Keywords

1 Introduction

Comparative effectiveness research (CER) aims to generate evidence on effectiveness of strategies for diagnosis, treating or preventing disease in the “real-world” compared to the existing “standard of care” so that informed decisions can be made to improve health care [1, 2]. CER studies thus often need to account for physician (surgeon), patient and system factors that affect outcomes observed when interventions are made in the real world, away from the controlled settings of a randomized clinical trial. As CER is by definition comparative, it can address questions on both clinical utility and added clinical value compared with standard of care. With an aging population, increasing cancer incidence, greater intensity of cancer care, and the need for continuing care of the growing number of cancer survivors, there is no doubt a growing economic burden associated with cancer and cancer treatment [3]. In this chapter, we will discuss CER issues for the surgeon with respect to the most common lethal malignancy—lung cancer. There are very few randomized clinical trials with respect to surgical interventions for lung cancer. Several factors have led to this situation including difficulties with accrual (including both provider and patient bias), and expense. With increasing need to answer questions on clinical care with respect to quality, outcomes and cost-effectiveness, opportunities for CER studies will emerge for this disease now and into the future. We will address 3 types of surgeon factors, patient factors and system factors that influence outcomes and can be the genesis of CER studies in the future.

In 1953, lung cancer became the most common cause of cancer death in men, while the same occurred for women in 1985. In 2012, there were 1.8 million patients with lung cancer globally, causing an estimated 1.6 million deaths [4]. In 2014, there will be an estimated 224,000 new cases of lung cancer and 159,000 deaths [5]. Lung cancer causes more deaths than the next three most lethal common cancers combined (colon, breast and pancreatic) [6].

There are several factors that can influence whether patients receive any type of treatment for lung cancer (Fig. 1). Evidence-based treatment options for lung cancer depend on histology, stage and patient specific factors (such as age, pulmonary function and comorbidities). Approximately 95 % of all lung cancers are classified as either small cell lung cancer (SCLC) or non-small cell lung cancer (NSCLC). SCLC is distinguished from NSCLC by its rapid doubling time, and early development of widespread metastases. In light of these factors, surgical intervention is not the standard of care for SCLC. Sadly, though SCLC is initially responsive to chemotherapy and radiation therapy, it typically relapses, with subsequent resistance to treatment modalities within a few months to a year. A two-stage classification is used for SCLC derived from the Veterans’ Affairs Lung Study Group (VALG) in the late 1950s [7]:

  • Limited disease (Tumor confined to ipsilateral hemithorax and regional nodes)

  • Extensive disease (Tumor beyond the borders of limited disease, including malignant pleural and pericardial effusions)

Fig. 1
figure 1

Non-disease factors that influence receipt of lung cancer treatment

Full size image

Limited disease is treated with chemotherapy and radiation therapy, while extensive disease is often treated with chemotherapy. Five-year survival rates for limited and extensive disease are poor (10–13 % vs. 1–2 %) with the median time of survival 15–20 months and 8–13 months respectively [8–10].

The remainder of this chapter will be focused on CER for NSCLC, where surgical intervention plays a role in early stage disease.

NSCLC accounts for the majority (approximately 85 %) of lung cancers. Localized disease is where the cancer is limited to one lobe of the lung, and does not involve the mediastinum. Stage I disease (small lesion, without involvement of any lymph nodes) and Stage II disease (larger lesion without lymph node involvement, involvement of structures that can be resected such as the chest wall or diaphragm, or involvement of hilar lymph nodes) makes up about a third of patients with NSCLC [11]. Although validation hasn’t occurred with randomized clinical trials, good results and long-term survival data have established surgery as the treatment of choice for localized disease in those patients who are medically operable [12]. Surgical resection has also found a role for select patients with stage IIIA disease as well (involvement of ipsilateral mediastinal nodes).

