Andrew M. Slavitt, Acting Administrator Centers for Medicare & Medicaid Services Attention: CMS-1655-P P.O. Box 8011 Baltimore, MD

June 17, 2016 Andrew M. Slavitt, Acting Administrator Centers for Medicare & Medicaid Services Attention: CMS-1655-P P.O. Box 8011 Baltimore, MD 21244-1850 RE: Medicare Program; Hospital Inpatient Prospective Payment Systems for Acute Care Hospitals and the Long-Term Care Hospital Prospective Payment System and Proposed Policy Changes and Fiscal Year 2017 Rates; Quality Reporting Requirements for Specific Providers; Graduate Medical Education; Hospital Notification Procedures Applicable to Beneficiaries Receiving Observation Services; and Technical Changes Relating to Costs to Organizations and Medicare Cost Reports Dear Mr. Slavitt: On behalf of the over 80,000 members of the American College of Surgeons (ACS), we appreciate the opportunity to submit comments to the proposed rule: Medicare Program; Hospital Inpatient Prospective Payment Systems for Acute Care Hospitals and the Long-Term Care Hospital Prospective Payment System and Proposed Policy Changes and Fiscal Year 2017 Rates; Quality Reporting Requirements for Specific Providers; Graduate Medical Education; Hospital Notification Procedures Applicable to Beneficiaries Receiving Observation Services; and Technical Changes Relating to Costs to Organizations and Medicare Cost Reports (Proposed Rule) published in the Federal Register on April 27, 2016. The ACS was founded in 1913 to improve the quality of care for the surgical patient by setting high standards for surgical education and practice. Since a large proportion of surgical care is provided in the inpatient hospital setting, we have a strong interest in the Centers for Medicare & Medicaid Services (CMS) Inpatient Prospective Payment System (IPPS) and related hospital quality improvement efforts, and can offer insight to CMS proposed modifications. QUALITY METHODOLOGY: GENERAL CONSIDERATIONS Across all quality programs discussed in this letter, the ACS continues to strongly support measures based on clinical data because clinical data typically provide more accurate and relevant information compared to claims-based data for the purposes of quality measurement. Studies have

demonstrated that claims-based data are inaccurate specifically for tracking surgical complications, i.e. there are more false-negative and false-positive rates, compared to clinical data. 1 Claims-based data do not accurately address the nuances of comorbidities, severity, conditions present on admission, complications, patient experience, and do not enable adequate risk adjustment. In addition, CMS notes that it continues to have concerns about holding hospitals to different standards for the outcomes of their patients of diverse sociodemographic status (SDS), because it would mask potential disparities or minimize incentives to improve the outcomes of disadvantaged populations. The Office of the Assistant Secretary for Planning and Evaluation (ASPE) is conducting research to examine the impact of SDS on quality measures, and the ACS agrees that CMS should examine the findings of the ASPE reports and related Secretarial recommendations, and consider how these can apply to quality programs. The ACS also recommends that CMS follow the work of the National Quality Forum (NQF) two-year pilot project titled Risk Adjustment for Sociodemographic Factors, 2 which aims to provide recommendations on the appropriate application of risk adjustments to performance measures data. During the trial, measure developers are expected to submit information such as analyses, interpretations, and performance scores with and without SDS factors in the risk adjustment model. The ACS currently supports SDS risk adjustment for measures used in accountability applications (e.g., public reporting and pay-for-performance) on a case-by-case basis. Without the use of appropriate risk adjustment for certain measures, clinical outcomes will be less reliable due to SDS confounding variables. Closely evaluating the appropriate measures for SDS confounding variables will lead to a deeper understanding of the relationship between these variables and clinical outcomes. However, only measures where the provider does not have control over the outcome should be considered for SDS. Until there are further findings on the appropriate application of risk adjustment, the ACS supports the following methodology, when appropriate: For purposes of accountability (e.g., public reporting, pay-forperformance), SDS factors should be included in risk adjustment of the performance score unless there are conceptual reasons or empirical evidence indicating that adjustment is unnecessary or inappropriate; and 1 Lawson, Elise H., Rachel Louie, David S. Zingmond, Robert H. Brook, Bruce L. Hall, Lein Han, Michael Rapp, and Clifford Y. Ko. "A Comparison of Clinical Registry Versus Administrative Claims Data for Reporting of 30-Day Surgical Complications." Annals of Surgery 256.6 (2012): 973-81. 2 National Quality Forum Socioeconomic Status (SES) Trial Period. Available at http://www.qualityforum.org/projectdescription.aspx?projectid=80124. 2

