Quality Indicators of Laryngeal Cancer Care in the Elderly

Transcription

1 The Laryngoscope VC 2014 The American Laryngological, Rhinological and Otological Society, Inc. Quality Indicators of Laryngeal Cancer Care in the Elderly Christine G. Gourin, MD, MPH; Kevin D. Frick, PhD; Amanda L. Blackford, ScM; Robert J. Herbert, BS; Harry Quon, MD; Arlene A. Forastiere, MD; David W. Eisele, MD; Sydney M. Dy, MD, MSc Objectives/Hypothesis: To examine associations between quality of care, survival, and costs in elderly patients treated for laryngeal squamous cell cancer (SCCA). Study Design: Retrospective analysis of Surveillance, Epidemiology, and End Results Medicare data. Methods: We evaluated 2,370 patients diagnosed with laryngeal SCCA from 2004 to 2007 using multivariate regression and survival analysis. Using quality indicators derived from guidelines for recommended care, summary measures of quality were calculated for diagnosis, initial treatment, surveillance, treatment of recurrence, end-of-life care, performance, and an overall summary measure of quality. Results: High-quality care was associated with significant differences in survival for diagnosis [HR , 95% CI ( )], initial treatment [HR ( )], surveillance [HR ( )], treatment of recurrence [HR ( )], end-of-life care [HR ( )], performance [HR ( )], and an overall summary measure of quality [HR ( )], which was significantly associated with lower mean incremental costs [2$24,958 (2$35,873 2$14,042)]. There was a significant survival advantage for initial treatment with surgery and postoperative radiation [HR ( )] and high-volume surgical care [HR ( )] after controlling for all other variables, including quality of care. Conclusions: High-quality larynx cancer care in elderly patients was associated with improved survival and reduced costs; however, high-quality care for treatment of recurrence was associated with poorer survival. These data suggest that survival outcomes in elderly patients with laryngeal cancer are not entirely explained by differences in the receipt of quality care using existing treatment and performance quality indicators and also suggest a need to develop sensitive and valid quality indicators of larynx cancer care in this population. Key Words: Laryngeal neoplasms, quality, squamous cell cancer, treatment, surgery, radiation, chemotherapy, survival, costs, SEER-Medicare, elderly. Level of Evidence: 2c. Laryngoscope, 124: , 2014 INTRODUCTION Quality in health care has been identified as a major target for health care reform following a series of landmark reports by the Institute of Medicine (IOM) that documented a large gap between ideal health care and the reality of the care that many Americans receive. 1 3 The IOM has defined quality as the degree to which health services for individuals and populations Additional Supporting Information may be found in the online version of this article. From the Department of Otolaryngology Head and Neck Surgery (C.G.G., D.W.E.); the Department of Oncology, Sidney Kimmel Comprehensive Cancer Center (A.L.B., A.A.F., S.M.D.); and the Department of Radiation Oncology and Molecular Radiation Sciences (H.Q.), Johns Hopkins Medical Institutions; the Johns Hopkins Carey Business School (K.D.F.); and the Department of Health Policy and Management, the Johns Hopkins Bloomberg School of Public Health (K.D.F., R.J.H., S.M.D.); Baltimore, Maryland, U.S.A. Editor s Note: This Manuscript was accepted for publication January 10, Supported by an American Academy of Otolaryngology Head and Neck Surgery Percy Memorial Research Award. The authors have no other funding, financial relationships, or conflicts of interest to disclose. Send correspondence to Christine G. Gourin, MD, MPH, FACS, Johns Hopkins Outpatient Center, Department of Otolaryngology Head and Neck Surgery, 601 N. Caroline Street Suite 6260, Baltimore, MD cgourin1@jhmi.edu DOI: /lary increase the likelihood of desired health outcomes and are consistent with current professional knowledge. 4 Cancer care has been identified as a priority area for quality improvement. The National Cancer Policy Board of the IOM found that the U.S. health care delivery system does not provide consistent, high-quality cancer care to all people, but instead provides care that is fragmented, costly, increasingly complex, and often not supported by evidence of effectiveness. 