STATE OF THE ART
The Theory, Practice, and Future of Process Improvement in General Thoracic Surgery Richard K. Freeman, MD, MBA* Process improvement, in its broadest sense, is the analysis of a given set of actions with the aim of elevating quality and reducing costs. The tenets of process improvement have been applied to medicine in increasing frequency for at least the last quarter century including thoracic surgery. This review outlines the theory underlying process improvement, the currently available data sources for process improvement and possible future directions of research. Semin Thoracic Surg 26:310–316 I 2015 Elsevier Inc. All rights reserved. Keywords: thoracic surgery, improvement, database, cost
Process improvement, in its broadest sense, is the analysis of a given set of actions with the aim of elevating quality and reducing costs. The origins of much of what have become the principles of process improvement had its origin within manufacturing and engineering. Essential to process improvement is the objective measure of performance characterized by the collection, analysis, and reporting of information pertaining to the performance of an individual, group, organization, or system. The tenets of process improvement have been applied to medicine in increasing frequency for at least the past quarter century. Health insurers seeking an estimation of the value of the health care, government organizations hoping to reduce the cost of health care, and clinicians with the desire to study their results have all applied the concepts of process improvement to various areas of medicine and surgery. However, such efforts, with notable exceptions, have lacked coordination, commonality, and meaning. Process improvement in health care is often referred to by the vernacular term “quality.” Medical quality, as defined by the Institute of Medicine in 1990, is the degree to which health care systems, services, and supplies for individuals and populations increase the likelihood for positive health outcomes and are consistent with current professional knowledge.1 The latter is described as “best *
Department of Thoracic and Cardiovascular Surgery, St Vincent Hospital, Indianapolis, Indiana. Dr Freeman reports equity ownership in St. Theresa. Address reprint requests to Richard K. Freeman, MD, MBA, 8433 Harcourt Road, Indianapolis, IN 46260. E-mail: [email protected]
310
practice” or “evidence-based practice” and may include the use of national guidelines constructed by specialty organizations, national cooperatives, and health systems. Although several quality of care frameworks exist including the World Health Organization–recommended Quality of Care Framework and the Bamako Initiative, the Donabedian model continues to be the most commonly encountered paradigm for assessing health care quality.2-4 Avedis Donabedian, a physician and health services researcher at the University of Michigan, developed this model in 1966. According to the Donabedian model, information about quality of care can be drawn from 3 categories of measures: structure, process, and outcomes. Structure describes the context in which care is delivered including providers, facilities, financing, and equipment. Process denotes the transactions between patients and providers throughout the delivery of health care. Finally, outcome refers to the effects of health care on the health status of patients and populations. A fourth category of metric termed intermediate outcomes has become increasingly more common in the field of medical process improvement.5 These outcomes are measures of clinical conditions that do not directly reflect patients’ quality or quantity of life. Examples are blood pressure or lipid control. Such variables are often more easily and timely measured than the true outcome in question such as coronary artery disease in these examples. Intermediate outcomes are also referred to as surrogate outcomes and require confirmation that they reliably relate to the true outcome being considered. Any discussion of metrics requires consideration of validity. In measurement theory, validity represents the extent to which a given measure actually 1043-0679/$-see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1053/j.semtcvs.2014.12.003
QUALITY IMPROVEMENT IN THORACIC SURGERY captures what it is supposed to measure.6 Within health care quality, 2 corollaries exist. First, process or structural metrics are valid only if improved performance on the given variable actually results in better health outcomes. The second is that the relationship between the measured, often surrogate, metric and the true outcome needs to be valid. Lastly, for a given measure of quality for a provider to be valid, it must be, in some significant amount, under the provider’s control.7 CURRENT ENVIRONMENT Within the specialty of noncardiac or general thoracic surgery, several ongoing national initiatives exist, which have been designed to assess and improve quality. Each of these systems relies on the underlying principles of the Donabedian quality framework in which surrogate metrics denoting the concept of quality for a given procedure or disease are identified, measured, and compared. The other major force in the specialty for quality assessment is scholarly investigations of specific procedures or disease processes using regional, national, and international administrative databases. It is important to the discussion of process improvement in general thoracic surgery to understand the attributes and limitations of these databases when reviewing the findings of what is a growing genre of literature. The following is a synopsis of the more common data sources relevant to general thoracic surgery.
Quality Reporting System and for the data to be submitted electronically. Criticisms of this database are that most participating institutions are tertiary referral hospitals or sites for cardiothoracic residency training programs or both and that few, if any, general surgeons participate. The former criticism may result in the database containing a population of patients with significant comorbidities or who require a complex surgical procedure, which could confound the conclusions of analysis. It is currently estimated that general surgeons perform most noncardiac surgeries in the United States.8 Hence, the latter criticism implies that a significant proportion of general thoracic surgery cases are not captured for analysis, and in the era of minimally invasive esophagectomy and lobectomy, the traditional procedures may be underrepresented.
