Preventable Readmissions ACMQ February, 2010

Transcription

1 Preventable Readmissions ACMQ February, 2010 Norbert Goldfield, M.D. Medical Director, 3M Health Information Systems

2 3M HIS Clinical Research Experience 3M HIS Experience in developing classification and payment/quality systems includes: Development of the first DRG Prospective Payment System (PPS) in NJ in 1980 Design and development of the first outpatient PPS for Iowa Medicaid Under contract with CMS, design, development and maintenance of acute long term care hospital PPS Design and development of ICD-10 PCS Design and development of Potentially Preventable Readmission (PPRs) and Potentially Preventable Complication (PPCs) using APR-DRGs Under contract to the Federal Government, development of Clinical Risk Groups (CRGs) and CRxGs (privately funded - using pharmaceutical data) for population profiling/ risk adjustment/ physician profiling Together with the State of Maryland and JHU/U of Maryland developing new payment system for inpatient mental health services with NIMH grant All classification tools including PPRs are developed jointly with NACHRI

3 Strategy to reduce readmissions and increase quality needs to have the following elements A tool to identify preventable readmissions Payment incentives Public reporting Quality Improvement Strategy

4 Potentially Preventable Readmissions (PPRs) - definition Return hospitalizations that may result from deficiencies in the process of care and treatment (readmission for a surgical wound infection) or lack of post discharge followup (prescription not filled) rather than unrelated events that occur post discharge (broken leg due to trauma), or readmissions that were planned at the time of the discharge from the initial admission.

5 Research Approach for Development of PPRs Define exclusion criteria for identifying initial admissions for which a subsequent readmission is excluded from consideration as a PPR (e.g. discharged against medical advice) Develop criteria for determining if a readmission is potentially preventable (i.e. a PPR) Identify chains of related readmissions Develop a method of determining the risk of a PPR occurring and develop a method for computing actual and expected hospital PPR rates Test methodology in large databases

6 PPR Exclusions: Initial Admissions If any of the following conditions apply to the initial admission, a subsequent readmission is excluded from consideration as a PPR Admissions for which follow-up care is intrinsically extensive and complex Major or metastatic malignancies treated medically Multiple trauma, burns Discharge status indicates limited hospital & provider control Left against medical advice Transferred to another acute care hospital Neonates Other exclusions Specific eye procedures and infections Cystic fibrosis with pulmonary diagnoses Died not included as candidate initial admissions (denominator)

7 Research Approach for Development of PPRs Define exclusion criteria for identifying initial admissions for which a subsequent readmission is excluded from consideration as a PPR Develop criteria for determining if a readmission is potentially preventable (i.e. a PPR) Identify chains of related readmissions Develop a method of determining the risk of a PPR occurring and develop a method for computing actual and expected hospital PPR rates Test methodology in large databases

8 General Guidelines for PPRs Readmission Initial Admission Medical Surgical Medical Surgical PPR except if clearly unrelated acute events PPR except conditions clearly unrelated Not PPR unless initial medical diagnosis clearly should have resulted in surgery PPR if related to prior surgery

9 Clinical Factors make a readmission not potentially preventable No clinical relationship to prior discharge Cholecystectomy two weeks after hip replacement Discharge status of prior discharge AMA and transferred to another acute care hospital Type of prior discharge Follow-up care is intrinsically complex and extensive Metatastic malignancies, Multiple trauma, Burns Longer interval between discharge and readmission Long time intervals (>30 days) reduce confidence that readmission is causally linked to the prior discharge

10 Clinical Relation Reasons 1 Medical readmission for a continuation or recurrence of the reason for the initial admission, or for a closely related condition. 2a Ambulatory care sensitive conditions as designated by ARHQ 2b All other readmissions for a chronic problem that may be related to care either during or after the initial admission 3 Medical readmission for an acute medical condition or complication that may be related to or may have resulted from care during the initial admission or in the post-discharge period after the initial admission. 4 Readmission for a surgical procedure to address a continuation or a recurrence of the problem causing the initial admission. 5 Readmission for surgical procedure to address a complication that may be related to or may have resulted from care during the initial admission.

11 Clinical Relation Reasons [continued] 6a 6b 6c Readmission for mental health reasons following an initial admission for a non-mental health, non-substance abuse reason Readmission for a substance abuse diagnosis reason following an initial admission for a non-mental health, non-substance abuse reason Mental health or substance abuse readmission following an initial admission for a substance abuse or mental health diagnosis

12 Research Approach for Development of PPRs Define exclusion criteria for identifying initial admissions for which a subsequent readmission is excluded from consideration as a PPR Develop criteria for determining if a readmission is potentially preventable (i.e. a PPR) Identify chains of related readmissions Develop a method of determining the risk of a PPR occurring and develop a method for computing actual and expected hospital PPR rates Test methodology in large databases