3 There is a lack of readily available data on quality of care for patients with head and neck cancer, which is a barrier to improving decision making for patients and clinicians. A quality indicator is an agreed-upon process or outcome measure that is used to assess quality of care by indicating the presence or absence of potentially poor care practices or outcomes. These are derived from evidence-based standards of care and are specified with a numerator and denominator to indicate the intended population, recommended care, and exclusions and to describe the performance that should occur, evaluate whether care is consistent, and identify problem areas. 5 Quality indicators identify areas that require further investigation but may not be relevant to all patients or settings. 6 Using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database, we have 2049

2 found a survival advantage for elderly laryngeal cancer patients treated with surgery and postoperative radiation after controlling for stage, demographic variables, and comorbidity; and surgical care provided by a highvolume hospital was associated with improved survival and lower costs. 7 These data suggest that the favorable outcomes of larynx preservation trials 8,9 may not apply to the elderly population. We undertook the present study to determine if differences in quality of care underlie the observed differences in survival, using quality indicators derived from the National Comprehensive Cancer Network (NCCN) guidelines for larynx cancer care, and established quality indicators for end-of-life care and performance to derive a set of summary quality measures that span the spectrum of cancer care. MATERIALS AND METHODS Data Source A cross-sectional analysis of patients diagnosed with larynx cancer in 2004 to 2007 was performed using discharge data from the SEER-Medicare linked database. The SEER data are a combination of population-based registries and are linked with Medicare data and data from the American Medical Association Masterfile through a collaborative effort that involves the National Cancer Institute, the SEER registries, and the Centers for Medicare and Medicaid Services. Study Population Patients with larynx (SEER site code 38) squamous cell cancer without a previous diagnosis of head and neck cancer (01 10, 37, 38, and 41) or lymphoma (68 69, 71-72) comprised the study population. Patients with in situ disease, distant metastatic disease at presentation, diagnosis by autopsy or death certificate, and less than 1 year of continuous claims were excluded from analysis, as were patients enrolled in Medicare health maintenance organizations, because health maintenance organizations do not generate claims for their care. Analyses were restricted to the group of Medicare part A and B enrollees in the SEER-Medicare data aged 66 and older to include patients who were enrolled continuously in Medicare s fee-forservice program from 365 days prior to the time of diagnosis to allow identification of prevalent comorbid conditions prior to diagnosis. Patients were followed through December 2009 until the end of data (ineligibility or end of claims) or death. Variables Primary site, American Joint Committee on Cancer stage, subsite, age, sex, ethnicity, U.S. Census tract median income, marital status, geographic region, urban/rural location, and hospital characteristics were obtained from the SEER Patient Entitlement and Diagnosis Summary File. Age, marital status, location, and income were examined as categorical variables, with census tract median income divided into quintiles. Preexisting chronic conditions identified in claims from 365 days prior to diagnosis were used to calculate a comorbidity index using the Charlson score, 10 as implemented by Deyo et al. 11 and modified by Klabunde et al. 12 Treatment was obtained from Medicare claims files including physician (national claims history, hospital, outpatient, and hospice claims files using International Classification of Disease, 9th revision (ICD-9) codes, current procedural terminology codes, and healthcare common procedure coding system codes 2050 for larynx cancer treatment, as previously defined. 