CLINICAL DATABASES Society of Thoracic Surgeons General Thoracic Surgery Database The most recognized database within the specialty is the Society of Thoracic Surgeons (STS) General Thoracic Surgery Database initiated in 2002. Based on the unprecedented success of the Society for Thoracic Surgery’s Adult Cardiac Surgery Database, this database currently represents more than 371,000 submitted cases. The database is administered through the Duke University Clinical Research Institute. The General Thoracic Surgery Database is organized around data sheets submitted for each major thoracic surgery procedure a program performs. Data are harvested and analyzed twice each year. The database is not only used to compare raw outcome data between programs but also to create models that allow the risk adjustment of morbidity and mortality for specific procedures such as lobectomy or esophagectomy. Site auditing of database submissions has also been initiated to validate submitted information. Efforts also continue to use data elements from the database for the Physician
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
STS Database Risk Models: Predictors of Mortality and Major Morbidity for Lung Cancer Resection9 Background The aim of this study is to create models for perioperative risk of lung cancer resection using the STS General Thoracic Database (GTDB). Methods The STS GTDB was queried for all patients treated with resection for primary lung cancer between January 1, 2002, and June 30, 2008. Three separate multivariable risk models were constructed (mortality, major morbidity, and composite mortality or major morbidity). Results There were 18,800 lung cancer resections performed at 111 participating centers. Perioperative mortality was 413 of 18,800 (2.2%). Composite major morbidity or mortality occurred in 1612 patients (8.6%). Predictors of mortality include the following: pneumonectomy (P o 0.001), bilobectomy (P o 0.001), American Society of Anesthesiology rating (P o 0.018), Zubrod performance status (P o 0.001), renal dysfunction (P ¼ 0.001), induction chemoradiation therapy (P ¼ 0.01), steroids (P ¼ 0.002), age (P o 0.001), urgent procedures (P ¼ 0.015), male gender (P ¼ 0.013), forced expiratory volume in 1 second (P o 0.001), and body mass index (P ¼ 0.015). Conclusions Thoracic surgeons participating in the STS GTDB perform lung cancer resections with a low mortality and morbidity. The riskadjustment models created have excellent
311
QUALITY IMPROVEMENT IN THORACIC SURGERY performance characteristics and identify important predictors of mortality and major morbidity for lung cancer resections. These models may be used to inform clinical decisions and to compare risk-adjusted outcomes for quality improvement purposes. (Reprinted with permission from Kozower et al.9) NATIONAL SURGICAL QUALITY IMPROVEMENT PROGRAM The National Surgical Quality Improvement Program (NSQIP) was initiated in 44 Veterans Administration (VA) hospitals in 1994. This program was the next evolution of its progenitor the National VA Surgical Risk Study, which, between 1991 and 1993, allowed a comparative measurement of the quality of surgical care in 9 specialties within these hospitals. In 1999, a pilot program established hospitals outside the VA system within the NSQIP. In partnership with the American College of Surgeons, enrollment for hospital outside the VA system in NSQIP was opened in 2004. NSQIP has continued to grow in number of affiliated hospitals and accrued cases and allowed the VA to report a reduction in surgical mortality of 47% between 1991 and 2006.10 The primary criticism of NSQIP is that it does not abstract consecutive patients. Rather it is a random sampling of patients undergoing a given procedure. Strengths of the program include a required data monitor, the ability to sample any provider performing a given procedure, and the advanced analytics available. Morbidity of Lung Resection After Prior Lobectomy: Results from the Veterans Affairs National Surgical Quality Improvement Program11 Background Lobectomy is the current standard operation for localized lung cancer. Patients who undergo lobectomy have a 1% to 2% chance per year of developing a second lung cancer. The risks of repeat lung resection have not been well quantified or analyzed. We used a national, prospectively recorded database to evaluate the complication rate and risk factors in this population. Methods The Veterans Affairs National Surgical Quality Improvement Program Database was queried for all patients who underwent lobectomy, followed by an additional lung resection, between 1994
312
and 2002. Preoperative variables, intraoperative variables, and complications were analyzed. Pulmonary function data were not collected. Results Excluding 17 patients who underwent repeat resection for complications of lobectomy, 186 patients underwent 191 repeat resections. The 30-day mortality was 11%; the complication rate was 19%. Mortality for pneumonectomy was 34%; lobectomy, 7%; segmentectomy, 0%; and wedge resection, 6%. The most frequent complications were pneumonia (9%), reintubation (8%), ventilator dependence (6%), cardiac arrest (3%), dysrhythmia (3%), and sepsis (3%). Multivariate analysis revealed that operative time exceeding 2 hours, preoperative dyspnea at rest or with minimal exertion, and white blood cell count of more than 10,000/ mm3 were predictors of complication. Presence of a contaminated or infected case, pneumonectomy, and intraoperative transfusion were predictors of death. Age, complications from prior lobectomy, time interval between lobectomy and repeat resection, smoking history, other comorbidities, and preoperative laboratory values were not independent predictors. Conclusions Repeat lung resection after lobectomy carries an 11% overall mortality predicted by the presence of a contaminated/infected case, need for intraoperative transfusion, and pneumonectomy versus a lesser resection. (Reprinted with permission from Linden et al.11)
NATIONAL CANCER REGISTRY The National Cancer Database is a joint program of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. It is a nationwide database of patients diagnosed with a malignancy at one of the more than 1500 commission-accredited cancer programs in the United States and Puerto Rico. Data for most cancers are tracked in an electronic database, which includes patient demographics, cancer staging, tumor histologic characteristics, treatment, and survival. This database captures approximately 75% of newly diagnosed cancers. The main limitation of this database is that cohorts are identified from hospitals where they present for diagnosis and/or treatment. This may limit the generalizability of the patient population and
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
QUALITY IMPROVEMENT IN THORACIC SURGERY management algorithms. Other weaknesses have included the lag time included in the retrospective patient updates, patients lost to follow-up, and absence of comprehensive auditing. However, this database can be especially powerful when local registry data can be linked to other clinical and nonclinical data sources. A Comparison of Quality and Cost Indicators by Surgical Specialty for Lobectomy of the Lung12 Objectives This investigation compared patients undergoing lobectomy for non–small cell lung cancer by either a general surgeon or a cardiothoracic surgeon across a geographically diverse system of hospitals to see whether a significant difference in quality or cost was present. Methods The Premiere administrative database and tumor registry data of a single health system’s hospitals was used to compare adherence to national treatment guidelines, patient outcomes, and charges for patients undergoing lobectomy for non–small cell lung cancer in a 5-year period. Surgeons performing lobectomy were designated as a general surgeon or cardiothoracic surgeon according to their national provider number and board certification status. Excluded from analysis were centers that performed fewer than 50 lobectomies during the study period. Results During the study period, 2823 lobectomies were performed by 46 general surgeons and 3653 lobectomies were performed by 29 cardiothoracic surgeons in 54 hospitals in a single health care system. Significant differences were found between general and cardiothoracic surgeons with respect to adherence to national guidelines in staging and treatment, mean length of stay, significant morbidity, and operative mortality. Mean charges for lobectomy of the lung were also found to differ significantly between general and cardiothoracic surgeons. Conclusions This review found that currently measurable indicators for quality of care were significantly superior and overall charges were significantly reduced when a lobectomy for non–small cell lung cancer was performed by a cardiothoracic surgeon rather than by a general surgeon. (Reprinted with permission from Freeman et al.12)
THE SURVEILLANCE, EPIDEMIOLOGY, AND END RESULTS PROGRAM The Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute collects information on cancer incidence and survival in the United States. The program collects and publishes cancer incidence and survival data from 17 population-based cancer registries covering nearly 30% of the United States population. Data collected includes information about cancer stage and treatment. Criticisms of this database include the fact that therapy completion and long-term outcomes other than death are not recorded. In addition, the SEER database population is predominantly Medicare or Medicaid based. This fact creates a potential bias toward older subjects and, among older records, toward white patients. Furthermore, SEER also lacks recurrence, radiation, and chemotherapy data.
