loss ofconsciousness. One commenter statedthat it has come to the attention of theAmerican Hospital Associations CentralOffice on ICD9CM that there have

been different interpretations of thePOA coding guidelines for the reportingof the following ICD9CM code

categories: 850 Concussions; 851 Cerebral laceration and

contusion; 852 Subarachnoid, subdural, and

extradural hemorrhage, followinginjury;

853 Other and unspecifiedintracranial hemorrhage followinginjury; and

854 Intracranial injury of other andunspecified nature.

The commenter pointed out that theabove mentioned ICD9CM codecategories require a fifth digit to specifywhether there was a loss of

consciousness, and the approximatelength of time that the patient wasunconscious. The commenter statedthat, currently, the POA guidelines stateto assign N if any part of thecombination code was not present onadmission. The commenter furtherindicated that, in some instances, codershave assigned N to these codes if thepatient lost consciousness afteradmission, even though the intracranialinjury occurred prior to admission. Thecommenter stated that loss ofconsciousness is a component ofintracranial injuries rather than a

separate condition. The commenterbelieved that this guideline has resultedin data implying that the intracranialinjuries were a result of traumasustained after admission to thehospital, when the injury occurred priorto admission.

The commenter stated that this POAguideline was discussed by the EditorialAdvisory Board for Coding Clinic forICD9CM. After review, the commenterstated that the Board determined thatthe POA guideline should be clarified sothat coders will understand that these

intracranial injury cases that have a lossof consciousness after admission should

be assigned a POA indicator ofYrather than a N. The commenter statedthat this advice will be provided in afuture issue ofCoding Clinic for ICD9CM. The commenter pointed out thatCMS collaborated in this decision.

Response: We agree that there appears

to be inconsistency in how codersinterpret and apply the official POAcoding guideline for these combinationcodes that include loss ofconsciousness. CMS participated as avoting member of the AmericanHospital Associations EditorialAdvisory Board for Coding Clinic forICD9CMto develop clarifications onthe POA reporting for combinationcodes that involve loss ofconsciousness. We agree that thisclarification will lead to greaterconsistency and accuracy in POAindicator reporting. CMS looks forward

to continuing its efforts as part of theAmerican Hospital AssociationsEditorial Advisory Board for CodingClinic for ICD9CMto provideguidance on accuracy of coding and thereporting of POA indicators. Hospitalslook to this publication to providedetailed guidance on ICD9CM codeand POA reporting. We encouragehospitals to send any other questionsabout ICD9CM codes or POAindicator selection to the AmericanHospital Association so that theEditorial Advisory Board can continueits role of providing instruction on theaccurate selection and reporting of both

ICD9CM codes and POA indicators.As described earlier, in the FY 2009

IPPS final rule (73 FR 48486 through48487), we adopted as final ourproposal to: (1) Pay the CC/MCC MSDRGs for those HACs coded with Yand W indicators; and (2) not pay theCC/MCC MSDRGs for those HACscoded with N and U indicators. Wealso discussed the comments wereceived urging CMS to stronglyconsider changing the policy and to payfor those HACs assigned a POAindicator ofU (documentation isinsufficient to determine if the

condition was present at the time ofadmission). We stated we wouldmonitor the extent to which and underwhat circumstances the U POAreporting option is used. In the FY 2010IPPS/RY 2010 LTCH PPS final rule, wealso discussed and responded tocomments regarding HACs coded withthe U indicator (74 FR 43784 and43785). As shown in Chart B above,RTIs analysis provides some data on atotal of 404 HAC-associated secondarydiagnoses reported with a POAindicator ofU. Of those diagnoses, 270

(0.2 percent) were assigned to the Fallsand Trauma HAC category.

In the FY 2011 IPPS/LTCH PPSproposed rule (75 FR 23885), we statedthat we continue to believe that betterdocumentation will result in moreaccurate public health data. Because theRTI analysis we summarized in the FY2011 IPPS/LTCH PPS proposed rule was

based on preliminary data, we did notpropose to change our policy underwhich CMS does not pay at the higherCC/MCC amount when a selected HACdiagnosis code is reported with a POAindicator ofU.

Comment: Several commenters askedCMS to change our policy under whichwe do not pay at the higher CC/MCCamount when a HAC code reported witha POA ofU. (A POA indicator ofUmeans that documentation wasinsufficient to determine if thecondition was present at the time of theinpatient admission.) The commentersstated that while hospitals arecontinuing to work on coding anddocumentation improvement issueswith physicians who practice in theirfacilities, in some cases, hospitals havenot been successful in obtaining cleardocumentation to clarify whether or nota condition was present on admission.They added that when physicians donot provide clear documentation in themedical record, a POA indicator ofUis assigned. The commenters asked thatCMS allow these cases with poordocumentation to result in a higherpayment if the HAC code is reportedwith a U.

Response: We are committed toimproving the accuracy of health caredata. Accurate and completedocumentation within the health recordis important for patient management,outcome measurement, and qualityimprovement, as well as paymentaccuracy. We believe that it would beinappropriate to pay a higher amount tohospitals based on incomplete or poordocumentation. If accurate informationis not available within the health recordfor a hospital to report a precise POAindicator, hospitals are encouraged toseek this additional documentation from

their physicians and/or other hospitalsif the hospital treated a patient who wastransferred. For these reasons, we

believe that reducing payment forconditions on the HAC list with poordocumentation is appropriate.Therefore, we did not propose to changeour approach to discounting the CC orMCC assignment for selected HACsreported with a POA indicator ofU.We will maintain our existing policyand not allow HACs with a POAindicator ofU to lead to the higherpayment.



