Methodology for Lung Cancer Evidence Review and Guideline Development* ACCP Evidence-Based Clinical Practice Guidelines (2nd Edition)

Transcription

1 Supplement DIAGNOSIS AND MANAGEMENT OF LUNG CANCER: ACCP GUIDELINES Methodology for Lung Cancer Evidence Review and Guideline Development* ACCP Evidence-Based Clinical Practice Guidelines (2nd Edition) Douglas C. McCrory, MD, MHS; Sandra Zelman Lewis, PhD; Julia Heitzer, MS; Gene Colice, MD, FCCP; and W. Michael Alberts, MD, FCCP Background: To assemble a geographically diverse panel of experts in the diagnosis and treatment of lung cancer, representative of multiple clinical specialties, with the intention of developing clinically relevant practice guidelines for chest medicine and primary care physicians, including recommendations covering the full spectrum of care of the patient with non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC). Methods: The Duke University Center for Clinical Health Policy Research was selected to review and summarize the current evidence in the treatment of NSCLC. The BlueCross BlueShield Association Technology Evaluation Center was chosen and funded by the Agency for Healthcare Research and Quality to review and synthesize the current evidence on treatment of SCLC. Other chapters received existing guidelines, systematic reviews, and metaanalyses that were published since the first edition of these guidelines, as collected by the Duke University Evidence-based Practice Center. The writing committees for these chapters conducted searches for the primary articles and additional evidence in their topic area. The expert panel established clinical recommendations founded on the synthesis of this evidence. Conclusions: This section describes the approach used to develop the guidelines, including identifying, evaluating, and synthesizing the evidence, assessing the strength of evidence and grading the individual recommendations, and suggestions for implementation of the guidelines. (CHEST 2007; 132:23S 28S) Key words: clinical practice guidelines; evidence-based medicine; non-small cell lung cancer; small cell lung cancer Abbreviations: ACCP American College of Chest Physicians; AHRQ Agency for Healthcare Research and Quality; EPC Evidence-based Practice Center; HSP Health and Science Policy; NSCLC non-small cell lung cancer; RCT randomized controlled trial; SCLC small cell lung cancer The American College of Chest Physicians (ACCP) through its Health and Science Policy (HSP) Committee develops evidence-based clinical *From the Department of Medicine and the Center for Clinical Health Policy Research (Dr. McCrory), Duke University Medical Center, Durham, NC; Center for Health Services Research in Primary Care (Dr. Durham), Department of Veterans Affairs Medical Center, Durham, NC; American College of Chest Physicians (Dr. Lewis and Ms. Heitzer), Northbrook, IL; Division of Pulmonary and Critical Care Medicine (Dr. Colice), Washington Hospital Center and The George Washington University School of Medicine, Washington, DC; and H. Lee Moffitt Cancer Center and the University of South Florida (Dr. Alberts), Tampa, FL. Dr. McCrory was supported by a contract from the ACCP. Drs. Lewis, Colice, Alberts, and Heitzer have no conflicts of interest to disclose. practice guidelines to assist the practicing clinician, patient, and researcher in the diagnosis and management of various cardiopulmonary diseases. In 2003, the ACCP developed the Diagnosis and Management of Lung Cancer: ACCP Evidence-Based Guidelines. 1 Since publication of this evidence-based guideline, the field of lung cancer has continued to Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Douglas McCrory, MD, MHS, Duke Center for Clinical Health Policy Research, 2200 W. Main St, Suite 220, Durham, NC 27705; douglas.mccrory@duke.edu DOI: /chest CHEST / 132 / 3/ SEPTEMBER, 2007 SUPPLEMENT 23S

2 expand and evolve. Consequently, the ACCP Board of Regents authorized the HSP Committee to undertake a revision of this guideline. This Second Edition seeks to both update evidence supporting the original guidelines, and to expand the scope of the guidelines to include additional relevant areas. Panel Composition The Chair and Co-Chair of the Lung Cancer Guideline Panel were selected by the HSP Committee based on their experience in lung cancer and guideline development. The Chair and Co-Chair selected an Executive Committee to assist in planning the chapter outlines and identifying chapter editors. Nominations for Writing Committee members for the individual chapters were solicited from the chapter editors, ACCP membership, and the ACCP cancer clinical network. The chapter editors and Writing Committee members were approved by the HSP Committee based on established criteria for membership in the guideline development process. The HSP Committee approval process included review by the HSP Subcommittee on Policies and Procedures, which vetted each nominee to ensure that they met the criteria established in the HSP Authorship Policy, and that they did not have conflicts of interest that would preclude their ability to participate or clash with the HSP policy on conflicts of interest. An international group of 90 experts, composed of pulmonologists, medical oncologists, radiation oncologists, thoracic surgeons, integrative medicine specialists, oncology nurses, pathologists, health-care researchers, and epidemiologists, was selected through this process to participate as Writing Committee members. In addition, medical and nursing specialty societies and patient advocacy organizations, which