1 Impact of Medicare Readmissions Penalties on Targeted Surgical Conditions Karan R. Chhabra, MD MSc National Clinician Scholars Program Center for Healthcare Outcomes and Policy University of Michigan Institute for Healthcare Policy and Innovation Department of Surgery Brigham and Women’s Hospital @krchhabra
Readmissions in the crosshairs 2 Common Costly Preventable Quality Marker 30d readmission $17B Associated with 1/5 to 1/3 rates: surgical in annual 21% (medical) potentially complications Medicare avoidable and mortality 15% (surgical) spending Jencks SF et al. NEJM 2009. Tsai TC et al. NEJM 2013. @krchhabra van Walraven C et al. CMAJ 2011.
The Hospital Readmissions Reduction Program 3 Medical Conditions Surgical Conditions CHF AMI PNA Hip Knee CABG Policy announced Medical penalties Joint replacement Joint replacement (medical penalties) implemented penalties penalties announced implemented March 2010 October 2012 August 2013 October 2014 @krchhabra
The Hospital Readmissions Reduction Program 4 Penalty: Up to 3% of base Average payment DRG rate based on “excess” adjustment = observed/expected -0.6% readmissions in prior 3 years. Policy announced Medical penalties Joint replacement Joint replacement (medical penalties) implemented penalties penalties announced implemented March 2010 October 2012 August 2013 October 2014 @krchhabra Boccuti C & Casillas G, KFF HRRP Issue Brief, 2017.
Impact on medical conditions 5 slope = -.005 slope = -.017 slope = -.103 ~3% reduction Medical penalties Medical penalties Policy announced Medical penalties announced implemented (medical penalties) implemented March 2010 March 2010 October 2012 October 2012 @krchhabra Zuckerman RB et al, NEJM 2016
Impact on medical conditions 6 slope = -.005 slope = -.017 slope = -.103 ~3% reduction Medical penalties Medical penalties Surgical Policy announced Medical penalties announced implemented penalties (medical penalties) implemented announced March 2010 March 2010 October 2012 October 2012 @krchhabra Zuckerman RB et al, NEJM 2016
Growing concern for unintended consequences 7 Increased mortality in Overstated benefits Penalizing medical conditions? due to upcoding? safety-net hospitals Chaiyachati KH et al, JAMA Network Open 2018. Wadhera RK et al, JAMA 2018. Ibrahim AM et al, JAMA Int Med 2017. @krchhabra Ody et al, Health Affairs 2019.
Upcoding and medical readmissions 8 Ody et al, Health Affairs 2019. @krchhabra
Upcoding and medical readmissions 9 Restricting to 9 diagnosis codes 48% (1 pp) smaller reduction Ody et al, Health Affairs 2019. @krchhabra
Our questions 10 Did targeted penalties reduce readmissions after joint replacement? @krchhabra
Our questions 11 Did targeted penalties reduce readmissions after joint replacement? Did the policy have unintended consequences? Episode Length of Observation Post-acute Upcoding spending stay status care utilization @krchhabra
Study Approach: Interrupted Time-Series Analysis 12 Outcomes Cohort 100% FFS Medicare claims Total 30-day episode for total hip and knee 30-day readmissions per payments : replacements CMS definitions • hospital • physician Episodes from 2008-2016 Risk-adjusted for: • post-acute care • age • readmissions • Exclusions: gender • • partial joint replacements race Price-standardized (for • • fractures socio-economic status intentional differences in • • malignancy Elixhauser comorbidities Medicare payments) and • • revisions season risk-adjusted • device complications @krchhabra
Patient Characteristics 13 Baseline Post-HRRP (2008) (2016) Age 74.8 74.0 Comorbidity count (Elixhauser): 0 11% 11% 1 29% 26% 2 61% 63% Discharge destination: Home 18% 30% Home health agency 34% 41% SNF/Rehab 47% 29% @krchhabra