5/17


In the FY 2011 IPPS/LTCH PPSproposed rule, we encouraged readers tofurther review the RTI detailed reportwhich demonstrates the frequency ofeach individual HAC-associateddiagnosis code within the HACcategories. For example, in the ForeignObject Retained After Surgery HACcategory, there are two unique ICD9

CM diagnosis codes to identify thatcondition: code 998.4 (Foreign bodyaccidentally left during a procedure)and code 998.7 (Acute reaction toforeign substance accidentally leftduring a procedure). In the updateddetailed RTI report, readers can viewthat code 998.4 was reported 428 timesand code 998.7 was reported 13 times,for a total of 441 times, as shown inChart B above. The RTI detailed reportis available at the following Web site:http://www.rti.org/reports/cms/.

d. RTI Analysis of Frequency ofDischarges and POA Indicator Reporting

for Current HACsRTI further analyzed the effect of the

HAC provision by studying thefrequency in which a HAC-associateddiagnosis was reported as a secondarydiagnosis with a POA indicator ofNor U and, of that number, how manyresulted in MSDRG reassignment. InChart C below, Column A shows thenumber of discharges for each HACcategory where the HAC-associateddiagnosis was reported as a secondarydiagnosis. For example, there were 33discharges that reported Air Embolismas a secondary diagnosis. Column C

shows the number of discharges for eachHAC reported with a POA indicator ofN or U. Continuing with the exampleof Air Embolism, the chart shows that,of the 33 reported discharges, 24discharges (72.73 percent) had a POAindicator ofN or U and wasidentified as a HAC discharge. Therewere a total of 24 discharges to whichthe HAC policy applies and that could,therefore, have had an MSDRGreassignment. Column E shows thenumber of discharges where an actualMSDRG reassignment occurred. Asshown in Column E, the number of

discharges with an Air Embolism thatresulted in actual MSDRGreassignments is 12 (50 percent of the 24discharges with a POA indicator ofNor U). Thus, while there were 24discharges (72.73 percent of the original33) with an Air Embolism reported witha POA indicator ofN or U identifiedas a HAC discharge that could havecaused MSDRG reassignment, the endresult was 12 (50 percent) actual MSDRG reassignments. There are a numberof reasons why a selected HAC reportedwith a POA indicator ofN or U will

not result in MSDRG reassignment.These reasons were illustrated with thediagram in section II.F.1.c. of thispreamble and will be discussed infurther detail in section II.F.3.e. of thispreamble.

Chart C below also shows that, of the264,810 discharges with a HAC-associated diagnosis as a secondary

diagnosis, 3,416 discharges ultimatelyresulted in MSDRG reassignment. Aswe discuss below, there were 15 claimsthat resulted in MSDRG reassignmentwhere two HACs were reported on thesame admission. The four HACcategories that had the most dischargesresulting in MSDRG reassignmentwere: (1) Falls and Trauma; (2)Pulmonary Embolism and DVTOrthopedic (Orthopedic PE/DVT); (3)Pressure Ulcer Stages III & IV; and (4)Catheter-Associated Urinary TractInfection (UTI). Codes falling under theFalls and Trauma HAC category were

the most frequently reported secondarydiagnoses with 126,078 discharges. Ofthese 126,078 discharges, 5,312 (4.21percent) were coded as not POA andidentified as HAC discharges. Thiscategory also contained the greatestnumber of discharges that resulted in anMSDRG reassignment. Of the 5,312discharges within this HAC categorythat were not POA, 1,577 (29.69percent) resulted in an MSDRGreassignment.

Of the 264,810 total dischargesreporting HAC-associated diagnoses as asecondary diagnosis, 3,110 dischargeswere coded with a secondary diagnosis

of Orthopedic PE/DVT. Of these 3,110discharges, 2,335 (75.08 percent) werecoded as not POA and identified asHAC discharges. This categorycontained the second greatest number ofdischarges resulting in an MSDRGreassignment. Of the 2,335 discharges inthis HAC category that were not POA,1,024 discharges (43.85 percent)resulted in an MSDRG reassignment.

The Pressure Ulcer Stages III & IVcategory had the second most frequentlycoded secondary diagnoses, with 99,656discharges. Of these discharges, 1,316(1.32 percent) were coded as not POA

and identified as HAC discharges. Thiscategory contained the third greatestnumber of discharges resulting in anMSDRG reassignment. Of the 1,316discharges in this HAC category thatwere not POA, 384 discharges (29.18percent) resulted in an MSDRGreassignment.

The Catheter-Associated UTI categoryhad the third most frequently codedsecondary diagnoses, with 14,089discharges. Of these discharges, 2,333(16.56 percent) were coded as not POAand identified as HAC discharges. This

category contained the fourth greatestnumber of discharges resulting in anMSDRG reassignment. Of the 2,333discharges in this HAC category thatwere not POA, 223 discharges (9.56percent) resulted in a MSDRGreassignment.

The remaining 6 HAC categories onlyhad 208 discharges that ultimately

resulted in MSDRG reassignment. Wenote that, even in cases where a largenumber of HAC-associated secondarydiagnoses were coded as not POA, thisfinding did not necessarily translate intoa large number of discharges thatresulted in MSDRG reassignment. Forexample, only 26 of the 2,573 VascularCatheter-Associated Infection secondarydiagnoses that were coded as not POAand identified as HAC dischargesresulted in a MSDRG reassignment.

There were a total of 417 dischargeswith a HAC-associated secondarydiagnosis reporting a POA indicator ofN or U that were excluded fromacting as a HAC discharge (subject toMSDRG reassignment) due to the CCExclusion List logic within theGROUPER. The CC Exclusion Listidentifies secondary diagnosis codesdesignated as a CC or MCC that aredisregarded by the GROUPER logicwhen reported with certain principaldiagnoses. For example, a claim withthe principal diagnosis code of 250.83(Diabetes with other specifiedmanifestations, type 1 [juvenile type],uncontrolled) and a secondary diagnosiscode of 250.13 (Diabetes with

ketoacidosis, type 1, [juvenile type],uncontrolled) with a POA indicator ofN would result in the HAC-associatedsecondary diagnosis code 250.13 beingignored as a CC. According to the CCExclusion List, code 250.13 is excludedfrom acting as a CC when code 250.83is the principal diagnosis. As a result,the HAC logic would not be applicableto that case. For a detailed discussion onthe CC Exclusion List, we refer readersto section II.G.9. of this preamble.