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
Invasive Adenocarcinoma with Bronchoalveolar Features: A Population-based Evaluation of the Extent of Resection in Bronchoalveolar Cell Carcinoma13 Objective We used a population-based data set to assess the association between the extent of pulmonary resection for bronchoalveolar carcinoma and survival. The reports thus far have been limited to small, institutional series. Methods Using the Surveillance, Epidemiology, and End Results database (1988-2007), we identified patients with bronchoalveolar carcinoma who had undergone wedge resection, segmentectomy, or lobectomy. The bronchoalveolar carcinoma histologic findings were mucinous, nonmucinous, mixed, not otherwise specified, and alveolar carcinoma. To adjust for potential confounders, we used a Cox proportional hazards regression model. Results A total of 6810 patients met the inclusion criteria. Compared with the sublobar resections (wedge resections and segmentectomies), lobectomy conferred superior 5-year overall (59.5% vs 43.9%) and cancer-specific (67.1% vs 53.1%) survival (P o .0001). After adjusting for potential confounding patient and tumor characteristics, we found that patients who underwent an anatomic resection had significantly better overall (segmentectomy: hazard ratio, 0.59; 95% confidence interval, 0.43-0.81;
313
QUALITY IMPROVEMENT IN THORACIC SURGERY lobectomy: hazard ratio, 0.50; 95% confidence interval, 0.44-0.57) and cancer-specific (segmentectomy: hazard ratio, 0.51; 95% confidence interval, 0.34-0.75; lobectomy: hazard ratio, 0.46; 95% confidence interval, 0.400.53) survival compared with patients who underwent wedge resection. Additionally, gender, race, tumor size, and degree of tumor dedifferentiation were negative prognostic factors. Our results were unchanged when we limited our analysis to early-stage disease. Conclusions Using a population-based data set, we found that anatomic resections for bronchoalveolar carcinoma conferred superior overall and cancer-specific survival rates compared with wedge resection. Bronchoalveolar carcinoma’s propensity for intraparenchymal spread might be the underlying biologic basis of our observation of improved survival after anatomic resection. (Reprinted with permission from Whitson et al.13)
This potential for the introduction of error in research when using administrative data is well illustrated in the investigation by Shahian et al14 in which administrative and clinical databases for isolated coronary artery bypass grafting in fiscal year 2003 were compared for the State of Massachusetts. Their investigation discovered a 27.4% disparity rate for case volume along with a 0.83% difference in observed in-hospital mortality between the administrative and clinical databases.
ADMINISTRATIVE DATABASES Over the past decade, outcomes research has grown exponentially with increased use of largevolume administrative and clinical databases. Administrative databases were not originally designed for clinical research. Rather they were constructed to track billing for hospitals, providers, and insurers. However, administrative databases offer several potential advantages to researchers. Compared with performing a large prospective study, access to administrative data is readily available and relatively inexpensive. Administrative databases also provide a mechanism to look at population-based data with large numbers of cases of patients collected over relatively long time intervals. Significant and appropriate concern exists regarding the use of administrative databases for medical process improvement research. The requirements for reimbursement mandate the nomenclature with which procedures are identified. Similarly, data input is accomplished to maximize revenue, but are relied upon in research efforts for the accuracy and validity of medical diagnoses. Because of factors not included in these databases such as tumor stage, quality of life, performance status, operative factors, and survival, administrative databases do not allow risk adjustments to be made when comparing outcomes, a concept considered an imperative when comparing patient outcomes following surgery.13
314
CENTERS FOR MEDICARE AND MEDICAID The United States Centers for Medicare and Medicaid Services collects data regarding reimbursements for services provided to beneficiaries. The files contain claims for inpatient, outpatient, and physician services for nearly 98% of the United States population 65 years and older. Patient demographic information is also collected. Criticisms of this database include the fact that the data are contained within separate files and require significant statistical manipulation to link different types of files for a given patient. Stratification of patients based on disease characteristics, such as stage of malignancy, is also not available. However, linkage to clinical registries such as the SEER database is possible. HEALTHCARE COST AND UTILIZATION PROJECT The Healthcare Cost and Utilization Project (HCUP) is a family of health care databases developed through a federal-state-industry partnership and sponsored by the Agency for Healthcare Research and Quality. The HCUP databases bring together the data collection efforts of state data organizations, hospital associations, private data organizations, and the federal government to create a national information resource of patient-level health care data. The HCUP databases facilitate research on issues including cost and accessibility of health services, practice patterns, and treatment outcomes at the regional and national levels. THE NATIONWIDE INPATIENT SAMPLE A member of the HCUP family, the National Inpatient Sample (NIS), is the largest all-payer inpatient care database in the United States, representing a 20% sample of US hospitals and approximately 8 million annual admissions. Patient discharge information is available by year, allowing national trends to be evaluated. The NIS’s large sample size enables analyses of rare conditions such as congenital anomalies, uncommon treatments such
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
QUALITY IMPROVEMENT IN THORACIC SURGERY as organ transplantation, and special patient populations such as the uninsured. Criticisms of the NIS database includes the fact that inpatient morbidity and mortality may be studied but is not comparable to 30-day morbidity and mortality available in other databases as only inpatient encounters are included. Although each unique admission has several diagnoses available to track variables such as comorbidities and demographics, it is often difficult to determine the specifics of an admission that are needed for outcome analysis. Additionally, no unique patient identifier is assigned in NIS making it impossible to track readmissions.