7 Initial treatment was categorized as surgery only, surgery with postoperative radiation, radiation only, or chemoradiation based on treatment claims provided within the first 180 days following diagnosis. Patients undergoing chemotherapy only, or no treatment were excluded from analysis. Treatment of recurrence was defined by codes for treatment that is usually given for recurrence (surgery, chemotherapy, or radiation) occurring more than 30 days after the end of initial treatment, including entry into hospice services. Volume Hospital volumes associated with initial treatment were categorized into tertiles using the number of patients with laryngeal squamous cell cancer treated by each hospital. We examined the distribution of the number of cases per hospital and categorized hospitals into low ( 6 cases), intermediate (7 11 cases), or high ( 12 cases) volume hospitals. Quality Indicators Quality indicators were derived from NCCN guidelines for larynx cancer 13 for diagnosis, treatment, surveillance, and management of recurrence, with additional indicators derived for end-of-life care and performance 18,19 based on previously defined markers for quality care (Table I). We calculated a summary measure of quality by phase of cancer care (diagnosis, initial treatment, surveillance, treatment of recurrence, end-of-life care, and performance) and an overall summary measure of quality. These measures have in the denominator all of the indicators for which a study subject is at risk, and in the numerator are all of the indicators for which the subject received recommended care, resulting in a proportion between 0% and 100%. Each score was then dichotomized at the median, indicating lower quality of care and higher quality of care. Costs Costs were evaluated using all Medicare paid amounts from all standard analytic files, including inpatient, outpatient, physician/supplier, hospice, home health, and durable medical equipment. Costs were categorized as inpatient, outpatient, and other, and combined into overall costs. Costs were calculated for each year after diagnosis through year 5 as well as total 5-year costs. Costs were adjusted for inflation, with results converted to 2012 USD and adjusted for differences in Medicare reimbursements by geographic region, as previously described. 20 Statistical Analysis Data were analyzed using Stata 12 (StataCorp, College Station, TX). Associations between variables were analyzed using cross-tabulations and multivariate regression modeling. Nonzero counts with fewer than 11 observations were masked in accordance with the SEER-Medicare data use agreement. The dichotomized individual quality indicators and overall summary measure of quality were examined as dependent variables using multiple logistic regression analysis. Independent variables included age, sex, race, comorbidity, marital status, median income quintile, primary site, stage, urban/rural location, SEER region, hospital volume, and initial treatment. Overall survival, defined as time from diagnosis to either last claim date or death, was analyzed using the Kaplan-Meier method and multivariate Cox proportional hazard analysis. Multivariate

3 Level of Evidence TABLE I. Quality Indicators. Quality Indicator Number With Indicator/ Number Eligible (%) 2a* Diagnosis 13 Histologic confirmation of disease 1,339/2,370 (56.5) Staging 2,200/2,370 (92.8) Pretreatment imaging (excluding Tis or T1 glottic) 777/1,398 (55.6) Pretreatment dental evaluation prior to RT /2,099 2a* Initial Treatment 13 Appropriate surgery (neck dissection if indicated based on stage or site with primary ablative surgery for N0 disease if not followed by postoperative radiation, or for N1 disease if primary ablative surgery performed; no surgery for T4b disease) 36/137 (26.3) Appropriate radiation 1,171/2,099 (55.8) Appropriate chemotherapy 176/427 (41.2) Time to start of postoperative RT 6 weeks after surgery 871/1,017 (85.6) 2a* Surveillance 13 Follow-up appointments Year 1: at least one visit at days following last initial treatment date 1,484/2,263 (65.6) days following last initial treatment date 1,579/2,263 (69.8) days following last initial treatment date 1,428/2,263 (63.1) days following last initial treatment date 1,281/2,263 (56.6) At least 1 visit 1,938/2,263 (85.6) 2 or more visits 1,690/2,263 (74.7) 3 or more visits 1,333/2,263 (58.9) All 4 visits 811/2,263(35.8) Year 2: at least one visit at days following last initial treatment date 1,288/1,838 (70.1) days following last initial treatment date 834/1,838 (45.4) At least 1 visit 1,366/1,838 (74.3) Both visits 756/1,838 (41.1) Year 3: at least one visit at days from last initial treatment date 654/1,142 (57.3) Year 4: at least one visit at 1,096 1,335 days from last initial treatment date 336/666 (50.5) Year 5: at least one visit at 1,460 1,700 days from last initial treatment date 86/275 (31.3) Posttreatment imaging (if T3/4 or N2/3) 171/309 (55.3) TSH screening if received RT Year 1 18/2,048 (0.9) Year 2 11/1,673 (0.7) Year 3 /1,036 Year 4 /606 Year 5 0/248 (0) Dental evaluation if received RT** Year 1 /2,048 Year 2 /1,673 Year 3 /1,036 Year 4 0/606 (0) Year 5 0/248 (0) 2a* Treatment of Recurrent Disease 13 Appropriate surgery (no surgery for distant metastatic disease) 439/664 (66.1) Appropriate radiation (if no previous RT) 31/43 (72.1) Appropriate chemotherapy 208/421 (53.9) Hospice for distant metastatic disease not treated with chemotherapy 132/565 (23.4) Time to start of postoperative RT 6 weeks after surgery 51/172 (29.7) 2051

4 Level of Evidence TABLE I. (Continued) Quality Indicator Number With Indicator/ Number Eligible (%) 1b, 2c End of Life Care Hospice care > 7 days before death from cancer 93/303 (30.7) No chemotherapy within 14 days of death from cancer 289/303 (95.4) Death from cancer not in acute setting 198/303 (65.4) No ICU care in last 30 days of life 240/303 (79.2) No acute care in last 30 days of life 165/303 (54.5) 2b, 4 Negative Performance Indicators 18,19 Surgical patients: Length of hospitalization 6 days 149/1,288 (11.6) Return to operating room within 7 days of surgery 26/1,288 (2.0) Readmission within 30 days 158/1,288 (12.3) Blood transfusion 12/1,288 (0.9) Wound infection within 30 days of surgery /1, day mortality 14/1,288 (1.1) Nonsurgical patients: Hospitalization within 30 days of treatment 100/1,082 (9.2) 30 day mortality 19/1,082 (1.8) *NCCN guideline. Less than 11 observations. Includes all patients who received chemotherapy with radiation, including chemotherapy given as postoperative chemoradiation. ICU 5 intensive care unit; NCCN 5 National Comprehensive Cancer Network; RT 5radiation therapy; TSH thyroid-stimulating hormone. generalized linear regression modeling with a log link was used to analyze incremental differences in costs by quality, as costs were not normally distributed. This protocol was reviewed and approved by the Johns Hopkins Medical Institutions Institutional Review Board. RESULTS There were 2,370 cases that met study criteria (Supplemental Table I). The distribution of patients by receipt of care, as specified by quality indicators, is described in Table I. Because Medicare coverage of dental services is very limited, it was excluded from subsequent analysis as a quality indicator. Descriptive statistics of the summary measures of quality by phase of cancer care and the overall summary measure of quality are shown in Table II, with demographic characteristics shown in Supplemental Table II. Patients with advanced age were less likely to receive higher-quality care for diagnostic measures, whereas patients with advanced comorbidity were less likely to receive higherquality care for initial treatment and to have highquality performance outcomes. Patients who were married were more likely to receive high-quality care for initial treatment and surveillance, and patients in the topincome quintile were more likely to receive high-quality care for diagnostic and end-of-life measures. Those patients with advanced stage disease were less likely to receive high-quality care for diagnostic, initial treatment, and overall quality measures, and were less likely to have high-quality performance outcomes but more likely to receive high-quality care for recurrent disease. By initial treatment category, patients treated with surgery and postoperative radiation were less likely to receive high-quality diagnostic care, but were more likely to receive high-quality care for initial treatment and to have high-quality performance outcomes. Patients treated with chemoradiation were more likely to receive high-quality diagnostic care, but were less likely to receive high-quality care for recurrent disease and to have high-quality performance outcomes. Highvolume hospital care was associated with higher-quality care for initial treatment. Mortality was lower for patients who received higher-quality care for each quality measure. Multivariate logistic regression analysis of variables associated with higher-quality