have a vested interest in lung cancer, were invited to send representatives to participate in the review of manuscripts and development of recommendations at the final conference. Those accepting the invitation to provide such representation include the American Association for Bronchology; American Cancer Society; American College of Surgeons Oncology Group; American Society of Clinical Oncology; American Society for Therapeutic Radiology and Oncology; American Thoracic Society; Canadian Thoracic Society; International Association for the Study of Lung Cancer; The Lung Cancer Alliance; National Comprehensive Cancer Network; Oncology Nurses Society; and Society of Thoracic Surgeons. Many of these organizations were also represented in the First Edition. The Executive Panel met for planning purposes on several occasions. The entire Expert Panel convened in July At this conference, the panel reviewed the content of all chapter recommendations and the rating of the quality of the evidence supporting these recommendations. The method used to grade the strength of the recommendations had been developed by the HSP Committee and was based on the source of the clinical data (eg, randomized controlled clinical trial vs case reports) and an estimation of the balance of benefits to harm for the patient population (see grading system, below). This conference was followed by a series of conference calls with chapter editors to ensure that the recommendations closely followed the evidence and did not contradict those of other chapters. Target Audience The ACCP promotes interdisciplinary coordination for patient-focused care. 2 The panel of experts provided clinically relevant recommendations synthesized from the results of the evidence review and targeted toward an audience of pulmonologists, oncologists, thoracic surgeons, and primary care physicians who manage patients with lung cancer. In addition, patients with lung cancer and their family members, oncology nurses, hospice workers, chaplains, social workers, and psychologists are expected to gain insight from this set of guidelines. Scope The Chair, Co-Chair, and Executive Committee chose to include in the Second Edition of the guidelines most topics from the First Edition. These chapters were not intended to replicate the evidence that was reviewed in the First Edition, but were expanded to include both additional evidence published since that First Edition and subareas of content that were not included in the first version. This edition of the lung cancer guidelines was intended to cover the full spectrum of care domains, from prevention and screening, to diagnosis and staging, treatment, follow-up, and surveillance, palliative, and end-of-life care. The executive committee also determined that selected areas, not included in the First Edition, should be incorporated into the Second Edition of the guidelines, including pathology, bronchoalveolar cell carcinoma, and integrative oncology. Other malignancies that may present in the lung (eg, mesothelioma, hamartoma, thymoma, and carcinoid and neuroendocrine tumors) were not included in this edition of the guidelines but might be considered for addition in future guidelines. 24S Diagnosis and Management of Lung Cancer: ACCP Guidelines

3 In the original guideline, evidence reviews were performed of all guidelines, systematic reviews, and metaanalyses pertinent to lung cancer for all topic areas. In addition, five topic areas were selected for complete systematic reviews of the literature. These five topic areas were screening, prevention, diagnosis, noninvasive staging, and invasive staging. For this edition of the guidelines, the Executive Committee was again provided sufficient resources to fund a general review of guidelines, systematic reviews, and metaanalyses pertinent to all topic areas. Also, adequate funds were available to conduct complete systematic reviews of the literature in five new topic areas: evaluations of solitary pulmonary nodules, stage I and II non-small cell lung cancer (NSCLC), stage IIIA NSCLC, stage IIIB NSCLC, and stage IV NSCLC. The evidence review for the treatment of small cell lung cancer (SCLC) was funded by the Agency for Healthcare Research and Quality (AHRQ). Some support was available to update the original indepth systematic literature reviews. In other individual chapters, literature reviews were performed by the chapter writing committee who conducted literature searches based on selected criteria pertinent to the specific topic areas. Funding and Conflicts of Interest Funding for both the evidence review and guideline development was supported by educational grants from AstraZeneca LP, Bristol-Myers Squibb Company, Eli Lilly and Company, Genentech, and sanofi-aventis. Representatives from these companies were neither granted the right of review, nor were they allowed participation in any portion of the guideline development process. This precluded participation in either conference calls or conferences. No panel members or ACCP reviewers were paid any honoraria for their participation in the development and review of these guidelines. The ACCP approach to the issue of potential or perceived conflicts of interest established clear firewalls to ensure that the guideline development process was not influenced by industry sources. This policy is published on the ACCP Web site at www. chestnet.org. All conflicts of interest within the preceding 5 years were required to be disclosed by all panelists, including those who did not have writing responsibilities, at all face-to-face