Discharges where the HAC logic wasnot applicable due to the CC ExclusionList occurred among the following 4HAC categories: Pressure Ulcer StagesIII and IV (44 cases), Falls and Trauma(311 cases), Catheter-Associated UTI (9cases), Vascular Catheter-AssociatedInfection (4 cases), and Manifestationsof Poor Glycemic Control (49 cases).Further information regarding thespecific number of cases that wereexcluded for each HAC-associatedsecondary diagnosis code within each ofthe above mentioned HAC categories isalso available. We refer readers to theRTI detailed report at the following Website: http://www.rti.org/reports/cms/.

http://www.rti.org/reports/cms/http://www.rti.org/reports/cms/http://www.rti.org/reports/cms/http://www.rti.org/reports/cms/http://www.rti.org/reports/cms/


6/17


In summary, Chart C belowdemonstrates that there were a total of264,810 discharges with a reportedHAC-associated secondary diagnosis. Ofthe total 264,810 discharges, 14,681

(5.68 percent) discharges includedHACs that were reported with a POAindicator ofN or U and wereidentified as a HAC discharge. Of these14,681 discharges, the number of

discharges resulting in MSDRGreassignments was 3,416 (22.72percent).BILLING CODE 412001P



7/17



An extremely small number ofdischarges had multiple HACs reportedduring the same stay. In reviewing the9.3 million claims, RTI found 60 casesin which two HACs were reported onthe same discharge. Chart D belowsummarizes these cases. There were 9cases in which a Falls and Trauma HACwas reported in addition to a PressureUlcer Stages III & IV HAC. Twenty of thecases with two HACs involved Pressure

Ulcer Stages III & IV and 24 casesinvolved Falls or Trauma. Othermultiple HAC cases included 10Catheter-Associated UTI cases and 6Vascular Catheter-Associated Infectioncases.

Some of these cases with multipleHACs reported had both HAC codesignored in the MSDRG assignment. Ofthese 60 claims, 15 did not receivehigher payments based on the presenceof one or both of these reported HACs

and we describe these claims below insection II.F.3.g.(2) of this preamble.Depending on the MSDRG to whichthe cases were originally assigned,ignoring the HAC codes would have ledto a MSDRG reassignment if there wereno other MCCs or CCs reported, if theMSDRG was subdivided into severitylevels, and if the case were not already

in the lowest severity level prior toignoring the HAC codes.



8/17


e. RTI Analysis of Circumstances WhenApplication of HAC Provisions WouldNot Result in MSDRG Reassignmentfor Current HACs

As discussed in section II.F.1. andillustrated in the diagram in sectionII.F.1.c. of this preamble, there areinstances when the MSDRGassignment does not change even whena HAC-associated secondary diagnosishas a POA indicator of either N or U.In analyzing our claims data, RTIidentified four main reasons why a MSDRG assignment would not changedespite the presence of a HAC. Thosefour reasons are described below andare shown in Chart E below. Column A

shows the frequency of discharges thatincluded a HAC-associated secondarydiagnosis. Column B shows thefrequency of discharges where the HAC-associated secondary diagnosis wascoded as not POA and identified as aHAC discharge. Column C shows thefrequency of discharges in which theHAC-associated secondary diagnosiscoded as not POA resulted in a changein MSDRG. Columns D, E, F, and Gshow the frequency of discharges inwhich the HAC-associated secondarydiagnosis coded as not POA did not

result in a change in MSDRGassignment. Columns D, E, F, and G areexplained in more detail below.

(1) Other MCCs/CCs PreventReassignment

Column D (Other MCC/CCs thatPrevent Reassignment) in Chart E belowindicates the number of cases reportinga HAC-associated secondary diagnosiscode that did not have a MSDRGreassignment because of the presence ofother secondary diagnoses on the MCCor CC list. A claim that is coded witha HAC-associated secondary diagnosesand a POA status of either N or Umay have other secondary diagnosesthat are classified as an MCC or a CC.In such cases, the presence of theseother MCC and CC diagnoses will stilllead to the assignment of a higherseverity level, despite the fact that theGROUPER software is disregarding theICD9CM code that identifies theselected HAC in making the MSDRGassignment for that claim. For example,there were 96 cases in which the ICD9CM codes for the Foreign ObjectRetained After Surgery HAC categorywere present, but the presence of othersecondary diagnoses that were MCCs or

CCs resulted in no change to the MSDRG assignment. Chart E shows that atotal of 8,208 cases did not have a

change in the MSDRG assignmentbecause of the presence of otherreported MCCs and CCs.