13 Readmission Chain A PPR chain is an initial discharge followed by a number of clinically related readmissions A PPR chain terminates if a readmission meets any of the following criteria: Is outside the x day window of time Is clinically unrelated to initial discharge Left against medical advice Is a transfer to another acute care hospital Meets discharge exclusion criteria Other trauma admission Died

14 Example of a PPR Chain Days between current admission & previous discharge DRG SOI Med Surg Type of Admit S DOR/LUMB FUS EXC CRV BCK Initial discharge M POST-OP INFECTION PPR M BACK PAIN PPR S DOR/LUMB FUS EXC CRV BCK New initial discharge M POST-OP INFECTION PPR

15 Research Approach for Development of PPRs Define exclusion criteria for identifying initial discharges for which a subsequent readmission is excluded from consideration as a PPR (e.g. discharged against medical advice) Develop criteria for determining if a readmission is potentially preventable (i.e. a PPR) Identify chains of related readmissions Develop a method of determining the risk of a PPR occurring and develop a method for computing actual and expected hospital PPR rates Test methodology in large databases

16 All Patient Refined DRGs (APR DRGs) APR DRGs are an extension of DRGs to account for severity of illness and risk of mortality Assignment to a Base APR-DRG based on: Principal Diagnosis, for Medical patients, or Most Important Surgical Procedure (performed in an O.R.) Each Base APR-DRG is divided into 4 subclasses Two types of Subclasses: Severity of Illness (SOI) Risk of Mortality (ROM) SOI and ROM assignment take into account the interaction among principal & secondary diagnoses, age, and, in some cases, procedures Both an admission APR DRG and discharge APR DRG are computed Admission APR DRG requires the present on admission indicator for secondary diagnoses

17 Patients With At Least One PPR in Selected APR-DRGs, by Severity Level (Wisconsin, ) Admission APR-DRG Admission Severity of Illness Level Surgical SOI 1 SOI 2 SOI 3 SOI 4 Total PPRs Stroke At Risk 1,601 7,386 2, ,995 Percent PPRs 350 1,774 1, ,693 Other Pneumonia At Risk 8,342 20,173 10, ,657 Percent PPRs 227 1, ,305 CABG without Catheterization At Risk 2,842 9,957 3, ,547 Percent PPRs ,771 Acute MI At Risk 2,338 5,995 2, ,955 Percent PPRs ,029 Major Large & Small Bowel Procedures At Risk 5,305 9,246 4, ,869 Percent

18 Liver trans &/or intest trans Heart &/or lung transplant Bone marrow transplant ECMO or trach w MV w extn proc Trach w MV wo exten proc Pancreas transplant Craniotomy for trauma Craniotomy exc for trauma Ventricular shunt proc Spinal procedures Extracranial vascular proc Oth nervous syst & relat proc Spinal disorders & injuries Nervous system malignancy Degeneratv nerv sys dis exc MS Mult sclerosis/oth demyelin dx Intracranial hemorrhage CVA w infarct Nonspec CVA & precereb occl Transient ischemia Periph, cranial, auton nerv dx Bact & tuberculous nerv infect Non-bact nerv infect exc VM Methodology for Selecting Potentially Preventable Readmissions (PPRs) A C D E F G H I J K L M N O P Q R S T U V W X Y Z AA AB M - Medical MS - Major Surgical OS - Other Surgical EM - Elective Medical ES - Elective Major Surgical EO - Elective Other Surgical XMA - Malignancy Exclusion XTB - Trauma Exclusion XNN - Neonatal Exclusion XOB - Obstectrical Exclusion XOG - Other Exclusion MS MS MS MS MS MS XTB MS MS MS MS MS M XMA M M M M M M M M M PPR MDC Type IA APR Desc RA APR IA APR 00 MS Liver trans &/or intest trans 1 4 X X 5 3 X G X X X X X X G X X X X X X X X X 00 MS Heart &/or lung transplant 2 X 4 X 5 1 X G X X X X X X G X X X X X X X X X 00 MS Bone marrow transplant 3 X X X G X X X X X X G X X X X X X X X X 00 MS ECMO or trach w MV w extn 4 X X X X X X G X X X X X X G X X X X X X X X X 00 MS Trach w MV wo exten proc 5 X X X 4 4 X G G 2B 2B B MS Pancreas transplant 6 X X X G X X X X X X G X X X X X X X X X 01 XTB Craniotomy for trauma 20 G G G G G G G G G G G G G G G G G G G G G G G 01 MS Craniotomy exc for trauma 21 X X X 5 3 X G 4 4 X X 4 X G 2B X B MS Ventricular shunt proc 22 X X X 5 3 X G 4 4 X X 4 X G 2B X MS Spinal procedures 23 X X X 5 3 X G X X 4 X 4 1 G 2B X X X X X 2B MS Extracranial vascular proc 24 X X X 5 3 X G 4 X X 4 4 X G 2B X X B X X 01 MS Oth nervous syst & relat proc 26 X X X 5 3 X G X 4 X G 2B X X X X 2B 2B M Spinal disorders & injuries 40 X X X X X X G X X X X X X G 2B X X X X X 3 X X 01 XMA Nervous system malignancy 41 G G G G G G G G G G G G G G G G G G G G G G G 01 M Degeneratv nerv sys dis exc M 42 X X X X X X G X X X X X X G 1 X X X X X 2B X X 01 M Mult sclerosis/oth demyelin dx 43 X X X X X X G X X X X X X G 1 1 X X X X 2B X X 01 M Intracranial hemorrhage 44 X X X X X X G X X X X X X G 2B X B X X 01 M CVA w infarct 45 X X X X X X G X X X X X X G 2B X B X X 01 M Nonspec CVA & precereb occ 46 X X X X X X G X X X X X X G 2B X B X X 01 M Transient ischemia 47 X X X X X X G X X X X X X G 2B X B X X 01 M Periph, cranial, auton nerv dx 48 X X X X X X G X X X X X X G 2B X X X X X 1 X X 01 M Bact & tuberculous nerv infec 49 X X X X X X G X X X X X X G 2B X X X X X 2B 1 X 01 M Non-bact nerv infect exc VM 50 X X X X X X G X X X X X X G 2B X X X X X 2B X 1