For purposes of identifying and reducing disparities, performance measures should be stratified on the basis of relevant SDS factors when used in analysis by individual providers, policymakers, researchers, and the public working to reduce disparities. 3 OTHER DECISIONS AND CHANGES TO THE IPPS FOR OPERATING COSTS AND INDIRECT MEDICAL EDUCATION COSTS Hospital Value-Based Purchasing Program Under the Hospital Value-Based Purchasing (VBP) Program, CMS calculates a VBP incentive payment percentage for a hospital based on its Total Performance Score for a specified performance period. The total amount available for value-based incentive payments for a fiscal year is equal to the total amount of the payment reductions for all participating hospitals for such fiscal year, as established by the Secretary. For fiscal year (FY) 2017, the available funding pool for value-based incentive payments is 2.0 percent. For each payment year, CMS specifies through rulemaking a VBP measure set, and a baseline and performance period for each measure. Measures available for inclusion in the VBP are those that are included in the Inpatient Quality Reporting (IQR) Program and have been included on the Hospital Compare website for at least one year prior to the start of the relevant VBP Program performance period. PSI-90 Measure in the FY 2018 Program and Future Program Years CMS proposes to change the performance period for the Agency for Healthcare Research and Quality (AHRQ) Patient Safety Indicator (PSI) 90 composite patient safety measure, which was previously adopted for FY 2018, to accommodate for the ICD-10 transition. The proposed period would measure performance for 15 months, July 1, 2014 through September 30, 2015, instead of the previously adopted 24 months. The base year period would not change and would include 24 months of data: July 1, 2010 through June 30, 2012. Using a cutoff date of September 30, 2015 will allow CMS to only use ICD-9 claims data, which aligns with AHRQ s recommendation against combining both ICD-9 and ICD-10 data. While we appreciate the consideration CMS gave in its proposal to shorten the FY 2018 performance period to accommodate for the ICD-10 transition, we 3 National Quality Forum. Risk Adjustment for Socioeconomic Status or Other Sociodemographic Factors Draft Report. March 18, 2014. Available at http://www.qualityforum.org/risk_adjustment_ses.aspx. 3

question the validity and reliability of a 15-month performance period. Overall, the ACS supports having base and performance year periods that are timely, long enough to garner sufficient and meaningful data, and consistent across program years to make accurate comparisons. And AHRQ, which is the measure steward of PSI-90, established 24 months as the appropriate performance period for this measure. The ACS recommends that CMS removes PSI-90 from the FY 2018 Hospital VBP Program until there are 24 months of performance data available using ICD-10 data. Additionally, AHRQ has not yet made available a risk-adjusted ICD-10 version of the PSI- 90 software that permits hospitals to track how this measure performs. Lastly, the ACS has concerns that PSI-90 relies on administrative or claims-based data, versus clinical data. There is limited clinical information included in claims data, which makes it difficult to address the nuances of comorbidities, severity, and complications. As stated above, this also affects the ability to perform adequate risk adjustment. Newly Proposed Measures and Measure Refinements for the FY 2021 Program Year and Subsequent Years Two new risk-standardized payment measures are proposed for inclusion in the VBP program, beginning with FY 2021 payment: Risk-Standardized Payment Associated with a 30-Day Episode of Care for Acute Myocardial Infarction (AMI) and Risk Standardized Payment Associated with a 30-Day Episode of Care for Heart Failure (HF). These new measures are categorized in the Efficiency and Cost Reduction domain. In general, the ACS supports riskstandardized, granular, and condition-specific Medicare payment measures and encourages CMS to continue including these types of measures in pay-for-performance programs. We agree with CMS that riskadjusted standardized Medicare payments are an appropriate indicator of efficiency because they allow for clinical comparisons without the influence of external factors such as age and comorbidities, and allow for specific actionable feedback to hospitals to implement improvements. We also support CMS effort to add measures to the Efficiency and Cost Reduction domain to create incentives for greater coordination between hospitals and physicians. CMS also proposes to include a modified version of the Hospital 30-Day, All- Cause, Risk-Standardized Mortality Rate Following Pneumonia Hospitalization measure in the VBP Program beginning with FY 2021 payment. The modification expands the measure cohort to include: 1) patients with a principal discharge diagnosis of aspiration pneumonia; and 2) patients with a principal discharge diagnosis of sepsis (excluding severe sepsis) and a secondary diagnosis of pneumonia present on admission as well as the original cohort of patients with a principal diagnosis of pneumonia. The ACS supports 4