2 Surgeon Factors

2.1 Extent of Resection

The surgical resection of a single lobe of the lung, or lobectomy, is the procedure of choice for early stage NSCLC. The lung cancer study group [13] conducted a prospective randomized clinical trial comparing limited resection (segmentectomy or wedge resection) versus lobectomy in the management of early stage (less than 3 cm, and absence of lymph node involvement) NSCLC. When the group first reported their results, of the 247 patients followed for a minimum of 4.5 years, patients undergoing limited resection had an observed 75 % increase in recurrence rates, 30 % increase in overall death rate, and an observed 50 % increase in death with cancer rate compared to patients undergoing lobectomy. Interestingly, errors in accounting for patients lost to follow-up were noted by Dr. Frank Lederle, and detailed in his letter to the journal [14]. This prompted a second review of the study that uncovered 12 additional recurrences and 3 additional deaths. Using the corrected data, there remained a survival benefit to the lobectomy group (5-year survival 73 % vs. 56 %), but a decrease in the rate of recurrence (5-year 63 % vs. 78 %). The rate of distant recurrences was the same in both groups, whereas limited resection patients experienced threefold higher rate of locoregional recurrence (5.4 % vs. 1.9 %). In the new multivariate analysis, weight loss replaced performance status as a significant predictor of overall survival. Though the numbers, graphs and multivariate findings changed slightly, the overall conclusions of the lung cancer study group study were not altered by the corrected data [15].

Lobectomy became the norm for those patients with adequate pulmonary reserve. Additional observational studies demonstrated improved survival in individuals with earlier stage disease who undergo resection (lobectomy) [16–18]. Most surgical series demonstrated five-year survival for stage I NSCLC in the range from 55 to 72 % [19, 20], with even more favorable results reported for those with small (less than 3 cm) peripheral lesions [21, 22]. In contrast, the 5-year survival of patients with Stage I lung cancer not treated surgically is reported to be from 4 to 14 % [23–25]. The role of surgery for Stage IIIA disease is limited. If previously unsuspected microscopic disease is found in the mediastinal nodes at the time of resection, then proceeding with lobectomy followed by adjuvant therapy is recommended [26]. In those patients where the mediastinal nodes are prospectively identified with sampling and imaging, then multimodality treatment is recommended, with concurrent chemotherapy and radiation therapy. Surgical resection can subsequently be an option in a subset of these patients after chemoradiation with low volume or microscopic mediastinal nodal disease involvement, where resection is technically feasible [26].

Proximal tumors with endobronchial involvement traditionally were treated with pneumonectomies. In the modern era, sleeve resection of the involved airway with lobectomy is preferred over pneumonectomy as a result of similar oncologic results, better preservation of pulmonary function, and avoidance of the complications associated with pneumonectomy [27].

There has been a renewed interest in the role of sublobar (limited) resection. Sublobar resection may be performed either as the removal of one or more anatomical segments (segmentectomy) or as a non-anatomical wedge resection. Sublobar resections can be an option for those patients with early stage disease who cannot tolerate a lobectomy due to decreased pulmonary reserve, or medical co-morbidities. Several prospective studies have reported favorable outcomes in those patients undergoing segmentectomies or wedge resections for peripheral <2 cm in size, N0 lung cancers [28–30]. There are two ongoing clinical trials focusing on this issue. The Cancer and Leukemia Group B (CALGB) trial 140503 [31] has had a difficult time accruing patients since its launch in 2007, and aims to directly compare lobectomy versus limited resection among patients with peripheral tumors measuring <2 cm in size. In Japan, the nearly accrued JCOG 0802/WJOG 4607 L 1,100 trial is comparing the outcomes of peripheral invasive adenocarcinomas of less than or equal to 2 cm in size treated by lobectomy or segmentectomy. Another Japanese study (JCOG 0804/WJOG 4507L) has completed accrual evaluating the role of limited resections in the management of non-invasive adenocarcinomas. Results from these Japanese studies are awaited [32, 33]. Even after findings from these studies are released, CER opportunities will be present (Table 1).

Table 1 Surgeon factors—extent of resection
Full size table

2.2 Video-Assisted Thoracoscopic Surgery (VATS) Versus Open Lobectomy

VATS is a minimally invasive thoracic surgical procedure that can be utilized for the diagnosis and treatment of intra-thoracic diseases. Many procedures that historically were performed with an open thoracotomy are now performed as VATS. Absolute contraindications for VATS are the same as for a thoracotomy including the inability to perform a complete (R0) resection with suitable residual cardiopulmonary reserve, lymph node metastasis beyond regional lymph nodes, and widely metastatic disease [34, 35].