meetings, the final conference, and prior to submission for publication. The most recent of these conflict of interests are documented in this guideline Supplement. Furthermore, the panel was instructed in this matter, verbally and in writing, prior to the deliberations of the final conference. Any disclosed memberships on speaker s bureaus, consultant fees, grants and other research monies, and any fiduciary responsibilities to industry were provided to the full panel in writing at the beginning of the conference and at submission for publication. Evidence Review The ACCP published a request for proposals intended to identify Evidence-based Practice Centers (EPCs) that would be capable of providing both the general and indepth reviews. After review of these proposals, the ACCP chose the Duke University Center for Clinical Health Policy Research, an AHRQ-sanctioned EPC, to perform formal systematic reviews of the current evidence in the five new NSCLC topic areas, as well as to provide a search for the existing guidelines, systematic reviews, and metaanalyses in all of the topics areas. In addition, the AHRQ agreed to fund the BlueCross BlueShield Association Technology Evaluation Center to perform the formal systematic review of literature on SCLC. The Health Outcomes Research Group of the Department of Epidemiology and Biostatistics at Memorial Sloan-Kettering Cancer Center conducted a full-scale review of the literature since the first set of guidelines in the area of screening for lung cancer to assist that particular writing group. The formal systematic reviews of the five new topic areas were guided by the appropriate chapter editors and their writing committees, in concert with the Executive Committee of the panel. The Executive Committee included Douglas McCrory, MD, MHS, methodologist and primary investigator for the formal systematic reviews. The writing committees for each chapter identified the important treatment issues for which clinical guidance is needed and expressed these as questions. These research questions were used as the bases for the formal systematic literature reviews. The writing committees for the other chapters were provided published guidelines, systematic reviews, and metaanalyses identified as part of the general review of this field by the Duke EPC. Additional computerized searches of literature databases were performed by the writing committees to supplement this material. The two EPC research teams conducted a variety of systematic computerized bibliographic database searches including the following: (1) a search for systematic reviews, guidelines, and metaanalyses published since the last ACCP lung cancer guideline (MEDLINE, The Cochrane Library, National Guidelines Clearinghouse); (2) targeted searches for reviews in each of five selected treatment sections (solitary pulmonary nodules, stage I and II, stage CHEST / 132 / 3/ SEPTEMBER, 2007 SUPPLEMENT 25S

4 IIIA, stage IIIB, stage IV); these searches, run in OVID version of MEDLINE, were performed in July and August 2005 and were limited to publication years since 1995, English language, and human subjects; and (3) searches related to SCLC are described in the evidence chapter on SCLC. Search terms included the medical subject heading terms lung neoplasms (exploded) and bronchial neoplasms for the lung cancer concept. Each topic search utilized key words specific to the key questions of interest (complete search strategies are available on request from the authors). The studies identified in this search were provided to the Treatment chapter authors and in some cases described in evidence Tables, or more extensive reports. Consistent with the most recent grading scheme for recommendations, individual studies were rated according to study design, and important limitations (inconsistent results, methodologic flaws, indirect, or imprecise) were noted in order to facilitate grading of the body of evidence supporting a recommendation (see below). Grading of the Evidence and Recommendations The lung cancer panel was provided guidelines on the grading of evidence and wording of recommendations by the HSP Committee. This guidance was based on work performed by the ACCP Task Force on Grading. This task force reviewed several grading systems that were in place as of March 2005, including the version that had been currently in use by the ACCP for the First Edition of the lung cancer guidelines, 1 with the goal of developing an improved system for future guidelines. The report 3 of this task force explained that the optimum approach would be to meld the best characteristics of several grading systems, creating a more user-friendly and transparent grading system, as described below and in Tables 1 4. This guideline panel was one of the first to use the new scheme. The ACCP grading system is composed of only two types of recommendations (strong and weak) and two dimensions (the ratio of benefit to harm and the quality of evidence). The benefit-to-harm ratio Table 2 Balance of Benefits to Risks/Burdens Scale Benefits clearly outweigh the risks and burdens Risks and burdens clearly outweigh the benefits The risks/burdens and benefits are closely balanced The balance of benefits to risks and burdens is uncertain Certainty of imbalance Certainty of imbalance Less certainty Uncertainty Table 3 Quality of Evidence Scale High RCTs without important limitations or overwhelming evidence from observational studies* Moderate RCTs with important limitations (inconsistent results, methodologic flaws, indirect, or imprecise) or exceptionally strong evidence