the American Hospital Association s separately submitted request that CMS not finalize this proposal until the updated version of this measure receives NQF endorsement. Previously Adopted and Newly Proposed Baseline and Performance Periods For the AMI and HF payment measures, CMS proposes a 36-month baseline and performance period, with the exception of the initial implementation year, FY 2021, for which it proposes a 24-month performance period. The full 36- month performance period would apply beginning in FY 2022. As stated above, the ACS supports having base year and performance year periods that are consistent across program years, to make accurate and fair comparisons. CMS should consider delaying the implementation of the AMI and HF measures for a year so that a 36-month performance period would apply. Hospital-Acquired Condition Reduction Program Section 3008 of the Affordable Care Act required CMS to implement a hospital-acquired conditions (HAC) payment adjustment beginning in FY 2015. This requires CMS to reduce hospital payments by one percent for hospitals that rank among the lowest performing 25 percent with regard to HACs specified under this program. The payment adjustment will result in the applicable hospitals receiving 99 percent of the payment that would otherwise apply. Implementation of the HAC Reduction Program for FY 2018 Proposed Adoption of Modified PSI-90: Patient Safety and Adverse Events Composite CMS proposes to adopt refinements to the AHRQ PSI-90 composite safety measure for the HAC Reduction Program beginning with FY 2018 payment determination and subsequent years. NQF undertook maintenance review of this measure and approved the updated measure for continued endorsement in December 2015. The PSI-90 measure s NQF maintenance review led to several changes to the measure, including: 1) the addition of PSI-9, 2) the removal of PSI-7, 3) specification changes to PSI-12 and PSI-15, and 4) the weighting of individual measures based on both volume of the adverse event and harm associated with the adverse event. The ACS offers comments on these changes, below. AHRQ PSI-7 (Central Line Related Bloodstream Infection Rate): The modified PSI-90 measure removes one of the original indicators, PSI-7, because of overlap with the National Healthcare Safety Network (NHSN) 5