THE FUTURE Understanding the tenets of process improvement previously discussed in general thoracic surgery and the attributes and limitations of currently available data sets, what should the specialty of general thoracic surgery be measuring? The potential best answer to that question would be a collection of disease-specific composite scores reflecting process measures identified from evidence-based, best practices, outcome measure such as National Quality Forum patient safety “never” events and variables tracked through the STS General Thoracic Surgery database such as operative mortality, morbidity, length of hospital stay, and readmission. Externally applied metrics such as patient satisfaction and cost could also be included. The theory behind composite score used as a process improvement metrics is well outlined in the report of the STS Quality Measurement Task Force from 2007.15 The methodology for metric selection, validation, and individual attribution is thoughtfully discussed. For the purposes of the remainder of this discussion, we use patients with non–small cell lung cancer to illustrate processes, but similar discussions could focus on many other general thoracic surgery diseases such as esophageal cancer, end-stage pulmonary disease, mesothelioma, hyperhidrosis, and tracheomalacia. The selection of process variables can be daunting. One unique approach was reported by Darling et al16 using a modified Delphi consensus process, which essentially abstracted disease-specific metrics from evidence-based guidelines or literature or both. Surveys of physicians were then used to narrow the metrics based on consensus. Detterbeck et al and Cassivi et al used a similar consensus method to identify metrics for review in patients with non– small cell lung cancer.17,18 Unique to the efforts of Cassivi et al was the inclusion of National Quality
Forum patient safety “never” events in their dashboard. Brunelli et al similarly developed an index of 4 measures to track quality in their thoracic surgery unit—30-day or inpatient mortality, cardiopulmonary morbidity, unplanned emergency intensive care unit admission, and length of stay greater than 14 days.19 The difference in their composite or index measure from the studies previously mentioned is that they risk adjusted each measure before the individual rates being integrated into their “combined outcome index.” This resulted in a reproducible model that would appear to be appropriate for tracking internal program performance as well as comparing different programs’ results in the postoperative period. In summary, the goals of process improvement or “quality” within the field of general thoracic surgery are as they are in every field of medicine—reducing harm and improving the overall well being of individual patients and patient populations while remaining cognizant of society’s resources. How to best achieve them, however, will be a solution unique to our specialty and will require a more coordinated and inclusive effort than we have all put forth in the past. Although a separate monograph would be required to fully outline such a process, a brief discussion of such efforts segregated at the local, national, and international levels is offered as a starting point for discussion. At the local level, all surgeons performing general thoracic surgical procedures, regardless of their board certification(s), case mix, or practice type, should track and review their outcome data. This could take many forms but, as will be discussed, would benefit the specialty in an exponential manner if accomplished as a member of the STS General Thoracic Surgery database. Regardless of the mechanism, individual surgeon outcome data should be compared with national benchmarks and among other local surgeons at regular intervals. This not only fulfills local and national requirements such as maintenance of certification, ongoing professional practice evaluation, and focused professional practice evaluation but, in our experience, can rapidly change surgeon behavior when necessary. At the national level, we need to empower our specialty organizations to build on the solid foundation of process improvement that exists. Our societies should continue to influence federal agencies so that the quality indicators we are task with measuring are meaningful, valid, and under our control. We should ensure that the committees governing our database efforts are more inclusive going forward
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
315
QUALITY IMPROVEMENT IN THORACIC SURGERY than they have ever been, soliciting representatives from academic, private practice, military, and VA thoracic surgeons as well as general surgeon representation through the American College of Surgeons. Consideration should also be given to the feasibility of linking the STS database in a more permanent way with other databases such as NSQIP, SEER, National Cancer Registry, and Centers for Medicare and Medicaid Services to allow the continuum of care to be evaluated rather than isolated inpatient encounters. Lastly, the next generation of thoracic surgeons should be offered opportunities for formal education in outcomes or population health research through fellowships supported by our specialty
1. Institute of Medicine: America’s Health in Transition: Protecting and Improving Quality. Washington, DC: The National Academies Press; 1994 2. Shaw C: How can hospital performance be measured and monitored? Health Evidence Network Report. WHO Regional Office for Europe, Copenhagen. Available at: http:// www.euro.who.int/document/e82975. pdf. Accessed August 2014. 3. Knippenberg R, Alihonou E, Soucat A, et al: Implementation of the Bamako initiative: Strategies in Benin and Guinea 12: S29-S47, 1997 (suppl 1)Int J Health Plann Manage 12:S29-S47, 1997 (suppl 1) 4. Donabedian, A: An introduction to Quality Assurance in Health Care. (ed 1, Vol. 1). New York, NY: Oxford University Press; 2003. 5. Physician Consortium For Performance Improvement.Measurement development, methodology and oversight advisory committee: Recommendations to PCPI workgroups on outcomes measures. Chicago, IL, American Medical Association, 2011. 6. Hays RD, Fayers P: Reliability and validity. In Assessing Quality of Life in Clinical Trials: Methods and Practice, ed 2. New York, NY: Oxford University Press; 25, 2005 7. Friedberg MW, Damberg CL: Methodological Considerations in Generating Provider Performance Scores for Use in Public Reporting:
316
8.
9.
10.
11.
12.
13.
organizations, ensuring that our leadership in these areas continues to grow. Internationally, our specialty should continue to work with other national and international organizations to more formally standardize data collection and analysis methods. Study design and database comparisons could be greatly simplified. Comparative effectiveness research could then include international cohorts allowing more diverse populations to be examined more easily. It is only through the efforts of continuous improvement that our specialty has and will continue to provide safe, effective, and responsible care for our patients.
A Guide for Community Quality Collaboratives. Rockville, MD: Agency for Healthcare Research and Quality, US Department of Health and Human Services; 2011 Shipper PH, Diggs BS, Unger eider RM, et al: The influence of surgeon specialty on outcomes in general thoracic surgery: A national sample 1996 to 2005. Ann Thorac Surg 88: 1566-1573, 2009 Kozower BD, Sheng S, O’Brien SM, et al: STS database risk models: Predictors of mortality and major morbidity for lung cancer resection. Ann Thorac Surg 90:875-883, 2010 Chen ME, Bilimoria KY, Ko CY, et al: Development of an American College of Surgeons national quality improvement program: Morbidity and mortality risk calculator for colorectal surgery. J Am Coll Surg 26: 1009-1016, 2009 Linden PA, Yeap BY, Chang MY, et al: Morbidity of lung resection after prior lobectomy: Results from the Veterans Affairs National Surgical Quality Improvement Program. Ann Thorac Surg 83:425-432, 2007 Freeman RK, Dilts JR, Ascioti AJ, et al: A comparison of quality and cost indicators by surgical specialty for lobectomy of the lung. J Thorac Cardiovasc Surg 145:68-74, 2013 Whitson BA, Groth SS, Andrade RS, et al: Invasive adenocarcinoma with bronchoalveolar features: A population-based evaluation of the
14.
15.
16.
17.
18.
19.