care revealed that for the overall summary measure of quality, patients with supraglottic tumors, age 80 years, advanced comorbidity, or advanced tumor stage were less likely to receive high-quality care (Supplemental Table III). Radiation and multimodality treatment were associated with an increased likelihood of high-quality initial treatment but lower odds of high-quality surveillance care. Highquality care outcomes as measured by performance indicators were significantly less likely for patients with advanced comorbidity, advanced age, and advanced stage, but were more likely for patients treated with surgery and postoperative radiation. Advanced comorbidity was significantly less likely to be associated with highquality end-of-life care. High-volume hospital care was significantly associated with an increased likelihood of high-quality care for initial treatment. However, overall there was no significant pattern of predictor variables associated with the receipt of high-quality care across the cancer care continuum. 2052

5 TABLE II. Descriptive Statistics Summarizing Quality Scores. Quartile Ranking, N (%) N Mean (SD) Median (range) 0% 25% 25% 50% 50% 75% 75% 100% Overall 2, (0.19) 0.75 (0,1) 30 (1) 206 (9) 788 (33) 1,346 (57) Diagnosis 2, (0.22) 0.67(0,1) 64 (3) 378 (16) 1,201 (51) 727 (31) Initial treatment 2, (0.39) 0.67 (0,1) 432 (20) 67 (3) 578 (27) 1,051 (49) Surveillance 2, (0.19) 0.38 (0,1) 648 (29) 1,085 (48) 477 (21) 53 (2) Treatment of recurrent disease (0.37) 0.50 (0,1) 233 (26) 111 (12) 320 (36) 229 (26) End of life (0.29) 0.80 (0,1) 46 (15) 61 (20) 44 (15) 152 (50) Performance 2, (0.13) 1 (0,1) * 134 (6) * 2,172 (92) *Less than 11 observations. SD 5 standard deviation. Cox proportional survival analysis demonstrated that high-quality care was associated with improved survival for each of the summary measures of quality, with the exception of high-quality care for recurrent disease, which was associated with poorer survival (Table III). The survival advantage of initial treatment with surgery and postoperative radiation remained significant in models controlling for high-quality care for diagnosis (HR 0.65; 95% CI ; P < 0.001), surveillance (HR 0.58; 95% CI ; P < 0.001), performance outcomes (HR 0.68; 95% CI ; P ), and the overall summary measure of quality (HR 0.66; 95% CI ; P <0.001). Overall, high-volume hospital care was not associated with survival, but in the subset of patients who underwent surgery as part of initial treatment, high-volume hospital care was associated with a TABLE III. Overall Survival Estimates by Categories of Receipt of Quality Indicators. N 1 year OS 2 years OS 5 Years OS HR All patients 2, (84, 87) 72 (70, 74) 49 (46, 52) Diagnosis low (75, 83) 66 (61, 70) 43 (37, 50) high 1, (85, 88) 73 (71, 75) 50 (47, 54) 0.80 (0.66, 0.97) Initial treatment low (72, 79) 60 (56, 64) 38 (32, 44) high 1, (87, 90) 75 (73, 78) 53 (49, 56) 0.75 (0.63, 0.88) Surveillance low 1, (84, 87) 71 (69, 73) 49 (46, 52) high (98, 100) 87 (84, 90) 59 (49, 68) 0.54 (0.44, 0.66) Treatment of recurrence low (92, 96) 74 (69, 78) 43 (36, 50) high (81, 87) 60 (55, 64) 27 (22, 33) 1.54 (1.26, 1.89) End-of-Life low (34, 53) 19 (12, 27) 0 high (53, 67) 27 (21, 33) (0.52, 0.92) Performance low (43, 60) 35 (27, 43) 16 (8, 27) high 2, (86, 89) 74 (72, 76) 51 (48, 54) 0.41 (0.33, 0.52) Overall low (63, 74) 56 (49, 61) 35 (28, 42) high 2, (86, 88) 74 (72, 76) 51 (47, 54) 0.66 (0.54, 0.80) Values for 1-, 2-, and 5-year survival probabilities are shown as percent alive, with a 95% confidence interval. HRs and corresponding confidence intervals and P values are estimated from Cox proportional hazards models and adjust for stage, age at diagnosis, site, sex, race, marital status, comorbidities, SEER region, initial treatment, and hospital volume. HRs are for high vs. low categories of quality, so values greater than 1 suggest high quality is associated with poorer survival; values less than 1 suggest high quality is associated with improved survival. BOLD 5 P < 0.05 vs. reference category of low-quality care. HR 5 hazard ratio; OS 5 overall survival; SEER 5 Surveillance, Epidemiology, and End Results. 2053