Central Line-Associated Bloodstream Infection (CLABSI) measure, which is also part of the HAC Reduction Program. We thank CMS for responding to the ACS comments to the FY 2016 IPPS rule, namely that these two measures address the same condition (central line infections) within the same program and across programs. We support the removal of PSI-7 from the composite PSI-90 measure. AHRQ PSI-9 (Postoperative Hemorrhage or Hematoma Rate): The ACS has concerns with the addition of PSI-9 to the revised PSI-90 composite measure, as PSI-9 unfairly penalizes hospitals with large liver transplant programs. Liver transplants result in significant blood loss in nearly every case, and poor performance on this measure can be driven by the number of liver transplants performed. Therefore we ask that liver transplant patients be excluded from the PSI-9 denominator. We also request that CMS investigate what other high-blood-loss operations may be confounding this measure. AHRQ PSI-12 (Perioperative Pulmonary Embolism (PE) or Deep Vein Thrombosis (DVT) Rate): The ACS appreciates the two changes made to PSI-12: removal of isolated calf vein DVTs from the numerator specification, and removal of patients with any diagnosis of acute brain and/or spinal injury from the denominator specification. 4 The ACS also appreciates that under the modified weighting methodology that accounts for patient harm, the weight of PSI-12 within the composite measure drops from 34 percent to 18 percent. 5 However, the ACS continues to have concerns for the measure steward, NQF, about the vulnerability of PSI-12 to surveillance bias. Studies have shown that hospitals with increasing numbers of structural quality characteristics (i.e. larger hospitals with more accreditations, more complex patients, and engagement in quality initiatives that typically suggest high quality care) have better venous thromboembolism (VTE) prophylaxis, but actually have higher VTE rates, or an increase in PSI-12 event rates. 6 In other words, hospitals with more sophisticated tools and technology used to track VTE show higher rates of VTE and are being penalized for doing a better job at detection. To this end, performance on PSI-12 may reflect differences in VTE imaging use rather than differences in quality of care, and the inclusion of PSI-12 could unfairly 4 Agency for Healthcare Research & Quality. Patient Safety and Adverse Events Composite (modified Version PSI 90). AHRQ, 2016. Web. 26 May 2016. 5 Agency for Healthcare Research & Quality. Patient Safety and Adverse Events Composite (modified Version PSI 90). AHRQ, 2016. Web. 26 May 2016. 6 Bilimoria, Karl Y., Jeanette Chung, Mila H. Ju, Elliott R. Haut, David J. Bentrem, Clifford Y. Ko, and David W. Baker. "Evaluation of Surveillance Bias and the Validity of the Venous Thromboembolism Quality Measure." JAMA 310.14 (2013): 1482-489. Web. 26 May 2016. 6

penalize hospitals with increased vigilance in VTE detection. 7 For these reasons, we recommend that PSI-12 be removed from the Hospital VBP Program. AHRQ PSI-15 (Unrecognized Abdominopelvic Accidental Puncture/Laceration Rate): The specifications for PSI-15 were refined so that the indicator focuses on the most serious intraoperative injuries due to an accidental puncture or laceration. The denominator is now limited to abdominal and pelvic surgery. The numerator is limited to accidental punctures or lacerations that require a return to the operating room at least one day after the index procedure. 8 The ACS appreciates these modifications to PSI-15, but asks that CMS update its guidance regarding the correct coding of PSI-15 to ensure that abdominopelvic punctures or lacerations inherent to a surgery are not incorrectly coded as accidental. Weighting of Component Measures: In the modified version of PSI-90, weighting of the individual component indicators is based on two concepts: the volume of adverse events and the disutility/severity of harm associated with adverse events. 9 The new weighting scheme, along with the addition of indicators and the removal of PSI-7, more equally distributes the component weights compared to earlier versions. The revised weighting approach offers a better measure of iatrogenic harms experienced by patients in U.S. hospitals, supporting performance comparisons based on hospitals success at keeping patients safe from these harms. 10 The ACS supports the re-weighting of component measures within PSI-90 for the reasons AHRQ cites. Maintenance of Technical Specifications for Quality Measures CDC NHSN CLABSI & CAUTI: Standardized Infection Ratios (SIRs) for the CLABSI and Catheter Association Urinary Track Infection (CAUTI) measures currently only have three variables for risk adjustment: facilityspecific patient location, hospital affiliation with a medical school, and bed size of the patient care location. These factors adjust for facility-level variables, but not individual patients. Therefore, hospitals with more complex case mixes are penalized more frequently. The ACS requests that the measure steward, 7 Bilimoria, Karl Y., Jeanette Chung, Mila H. Ju, Elliott R. Haut, David J. Bentrem, Clifford Y. Ko, and David W. Baker. "Evaluation of Surveillance Bias and the Validity of the Venous Thromboembolism Quality Measure." JAMA 310.14 (2013): 1482-489. Web. 26 May 2016. 8 Agency for Healthcare Research & Quality. Patient Safety and Adverse Events Composite (modified Version PSI 90). AHRQ, 2016. Web. 26 May 2016. 9 Agency for Healthcare Research & Quality. Patient Safety and Adverse Events Composite (modified Version PSI 90). AHRQ, 2016. Web. 26 May 2016. 10 Agency for Healthcare Research & Quality. Patient Safety and Adverse Events Composite (modified Version PSI 90). AHRQ, 2016. Web. 26 May 2016. 7