(2) Two Severity Levels Where HACDoes Not Impact MSDRG Assignment

Column E (Number of MSDRGs withTwo Severity Levels Where HAC DoesNot Impact MSDRG Assignment)shows the frequency with whichdischarges with a HAC as a secondarydiagnosis coded as not POA did notresult in an MSDRG change becausethe MSDRG is subdivided solely by thepresence or absence of an MCC. A claim

with a HAC and a POA indicator ofeither N or U may be assigned to anMSDRG that is subdivided solely bythe presence or absence of an MCC. Insuch cases, removing a HAC ICD9CMCC code will not lead to further changesin the MSDRG assignment. Examplesof these MSDRG subdivisions areshown in the footnotes to the chart andinclude the following examples:

MSDRGs 100 and 101 (Seizureswith or without MCC, respectively)

MSDRGs 102 and 103 (Headacheswith or without MCC, respectively)



9/17


The codes that fall under the HACcategory of Foreign Object RetainedAfter Surgery are CCs. If this case wereassigned to a MSDRG with an MCCsubdivision such as MSDRGs 100 and101, the presence of the HAC codewould not affect the MSDRG severitylevel assignment. In other words, if theForeign Object Retained After Surgery

code was the only secondary diagnosisreported, the case would be assigned toMSDRG 101. If the POA indicator wasN, the HAC Foreign Object RetainedAfter Surgery code would be ignored inthe MSDRG assignment logic. Despitethe fact that the code was ignored, thecase would still be assigned to the same,lower severity level MSDRG.Therefore, there would be no impact onthe MSDRG assignment.

Column E in Chart E below showsthat there were a total of 1,793 caseswhere the HAC code was N or U andthe MSDRG assignment did not change

because the case was already assigned tothe lowest severity level.

(3) No Severity Levels

Column F (Number of MSDRGs withNo Severity Levels) shows the frequencywith which discharges with an HAC asa secondary diagnosis coded as not POAdid not result in an MSDRG change

because the MSDRG is not subdividedby severity levels. A claim with a HACand a POA ofN or U may beassigned to a MSDRG with no severitylevels. For instance, MSDRG 311(Angina Pectoris) has no severity levelsubdivisions; this MSDRG is not split

based on the presence of an MCC or aCC. If a patient assigned to this MSDRG develops a secondary diagnosissuch as a Stage III pressure ulcer afteradmission, the condition would beconsidered to be a HAC. The code forthe Stage III pressure ulcer would beignored in the MSDRG assignment

because the condition developed afterthe admission (the POA indicator wasN). Despite the fact that the ICD9CM

code for the HAC Stage III pressureulcer was ignored, the MSDRGassignment would not change. The casewould still be assigned to MSDRG 311.Chart E below shows that 1,255 casesreporting a HAC-associated secondarydiagnosis did not undergo a change inthe MSDRG assignment based on thefact that the case was assigned to a MS

DRG that had no severity subdivisions(that is, the MSDRG is not subdivided

based on the presence or absence of anMCC or a CC, rendering the presence ofthe HAC irrelevant for paymentpurposes).

(4) MSDRG Logic

Column G (MSDRG Logic Issues)shows the frequency with which a HACas a secondary diagnosis coded as notPOA did not result in an MSDRGchange because of MSDRG assignmentlogic. There were nine discharges wherethe HAC criteria were met and the HAClogic was applied, however, due to thestructure of the MSDRG logic, thesecases did not result in MSDRGreassignment. These cases may appearsimilar to those discharges where theMSDRG is subdivided into twoseverity levels by the presence orabsence of an MCC and did not resultin MSDRG reassignment; however,these discharges differ slightly in thatthe MSDRG logic also considersspecific procedures that were reportedon the claim. In other words, for certainMSDRGs, a procedure may beconsidered the equivalent of an MCC orCC. The presence of the procedure code

dictates the MSDRG assignmentdespite the presence of the HAC-associated secondary diagnosis codewith a POA indicator ofN or U.

For example, a claim with theprincipal diagnosis code of 441.1(Thoracic aneurysm, ruptured) withHAC-associated secondary diagnosiscode of 996.64 (Infection andinflammatory reaction due to indwellingurinary catheter) and diagnosis code

599.0 (Urinary tract infection, site notspecified), having POA indicators ofY, N, N, respectively, andprocedure code 39.73 (Endovascularimplantation of graft in thoracic aorta),results in an assignment to MSDRG237 (Major Cardiovascular Procedureswith MCC or Thoracic Aortic AneurysmRepair). In this case, the thoracic aortic

aneurysm repair is what dictated theMSDRG assignment and the presenceof the HAC-associated secondarydiagnosis code, 996.64, did not affectthe MSDRG assigned. Other examplesof MSDRGs that are subdivided in thissame manner are as follows:

MSDRG 029 (Spinal procedureswith CC or Spinal Neurostimulators)

MSDRG 129 (Major Head & NeckProcedures with CC/MCC or MajorDevice)

MSDRG 246 (PercutaneousCardiovascular Procedure with Drug-

Eluting Stent with MCC or 4+ Vessels/Stents)

Column G in the chart below showsthat four of the nine cases that did notresult in MSDRG reassignment due tothe MSDRG logic were in the CatheterAssociated UTI HAC category, threecases were in the Falls and Trauma HACcategory, one case was in the ForeignBody Retained After surgery HACcategory, and one case was in theVascular Catheter-Associated InfectionHAC Category.

In conclusion, a total of 11,265 cases(8,208 + 1,793 + 1,255 + 9) did not have

a change in MSDRG assignment,regardless of the presence of a HAC. Thereasons described above explain whyonly 3,416 cases had a change in MSDRG assignment despite the fact thatthere were 14,681 HAC cases with aPOA ofN or U. We refer readers tothe RTI detailed report at the Web site:http://www.rti.org/reports/cmsforfurther information.BILLING CODE 412001P

http://www.rti.org/reports/cmshttp://www.rti.org/reports/cms


10/17




11/17



f. RTI Analysis of Coding Changes forHAC-Associated Secondary Diagnosesfor Current HACs

In the FY 2011 IPPS/LTCH PPSproposed rule (75 FR 23892), wediscussed RTIs preliminary analysis oncoding changes using 9 months ofclaims data from October 2008 through

June 2009. We noted that, in addition tostudying claims from October 2008through June 2009, RTI evaluated claimsdata from 2 years prior to determine ifthere were significant changes in thenumber of discharges with a HAC beingreported as a secondary diagnosis. Forthis FY 2011 IPPS/LTCH PPS final rule

analysis, RTI examined an additional 3months of claims data for each fiscalyear (FY 2007 and FY 2008), andcompared these data to the updated FY2009 data. Below we summarize theresults of the fiscal year to fiscal yearcomparison using 12 months of claimsdata.