from observational studies* Low or very low Observational studies or case series *Although the determination of magnitude of the effect based on observational studies is often a matter of judgment, we offer the following suggested rule to assist this decision: a large effect would be a relative risk 2 (risk ratio 0.5) which would justify moving from weak to moderate, and a very large effect is a relative risk 5 (risk ratio 0.2) which would justify moving from weak to strong. There is some theoretical justification in the statistical literature for these thresholds (the magnitude of effect that is unlikely or very unlikely to be due to residual confounding after adjusted analysis). However, once the decision is made, authors should be explicit in justifying their decisions. includes consideration of the clinical improvements in health and quality of life, as well as the burdens, risks, and costs, when applicable, identifiable, and determinable (Table 2). If the advantages of the recommended procedure, service, test, or treatment are greater than the disadvantages or if the disadvantages outweigh the advantages, the benefit-to-harm ratio is said to be imbalanced. Either way, it is clear that there is a direction to the recommendation, positive or negative. When the benefits and harms are more evenly balanced, recommendations are not as strong and patient preferences play a larger role. The same holds true when the balance of benefits and burdens is not clear. Quality of evidence is the second dimension of the Table 1 Relationship of Strength of the Supporting Evidence to the Balance of Benefits to Risks and Burdens Balance of Benefits to Risks and Burdens Quality of Evidence Benefits Outweigh Risks/Burdens Risks/Burdens Outweigh Benefits Evenly Balanced Uncertain High 1A 1A 2A Moderate 1B 1B 2B Low or very low 1C 1C 2C 2C 26S Diagnosis and Management of Lung Cancer: ACCP Guidelines

5 Grade 1A 1B 1C 2A 2B 2C Table 4 Grade of Recommendations Scale Recommendation Strong Strong Strong Weak Weak Weak grading chart. Quality of evidence is scored in three categories with high-quality evidence obtained from randomized controlled trials (RCTs) without important methodologic limitations based on the study design, the consistency of the results, and the directness of the evidence. In extraordinary circumstances, significant and consistent evidence from observational studies could also be ranked as high quality. RCTs with important methodologic limitations or flaws, inconsistent results, or indirect or imprecise results would be scored as medium quality, as well as exceptionally strong evidence from observational studies. Other observational studies or case-series data would fall into the low quality of evidence category. It is the interface of the quality of the evidence and the balance of benefits to harms or burdens that determines the strength of the recommendation, with a 1A recommendation being the strongest and 2C the weakest, according to the schema in Table 1, which is further clarified in Tables 2 4. Guideline Writing and Validation Process Writing committees studied the evidence and summary tables or reviewed the literature for their assigned topics, developing their arguments for the recommendations and suggested grading of those recommendations that were put forth for early drafts. The Executive Committee of the panel, composed of the Chair, Vice-Chair, methodologist, and both project managers, reviewed drafts of each chapter of the manuscript during the writing process. Sections that were determined to be potentially overlapping were shared among the appropriate chapter editors, and conference calls were organized to coordinate the placement of these sections and to confirm that there would be no conflicting information or recommendations. A conference of the panel was convened in July 2006, prior to which time all panelists, including representatives from the invited organizations, were requested to review the complete manuscript and identify recommendations for which the proposal, wording, or grading were determined to be controversial or could be interpreted as controversial by others, incorrectly evolved from the evidence, disagreement existed with regard to the proposal or the grading, or required full panel discussion and further review for any reason. When the panelists who were present were not in unanimous agreement with the proposed recommendations or the grading of the recommendations, informal group consensus techniques were employed. After the meeting, a series of conference calls were convened to finish the discussions and finalize the recommendations. There were a few chapters for which there was insufficient time for full dialogue during the meeting; in the interest of ensuring that the recommendations followed the evidence, the conference calls were necessary. This process ensured the buy-in of the panelists and was deemed to be a worthwhile effort. Following final chapter revisions and incorporation of these ultimate recommendations and grading, a concluding review was conducted by the guideline panel Executive Committee. The guidelines were then submitted for review and approval to the ACCP HSP Committee, as well as the Thoracic Oncology Network of the college. The HSP Committee review of the chapters was divided among the members of the committee; in addition, two members were assigned to read all 26 chapters to identify any inconsistencies. HSP members were charged with verifying if the methodology was according to ACCP standards, the literature was well described, the recommendations