Search

The mission of American Society of Clinical Oncology (ASCO) is to conquer cancer through research, education, and promotion of the highest quality patient care. Toward fulfillment of this goal and at the direction of its board of directors, the ASCOValue in Cancer Care Task Force set out to develop a framework that would enable a physician and patient to assess the value of a particular cancer treatment regimen given the patient’s individual preferences and circumstances. The rationale that served as the impetus for this initiative is many faceted. Substantial progress has been made in translating our knowledge of the biologic characteristics of cancer into novel therapies. Some of these therapies have led to major improvements in outcomes for specific diseases, and others have produced only modest advances. There is now a wide array of choices for treating many cancer types, and these treatment choices often differ by only small degrees in clinical effectiveness and toxicity. Yet, there is often a wide disparity in cost to patients and payers. Because patients are often confronted with enormous expenses when receiving cancer care, the goal of describing a relationship between the cost of an agent or regimen and the clinical benefits it delivers takes on great importance. As the primary advisor to the patient, the oncologist has an important role in providing a comparative assessment of the various treatment options available; in the spirit of shared decision making, the patient should have transparent information about the clinical impact that can be expected from the different options presented and their relative financial implications. The value framework has been constructed as a conceptual model that incorporates the elements of clinical benefit, toxicity, and symptom palliation as derived from a comparative clinical trial and combines these elements into a score termed the net health benefit (NHB). Ultimately, deployment of the framework as a software application is planned, enabling a patient to modify the weight attributed to any of the elements included in the NHB depending on his or her personal preferences and circumstances. The final NHB will therefore reflect the priorities that are most important to the patient and will be arrived at through guidance from the physician. Information on the cost of the regimens will also be presented so the patient can consider the relative financial impact of his or her treatment options. Two versions of the framework have been created: one for advanced disease and the other for potentially curable (adjuvant therapy) clinical presentations. The original framework versions are shown in Appendix Tables A1 and A2 (online only). The key elements included in the framework— namely, clinical benefit and toxicity—are also those that are regularly reported in the scientific literature when discussing the outcome of a clinical trial that compares two or more therapies. The importance of relying on high-quality, quantifiable evidence cannot be overstated, and this is most often provided by a well-designed, well-conducted prospective randomized trial. The task force recognizes that a limitation of this approach is that it does not readily permit cross-trial comparisons. Such analyses are important to patients and remain a goal for future versions of the value framework. The task force is well aware that there are many elements that might be important to individual patients in assessing the relative value of their treatment options that are not taken into account in our model. These include the convenience of receiving therapy, the avoidance of interrupting the flow of activities of daily living, and the impact of a treatment on quality of life

4643 Background: Observational evidence associates energy balance factors, particularly diet, with survival in patients with colorectal cancer (CRC). Consumption of a diet with high glycemic indices has been associated with inferior cancer-specific outcomes, but there is limited prospective evidence that alterations in dietary habits improve cancer outcomes. This was a pilot study to determine the feasibility and acceptability of following a low glycemic load (GL) diet in patients with stage I-III CRC and to assess the nutritional resources necessary to follow the diet. Methods: 18 patients with stage I-III CRC, who completed definitive cancer therapy and consumed an avg daily GL > 150 participated in a 12 week, tailored, in-person dietary intervention with a target GL of ≤102. Compliance was assessed using 24 hour telephone recalls. Acceptability of the diet was assessed using a food acceptability questionnaire, and exploratory correlative laboratories were assessed monthly. Results: 67% of patients were compliant with a low GL diet ≥ 75% of the time, over a 12 week time period. Majority of participants experienced a decrease in BMI and waist circumference, 28% experienced meaningful weight loss defined as ≥ 5%. The nutritionist spent an avg of 6.97 hours (SD 2.18) in-person and 1.58 hours (SD 0.68) by phone with each participant. In the overall group, significant decreases were seen in total cholesterol (7.2% decrease; t = -2.33, p = 0.03), VLDL (26.8% decrease; t = -2.33, p = 0.03) and triglycerides (26.6% decrease; t = -2.29; p = 0.04). All participants were satisfied with the diet; 43% were extremely satisfied. 75% of participants liked the foods they were able to eat “very much” or “extremely”. All participants felt the in-person meetings were helpful. 77% did not feel an online video could replace the in-person meetings. 62% of participants did not feel a virtual meeting (e.g skype, etc.) could replace the in- person meeting while 38% felt it could. Conclusions: Patients with stage I-III CRC are able to follow a low GL diet with an in-person dietary intervention. Significant decreases in laboratory measures confirm the efficacy of the diet in altering metabolic indices. All participants who completed the study were satisfied with the diet, the majority of whom enjoyed the foods and planned to continue to follow the diet after study completion. The majority felt in-person contact with the nutritionist was essential to their success. This study was an essential step in designing a larger scale trial to evaluate the impact of low GL diet on cancer outcomes. Clinical trial information: NCT02129218 .