6 Fig. 1. Generalized linear regression of mean incremental costs of larynx cancer care for patients receiving high-quality care compared to the reference group of low-quality care. High-quality care was associated with significantly lower mean costs for surveillance, treatment of recurrent disease, and the overall summary measure of quality care, after controlling for stage, age at diagnosis, site, sex, race, marital status, comorbidities, SEER region, initial treatment, and hospital volume. survival advantage in a model controlling for highquality initial treatment (HR 0.64; 95% CI ; P ), but 5 was not significant in models controlling for the other summary measures of quality and the overall summary measure of quality. The results of generalized linear regression analysis of mean incremental costs of larynx cancer care for patients receiving high-quality care compared to the reference group of low-quality care are shown in Figure 1. High-quality care was associated with significantly lower mean incremental costs for surveillance, treatment of recurrent disease, and the overall summary measure of quality care, after controlling for stage, age at diagnosis, site, sex, race, marital status, comorbidities, SEER region, initial treatment, and hospital volume. Costs were significantly higher for patients who underwent initial treatment with chemoradiation after controlling for all other summary measures of quality and the overall summary measure of quality (data not shown). Overall, high-volume hospital care was not associated with costs, but in the subset of patients who underwent primary surgery as part of initial treatment, high-volume hospital care was associated with lower mean incremental costs for high-quality initial treatment (mean 2$16,364; 95% CI 2$28,620 2$4,109; P ), surveillance (mean 2$12,754; 95% CI 2$24,487 2$1,022; P ), performance outcomes (mean 2$12,863; 95% CI 2$24,473 2$1,254; P ), and the overall summary measure of quality (mean 2$13,345; 95% CI 2$24,590 2$2,099; P ), but was not significantly associated with cost differences for high-quality diagnosis, treatment of recurrent disease, or end-of-life care DISCUSSION These data demonstrate that high-quality care, as defined by quality indicators derived from NCCN guidelines, performance indicators, and end-of-life quality indicators, was associated with improved survival and lower costs of care in elderly patients with laryngeal cancer across the cancer care continuum. However, highquality care for recurrence was associated with worse survival, and there remained a survival advantage for elderly larynx cancer patients treated with surgery and postoperative radiation that persisted after controlling for quality of care as well as a survival advantage and cost savings for surgical patients who received highvolume surgical care after controlling for quality of care associated with initial treatment. These data suggest that survival in elderly patients with laryngeal cancer is not entirely explained by differences in the receipt of quality care, as measured by existing treatment and performance quality indicators. Outside of the clinical trial setting, elderly patients with laryngeal cancer do not appear to have similar outcomes to those reported in chemoradiation clinical trials, and aggressive cancerdirected treatment of recurrence is associated with increased mortality in this population. These observations suggest that current treatment guidelines based on clinical trial data may not accurately reflect outcomes in the elderly. We did not find any consistent pattern of predictor variables associated with the receipt of high-quality care using the different quality indicators. Advanced age, advanced stage, and advanced comorbidity were significantly associated with a reduced likelihood of highquality care, as measured by the overall summary

7 measure of quality and performance indicators. Initial treatment with chemoradiation, or surgery and postoperative radiation as well as high-volume hospital care, were associated with an increased likelihood of NCCNguideline compliance with initial treatment recommendations. However, only patients treated with surgery and postoperative radiation were more likely to receive high-quality care as measured by performance indicators; whereas patients treated with chemoradiation were less likely to receive high-quality care for performance measures and management of recurrent disease. Among patients who underwent surgery as part of initial treatment, high-volume hospital care was associated with better survival and lower costs after controlling