19 PPR Matrix Florida Results 98,596 cells in 314 Initial Admission (IA) APR DRG x 314 Readmission (RA) APR DRG Matrix 33% clinically related IA by RA APR DRG combination 42% globally excluded IA by RA APR DRG combination 25% DRG specific clinical exclusion IA by RA APR DRG combination Percentage of readmissions considered potentially preventable: 61%

20 Example: Reasons for Readmission CORONARY BYPASS W/CARD CATH 1,386 (13.7%) HEART FAILURE 165 POST-OP/POST-TRAUM INFEC 134 OTHER RESPIRATORY DIAGNOSES 118 ANGINA PECT & CORONARY ATH 90 CARD ARRHYTHMIA & 90

21 Reasons for Readmission Florida, All Patients, Reason Reason Description Count Percent 1 Medical continuation of initial problem 57, A Medical decompensation of chronic problem (ambulatory care sensitive condition) 39, B Medical other decompensation 18, Medical complication of previous admission 76, Surgical continuation of initial problem 8, Surgical complication of initial admission 6, A Mental health, non-mh initial admission 6, B Substance abuse, non SA initial admission 1, C Mental health or substance abuse readmission, initial admission for MH or SA 18, Total 233,

22 Readmission Analysis Role of PPRs Identify admissions that are potentially preventable Identify admissions that are at risk for being followed by a potentially preventable readmission Identify chains of readmissions that are all related to the same prior admission Determine numerator and denominator for computing readmission rates (dependent variable)

23 Role of APR DRG Readmission Analysis [continued] Used as basis of risk adjustment for computing expected readmission rates Indirect rate standardization Additional adjustments for mental health, drug abuse and age Specifies independent variables for risk adjustment Alternatively, risk adjustment could be based on a regression analysis with APR DRGs as one of the factors

24 Readmission Payment System Reform Principals Key principals that you should expect to see and/or want to see in a well thought out payment system reform for readmissions 1) Should initially focus on the elimination of payment incentives that result in increased payment for poor quality outcomes 2) Financial incentives should be substantial enough to induce hospital behavior change 3) Financial incentives should be linked to quality outcomes and not to adherence to externally imposed processes 4) Financial penalties for poor quality outcomes should be rate based and not applied on an individual case-by-case basis 5) Quality standards should be based on the outcomes consistently achieved by the best performing hospitals 6) Methodologies must be transparent, clinically precise and comprehensive with a uniform and consistent structure

25 Current Applications of PPRs Medicare Payment Advisory Commission (MedPAC) used PPRs to evaluate potentially preventable Medicare readmissions in its 2007 Report to Congress Florida implemented public reporting for hospital readmissions using PPRs in June 2008 New York DOH added PPRs to PPCs in its confidential hospital performance reporting program in Utah, Hawaii will implement PPR reporting in Many state agencies and commercial payers evaluating PPRs (including six BCBS plans) Maryland HSCRC modeling the inclusion of PPRs into their quality-based payment initiative for updating FY2011 annual hospital rates.

26 Discussion/Validation/Controversies Discharge severity of illness. Hospital MUST be able to replicate the data if we wish improvement. Readmission window of time Fifteen day window for the hospital Starting at day sixteen upside risk potential for increased funding of the medical home Readmission to same hospital or any hospital Outlier chains Computation of expected value for beneficiaries with mental illness and/or substance abuse disorders Age specific groups; other socioeconomic variables? Public Comparison Perspective: based on rates (as proposed by Medpac) at the hospital level not on specific cases.

27 Summary of PPRs The APR DRG Readmission Module: Determines if there is a probable clinical relationship between an initial hospitalization and a readmission Determines expected rates of readmissions based on the characteristics of the previous admission Compares risk-adjusted rates of readmissions across hospitals The readmission train has left the station. Now our responsibility is to act and provide payment incentives to decrease readmissions