NHSN, includes proper risk-adjustment with the addition of patient-level variables, such as patient-specific demographics and comorbidities. The Centers for Disease Control and Prevention (CDC) has also received reports from NHSN users indicating that in some healthcare facilities, decisions about which infections to report are made by individuals who disregard CDC protocols, definitions, criteria, and who are not thoroughly familiar with NHSN specifications. 11 For example, it has been reported that in some instances, diagnostic microbiology tests are ordered even in the absence of clinical indications for testing, such as obtaining urine specimens for culture and sensitivity testing from patients who have no symptoms of a urinary tract infection. Many negative culture results are generated by this practice subjecting the patient to potentially unnecessary tests. 12 Audit functions are needed to ensure data validity and inter-hospital reliability in clinical data abstraction. The ACS requests that training and re-certification requirements for clinical data abstractors, or those obtaining data from medical records, are put in place. The ACS also recommends that stronger audit mechanisms are put in place. Lastly, the surgical community has noted that many central line infections after abdominal surgery are due to transient gram negative bacteremia, and not a true CLABSI. 13 The ACS requests that patients with a fever after abdominal surgery, that have transient gram negative bacteremia without a clear intraabdominal source, be excluded from the CLABSI measure. CDC NHSN CDI & MRSA: Clostridium difficile Infection (CDI) and methicillin-resistant Staphylococcus aureus (MRSA) use SIRs that are adjusted for the following risk factors: facility bed size, hospital affiliation with a medical school, the number of patients who have CDI or MRSA present on arrival at the hospital, and the type of test the hospital laboratory uses to identify CDI from patient specimens for hospital-onset CDI. We again request that NHSN include proper risk-adjustment with the addition of patientlevel variables, such as patient-specific demographics and comorbidities. QUALITY DATA REPORTING REQUIREMENTS FOR SPECIFIC PROVIDERS AND SUPPLIERS Hospital Inpatient Quality Reporting Program 11 "Adherence to CDC's Infection Definitions and Criteria Is Needed to Ensure Accuracy, Completeness, and Comparability of Infection Information." Letter. Oct. 2015. Centers for Disease Control & Prevention, n.d. Web. 26 May 2016. Available at: http://www.cdc.gov/nhsn/cms/cms-reporting.html. 12 "Adherence to CDC's Infection Definitions and Criteria Is Needed to Ensure Accuracy, Completeness, and Comparability of Infection Information." Letter. Oct. 2015. Centers for Disease Control & Prevention, n.d. Web. 26 May 2016. Available at: http://www.cdc.gov/nhsn/cms/cms-reporting.html. 13 Bilimoria, Karl. "IPPS Quality Measures." Message to the author. 26 May 2016. E-mail. 8

Under the Hospital IQR program, hospitals must meet the requirements for reporting specific quality information to receive the full market basket update for that year, and hospitals that do not will receive a two percentage point reduction in that year s inpatient hospital payment update factor. Removal and Suspension of Hospital IQR Program Measures Proposed Removal of Hospital IQR Program Measures for the FY 2019 Payment Determination and Subsequent Years CMS proposes to remove 15 measures from the IQR Program beginning with the FY 2019 payment determination, including the structural measure Participation in a Systematic Clinical Database Registry for General Surgery. Hospital participation in clinical registries continues to be valuable because it encourages hospital participation in risk-adjusted, audited clinical data registries. It is also important that CMS ensures that hospital and physician programs are in alignment; under the Merit-Based Incentive Payment System (MIPS), CMS proposed multiple options for physician participation in registries, including the Qualified Clinical Data Registry (QCDR) reporting option. Therefore, the ACS believes it is important to include the Participation in a Systematic Clinical Database Registry for General Surgery measure in the Hospital IQR Program. For the remaining measures CMS proposes to remove from the IQR Program, the ACS requests that even if measures are removed because they have high levels of performance with little variation, or are topped out, they be kept on reserve so that performance can be monitored as necessary to ensure that performance and/or adherence to best practices do not decline. We also ask CMS to consider other uses of these topped out measures, such as including it as part of a composite measure. Proposed Additional Hospital IQR Program Measures for the FY 2019 Payment Determination and Subsequent Years CMS proposes to add four new measures to the Hospital IQR Program for FY 2019 and subsequent years, including the clinical episode-based payment measure Cholecystectomy (Chole) and Common Duct Exploration (CDE). This measure includes a set of medical services related to a hospital admission for Chole and CDE, including treatment, follow-up, and post-acute care. This measure assesses the payment for services initiated during an episode that spans the period immediately prior to, during, and following a beneficiary s hospital stay. The Measure Applications Partnership (MAP) did not support this measure because although cost is important to measure, data supporting variation in costs for this procedure were not provided. The ACS agrees with the MAP and makes the additional note that it is very difficult to define the 9