for the quality of initial treatment, but volume was not significant in models evaluating the other summary measures of quality including performance or the overall summary measure of quality. These data suggest that elderly patients with larynx cancer may respond better to primary surgical management followed by radiation, with reduced morbidity and mortality compared to nonoperative treatment. Furthermore, the survival advantage associated with highvolume surgical care appears be due to differences in the quality of surgical care delivered at high-volume hospitals, with volume-based differences in process measures for patient selection and management during initial treatment that are not entirely explained by our analysis using published guidelines and performance indicators as summary measures of quality. Volume-based standards for larynx cancer surgical care appear to be a candidate marker for quality measurement and improvement, 21,22 particularly as the survival advantage of high-volume surgical care is associated with lower costs. However, quality indicators for laryngeal cancer care will require additional development and evaluation to determine which indicators are critical to efficient and effective quality measurement and correlate most strongly with patient outcomes. The observation that guideline-directed aggressive treatment of recurrence is associated with increased mortality suggests that the observed increase in mortality may be treatment-related. Future research should focus on the development, testing, and implementation of quality indicators that have good reliability and validity, variation among settings, and sensitivity to change to perform as valid measures of quality of laryngeal cancer care that patients, payers, and clinicians can use to evaluate and deliver high-quality care. 6 There are several limitations to the use of administrative databases in risk adjustment and documentation that may impact the observed differences in survival based on quality of care, initial treatment, and surgical volume. While comorbidity scores were used for risk classification, the ability to adequately control for case mix is limited when discharge diagnoses from claims data are used. Poor outcomes may reflect lower quality or may reflect unobserved patient severity. There are wellrecognized limitations of Medicare data in capturing coexisting illnesses, and differences in comorbidity prevalence that are not adequately captured and resultant noncancer mortality may underlie disparities in guideline compliance across the cancer-care continuum and in survival following cancer treatment. 23 The treatment of comorbid illnesses may differ between treatment groups and between centers of differing volumes that are not accurately reflected in claims data and may influence compliance with quality guidelines. Poor documentation of care itself may constitute poor quality of care and may limit indicator reliability. 6 Evaluation of the appropriateness of surgery was limited by NCCN guideline definitions, and the appropriateness of the extent of primary site surgery cannot be determined. Medicare volumes may be an imprecise surrogate for true hospital volumes, although previous analyses have shown a strong correlation between Medicare and all-payer volumes. 23 Finally, the summary measures used in this study may not necessarily serve as effective measurements of overall quality in the absence of testing, demonstrating the feasibility, reliability, and validity of these measures as summary indicators. Nevertheless, these data demonstrate that highquality care, as defined by NCCN guidelines, performance indicators, and end-of-life quality indicators, is associated with improved survival and lower costs in elderly patients with laryngeal cancer across the cancer care continuum. However, high-quality care for management of recurrent disease was associated with worse survival, suggesting a need to reevaluate guidelines for management of recurrent disease in the elderly in whom aggressive treatment may negatively impact survival. There remains a survival advantage for elderly larynx cancer patients who undergo surgery with postoperative radiation that persists after controlling for quality of care, and high-volume surgical care is associated with improved survival and lower costs, even after controlling for treatment-related differences