VATS procedures use ports that for the most part are <2 cm in length, except for one access or utility incision that ranges from 4 to 8 cm in length through which instruments can be inserted and allow for removal of the resected lung at the completion of the case. The important principle that is maintained during a VATS resection procedure is that a rib-spreader is not used. By avoiding muscle splitting, and rib-spreading with a retractor, VATS is believed to result in less pain, earlier ambulation, and fewer postoperative complications [36]. VATS lobectomy was first performed in 1992, and has been rapidly adopted in the surgical management of lung cancer. Several studies have shown lower rates of postoperative complications associated with VATS when compared with open lobectomy for early stage lung cancer [37–41]. Further studies have shown improved functional outcomes and equivalent oncologic efficacy as well [42–45].

Interestingly, cost effectiveness analyses have shown variable results. A study using the Surveillance, Epidemiology, and End-Results Medicare database found that VATS lobectomy was associated with a shorter length of stay (LOS) but not with differences in costs [36]. One explanation offered in the study was that payer reimbursement was linked to episodes of care rather than LOS. Other studies using the all-payer Nationwide Inpatient Sample [44, 45] described lower complication rates and LOS for VATS lobectomy, but no differences in costs. In contrast, an all-payer Premier Perspective database study [41] found that adjusted inpatients costs were $700 lower for VATS. A study from our group using the MarketScan database examined whether VATS lobectomy was associated with lower 90-day costs—thus assessing for costs and complications beyond the index hospitalization [46]. We found that the biggest driver of cost was prolonged LOS (PLOS) at the index hospitalization, VATS lobectomy was associated with lower rates of PLOS, and that the cost difference of the VATS approach extends only minimally into the period after discharge (i.e., health care use after discharge was only minimally different between the VATS and open thoracotomy groups). Outpatient use and readmissions accounted for approximately 16 % of the total 90-day costs of care after lobectomy. Emerging evidence suggests that the pressure to discharge patients earlier after lung resection may be driving readmission rates, independent of the surgical approach [47, 48].

Although use of VATS has increased over time, open thoracotomy is still the most widely used procedure for lobectomy with less than 50 % of lobectomies in the US performed using VATS [49]. Factors that have been proposed to explain this finding include insufficient training or experience, difficulty to achieve competency with minimally invasive approaches in low volume centers, and a belief that there remains an insufficient level of evidence for safety and efficacy. However, there are now two large institutional prospective cohort studies involving 1,100 and 500 patients respectively [50, 51], a randomized clinical trial [52] and two meta-analyses [53, 54] that have demonstrated safety and efficacy of the approach. The Cancer and Leukemia Group B (CALGB) 39802 trial [55] was a novel multi-institutional safety and feasibility study that not only standardized the definition of VATS lobectomy, but also standardized surgeon credentialing. The rigorous credentialing process had participating surgeons attend a course to review technique; submission of an unedited video tape, operative and pathology reports from a VATS lobectomy for central review; and participation in an animal laboratory. Surgeons were required to perform at least five VATS lobectomies before being credentialed. Eleven surgeons at six centers underwent credentialing, and the subsequent success rate, morbidity and mortality observed in the study achieved or surpassed prior levels cited in the literature. The study demonstrated a method for surgeons who had no prior experience with the technique to attain sufficient training and expertise in a supervised real-world environment.

Robotic-assisted thoracoscopic surgery (RATS) approaches for lobectomy have recently emerged as an alternative to traditional VATS. Initial results of robotic lobectomy have shown the same benefits achieved with VATS approaches as compared to open thoracotomy can be maintained [56–58]. Additionally, advocates for RATS cite benefits of improved ergonomics, three-dimensional optics, and wristed instrument motions. On the other hand, opponents of robotic surgery have cited increased costs and longer procedure times [59]. It is still to early to tell whether RATS is truly an advance in surgical technique, or as nay-sayers are fond of stating—a marketing gimmick. Cost analyses can be challenging for robotic programs as the cost of the robotic platform varies between institutions (purchase of a new robot versus incorporation of RATS into a facility with an existing platform), and that theoretic costs include the price of the robotic system divided by the total number of robotic cases (all specialties) performed. However, this shouldn’t deter us from assessing the impact of RATS into practice, as similar issues are raised when assessing the introduction of any new surgical technology. Even if one is not convinced that RATS is a significant advance as compared to VATS lobectomy, there may be other reasons for surgeons to transition to robotic surgery, including the performance of other thoracic surgical procedures such as a thymectomy, robotic technology will continue to evolve, and more clearly delineated benefits of robotic surgery may become the reality in the future. What is missing till date is a standardized detailed credentialing process for RATS similar to that which was outlined in CALGB 39802. CER opportunities with respect to minimally invasive approaches are outlined in Table 2.