extent of resection in bronchoalveolar cell carcinoma. J Thorac Cardiovasc Surg. 143: 591-600, 2012 Shahian DM, Torchiana DF, Shemin RJ, et al: Massachusetts cardiac surgery report card: Implications of statistical methodology. Ann Thorac Surg 80:2106-2113, 2005 O’Brien SM, Shahian DM, DeLong ER, et al: Quality measurement in adult cardiac surgery: Part 2–statistical considerations in composite measure scoring and provider rating. Ann Thorac Surg 83:S13-S26, 2007 (suppl 4) Darling G, Malthaner R, Dickie J, et al: Quality indicators for non-small cell lung cancer operations with use of a modified Delphi consensus process. Ann Thorac Surg 98:183-190, 2014 Mazzone PJ, Vachani A, Chang A, et al: Quality indicators for the evaluation of patients with lung cancer. Chest http://dx.doi.org/10.1378/ chest.13-2900 Brunelli A, Refai M, Salati M, et al: Standardized combined outcome index as an instrument for monitoring performance after pulmonary resection. Ann Thorac Surg 92:272-277, 2011 Cassivi SD, Allen MS, Vanderwaerdt GD, et al: Patient-centered quality indicators for pulmonary resection. Ann Thorac Surg 86: 927-932, 2008
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
The Theory, Practice, and Future of Process Improvement in General Thoracic Surgery Richard K. Freeman, MD, MBA* Process improvement, in its broadest sense, is the analysis of a given set of actions with the aim of elevating quality and reducing costs. The tenets of process improvement have been applied to medicine in increasing frequency for at least the last quarter century including thoracic surgery. This review outlines the theory underlying process improvement, the currently available data sources for process improvement and possible future directions of research. Semin Thoracic Surg 26:310–316 I 2015 Elsevier Inc. All rights reserved. Keywords: thoracic surgery, improvement, database, cost
Process improvement, in its broadest sense, is the analysis of a given set of actions with the aim of elevating quality and reducing costs. The origins of much of what have become the principles of process improvement had its origin within manufacturing and engineering. Essential to process improvement is the objective measure of performance characterized by the collection, analysis, and reporting of information pertaining to the performance of an individual, group, organization, or system. The tenets of process improvement have been applied to medicine in increasing frequency for at least the past quarter century. Health insurers seeking an estimation of the value of the health care, government organizations hoping to reduce the cost of health care, and clinicians with the desire to study their results have all applied the concepts of process improvement to various areas of medicine and surgery. However, such efforts, with notable exceptions, have lacked coordination, commonality, and meaning. Process improvement in health care is often referred to by the vernacular term “quality.” Medical quality, as defined by the Institute of Medicine in 1990, is the degree to which health care systems, services, and supplies for individuals and populations increase the likelihood for positive health outcomes and are consistent with current professional knowledge.1 The latter is described as “best *
Department of Thoracic and Cardiovascular Surgery, St Vincent Hospital, Indianapolis, Indiana. Dr Freeman reports equity ownership in St. Theresa. Address reprint requests to Richard K. Freeman, MD, MBA, 8433 Harcourt Road, Indianapolis, IN 46260. E-mail: [email protected]
310
practice” or “evidence-based practice” and may include the use of national guidelines constructed by specialty organizations, national cooperatives, and health systems. Although several quality of care frameworks exist including the World Health Organization–recommended Quality of Care Framework and the Bamako Initiative, the Donabedian model continues to be the most commonly encountered paradigm for assessing health care quality.2-4 Avedis Donabedian, a physician and health services researcher at the University of Michigan, developed this model in 1966. According to the Donabedian model, information about quality of care can be drawn from 3 categories of measures: structure, process, and outcomes. Structure describes the context in which care is delivered including providers, facilities, financing, and equipment. Process denotes the transactions between patients and providers throughout the delivery of health care. Finally, outcome refers to the effects of health care on the health status of patients and populations. A fourth category of metric termed intermediate outcomes has become increasingly more common in the field of medical process improvement.5 These outcomes are measures of clinical conditions that do not directly reflect patients’ quality or quantity of life. Examples are blood pressure or lipid control. Such variables are often more easily and timely measured than the true outcome in question such as coronary artery disease in these examples. Intermediate outcomes are also referred to as surrogate outcomes and require confirmation that they reliably relate to the true outcome being considered. Any discussion of metrics requires consideration of validity. In measurement theory, validity represents the extent to which a given measure actually 1043-0679/$-see front matter ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1053/j.semtcvs.2014.12.003
QUALITY IMPROVEMENT IN THORACIC SURGERY captures what it is supposed to measure.6 Within health care quality, 2 corollaries exist. First, process or structural metrics are valid only if improved performance on the given variable actually results in better health outcomes. The second is that the relationship between the measured, often surrogate, metric and the true outcome needs to be valid. Lastly, for a given measure of quality for a provider to be valid, it must be, in some significant amount, under the provider’s control.7 CURRENT ENVIRONMENT Within the specialty of noncardiac or general thoracic surgery, several ongoing national initiatives exist, which have been designed to assess and improve quality. Each of these systems relies on the underlying principles of the Donabedian quality framework in which surrogate metrics denoting the concept of quality for a given procedure or disease are identified, measured, and compared. The other major force in the specialty for quality assessment is scholarly investigations of specific procedures or disease processes using regional, national, and international administrative databases. It is important to the discussion of process improvement in general thoracic surgery to understand the attributes and limitations of these databases when reviewing the findings of what is a growing genre of literature. The following is a synopsis of the more common data sources relevant to general thoracic surgery.
Quality Reporting System and for the data to be submitted electronically. Criticisms of this database are that most participating institutions are tertiary referral hospitals or sites for cardiothoracic residency training programs or both and that few, if any, general surgeons participate. The former criticism may result in the database containing a population of patients with significant comorbidities or who require a complex surgical procedure, which could confound the conclusions of analysis. It is currently estimated that general surgeons perform most noncardiac surgeries in the United States.8 Hence, the latter criticism implies that a significant proportion of general thoracic surgery cases are not captured for analysis, and in the era of minimally invasive esophagectomy and lobectomy, the traditional procedures may be underrepresented.