2061 Background: Increasing enrollment in clinical trials remains a national priority, yet there are limited data from COCs on the degree to which common trial exclusion criteria (EC) and socioeconomic factors play a role in low enrollment rates. Methods: We analyzed data from the nationwide Flatiron Health electronic health record (EHR) derived de-identified database. COC were eligible if they had given a clinical trial study drug to ≥2 patients (pts)/year. We included pts with one of eight advanced or metastatic solid tumors who received ≥1 line of systemic anticancer therapy between 1/1/2014 and 11/30/2019. We defined EC as either: creatinine > 1.5 mg/dl or Ccl < 45 ml/min, Hb < 9 g/dL, ANC < 1500/ul, plts < 100,000/ul, bilirubin > 1.5 upper limit of normal (uln) or AST/ALT > 2.5 uln within 30 days or ECOG performance status (PS) ≥ 2 within 60 days prior to start of therapy. We calculated the percentage of pts with ≥1 EC relative to the group of candidate pts, stratified by therapy line (1L, 2L, 3L+). We used multivariate logistic regression models to evaluate the effect of EC and socioeconomic factors (age, race, Medicaid) on the likelihood of receiving a clinical study drug for each line of therapy. Results: In this sample of 35 COCs, 26,988 pts received ≥1 systemic therapy. Pts with ≥ 1 EC: 28.4% in 1L, 34.2% in 2L, 37.4% in 3L. Percentages of pts with an ECOG PS ≥ 2 were: 15.6% (1L), 18.2% (2L), 19.8% (3L). Pts receiving a clinical study drug: 1.7% of 26,988 in 1L, 2.0% of 12,738 in 2L, 2.9% of 5,333 in 3L+, and 3.1% in any line. Excluding pts with ≥1 EC from the denominator modestly improved overall accrual: 2.0% of 19,729 in 1L, 2.3% of 8,588 in 2L, 3.7% of 3,470 in 3L+. In multivariate logistic regression, ECOG PS ≥ 2 was strongly associated with not receiving a study drug [odds ratio (95% CI); 1L: 0.25 (0.16-0.4); 2L: 0.28 (0.17-0.49); 3L: 0.21 (0.1-0.44)]. The likelihood of receiving a clinical study drug (any line) was lower for pts who are Black [0.63 (0.48-0.82)], Latino [0.49 (0.32-0.75)], and pts older than 70 years [0.63 (0.54-0.72)]. Medicaid pts were not significantly less likely to receive study drug [0.83 (0.64-1.07)]. Conclusions: In COC, common trial EC reduce pt availability for trials by > 25%. Poor PS is highly prevalent and influential. These EC and complex trial requirements challenge COC’s ability to recruit representative pt populations. Future efforts to increase enrollment in trials must consider common EC along with well known barriers to enrollment of unrepresented groups.

3634 Background: Phase 1 clinical trials are the first step in developing new cancer therapeutics. Patients enrolled in these studies have typically exhausted standard therapies and are at a point in their disease trajectory where they often are choosing between a phase 1 clinical trial and hospice care. These patients may have significant symptom burden, which can result in early trial discontinuation and confound phase 1 trial outcomes, including toxicity profiles, which may influence further drug development. This study aimed to determine the palliative care needs of patients enrolled on phase 1 clinical trials and their caregivers (CGs), along with differences in study duration, adverse event (AE) and symptom profiles, and quality of life (QOL) between those receiving structured palliative care vs usual supportive care. Methods: 68 patients enrolled on phase 1 clinical trials were randomly assigned to receive structured palliative care or usual supportive care. 39 of their CGs were enrolled and assigned to the same arm as the patient. Quality of life metrics were obtained monthly: the Functional Assessment of Cancer Therapy-General and Memorial Symptom Assessment Scale -Short Form for patients, and The Quality of Life in Life Threatening Illness - Family Carer Version and Caregiver Reaction Assessment for CGs. Palliative care resources utilized were assessed for those in the palliative care arm, and referrals to supportive care services assessed in those in the usual care arm. AEs recorded on the Phase 1 trials were evaluated & compared between arms. Results: Mean duration on phase 1 study was 132 days in the palliative care arm vs 114 days in the usual care arm (p = 0.55). Total weighted AE rate (# of AE [x] AE grade per month) was 26.9 in the palliative care arm vs 34.0 in the usual care arm (p = 0.53). Patients in the palliative care arm experienced better QOL and lower symptom burden, as did their CGs, compared to those in the usual care arm. While the differences in outcomes were not statistically significant, all results favored structured palliative care. Conclusions: Phase 1 patients and their CGs have physical and psychosocial needs which warrant palliative care services. Preliminary results suggest structured palliative care is associated with increased duration on study (by nearly 3 weeks), improved patient and CG QOL, and reduced patient symptom and CG burden. A larger study is warranted to confirm these results and further develop the ideal palliative care intervention in these populations. Clinical trial information: NCT02543541 .