severity of illness for Chole and CDE thereby making it a poor proxy for quality. Lastly, this measure is a claims-based measure, using Medicare Part A and B administrative claims data. As noted earlier, the ACS supports measures based on clinical data compared to claims-based data. Therefore, the ACS does not support the inclusion of Chole and CDE Payment in the IQR. We urge CMS to bring up the above issues during the NQF endorsement process. The ACS will also track this measure and raise these issues during the NQF review process. CMS also proposes to adopt the outcome measure, Excess Days in Acute Care after Hospitalization for Pneumonia, within the IQR program beginning in FY 2019. This is a risk-standardized outcome measure that compares the number of days that patients discharged from a hospital for pneumonia are predicted to spend in acute care, across the full spectrum of possible events (hospital readmissions, observation stays, and emergency department (ED) visits), to the days that patients are expected to spend in acute care based on their degree of illness as defined using principal diagnosis and comorbidity data from administrative claims. The MAP conditionally supports this measure pending NQF review and endorsement. The ACS supports the MAP s recommendation and asks CMS to clarify the risk-adjustment methodology, provide additional study on the observed-to-expected ratio of infection, and consider capturing hospital readmissions, observation stays, and ED visits individually, instead of combining them to represent one number. Possible New Quality Measures and Measure Topics for Future Years CMS is considering proposing the NHSN Antimicrobial Use measure to advance national efforts to reduce the emergence of antibiotic resistance, by enabling hospitals and CMS to assess national trends of antibiotic use. The measure assesses antibiotic use in hospitals based on medication administration data, and compares this with predicted antibiotic use based on national data. CMS submitted this measure to the MAP in December, and the MAP gave it conditional support pending CDC s recommendation that the measure is ready for use in public reporting. The ACS supports efforts to encourage the appropriate use of antibiotics, but does not support the NHSN Antimicrobial Use measure for inclusion in the IQR because the measure is too broad and the measure calculation is unclear. Furthermore, NHSN, the measure steward, stated that it requires more experience and information before using this measure for payment or reporting. Finally, CMS seeks comment on the possibility of including IQR Program measure data that is stratified by race, ethnicity, sex, and disability on the 10

Hospital Compare website. That is, data would be separately reported by category, as well as at the overall un-stratified rate. Again, the ACS recommends that CMS follow the work of the NQF two-year pilot project titled Risk Adjustment for Sociodemographic Factors, 14 for determining the appropriate risk adjustment methodology for a pay-for-reporting program. We appreciate the opportunity to comment on this proposed rule. The ACS looks forward to continuing dialogue with CMS on these important issues. If you have any questions about our comments, please contact Vinita Ollapally, Regulatory Affairs Manager, in our Division of Advocacy and Health Policy. She may be reached at vollapally@facs.org or at (202) 672-1510. Sincerely, David B. Hoyt, MD, FACS Executive Director 14 National Quality Forum Socioeconomic Status (SES) Trial Period. Available at http://www.qualityforum.org/projectdescription.aspx?projectid=80124. 11