CLINICAL DATABASES Society of Thoracic Surgeons General Thoracic Surgery Database The most recognized database within the specialty is the Society of Thoracic Surgeons (STS) General Thoracic Surgery Database initiated in 2002. Based on the unprecedented success of the Society for Thoracic Surgery’s Adult Cardiac Surgery Database, this database currently represents more than 371,000 submitted cases. The database is administered through the Duke University Clinical Research Institute. The General Thoracic Surgery Database is organized around data sheets submitted for each major thoracic surgery procedure a program performs. Data are harvested and analyzed twice each year. The database is not only used to compare raw outcome data between programs but also to create models that allow the risk adjustment of morbidity and mortality for specific procedures such as lobectomy or esophagectomy. Site auditing of database submissions has also been initiated to validate submitted information. Efforts also continue to use data elements from the database for the Physician
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
STS Database Risk Models: Predictors of Mortality and Major Morbidity for Lung Cancer Resection9 Background The aim of this study is to create models for perioperative risk of lung cancer resection using the STS General Thoracic Database (GTDB). Methods The STS GTDB was queried for all patients treated with resection for primary lung cancer between January 1, 2002, and June 30, 2008. Three separate multivariable risk models were constructed (mortality, major morbidity, and composite mortality or major morbidity). Results There were 18,800 lung cancer resections performed at 111 participating centers. Perioperative mortality was 413 of 18,800 (2.2%). Composite major morbidity or mortality occurred in 1612 patients (8.6%). Predictors of mortality include the following: pneumonectomy (P o 0.001), bilobectomy (P o 0.001), American Society of Anesthesiology rating (P o 0.018), Zubrod performance status (P o 0.001), renal dysfunction (P ¼ 0.001), induction chemoradiation therapy (P ¼ 0.01), steroids (P ¼ 0.002), age (P o 0.001), urgent procedures (P ¼ 0.015), male gender (P ¼ 0.013), forced expiratory volume in 1 second (P o 0.001), and body mass index (P ¼ 0.015). Conclusions Thoracic surgeons participating in the STS GTDB perform lung cancer resections with a low mortality and morbidity. The riskadjustment models created have excellent
311
QUALITY IMPROVEMENT IN THORACIC SURGERY performance characteristics and identify important predictors of mortality and major morbidity for lung cancer resections. These models may be used to inform clinical decisions and to compare risk-adjusted outcomes for quality improvement purposes. (Reprinted with permission from Kozower et al.9) NATIONAL SURGICAL QUALITY IMPROVEMENT PROGRAM The National Surgical Quality Improvement Program (NSQIP) was initiated in 44 Veterans Administration (VA) hospitals in 1994. This program was the next evolution of its progenitor the National VA Surgical Risk Study, which, between 1991 and 1993, allowed a comparative measurement of the quality of surgical care in 9 specialties within these hospitals. In 1999, a pilot program established hospitals outside the VA system within the NSQIP. In partnership with the American College of Surgeons, enrollment for hospital outside the VA system in NSQIP was opened in 2004. NSQIP has continued to grow in number of affiliated hospitals and accrued cases and allowed the VA to report a reduction in surgical mortality of 47% between 1991 and 2006.10 The primary criticism of NSQIP is that it does not abstract consecutive patients. Rather it is a random sampling of patients undergoing a given procedure. Strengths of the program include a required data monitor, the ability to sample any provider performing a given procedure, and the advanced analytics available. Morbidity of Lung Resection After Prior Lobectomy: Results from the Veterans Affairs National Surgical Quality Improvement Program11 Background Lobectomy is the current standard operation for localized lung cancer. Patients who undergo lobectomy have a 1% to 2% chance per year of developing a second lung cancer. The risks of repeat lung resection have not been well quantified or analyzed. We used a national, prospectively recorded database to evaluate the complication rate and risk factors in this population. Methods The Veterans Affairs National Surgical Quality Improvement Program Database was queried for all patients who underwent lobectomy, followed by an additional lung resection, between 1994
312
and 2002. Preoperative variables, intraoperative variables, and complications were analyzed. Pulmonary function data were not collected. Results Excluding 17 patients who underwent repeat resection for complications of lobectomy, 186 patients underwent 191 repeat resections. The 30-day mortality was 11%; the complication rate was 19%. Mortality for pneumonectomy was 34%; lobectomy, 7%; segmentectomy, 0%; and wedge resection, 6%. The most frequent complications were pneumonia (9%), reintubation (8%), ventilator dependence (6%), cardiac arrest (3%), dysrhythmia (3%), and sepsis (3%). Multivariate analysis revealed that operative time exceeding 2 hours, preoperative dyspnea at rest or with minimal exertion, and white blood cell count of more than 10,000/ mm3 were predictors of complication. Presence of a contaminated or infected case, pneumonectomy, and intraoperative transfusion were predictors of death. Age, complications from prior lobectomy, time interval between lobectomy and repeat resection, smoking history, other comorbidities, and preoperative laboratory values were not independent predictors. Conclusions Repeat lung resection after lobectomy carries an 11% overall mortality predicted by the presence of a contaminated/infected case, need for intraoperative transfusion, and pneumonectomy versus a lesser resection. (Reprinted with permission from Linden et al.11)
NATIONAL CANCER REGISTRY The National Cancer Database is a joint program of the Commission on Cancer of the American College of Surgeons and the American Cancer Society. It is a nationwide database of patients diagnosed with a malignancy at one of the more than 1500 commission-accredited cancer programs in the United States and Puerto Rico. Data for most cancers are tracked in an electronic database, which includes patient demographics, cancer staging, tumor histologic characteristics, treatment, and survival. This database captures approximately 75% of newly diagnosed cancers. The main limitation of this database is that cohorts are identified from hospitals where they present for diagnosis and/or treatment. This may limit the generalizability of the patient population and
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
QUALITY IMPROVEMENT IN THORACIC SURGERY management algorithms. Other weaknesses have included the lag time included in the retrospective patient updates, patients lost to follow-up, and absence of comprehensive auditing. However, this database can be especially powerful when local registry data can be linked to other clinical and nonclinical data sources. A Comparison of Quality and Cost Indicators by Surgical Specialty for Lobectomy of the Lung12 Objectives This investigation compared patients undergoing lobectomy for non–small cell lung cancer by either a general surgeon or a cardiothoracic surgeon across a geographically diverse system of hospitals to see whether a significant difference in quality or cost was present. Methods The Premiere administrative database and tumor registry data of a single health system’s hospitals was used to compare adherence to national treatment guidelines, patient outcomes, and charges for patients undergoing lobectomy for non–small cell lung cancer in a 5-year period. Surgeons performing lobectomy were designated as a general surgeon or cardiothoracic surgeon according to their national provider number and board certification status. Excluded from analysis were centers that performed fewer than 50 lobectomies during the study period. Results During the study period, 2823 lobectomies were performed by 46 general surgeons and 3653 lobectomies were performed by 29 cardiothoracic surgeons in 54 hospitals in a single health care system. Significant differences were found between general and cardiothoracic surgeons with respect to adherence to national guidelines in staging and treatment, mean length of stay, significant morbidity, and operative mortality. Mean charges for lobectomy of the lung were also found to differ significantly between general and cardiothoracic surgeons. Conclusions This review found that currently measurable indicators for quality of care were significantly superior and overall charges were significantly reduced when a lobectomy for non–small cell lung cancer was performed by a cardiothoracic surgeon rather than by a general surgeon. (Reprinted with permission from Freeman et al.12)
THE SURVEILLANCE, EPIDEMIOLOGY, AND END RESULTS PROGRAM The Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute collects information on cancer incidence and survival in the United States. The program collects and publishes cancer incidence and survival data from 17 population-based cancer registries covering nearly 30% of the United States population. Data collected includes information about cancer stage and treatment. Criticisms of this database include the fact that therapy completion and long-term outcomes other than death are not recorded. In addition, the SEER database population is predominantly Medicare or Medicaid based. This fact creates a potential bias toward older subjects and, among older records, toward white patients. Furthermore, SEER also lacks recurrence, radiation, and chemotherapy data.
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
Invasive Adenocarcinoma with Bronchoalveolar Features: A Population-based Evaluation of the Extent of Resection in Bronchoalveolar Cell Carcinoma13 Objective We used a population-based data set to assess the association between the extent of pulmonary resection for bronchoalveolar carcinoma and survival. The reports thus far have been limited to small, institutional series. Methods Using the Surveillance, Epidemiology, and End Results database (1988-2007), we identified patients with bronchoalveolar carcinoma who had undergone wedge resection, segmentectomy, or lobectomy. The bronchoalveolar carcinoma histologic findings were mucinous, nonmucinous, mixed, not otherwise specified, and alveolar carcinoma. To adjust for potential confounders, we used a Cox proportional hazards regression model. Results A total of 6810 patients met the inclusion criteria. Compared with the sublobar resections (wedge resections and segmentectomies), lobectomy conferred superior 5-year overall (59.5% vs 43.9%) and cancer-specific (67.1% vs 53.1%) survival (P o .0001). After adjusting for potential confounding patient and tumor characteristics, we found that patients who underwent an anatomic resection had significantly better overall (segmentectomy: hazard ratio, 0.59; 95% confidence interval, 0.43-0.81;
313
QUALITY IMPROVEMENT IN THORACIC SURGERY lobectomy: hazard ratio, 0.50; 95% confidence interval, 0.44-0.57) and cancer-specific (segmentectomy: hazard ratio, 0.51; 95% confidence interval, 0.34-0.75; lobectomy: hazard ratio, 0.46; 95% confidence interval, 0.400.53) survival compared with patients who underwent wedge resection. Additionally, gender, race, tumor size, and degree of tumor dedifferentiation were negative prognostic factors. Our results were unchanged when we limited our analysis to early-stage disease. Conclusions Using a population-based data set, we found that anatomic resections for bronchoalveolar carcinoma conferred superior overall and cancer-specific survival rates compared with wedge resection. Bronchoalveolar carcinoma’s propensity for intraparenchymal spread might be the underlying biologic basis of our observation of improved survival after anatomic resection. (Reprinted with permission from Whitson et al.13)
This potential for the introduction of error in research when using administrative data is well illustrated in the investigation by Shahian et al14 in which administrative and clinical databases for isolated coronary artery bypass grafting in fiscal year 2003 were compared for the State of Massachusetts. Their investigation discovered a 27.4% disparity rate for case volume along with a 0.83% difference in observed in-hospital mortality between the administrative and clinical databases.
ADMINISTRATIVE DATABASES Over the past decade, outcomes research has grown exponentially with increased use of largevolume administrative and clinical databases. Administrative databases were not originally designed for clinical research. Rather they were constructed to track billing for hospitals, providers, and insurers. However, administrative databases offer several potential advantages to researchers. Compared with performing a large prospective study, access to administrative data is readily available and relatively inexpensive. Administrative databases also provide a mechanism to look at population-based data with large numbers of cases of patients collected over relatively long time intervals. Significant and appropriate concern exists regarding the use of administrative databases for medical process improvement research. The requirements for reimbursement mandate the nomenclature with which procedures are identified. Similarly, data input is accomplished to maximize revenue, but are relied upon in research efforts for the accuracy and validity of medical diagnoses. Because of factors not included in these databases such as tumor stage, quality of life, performance status, operative factors, and survival, administrative databases do not allow risk adjustments to be made when comparing outcomes, a concept considered an imperative when comparing patient outcomes following surgery.13
314
CENTERS FOR MEDICARE AND MEDICAID The United States Centers for Medicare and Medicaid Services collects data regarding reimbursements for services provided to beneficiaries. The files contain claims for inpatient, outpatient, and physician services for nearly 98% of the United States population 65 years and older. Patient demographic information is also collected. Criticisms of this database include the fact that the data are contained within separate files and require significant statistical manipulation to link different types of files for a given patient. Stratification of patients based on disease characteristics, such as stage of malignancy, is also not available. However, linkage to clinical registries such as the SEER database is possible. HEALTHCARE COST AND UTILIZATION PROJECT The Healthcare Cost and Utilization Project (HCUP) is a family of health care databases developed through a federal-state-industry partnership and sponsored by the Agency for Healthcare Research and Quality. The HCUP databases bring together the data collection efforts of state data organizations, hospital associations, private data organizations, and the federal government to create a national information resource of patient-level health care data. The HCUP databases facilitate research on issues including cost and accessibility of health services, practice patterns, and treatment outcomes at the regional and national levels. THE NATIONWIDE INPATIENT SAMPLE A member of the HCUP family, the National Inpatient Sample (NIS), is the largest all-payer inpatient care database in the United States, representing a 20% sample of US hospitals and approximately 8 million annual admissions. Patient discharge information is available by year, allowing national trends to be evaluated. The NIS’s large sample size enables analyses of rare conditions such as congenital anomalies, uncommon treatments such
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
QUALITY IMPROVEMENT IN THORACIC SURGERY as organ transplantation, and special patient populations such as the uninsured. Criticisms of the NIS database includes the fact that inpatient morbidity and mortality may be studied but is not comparable to 30-day morbidity and mortality available in other databases as only inpatient encounters are included. Although each unique admission has several diagnoses available to track variables such as comorbidities and demographics, it is often difficult to determine the specifics of an admission that are needed for outcome analysis. Additionally, no unique patient identifier is assigned in NIS making it impossible to track readmissions.
THE FUTURE Understanding the tenets of process improvement previously discussed in general thoracic surgery and the attributes and limitations of currently available data sets, what should the specialty of general thoracic surgery be measuring? The potential best answer to that question would be a collection of disease-specific composite scores reflecting process measures identified from evidence-based, best practices, outcome measure such as National Quality Forum patient safety “never” events and variables tracked through the STS General Thoracic Surgery database such as operative mortality, morbidity, length of hospital stay, and readmission. Externally applied metrics such as patient satisfaction and cost could also be included. The theory behind composite score used as a process improvement metrics is well outlined in the report of the STS Quality Measurement Task Force from 2007.15 The methodology for metric selection, validation, and individual attribution is thoughtfully discussed. For the purposes of the remainder of this discussion, we use patients with non–small cell lung cancer to illustrate processes, but similar discussions could focus on many other general thoracic surgery diseases such as esophageal cancer, end-stage pulmonary disease, mesothelioma, hyperhidrosis, and tracheomalacia. The selection of process variables can be daunting. One unique approach was reported by Darling et al16 using a modified Delphi consensus process, which essentially abstracted disease-specific metrics from evidence-based guidelines or literature or both. Surveys of physicians were then used to narrow the metrics based on consensus. Detterbeck et al and Cassivi et al used a similar consensus method to identify metrics for review in patients with non– small cell lung cancer.17,18 Unique to the efforts of Cassivi et al was the inclusion of National Quality
Forum patient safety “never” events in their dashboard. Brunelli et al similarly developed an index of 4 measures to track quality in their thoracic surgery unit—30-day or inpatient mortality, cardiopulmonary morbidity, unplanned emergency intensive care unit admission, and length of stay greater than 14 days.19 The difference in their composite or index measure from the studies previously mentioned is that they risk adjusted each measure before the individual rates being integrated into their “combined outcome index.” This resulted in a reproducible model that would appear to be appropriate for tracking internal program performance as well as comparing different programs’ results in the postoperative period. In summary, the goals of process improvement or “quality” within the field of general thoracic surgery are as they are in every field of medicine—reducing harm and improving the overall well being of individual patients and patient populations while remaining cognizant of society’s resources. How to best achieve them, however, will be a solution unique to our specialty and will require a more coordinated and inclusive effort than we have all put forth in the past. Although a separate monograph would be required to fully outline such a process, a brief discussion of such efforts segregated at the local, national, and international levels is offered as a starting point for discussion. At the local level, all surgeons performing general thoracic surgical procedures, regardless of their board certification(s), case mix, or practice type, should track and review their outcome data. This could take many forms but, as will be discussed, would benefit the specialty in an exponential manner if accomplished as a member of the STS General Thoracic Surgery database. Regardless of the mechanism, individual surgeon outcome data should be compared with national benchmarks and among other local surgeons at regular intervals. This not only fulfills local and national requirements such as maintenance of certification, ongoing professional practice evaluation, and focused professional practice evaluation but, in our experience, can rapidly change surgeon behavior when necessary. At the national level, we need to empower our specialty organizations to build on the solid foundation of process improvement that exists. Our societies should continue to influence federal agencies so that the quality indicators we are task with measuring are meaningful, valid, and under our control. We should ensure that the committees governing our database efforts are more inclusive going forward
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
315
QUALITY IMPROVEMENT IN THORACIC SURGERY than they have ever been, soliciting representatives from academic, private practice, military, and VA thoracic surgeons as well as general surgeon representation through the American College of Surgeons. Consideration should also be given to the feasibility of linking the STS database in a more permanent way with other databases such as NSQIP, SEER, National Cancer Registry, and Centers for Medicare and Medicaid Services to allow the continuum of care to be evaluated rather than isolated inpatient encounters. Lastly, the next generation of thoracic surgeons should be offered opportunities for formal education in outcomes or population health research through fellowships supported by our specialty
1. Institute of Medicine: America’s Health in Transition: Protecting and Improving Quality. Washington, DC: The National Academies Press; 1994 2. Shaw C: How can hospital performance be measured and monitored? Health Evidence Network Report. WHO Regional Office for Europe, Copenhagen. Available at: http:// www.euro.who.int/document/e82975. pdf. Accessed August 2014. 3. Knippenberg R, Alihonou E, Soucat A, et al: Implementation of the Bamako initiative: Strategies in Benin and Guinea 12: S29-S47, 1997 (suppl 1)Int J Health Plann Manage 12:S29-S47, 1997 (suppl 1) 4. Donabedian, A: An introduction to Quality Assurance in Health Care. (ed 1, Vol. 1). New York, NY: Oxford University Press; 2003. 5. Physician Consortium For Performance Improvement.Measurement development, methodology and oversight advisory committee: Recommendations to PCPI workgroups on outcomes measures. Chicago, IL, American Medical Association, 2011. 6. Hays RD, Fayers P: Reliability and validity. In Assessing Quality of Life in Clinical Trials: Methods and Practice, ed 2. New York, NY: Oxford University Press; 25, 2005 7. Friedberg MW, Damberg CL: Methodological Considerations in Generating Provider Performance Scores for Use in Public Reporting:
316
8.
9.
10.
11.
12.
13.
organizations, ensuring that our leadership in these areas continues to grow. Internationally, our specialty should continue to work with other national and international organizations to more formally standardize data collection and analysis methods. Study design and database comparisons could be greatly simplified. Comparative effectiveness research could then include international cohorts allowing more diverse populations to be examined more easily. It is only through the efforts of continuous improvement that our specialty has and will continue to provide safe, effective, and responsible care for our patients.
A Guide for Community Quality Collaboratives. Rockville, MD: Agency for Healthcare Research and Quality, US Department of Health and Human Services; 2011 Shipper PH, Diggs BS, Unger eider RM, et al: The influence of surgeon specialty on outcomes in general thoracic surgery: A national sample 1996 to 2005. Ann Thorac Surg 88: 1566-1573, 2009 Kozower BD, Sheng S, O’Brien SM, et al: STS database risk models: Predictors of mortality and major morbidity for lung cancer resection. Ann Thorac Surg 90:875-883, 2010 Chen ME, Bilimoria KY, Ko CY, et al: Development of an American College of Surgeons national quality improvement program: Morbidity and mortality risk calculator for colorectal surgery. J Am Coll Surg 26: 1009-1016, 2009 Linden PA, Yeap BY, Chang MY, et al: Morbidity of lung resection after prior lobectomy: Results from the Veterans Affairs National Surgical Quality Improvement Program. Ann Thorac Surg 83:425-432, 2007 Freeman RK, Dilts JR, Ascioti AJ, et al: A comparison of quality and cost indicators by surgical specialty for lobectomy of the lung. J Thorac Cardiovasc Surg 145:68-74, 2013 Whitson BA, Groth SS, Andrade RS, et al: Invasive adenocarcinoma with bronchoalveolar features: A population-based evaluation of the
14.
15.
16.
17.
18.
19.
extent of resection in bronchoalveolar cell carcinoma. J Thorac Cardiovasc Surg. 143: 591-600, 2012 Shahian DM, Torchiana DF, Shemin RJ, et al: Massachusetts cardiac surgery report card: Implications of statistical methodology. Ann Thorac Surg 80:2106-2113, 2005 O’Brien SM, Shahian DM, DeLong ER, et al: Quality measurement in adult cardiac surgery: Part 2–statistical considerations in composite measure scoring and provider rating. Ann Thorac Surg 83:S13-S26, 2007 (suppl 4) Darling G, Malthaner R, Dickie J, et al: Quality indicators for non-small cell lung cancer operations with use of a modified Delphi consensus process. Ann Thorac Surg 98:183-190, 2014 Mazzone PJ, Vachani A, Chang A, et al: Quality indicators for the evaluation of patients with lung cancer. Chest http://dx.doi.org/10.1378/ chest.13-2900 Brunelli A, Refai M, Salati M, et al: Standardized combined outcome index as an instrument for monitoring performance after pulmonary resection. Ann Thorac Surg 92:272-277, 2011 Cassivi SD, Allen MS, Vanderwaerdt GD, et al: Patient-centered quality indicators for pulmonary resection. Ann Thorac Surg 86: 927-932, 2008
Seminars in Thoracic and Cardiovascular Surgery Volume 26, Number 4