were based on the evidence, and the grading was accurate. All but four of the chapters were approved on the first review by the HSP Committee. The remaining four chapters were returned to the authors, who responded to comments from the reviewers and the full committee. Written responses and revised chapters were returned to the two designated reviewers of the HSP Committee for final approval, which was received on the second round. Because two of the revised chapters included additional evidence, even though the recommendations did not change, the chapter editors were requested to review and approve the four revised chapters. A conference call was convened to discuss and vote on the acceptance of the recommendations. After approval, the guidelines were forwarded for additional reviews by the Board of Regents. In addition, the journal CHEST contracted for outside independent reviews. All reviewers comments followed the manuscript through every stage of the review process, including submission for publication. These guidelines have not been field tested, although informal feedback on the First Edition guidelines was used to inform the revised scope and CHEST / 132 / 3/ SEPTEMBER, 2007 SUPPLEMENT 27S

6 foci of the Second Edition. Those organizations sending representatives to the final conference were requested to review the final manuscript for endorsement. Endorsements were obtained from the following organizations: American Association for Bronchology, American Association for Thoracic Surgery, American College of Surgeons Oncology Group, American Society for Therapeutic Radiology and Oncology, Asian Pacific Society of Respirology, Oncology Nurses Society, Society of Thoracic Surgeons, and World Association of Bronchology. Dissemination and Implementation The publication of the Diagnosis and Management of Lung Cancer: ACCP Evidence-Based Clinical Practice Guidelines; Second Edition in CHEST is the first of two dissemination vehicles. The circulation of the journal is 23,000 subscribers and libraries, including six translations and distribution to 107 countries. All subscribers received a copy of this full-text guideline. The ACCP Clinical Resource on Lung Cancer is composed of a printed publication and an accompanying CD-ROM, containing a quick reference guide for physicians and other health-care providers, patient-targeted educational materials, and a set of slides for use in educational or clinical contexts. In addition, the recommendations and grading are personal digital assistant downloadable from the clinical resource. This product is available for purchase from the ACCP. The patient education materials are accessible free of charge on The implementation and translation of evidencebased clinical practice guidelines facilitates knowledge uptake, critical for practice change, and should ultimately lead to better patient-focused care. The HSP Subcommittee on Implementation has proposed to collaborate with the Governors, Thoracic Oncology Network, and other groups within the ACCP to disseminate and implement the guidelines in their local communities. Residency and specialty training programs are encouraged to use the guidelines in journal clubs and grand rounds. Other organizations that were invited to send representatives to the final conference and review the proposed drafts were also requested to endorse the guidelines and market them to their membership through their own communication channels. As with all ACCP guidelines, these guidelines were submitted to the National Guideline Clearinghouse for posting on their Web site, Performance Measures The ACCP Quality Improvement Committee in partnership with the HSP Committee has been charged with the selection of ACCP clinical practice guideline recommendations for proposed development of performance measures and the fostering of the development, endorsement, and implementation of these performance measures. The panel, with guidance from the Executive Committee, is requested to consider which of the recommendations could be put forth for development into performance measures and which should not. Criteria for selecting those deemed appropriate for this use are two tiered. First, the evidence and benefits need to be sufficiently strong for the recommendations to have 1A grades. The second tier of criteria includes the following: (1) practicality for ACCP members and their patients, (2) importance, (3) scientific acceptability, (4) usability, and (5) feasibility. These recommendations, once identified, are transmitted to the American Medical Association Physicians Consortium for Quality Improvement for consideration for development into performance measures and, eventually, forwarded to the National Quality Forum for potential endorsement. Conclusion The goal of this project was to produce updated, evidence-based, clinically relevant guidelines for physicians and other health-care providers managing the care of patients with lung cancer or those who are at risk of lung cancer. The methods employed followed the policies and standards of the ACCP, as described in this chapter and on References 1 Diagnosis and management of lung cancer: ACCP evidencebased guidelines. Chest 2003; 123(suppl):S1 S326 2 Multidisciplinary approach to thoracic oncology. Northbrook, IL: American College of Chest Physicians, Guyatt G, Gutterman D, Baumann MH, et al. Grading strength of recommendations and quality of evidence in clinical guidelines: report from an American College of Chest Physicians Task Force. Chest 2006; 129: S Diagnosis and Management of Lung Cancer: ACCP Guidelines