Patient Characteristics 14 Baseline Post-HRRP (2008) (2016) Age 74.8 74.0 Comorbidity count (Elixhauser): 0 11% 11% 1 29% 26% 2 61% 63% Discharge destination: Home 18% 30% Home health agency 34% 41% SNF/Rehab 47% 29% @krchhabra
Changes in Readmissions Rates 15 ~2% @krchhabra
Impact of Upcoding 16 0.4 pp smaller reduction @krchhabra
No Evidence of Unintended Consequences 17 After Medical After Surgical Baseline Penalties Penalties (2008-2010) (2010-2013) (2013-2016) Length of stay 3.5 3.1 2.6 (days) 30-day episode $20,827 $19,895 $17,618 spending ($) Observation status 0.8 1.0 1.2 (%) @krchhabra
No Evidence of Unintended Consequences 18 After Medical After Surgical Baseline Penalties Penalties (2008-2010) (2010-2013) (2013-2016) Length of stay 3.5 3.1 2.6 (days) 30-day episode $20,827 $19,895 $17,618 spending ($) Observation status 0.8 1.0 1.2 (%) Observation trend stable/ .02 .016 .01 decreasing (% / quarter) @krchhabra
No Evidence of Unintended Consequences 19 After Medical After Surgical Baseline Penalties Penalties (2008-2010) (2010-2013) (2013-2016) Length of stay 3.5 3.1 2.6 (days) 30-day episode $20,827 $19,895 $17,618 spending ($) Observation status 0.8 1.0 1.2 (%) Observation trend stable/ .02 .016 .01 decreasing (% / quarter) @krchhabra
Possible Mechanisms for Findings 20 Spillover Effects Anticipatory Effects Floor Effects Medical penalties led Hospitals predicted Some readmissions to broad, non- program would expand inevitable; high- condition-specific after initial performing hospitals delivery changes implementation improved least Zuckerman RB et al, NEJM 2016. Ibrahim A et al, Ann Surg 2017. @krchhabra Desai N et al, JAMA 2016.
Limitations 21 Claims-based Other policies Observational; risk adjustment (BPCI, CJR, etc) no control @krchhabra
Summary of findings 22 Did targeted penalties reduce readmissions after joint replacement? Did the policy have unintended consequences? • • Spending Observation • Length of stay status use • • Post-acute Upcoding care use @krchhabra
Summary of findings 23 No . Readmissions dropped quickly Did targeted penalties after medical penalties via reduce readmissions spillover effects. After targeted after joint replacement? penalties, reductions slowed . Did the policy have unintended consequences? • • Spending Observation • Length of stay status use • • Post-acute Upcoding care use @krchhabra
Summary of findings 24 No . Readmissions dropped quickly Did targeted penalties after medical penalties via reduce readmissions spillover effects. After targeted after joint replacement? penalties, reductions slowed . Not for these outcomes: Did the policy have • Spending, LOS, unintended post-acute care consequences? • Observation: trend • • • Spending Observation Upcoding: minimal • Length of stay status use • • Post-acute Upcoding care use @krchhabra
Do side effects justify the benefits? 25 Possible harms: Benefits approaching a floor... Penalizing safety- risk net hospitals …while risks remain constant Time for new policy? benefit Increased mortality from CHF, PNA? Chaiyachati KH et al, JAMA Network Open 2018. @krchhabra Wadhera RK et al, JAMA 2018.
Thank you 26 Andrew Ibrahim MD MSc Andy Ryan PhD Justin Dimick MD MPH @AndrewMIbrahim @Andy_Ryan_dydx @jdimick1 Center for Healthcare Outcomes & Policy National Clinician Scholars Program Institute for Healthcare Policy and Innovation University of Michigan Department of Surgery Brigham and Women’s Hospital @krchhabra
Effect of Upcoding on Readmissions Reductions 27 After Medical After Surgical Before HRRP Penalties Penalties Difference Announcement Announcement Announcement Readmissions rate without 7.3% 6.6% 5.6% -1.7% comorbidity adjustment Readmissions rate with 7.6% 6.6% 5.5% -2.1% comorbidity adjustment @krchhabra
Recommend
More recommend