RTIs analysis found that there was anoverall increase in the reporting ofsecondary diagnoses that are currentlydesignated as HACs from FY 2007 to FY2008. The most significant increase wasin the Catheter-Associated UTI HACcategory, with 12,459 discharges beingreported in FY 2007, while 15,408discharges were reported in FY 2008, anincrease of 2,949 cases. The next

significant increase was in the Falls andTrauma HAC category with 151,321discharges being reported in FY 2007,while 153,600 discharges were reportedin FY 2008, an increase of 2,279 cases.

However, the analysis also found thatthere was an overwhelming decrease inthe HAC-associated secondarydiagnoses reported from FY 2008 to FY2009. The most significant decrease wasin the Falls and Trauma HAC category,with 153,600 discharges being reportedin FY 2008, while 125,505 dischargeswere reported in FY 2009, a decrease of28,095 cases. We point out that becausediagnosis codes for the Pressure UlcerStages III & IV HAC did not becomeeffective until October 1, 2008, there are



12/17


no data available for FY 2007 or FY2008.

We refer readers to the RTI detailedreport for all the conditions in eachfiscal year (FY 2007 through FY 2009)as described above at the following Website: http://www.rti.org/reports/cms/.

g. RTI Analysis of Estimated Net

Savings for Current HACsRTI estimated the net savings

generated by the HAC payment policybased on 12 months of MedPAR claimsfrom October 2008 through September2009.

(1) Net Savings Estimation Methodology

The payment impact of a HAC is thedifference between the IPPS paymentamount under the initially assignedMSDRG and the amount under thereassigned MSDRG. The amount forthe reassigned MSDRG appears on theMedPAR files. To construct this, RTImodeled the IPPS payments for eachMSDRG following the same approachthat we use to model the impact of IPPSannual rule changes. Specifically, RTIreplicated the payment computationscarried out in the IPPS PRICER programusing payment factors for IPPSproviders as identified in various CMSdownloaded files. The files used are asfollows:

Version 26 of the Medicare SeverityGROUPER software (applicable todischarges between October 1, 2008 andSeptember 30, 2009). IPPS MedPARclaims were run through this file toobtain needed HACPOA output

variables. The FY 2009 MSDRG payment

weight file. This file includes theweights, geometric mean length of stay(GLOS), and the postacute transferpayment indicators.

CMS standardized operating andcapital rates. Tables 1A through 1C, asdownloaded from the Web site at:http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009,includethe full update and reduced updateamounts, as well as the informationneeded to compute the blended amountfor providers located in Puerto Rico.

The IPPS impact files for FY 2009,also as downloaded from the Web siteat: http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/.This fileincludes the wage index and geographicadjustment factors, plus the providertype variable to identify providersqualifying for alternative hospital-specific amounts and their respectiveHSP rates.

The IPPS impact files for FY 2010,as downloaded from the Web site at:http://www.cms.hhs.gov/AcuteInpatientPPS/10FR/.This file includes indirect

medical education (IME) anddisproportionate share (DSH) percentadjustments that were in effect as ofMarch 2009.

CMS historical provider-specificfiles (PSF). This includes the indicatorto identify providers subject to the fullor reduced standardized rates and theapplicable operating and capital cost-to-

charge ratios. A SAS version wasdownloaded from the Web site at:http://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asp.

There were 50 providers withdischarges in the final HAC analysis filethat did not appear in the FY 2009impact file, of which 11 also did notappear in the FY 2010 impact file. Forthese providers, we identified thegeographic CBSA from the historicalPSF and assigned the wage index usingvalues from Tables 4A and 4C asdownloaded from the Web site at:http://www.cms.hhs.gov/

AcuteInpatientPPS/IPPS2009/.Forproviders in the FY 2010 file but not theFY 2009 file, we used IME and DSHrates from FY 2010. The 11 providers inneither impact file were identified asnon-IME and non-DSH providers in thehistorical PSF file.

The steps for estimating the HACpayment impact are as follows:

Step 1: Rerun the Medicare SeverityGrouper on all records in the analysisfile. This is needed to obtaininformation on actual HAC-related MSDRG reassignments in the file, and toidentify the CCs and MCCs that

contribute to each MSDRG assignment.Step 2: Model the base payment andoutlier amounts associated with theinitial MSDRG if the HAC wereexcluded using the computations laidout in the CMS file Outlier Example FY2007 new.xls, as downloaded from theWeb site at: http://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPage,and modified toaccommodate FY 2009 factors.

Step 3: Model the base payment andoutlier amounts associated with thefinal MSDRG where the HAC wasexcluded using the computations laidout in the CMS file Outlier Example FY2007 new.xls, as downloaded from theWeb site at: http://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPageand modified toaccommodate FY 2009 factors.

Step 4: Compute MSDRG basesavings as the difference between thenonoutlier payments for the initial andfinal MSDRGs. Compute outlieramounts as the difference in outlieramounts due under the initial and finalreassigned MSDRG. Compute netsavings due to HAC reassignment as the

sum of base savings plus outlieramounts.

Step 5: Adjust the model toincorporate short-stay transfer paymentadjustments.

Step 6: Adjust the model toincorporate hospital-specific paymentsfor qualifying rural providers receivingthe hospital-specific payment rates.

It is important to mention that usingthe methods described above, the MSDRG and outlier payments amounts thatare modeled for the final assigned MSDRG do not always match the DRG priceand outlier amounts that appear in theMedPAR record. There are severalreasons for this. Some discrepancies arecaused by using single wage index, IMEand DSH factors for the full periodcovered by the discharges, when inpractice these payment factors can beadjusted for individual providers duringthe course of the fiscal year. In addition,RTIs approach disregards any Part Acoinsurance amounts owed byindividual beneficiaries with greaterthan sixty covered days in a spell ofillness. Five percent of all HACdischarges showed at least some Part Acoinsurance amount due from the

beneficiary, although less than twopercent of reassigned discharges (55cases in the analysis file) showed PartA coinsurance amounts due. Any Part Acoinsurance payments would reduce theactual savings incurred by the Medicareprogram.

There are also a number of lesscommon special IPPS paymentsituations that are not factored into

RTIs modeling. These could includenew technology add-on payments,payments for blood clotting factors,reductions for replacement medicaldevices, adjustments to the capital ratefor new providers, and adjustments tothe capital rate for certain classes ofproviders who are subject to a minimumpayment level relative to capital cost.

(2) Net Savings Estimate

Chart F below summarizes theestimated net savings of current HACs

based on MedPAR claims from October2008 through September 2009, based onthe methodology described above.Column A shows the number ofdischarges where a MSDRGreassignment for each HAC categoryoccurred. For example, there were 12discharges with an Air Embolism thatresulted in an actual MSDRGreassignment. Column B shows the totalnet savings caused by MSDRGreassignments for each HAC category.Continuing with the example of AirEmbolism, the chart shows that the 12discharges with an MSDRGreassignment resulted in a total net

http://www.rti.org/reports/cms/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/10FR/http://www.cms.hhs.gov/AcuteInpatientPPS/10FR/http://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.rti.org/reports/cms/http://www.cms.hhs.gov/AcuteInpatientPPS/10FR/http://www.cms.hhs.gov/AcuteInpatientPPS/10FR/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/AcuteInpatientPPS/IPPS2009/http://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/ProspMedicareFeeSvcPmtGen/04_psf_SAS.asphttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPagehttp://www.cms.hhs.gov/AcuteInpatientPPS/04_outlier.asp#TopOfPage


13/17


savings of $148,394. Column C showsthe net savings per discharge for eachHAC category. For the Air Embolism

HAC category, the net savings perdischarge is $12,366.BILLING CODE 412001P



14/17



As shown in Chart F above, the totalnet savings calculated for the 12-monthperiod from October 2008 throughSeptember 2009 was roughly $18.78million. The three HACs with the largestnumber of discharges resulting in MSDRG reassignment, Falls and Trauma,Orthopedic PE/DVT, and Pressure UlcerStages III & IV, generated $17.17 millionof net savings for the 12 month period.Estimated net savings for the 12-monthperiod associated with the Falls andTrauma category were $8.09 million.Estimated net savings associated withOrthopedic PE/DVT for the 12-monthperiod were $6.92 million. Estimatednet savings for the 12-month periodassociated with Pressure Ulcer Stages III& IV were $2.16 million.

The mean net savings per dischargecalculated for the 12-month period fromOctober 2008 through September 2009was roughly $5,522. The HAC categoriesof Air Embolism; SSI, Mediastinitis,Following Coronary Artery Bypass Graft(CABG); and SSI Following CertainOrthopedic Procedures had the highestnet savings per discharge, butrepresented a small proportion of totalnet savings because the number ofdischarges that resulted in MSDRGreassignment for these HACs was low.With the exception of BloodIncompatibility, where no savingsoccurred because no discharges resultedin MSDRG reassignment, SSIFollowing Bariatric Surgery for Obesityand Catheter-Associated UTI had thelowest net savings per discharge.

We refer readers to the RTI detailedreport available at the following Website: http://www.rti.org/reports/cms/.

As mentioned previously, anextremely small number of cases in the12-month period of FY 2009 analyzed

by RTI had multiple HACs during thesame stay. In reviewing our 9.3 millionclaims, RTI found 60 cases where two

HACs were reported on the sameadmission as noted in section II.F.3. d.of this preamble. Of these 60 claims, 15resulted in MSDRG reassignment.Chart G below summarizes these cases.There were 15 cases that had two HACsnot POA that resulted in an MSDRGreassignment. Of these, 5 dischargesinvolved Pressure Ulcer Stages III & IVand Falls and Trauma and 4 dischargesinvolved Orthopedic PE/DVT and Fallsand Trauma.

As we discuss in section II.F.1.b. ofthis preamble, implementation of thispolicy is part of an array of MedicareVBP tools that we are using to promoteincreased quality and efficiency of care.We again point out that a decrease overtime in the number of discharges wherethese conditions are not POA is adesired consequence. We recognize thatestimated net savings should likelydecline as the number of suchdischarges decline. However, we believethat the sentinel effect resulting fromCMS identifying these conditions iscritical. (We refer readers to sectionIV.A. of this preamble for a discussionof the inclusion of the incidence ofthese conditions in the RHQDAPUprogram.) It is our intention to continueto monitor trends associated with thefrequency of these HACs and theestimated net payment impact through

RTIs program evaluation and possiblybeyond.

h. Previously Considered CandidateHACsRTI Analysis of Frequency ofDischarges and POA Indicator Reporting

RTI evaluated the frequency ofconditions previously considered, but

not adopted as HACs in priorrulemaking, that were reported assecondary diagnoses (across all 9.3million discharges) as well as the POAindicator assignments for theseconditions. Chart H below indicates thatthe three previously consideredcandidate conditions most frequentlyreported as a secondary diagnosis were:(1) Clostridium Difficile-AssociatedDisease (CDAD), which demonstratedthe highest frequency, with a total of85,096 secondary diagnoses codes beingreported for that condition, of which

28,844 reported a POA indicator ofN;(2) Staphylococcus aureus Septicemia,with a total of 22,433 secondarydiagnoses codes being reported for thatcondition, with 5,004 of those reportinga POA indicator ofN; and (3)Iatrogenic Pneumothorax, with a total of20,673 secondary diagnoses codes beingreported for that condition, with 17,602of those reporting a POA indicator ofN. As these three conditions had themost significant impact for reporting aPOA indicator ofN, it is reasonable to

believe that these same three conditionswould have the greatest number ofpotential MSDRG reassignments. Thefrequency ofdischarges for thepreviously considered HACs that couldlead to potential changes in MSDRGassignment is discussed in the nextsection. We take this opportunity toremind readers that because more than

http://www.rti.org/reports/cms/http://www.rti.org/reports/cms/


15/17


one previously considered HACdiagnosis code can be reported perdischarge (on a single claim) that the

frequency of these diagnosis codes maybe more than the actual number ofdischarges with a previously considered

candidate condition reported as asecondary diagnosis.

In Chart I below, Column A shows thenumber of discharges for eachpreviously considered candidate HACcategory when the condition wasreported as a secondary diagnosis. Forexample, there were 85,096 dischargesthat reported CDAD as a secondarydiagnosis. Previously consideredcandidate HACs reported with a POAindicator ofN or U may cause MSDRG reassignment (which would resultin reduced payment to the facility).Column C shows the discharges for eachpreviously considered candidate HAC

reported with a POA indicator ofN

or

U. Continuing with the example ofCDAD, Chart I shows that, of the 85,096

discharges, 29,296 discharges (34.43percent) had a POA indicator ofN orU. Therefore, there were a total of29,296 discharges that could potentiallyhave had an MSDRG reassignment.Column E shows the number ofdischarges where an actual MSDRGreassignment could have occurred; thenumber of discharges with CDAD thatcould have resulted in actual MSDRGreassignments is 896 (3.06 percent).Thus, while there were 29,296discharges with CDAD reported with aPOA indicator ofN or U that could

potentially have had an MSDRGreassignment, the result was 896 (3.06percent) potential MSDRG

reassignments. As discussed above,there are a number of reasons why acondition reported with a POA indicatorofN or U would not result in a MSDRG reassignment.

In summary, Chart I belowdemonstrates there were a total of203,844 discharges with a previouslyconsidered candidate HAC reported as asecondary diagnosis. Of those, 57,902discharges were reported with a POAindicator ofN or U. The totalnumber of discharges that could have

resulted in MSDRG reassignments is3,527.



16/17


i. Current and Previously ConsideredCandidate HACsRTI Report onEvidence-Based Guidelines

The RTI program evaluation includesan updated report that providesreferences for all evidence-basedguidelines available for each of theselected and previously consideredcandidate HACs that providerecommendations for the prevention of

the corresponding conditions.Guidelines were primarily identifiedusing the AHRQ National GuidelinesClearing House (NGCH) and the CDC,along with relevant professionalsocieties. Guidelines published in theUnited States were used, if available. Inthe absence of U.S. guidelines for aspecific condition, internationalguidelines were included.

Evidence-based guidelines thatincluded specific recommendations forthe prevention of the condition wereidentified for each of the 10 selected

conditions. In addition, evidence-basedguidelines were also found for thepreviously considered candidateconditions.

Comment: Several commenters statedthat CMS should not pay for HACs onlywhen evidence-based guidelinesindicate that the occurrence of an eventcan be reduced to zero, or near zero. Thecommenters stated that some patients,particularly high-risk, co-morbidindividuals, may still developconditions on the HAC list even thoughprotocols have been strictly followed.

Response: We thank the commentersfor this comment. The statute requiresthat CMS only choose conditions to beselected HACs if they couldreasonablybe prevented through theapplication of evidence-basedguidelines. We noted in the FY 2008IPPS final rule that we only selectedthose conditions where, if hospitalpersonnel are engaging in good medical

practice, the additional costs of thehospital-acquired condition will, inmost cases, be avoided (72 FR 47201).

RTI prepared a final report tosummarize its findings regardingevidence-based guidelines, which can

be found on the Web site at: http://www.rti.org/reports/cms.

j. Final Policy Regarding Current HACsand Previously Considered Candidate

HACsWe believe that the updated RTI

analysis summarized above does notprovide additional information thatwould require us to change our previousdeterminations regarding either currentHACs (as described in section II.F.2. ofthis preamble) or previously consideredcandidate HACs in the FY 2008 IPPSfinal rule with comment period and FY2009 IPPS final rule (72 FR 47200through 47218 and 73 FR 48471 through48491, respectively). Accordingly, in the

http://www.rti.org/reports/cmshttp://www.rti.org/reports/cmshttp://www.rti.org/reports/cmshttp://www.rti.org/reports/cms


17/17


FY 2011 IPPS/LTCH PPS proposed rule,we did not propose to add or removecategories of HACs, although weproposed to revise the BloodIncompatibility HAC category asdiscussed and finalized in section II.F.2.of this preamble. We also note that insection II.F.3.b. of this preamble, wediscuss our current policy regarding the

treatment of the U POA indicator.However, we continue to encouragepublic dialogue about refinements to theHAC list.

We refer readers to section II.F.6. ofthe FY 2008 IPPS final rule withcomment period (72 FR 47202 through47218) and to section II.F.7. of the FY2009 IPPS final rule (73 FR 48474through 48491) for detailed discussionsupporting our determination regardingeach of these conditions.

G. Changes to Specific MSDRGClassifications

In the FY 2011 IPPS/LTCH PPSproposed rule (75 FR 23898 through23910), we invited public comment oneach of the MSDRG classificationproposed changes described below, aswell as our proposals to maintaincertain existing MSDRG classifications,which are also discussed below. In somecases, we proposed changes to the MSDRG classifications based on ouranalysis of claims data. In other cases,we proposed to maintain the existingMSDRG classification based on ouranalysis of claims data. Below, we alsosummarize the public comments that wereceived, if any, on our proposals,present our responses to thesecomments, and state our final policies.