Table 2 Surgeon factors—type of surgical approach
Full size table

2.3 Specialty Training

The issue of who should provide care has led to considerable debate about access, and healthcare disparities. Surgeon specialty has been shown to be associated with better post-operative outcomes among high-risk operations [60–62]. In the United States, the majority of lung resections are performed by general surgeons [63]. However, there have been several studies demonstrating that board-certified thoracic surgeons have lower rates of operative mortality with lung resections compared to general surgeons [64–66]. In a study conducted by our group in the SEER-Medicare population [67], we extended the analysis to compare the results of board-certified general thoracic surgeons (GTS), board-certified cardiothoracic surgeons (CTS) (those who performed both cardiac and thoracic procedures as part of their practice), and those treated by general surgeons (GS). After adjustment for several well-known prognostic factors for survival, patients under the care of GTS had an 11 % lower risk of death compared with those treated by GS. General thoracic surgeons used preoperative and intraoperative staging procedures more often than GS or CTS, more often applied video-assisted thoracoscopic techniques, and less often performed bi-lobectomy (precision of resection).

Common themes in these studies were influence of provider volume on the overall effect, as well as more consistent process-of-care measures by specialty surgeons. As there is a trend towards increasing specialization amongst surgeons, other factors that may have influenced decision-making include training in the modern era, with inclusion of evidence-based protocols, and multi-disciplinary participation in tumor boards amongst specialists. The results however of all these studies are relevant to the overriding issue of how best to improve the quality of thoracic surgical care. One option might be to encourage referral of potentially resectable lung cancer patients to GTS. This of course would need to be accompanied by the addressing of issues such as barriers to access, especially amongst low-income patients and those that live in rural areas, as well as workforce issues. Attempting to improve outcomes by selective referral without addressing infrastructure and workforce issues could lead to vast segments of the population without access to surgical resection for an otherwise uniformly fatal disease. Opportunities for CER are detailed in Table 3.

Table 3 Surgeon factors—specialist care
Full size table

3 Patient Factors

As overall outcomes for lung cancer treatment are poor, opportunities exist to explore the impact of treatment choices on the patient. When one traditionally thinks of patient factors, the focus has been on co-morbidities. In surgery, the decision-making process is often situational. Patient autonomy and participation can be influenced by medical condition, surgeon factors, patient educational level, and availability of evidence-based information on the particular condition. The degree of decisional authority assumed by patients can lead to three types of surgeon-patient relationships—the surgeon as agent, shared decision-making, and informed decision-making. Large-scale studies on decision-making have not been performed till date for lung cancer, but are anticipated to increase in the years ahead alongside the patient-centered movement.

The surgeon as agent occurs when the surgeon acts as an expert adviser who incorporates the values of the patient when making a treatment recommendation. In this model, the patient role is passive, as the surgeon assumes the values of the patient, and has total command over the decision-making process. Patients may be subjected to biased treatment if the surgeon only gives or delivers limited treatment options. In contrast, the informed decision-making model is one where the surgeon is recognized as the one who has technical expertise, but the patient plays an active role eliciting and understanding information about their treatment choices. The surgeon in this model doesn’t volunteer their opinions, but rather presents the patient with various treatment options, allowing the patients to arrive at their own conclusions. In between these two extremes is shared decision-making. Here the surgeon and patient are equal partners, where each freely exchanges information and preferences about treatment options to arrive at a mutually acceptable decision. This works especially in those situations where there is ambiguity in treatment of choice, and helps to align the decision-making with patient’s preferences and values.