in quality of care. These data suggest that current guidelines for laryngeal cancer care may not accurately reflect outcomes in the elderly, lending support to the use of volumebased standards for larynx cancer surgical care as a candidate marker for quality measurement and improvement, and supporting the need for further study to determine the underlying causes for these observations to develop sensitive and valid quality indicators of larynx cancer care in this population. CONCLUSION High-quality larynx cancer care in elderly patients was associated with improved survival and reduced costs; however, high-quality care for treatment of recurrence was associated with poorer survival. These data suggest that survival outcomes in elderly patients with laryngeal cancer are not entirely explained by differences in the receipt of quality care using existing treatment and performance quality indicators, and suggest a need to develop sensitive and valid quality indicators of larynx cancer care in this population. BIBLIOGRAPHY 1. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press,

8 2. National Cancer Policy Board, Institute of Medicine and National Research Council. Ensuring quality cancer care. Hewitt M, Simone JV (eds). Washington, DC: National Academy Press, Institute of Medicine. Delivering high-quality cancer care. Levit LA, Balogh EP, Nass SJ, Ganz PA (eds). Washington, DC: National Academy Press, Crossing the quality chasm: the IOM Health Care Quality Initiative. Chasm-The-IOM-Health-Care-Quality-Initiative.aspx. Accessed June 9, Seow H, Snyder CF, Mularski RA, et al. A framework for assessing quality indicators for cancer care at the end of life. J Pain Symptom Manage 2009;38: Dy SM, Lorenz KA, ONeill SM, et al. Cancer Quality-ASSIST supportive oncology quality indicator set. Cancer 2010;116: Gourin CG, Dy SM, Herbert RJ, et al. Treatment, survival and costs of laryngeal cancer care in the elderly. Laryngoscope. Epub ahead of print. 8. The Department of Veterans Affairs Laryngeal Cancer Study Group. Induction chemotherapy plus radiation compared with surgery plus radiation in patients with advanced laryngeal cancer. N Eng J Med, 1991; 324: Forastiere A, Goepfert H, Maor M, et al. Concurrent chemotherapy and radiotherapy for organ preservation in advanced laryngeal cancer. N Eng J Med 2003;349: Charlson ME, Pompei P, Ales KL, Mackenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. Journal of Chronic Disease 1987;40: Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol 1992;45: Klabunde CN, Potosky AL, Legler JM, et al. Development of a comorbidity index using physician claims data. JClinEpidemiol2000;53: National Comprehensive Cancer Network. NCCN Guidelines for head and neck cancer care. head-and-neck.pdf. Accessed June 9, Earle CC, Park ER, Lai B, et al. Identifying potential indicators of the quality of end-of-life cancer care from administrative data. J Clin Oncol 2003;21: Earle CC, Neville BA, Landrum MB, et al. Trends in the aggressiveness of cancer care near the end of life. J Clin Oncol 2004;22: Earle CC, Neville BA, Landrum MB, et al. Evaluating claims-based indicators of the intensity of end-of-life cancer care. Int J Qual Health Care 2005;17: Mack JW, Cronin A, Keating NL, et al. Associations between end-of-life discussion characteristics and care received near death: a prospective cohort study. J Clin Oncol 2012;30: Weber RS, Lewis CM. Eastman SD, et al. Quality and performance indicators in an academic department of head and neck surgery. Arch Otolaryngol Head Neck Surg 2010;136: Shellenberger TD, Madero-Vishal R, Weber RS. Quality indicators in head and neck operations. A comparison with published benchmarks. Arch Otolaryngol Head Neck Surg 2011;137: Snyder CF, Frick KD, Blackford AL, et al. How does initial treatment choice affect short and long-term costs for clinically localized prostate cancer? Cancer 2010;116: Gourin CG, Forastiere AA, Marur S, Sanguineti G, Koch WM, Bristow RE. Impact of surgeon and hospital volume on short-term outcomes and cost of laryngeal cancer surgical care. Laryngoscope 2011;121: Gourin CG, Frick KD. National trends in laryngeal cancer surgery and the effect of surgeon and hospital volume on short-term outcomes and cost of care. Laryngoscope 2012;122: Birkmeyer JD, Sun Y, Wong SL, Stukel TA. Hospital volume and late survival after cancer surgery. Ann Surg 2007;245: