Performance Measurement Work Group April 20, 2016
Readmission Reduction Incentive Program (RRIP) Update Considerations
ICD-10 issue has been identified in CMMI Medicare Readmission Trend Cumulative Readmission Rate Change by Month CY15 vs CY14: Maryland vs Nation 0.00% -0.19% -0.24% -0.26% -0.28% -0.34% -0.37% -0.37% -0.41% -0.41% -0.55% -0.50% -0.75% -1.01% -1.00% -1.50% -2.18% -2.00% -2.26% -2.50% -2.72% -2.75% -2.84% -3.10% -3.18% -3.00% -3.27% -3.30% -3.42% -3.48% -3.50% -3.94% -3.94% -4.00% -4.05% -4.07% -4.10% -4.11% -4.14% -4.15% -4.17% -4.00% -4.34% -4.39% -4.59% -4.50% -5.00% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec National Medicare CMMI MD Medicare HSCRC MD Medicare 3
Medicare Readmission Rates- 2013 - Present 19.50% 18.50% 17.50% 16.50% 15.50% 14.50% 13.50% Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 201320132013201320132013201320132013201320132013201420142014201420142014201420142014201420142014201520152015201520152015201520152015201520152015 "CMMI National Rate" CMMI MD Rate HSCRC Medicare Rate 4
CMMI Target Calculation Base Year Statistics CY 2013 National Medicare Readmission Rate A 15.39% CY 2013 MD Medicare Readmission Rate B 16.61% MD vs National Difference C=B-A 1.22% Annual Requirement to Close the Gap D=C/5 0.24% CY 2014 Results and CY 2015 Projections MD- National % MD % MD % National MD MD Annual Annual Actual National Difference Target Actual Change Target Change CY14 15.50% 0.98% 16.48% 16.47% 0.71% -0.81% -0.84% CY15-November Trend 15.41% 0.73% 16.15% 16.10% -0.55% -2.00% -2.26% CY 15-December Trend 15.34% 0.73% 16.08% 16.12% -1.01% -2.43% -2.18% 5
-30.00% -25.00% -20.00% -15.00% -10.00% 10.00% 15.00% 20.00% -5.00% 0.00% 5.00% FY 2017 RRIP Results UNION OF CECIL COUNTY 6 UMROI UMMS SHORE AT EASTON PRINCE GEORGES MERITUS WASHINGTON ADVENTIST HOLY CROSS MEDSTAR HARBOR HOWARD COUNTY GENERAL GARRETT COUNTY MEMORIAL WESTERN MARYLAND REGIONAL CARROLL PENINSULA REGIONAL MEDSTAR SOUTHERN MARYLAND CY 15 vs CY13 Improvement Rates SHADY GROVE GREATER BALTIMORE FREDERICK MEMORIAL LAUREL REGIONAL SUBURBAN JOHNS HOPKINS DOCTORS' COMMUNITY DORCHESTER ANNE ARUNDEL MEDSTAR FRANKLIN SQUARE UMMS UPPER CHESAPEAKE UMMS MIDTOWN CAMPUS MEDSTAR MONTGOMERY JOHNS HOPKINS BAYVIEW UNIVERSITY OF MARYLAND UMMS BALTO WASHINGTON UMMS CHARLES REGIONAL SAINT AGNES UMMS ST JOSEPH UMMS HARFORD MEMORIAL CALVERT MEMORIAL MEDSTAR GOOD SAMARITAN SINAI MEDSTAR SAINT MARY'S UMMS SHORE AT CHESTERTOWN MEDSTAR UNION MEMORIAL NORTHWEST FORT WASHINGTON MERCY BON SECOURS ATLANTIC GENERAL
% Change from CY13 All Payer vs. Medicare Improvement -30.00% -20.00% -10.00% 10.00% 20.00% 30.00% 0.00% 7 UNION OF CECIL COUNTY UMROI UMMS SHORE AT EASTON PRINCE GEORGES MERITUS WASHINGTON ADVENTIST % Change CM-Adj All-Payer CY15 to CY13 HOLY CROSS MEDSTAR HARBOR HOWARD COUNTY GENERAL GARRETT COUNTY MEMORIAL WESTERN MARYLAND REGIONAL CARROLL PENINSULA REGIONAL MEDSTAR SOUTHERN MARYLAND SHADY GROVE GREATER BALTIMORE FREDERICK MEMORIAL LAUREL REGIONAL SUBURBAN JOHNS HOPKINS DOCTORS' COMMUNITY % Change in Adj Readmission Rate for Medicare FFS (IP only) DORCHESTER ANNE ARUNDEL MEDSTAR FRANKLIN SQUARE UMMS UPPER CHESAPEAKE UMMS MIDTOWN CAMPUS MEDSTAR MONTGOMERY JOHNS HOPKINS BAYVIEW UNIVERSITY OF MARYLAND UMMS BALTO WASHINGTON UMMS CHARLES REGIONAL SAINT AGNES UMMS ST JOSEPH UMMS HARFORD MEMORIAL CALVERT MEMORIAL MEDSTAR GOOD SAMARITAN SINAI MEDSTAR SAINT MARY'S UMMS SHORE AT CHESTERTOWN MEDSTAR UNION MEMORIAL NORTHWEST FORT WASHINGTON MERCY BON SECOURS ATLANTIC GENERAL
-30.00% -25.00% -20.00% -15.00% -10.00% 10.00% 15.00% 20.00% -5.00% 0.00% 5.00% Improvement Rates vs Base Rate UNION OF CECIL COUNTY 8 UMROI UMMS SHORE AT EASTON PRINCE GEORGES MERITUS WASHINGTON ADVENTIST HOLY CROSS MEDSTAR HARBOR HOWARD COUNTY GENERAL GARRETT COUNTY MEMORIAL WESTERN MARYLAND REGIONAL CARROLL PENINSULA REGIONAL MEDSTAR SOUTHERN MARYLAND Improvement SHADY GROVE GREATER BALTIMORE FREDERICK MEMORIAL LAUREL REGIONAL SUBURBAN JOHNS HOPKINS CY13 Rate DOCTORS' COMMUNITY DORCHESTER ANNE ARUNDEL MEDSTAR FRANKLIN SQUARE UMMS UPPER CHESAPEAKE UMMS MIDTOWN CAMPUS Cy13_Adjusted Out of State MEDSTAR MONTGOMERY JOHNS HOPKINS BAYVIEW UNIVERSITY OF MARYLAND UMMS BALTO WASHINGTON UMMS CHARLES REGIONAL SAINT AGNES UMMS ST JOSEPH UMMS HARFORD MEMORIAL CALVERT MEMORIAL MEDSTAR GOOD SAMARITAN SINAI MEDSTAR SAINT MARY'S UMMS SHORE AT CHESTERTOWN MEDSTAR UNION MEMORIAL NORTHWEST FORT WASHINGTON MERCY BON SECOURS ATLANTIC GENERAL 0.0% 2.0% 4.0% 6.0% 8.0% 10.0% 12.0% 14.0% 16.0% 18.0% 20.0% 22.0%
De Development of lopment of a a Risk-Adjusted Risk-Adjusted Readmission Ra eadmission Rate te Pr Preliminar eliminary R y Results sults April 20, 2016 Matthew Sweeney
Overview of recent work Develops regression-based adjustment model Converts current approach to use regression-based approach APR-DRG SOI fixed effects model Assesses model fit and predictive properties Tests whether simpler model yields similar results Reduces the number of variables needed in the model Tests impacts of adding covariates to the model Impacts on model fit Impacts on hospital rates, and improvement from CY2013 to CY2015 Covariates tested: Age Gender Elixhauser co-morbidities Primary payer ADI 10 10
Converting Current Approach Indirect standardization Calculate statewide readmission norms for each APR-DRG SOI category Calculate hospital-level predicted readmission, based on relative frequency of APR-DRG SOI categories Fixed effects regression Mathematically, yields identical number of predicted readmissions Stay-level regression Dependent variable: 0/1 indicator for 30-day readmission Independent variables: 0/1 indicator for each of the ~1100 APR-DRG SOI categories Pros: Facilitates assessment of explanatory power and predictive ability Easy to measure impact of additional covariates Con: Computationally intensive 11 11
Alternate Models “Norms” – based regression Replace APR-DRG SOI indicators with CY 2013 norms (single variable) Proxy for a readmission-based APR-DRG weight Log-transformation improves model fit Test impact of additional covariates Patient age and gender Elixhauser co-morbidities 31 indictors for various conditions Calculated based on information from the index stay Primary payer Medicare FFS Medicare Managed Care Medicaid Commercial Self pay Other ADI Indicators for each of the 20 quantiles of the ADI distribution 12 12
Data and Methods Data: CY 2013 and CY 2015 inpatient data Methods: Regressions Estimate logistic model on CY 2013 stays Calculate predicted probability of readmission for both CY 2013 and CY 2015 stays CY 2015 predicted values are benchmarked to CY 2013, similar to current approach Measure R-square and c-statistic R-square: how much variation is explained by the model? C- statistic: how well does model predict readmission? Hospital-level rates Calculate sum of predicted probabilities for each hospital Calculate O/E ratio (where E = sum of predicted probabilities) O/E x State Rate in CY 2013 = risk-adjusted rate Calculate percent improvement between CY 2013 and CY 2015 for each hospital 13 13
Summary of Models APR- SOI CY 2013 Age and Elixhauser Model Payer ADI Fixed Norms Gender Comorbidities Effects Baseline Yes No No No No No 15 No Yes No No No No 18 No Yes Yes Yes No No 19 No Yes Yes Yes Yes No 20 No Yes Yes Yes Yes Yes 14 14
Model Fit Statistics Number of Max-rescaled Model Controls c-statistic Observations R square APR-DRG SOI Baseline 561,903 0.712 0.128 Fixed Effects 561,903 15 CY 2013 Norms 0.712 0.127 Model 15 Plus: 561,903 18 Age, Gender, 0.726 0.142 Comorbidities Model 18 Plus: 561,903 19 Primary Payer 0.730 0.147 Model 19 Plus: 561,903 20 ADI 0.731 0.148 15 15
Recommend
More recommend