Though surgeons may profess to always include patient-centered values in their discussions, there are plenty of opportunities for improvement. Some of these issues and opportunities for CER are detailed in Table 4.

Table 4 Patient factors
Full size table

4 Healthcare System Factors Related to Clinical Decision Making

4.1 Impact of Practice Environment

In principle, the Patient Protection and Affordable Care Act (ACA) signed into law in March 2010 seeks to improve health care delivery. The goal of ACA is to create a movement of payment reforms, in which private insurance companies would follow the lead of successful government payment reforms, such as bundled payments, and ultimately create system-wide changes for reimbursement [68]. Changing the reimbursement structure for providers will inevitably create new issues for surgeons who are making decisions for their patients. Payment reforms began in 2011 and 2012, and will continue through 2016. Two programs designed to restructure the way health care is delivered have been proposed under ACA, namely Patient-Centered Medical Homes (PCMHs) and Accountable Care Organizations (ACOs). These programs are designed to improve care coordination by encouraging use of electronic medical records, changing providers’ financial incentives by including quality measures in reimbursement, and ultimately moving away from a fee-for-service to one where quality of care is valued [69].

The ACO movement has led to increased consolidation and integration in the medical marketplace. Hospitals are buying practices to keep their market share intact and to have access to electronic record systems and other infrastructure that are expensive to capitalize. Awareness emerges for surgeons that their medical decisions can potentially negatively influence their income. This is not necessarily unethical, as cost containment has been recognized as an important circumstance in good decision-making [70]. There will be the introduction penalties for lack of delivery of quality care, which could affect physician reimbursement. Adoption of rigid guidelines for the treatment of patients is anticipated in years to come, as well as expansion of care plans. All of these are attempts at decreasing variation in care, decreasing length of stay, and reducing use of resources.

A method of performance with a powerful impact on outcomes is participation in large national or regional databases with inclusion of all patients and frequent provider feedback with comparisons to peer norms. Surgeons in the Veterans Administration hospital system have participated for more than a decade in a systematic data-gathering and feedback system of outcomes for major surgery [71]. The National Surgical Quality Improvement Project (NSQIP) works to decrease variation in clinical outcomes by demonstrating to surgeons when their center is an “outlier” in performance. This system allows hospitals to target QI activities that may influence components of care, and subsequently decision-making. The Society of Thoracic Surgeons (STS) in 1989 created a national voluntary cardiac surgery database as a means of supporting national quality improvement efforts. In 2003, a separate database was launched by the STS encompassing procedures specific to general thoracic surgery—the General Thoracic Surgery Database (GTSD) [72]. GTSD provides participants risk-adjusted benchmarks as well as data for research that can be used to improve patient care processes and clinical outcomes. On 30 July 2008, the National Quality Forum endorsed participation in a systematic national database for general thoracic surgery [73]. Unfortunately in 2012, only 8 % of all lung cancer resection cases in the US were accounted for in GTSD, with most of the participants high-volume centers with dedicated general thoracic surgeons [74].

4.2 Impact of Political Environment

The reporting of surgeon-specific outcome data is another example of the influence of the political environment. Outcome data were rarely reported prior to the mid-1980s [75]. The first release of hospital open heart surgery risk-adjusted mortality rates in December 1990 [76] and the first formal public report in December 1992 [77], marked the start of a new era. These performance reports, or physician report cards, have increased in recent years [78], and many believe will increase in frequency and across specialties in the years to come. Advocates believe that increased transparency of information on the quality of care help consumers, employers, and health plans to improve their decision-making and to stimulate quality improvement among providers. However, physicians are concerned that risk adjustment strategies in these reports are not adequate. Without this confidence, publication of procedural mortality rates may result in physicians withholding procedures in high-risk patients. Unintended consequence of scorecards might be to adversely affect healthcare decisions for especially high-risk patients. Scorecards may also impair the development of new treatments because of the more restrictive clinical practice environment.

In light of these drawbacks, many have proposed revamping the current system to facilitate rapid and accurate access to outcome data in the local practice environment. Adoption of these efforts is often embraced as this occurs on a voluntary basis rather than in response to punitive restrictions. Examples of such grass-roots initiatives on a state level that are surgeon-led include those in the states of Michigan [79] and Washington [80, 81]. On a national level, data from the STS database has been used for public reporting, and thus have an impact on risk-stratification and outcomes in cardiothoracic surgery [82, 83].

Regional quality improvement (QI) efforts have succeeded in cardiac surgery, and a regional QI initiative in thoracic surgery has been proposed in Washington State [84]. Fifteen thoracic surgeons from five institutions examined the landscape of care in the state, as well as to discuss standards for a regional QI effort in lung cancer surgery. Consensus standards endorsed in this initiative include:

  • GTSD participation across all hospitals performing lung resection in the state

  • Limited enhancements to data collection efforts to address local concerns and survey ongoing interventions

  • Quarterly performance reports

  • Surgeon-led QI interventions for addressing performance gaps, quality improvement and value optimization

  • and leveraging existing QI infrastructure and relationships within Washington state to rapidly and successfully implement a regional QI effort for lung cancer surgery.

Success of the regional QI initiative will likely lead to performance of several CER studies, and can ultimately serve as a model for other regional and national efforts.

An example of a national effort to reduce unnecessary tests across specialties is the Choosing Wisely® campaign. The Choosing Wisely® initiative helps physicians and patients have important conversations necessary to ensure that timely and optimal care is delivered. Launched by the American Board of Internal Medicine (ABIM) Foundation, Choosing Wisely® enables physicians and patients to engage in conversation about the overuse of tests and procedures, and helps patients make smart and effective care choices [85]. The original campaign has evolved into a multi-year initiative where the ABIM Foundation has reached out to specialty societies to identify a list of five tests or procedures that may be overused or misused. Criteria for developing these lists include limiting to items that fall within the specialty; supported by evidence; documented and publicly available upon request; frequently ordered/costly; easy for a lay person to understand; and measurable/accountable. The STS participated in the February 2013 phase II release (Table 5) [86]. Two of the five proposed measures are directly applicable to lung cancer surgery. These specialty generated lists help to empower physician-patient conversations and to avoid unnecessary procedures that may harm patients while driving up health care costs. 63 specialty societies have joined the campaign since its inception in 2012.

Table 5 Society of thoracic surgeons choosing Wisely® list
Full size table

4.3 Lung-Cancer Screening

An example of how politics can influence standards of care recently arose with respect to lung cancer screening. A large prospective randomized clinical trial, the National Lung Screening Trial (NLST), demonstrated the potential of low-dose computed tomography (LDCT) to detect lung cancer at earlier stages, thereby decreasing mortality [87]. NLST demonstrated that annual lung cancer screening in a high-risk patient population for three years with LDCT resulted in 20 % fewer lung cancer deaths as a result of early detection and treatment. Using the strict NLST criteria, this translates to 8.6 million people eligible for screening in the US, and 12,000 averted lung cancer deaths if all those individuals are screened [88]. In light of these studies, a number of non-profit, professional and federal organizations have recommended evidence-based annual LDCT screening for lung cancer in high-risk patients, including the National Comprehensive Cancer Network (NCCN), the US Preventive Services Task Force (USPSTF), the STS, American Society of Clinical Oncology, and the American Cancer Society [89]. The USPSTF decision in the setting of the Affordable Care act resulted in a mandate for coverage by private insurers for lung cancer screening. However, controversy has arisen as despite the evidence, Medicare till date has not endorsed coverage for lung cancer screening [90]. The implications are especially pertinent as nearly 70 % of lung cancers occur in the Medicare population. Time will tell if the evidence and advocacy of several specialty societies and patient advocates will correct this, but the politics surrounding the issue are puzzling to say the least. CER opportunities will arise in the coming years assessing screening practices, access to LDCT, and impact of early referral for surgical interventions for cancers detected by LDCT.

5 Summary

Although the ideal is to practice evidence-based medicine at all times, there are many factors that influence the care that we provide. There is growing interest in assessing and improving the value of health care delivery, defined as health benefits per dollar spent. Value can be increased by improving clinical outcomes, decreasing costs, or ideally doing both. CER for surgical interventions in lung cancer should be viewed through this lens, and opportunities for health services researchers and surgeons in practice will continue to emerge in the years ahead.