1. Pre-Major Diagnostic Categories(MDCs)

a. Postsurgical Hypoinsulinemia (MSDRG 008 (Simultaneous Pancreas/Kidney Transplant))

Diabetes mellitus is a pancreaticdisorder in which the pancreas fails toproduce sufficient insulin, or in whichthe body cannot process insulin. Manypatients with diabetes will eventuallyexperience complications of the disease,

including poor kidney function. Whenthese patients show signs of advancedkidney disease, they are usually referredfor transplant evaluation. Currently,many doctors recommend thatindividuals with diabetes beingevaluated for kidney transplantationalso be considered for pancreastransplantation. A successful pancreastransplant may prevent, stop, or reversethe complications of diabetes.

Occasionally, secondary diabetes maybe surgically induced following apancreas transplant. This condition

would be identified by using ICD9CMdiagnosis code 251.3 (Postsurgicalhypoinsulinemia). However, currentlythe list of principal diagnosis codesassigned to surgical MSDRG 008(Simultaneous Pancreas/KidneyTransplant) does not include diagnosiscode 251.3. Therefore, when diagnosiscode 251.3 is assigned to a case as a

principal diagnosis, the case is notassigned to MSDRG 008. Instead, thesecases are grouped to MSDRG 652(Kidney Transplant) under MDC 11(Diseases and Disorders of the Kidneyand Urinary Tract). The use of diagnosiscode 251.3 as a principal diagnosiswithout a secondary diagnosis ofdiabetes mellitus and with a procedurecode for pancreas transplant only duringthat admission results in assignment ofthe case to MSDRG 628, 629, or 630(Other Endocrine, Nutritional &Metabolic Operating Room Procedureswith MCC, with CC, and without CC/

MCC, respectively). These MSDRGs areassigned to MDC 10 (Endocrine,Nutritional and Metabolic Diseases andDisorders).

As we stated in the FY 2011 IPPS/LTCH PPS proposed rule (75 FR 23898),we believe that the exclusion ofdiagnosis code 251.3 from the list ofprincipal diagnosis codes assigned tosurgical MSDRG 008 is an error ofomission. Therefore, in that proposedrule, we proposed to add diagnosis code251.3 to the list of principal orsecondary diagnosis codes assigned toMSDRG 008. As a conforming change,we also proposed to add diagnosis code

251.3 to the list of principal orsecondary diagnosis codes assigned toMSDRG 010 (Pancreas Transplant).

Comment: Commenters concurredwith CMS proposal to add diagnosiscode 251.3 to the list of principal orsecondary diagnosis codes assigned toMSDRG 008. In addition, thecommenters concurred with theproposal to add diagnosis code 251.3 tothe list of principal or secondarydiagnosis codes assigned to MSDRG010.

Response: We appreciate the supportfor our proposals.

We are adopting as final withoutmodification our proposals to adddiagnosis code 251.3 to the list ofacceptable principal diagnoses in MSDRG 008 and, as a conforming change,to add diagnosis code 251.3 to the listof acceptable principal or secondarydiagnoses in MSDRG 010.

b. Bone Marrow Transplants

As we discussed in the FY 2011 IPPS/LTCH PPS proposed rule (75 FR 23898),we received two requests to reviewwhether cost differences between an

autologous bone marrow transplant(where the patients own bone marrowor stem cells are used) and an allogeneic

bone marrow transplant (where bonemarrow or stem cells come from eithera related or unrelated donor) necessitatethe creation of separate MSDRGs tomore appropriately account for theclinical nature of the services being

rendered as well as the costs. One of therequestors stated that there are dramaticdifferences in the costs between the twotypes of transplants where allogeneiccases are significantly more costly.

Bone marrow transplantation andperipheral blood stem celltransplantation are used in thetreatment of certain cancers and bonemarrow diseases. These proceduresrestore stem cells that have beendestroyed by high doses ofchemotherapy and/or radiationtreatment. Currently, all bone marrowtransplants are assigned to MSDRG 009

(Bone Marrow Transplant).For the FY 2011 IPPS/LTCH PPSproposed rule, we performed ananalysis of the FY 2009 MedPAR dataand found 1,664 total cases assigned toMSDRG 009 with average costs ofapproximately $43,877 and an averagelength of stay of approximately 21 days.Of these MSDRG 009 cases, 395 ofthem were allogeneic bone marrowtransplant cases reported with one ofthe following ICD9CM procedurecodes: 41.02 (Allogeneic bone marrowtransplant with purging); 41.03(Allogeneic bone marrow transplantwithout purging); 41.05 (Allogeneic

hematopoietic stem cell transplantwithout purging); 41.06 (Cord bloodstem cell transplant); or 41.08(Allogeneic hematopoietic stem celltransplant). The average costs of theseallogeneic cases, approximately$64,845, were higher than the overallaverage costs of all cases in MSDRG009, approximately $43,877. Theaverage length of stay for the allogeneiccases, approximately 28 days, wasslightly higher than the average lengthof stay for all cases assigned to MSDRG009, approximately 21 days.

We found 1,269 autologous bone

marrow transplant cases reported withone of the following ICD9CMprocedure codes: 41.00 (Bone marrowtransplant, not otherwise specified);41.01 (Autologous bone marrowtransplant without purging); 41.04(Autologous hematopoietic stem celltransplant without purging); 41.07(Autologous hematopoietic stem celltransplant with purging); or 41.09(Autologous bone marrow transplantwith purging). The average costs ofthese cases, approximately $37,350, wasless than the overall average costs of all

HACsDiscussion.3RTIEvaluation

Documents