209 Background: Despite national guideline recommendations for universal biomarker testing (KRAS, NRAS, BRAF, and mismatch repair/microsatellite instability [MMR/MSI]) in all patients with metastatic colorectal cancer (mCRC), little is known regarding adherence to these recommendations in the community. Methods: We evaluated 20,333 patients aged >18 years with mCRC diagnosed between 1/1/2013-12/27/18 from the nationwide de-identified Flatiron Health electronic health record-derived database. Changes in rates of biomarker testing based upon chart abstraction according to date of mCRC diagnosis were assessed in patients with ≥ 6 months of follow-up using Cochran-Armitage trend tests. We also assessed whether testing differed by patient characteristics using chi-square tests. Results: Biomarker testing rates increased between 2013 and 2018 (15.3 to 65.7% for NRAS, 23.9 to 56.6% for BRAF, and 34 to 76.6% for MMR, all p < 0.0001). There was no change in rate of KRAS testing (63.8 to 68.3% p= 0.1695) over this period of time. Univariate analysis showed that patients with, worse performance status, and Medicare/Medicaid insurance were less likely to be tested (for all variables, p

As more clinical trials of molecularly targeted agents evolve, the number of eligibility criteria seems to be increasing. The importance and utility of eligibility criteria must be considered in the context of the fundamental goal of a clinical trial: to understand the risks and benefits of a treatment in the intended-use patient population. Although eligibility criteria are necessary to define the population under study and conduct trials safely, excessive requirements may severely restrict the population available for study, and often, this population is not reflective of the general population for which the drug would be prescribed. The American Society of Clinical Oncology Cancer Research Committee, which comprises academic faculty, industry representatives, and patient advocates, evaluated this issue. Evaluation results were mixed. Most physicians agreed that excessive eligibility criterias slow study enrollment rates and prolong the duration of enrollment; however, this hypothesis was difficult to validate with the data examined. We propose the organization of a public workshop, with input from regulatory bodies and key stakeholders, with the goal of developing an algorithmic approach to determining eligibility criteria for individual study protocols, which may help guide future investigators and companies in streamlining eligibility criteria in the era of molecularly driven therapy.

Health care costs in the United States present a major challenge to the national economic well being. The Centers for Medicare and Medicaid Services (CMS) has projected that US health care spending will reach $4.3 trillion and account for 19.3% of the national gross domestic product by 2019.1 This growth in spending—both in absolute terms and as a proportion of our gross domestic product—has not been accompanied by commensurate improvements in health outcomes, despite expenditures far exceeding those of other countries.2–4 One of the fastest growing components of US health care costs is cancer care, the cost of which is now estimated to increase from $125 billion in 2010 to $158 billion in 2020.1 Although cancer care represents a small fraction of overall health care costs, its contribution to health care cost escalation is increasing faster than those of most other areas because of several factors: the increasing prevalence of cancer due to the overall aging of the population and better control of some causes of competing mortality; the introduction of costly new drugs and techniques in radiation therapy and surgery; and the adoption of more expensive diagnostic tests. In some cases, the adoption of newer, more expensive diagnostic and therapeutic interventions may not be well supported by medical evidence, thereby raising costs without improving outcomes.5 Coupled with, or even driving, some of these rising costs are sometimes unrealistic patient and family expectations that lead clinicians to offer or recommend some of these services, despite the lack of supporting evidence of utility or benefit.6 Historically, most individuals in the United States were shielded from the acute economic impact of expensive care because they had health insurance. However, current trends suggest that patients will find themselves increasingly responsible for a greater proportion of the cost of their health care. Cost shifting or sharing can occur through the increased use of high-deductible policies and larger copayments. These increased costs are already commonplace and may not be affordable for many families. Indeed, health care expenditures are cited as a major cause of personal bankruptcy,7 and the term financial toxicity has entered the vernacular as a means of describing the financial distress that now often accompanies cancer treatment.8 Like other toxicities of cancer treatment, financial toxicity resulting from out-of-pocket treatment expenses can reduce quality of life and impede delivery of high-quality care.9,10 Patients experiencing high out-of-pocket costs have reported reducing their spending on food and clothing, reducing the frequency with which they take prescribed medications, avoiding recommended procedures, and skipping physician appointments to save money.10,11 These unintended consequences risk an increase in health disparities, which runs counter to some of the key goals of health care reform. In many communities, the high costs associated with cancer care have created a difficult situation for patients and the oncologists who care for them. Addressing this situation will require greater understanding of all the risks and benefits of various treatment options as well as the consequences of specific choices. In this regard, studies have shown that patients specifically want financial information about treatment alternatives along with information about medical effectiveness and treatment toxicity. However, they often do not receive it. Closing this knowledge gap will require educated providers who are able to sensitively initiate a dialogue about the cost of care with their patients when appropriate.12,13 Patients with cancer are often surprised by and unprepared for the high out-of-pocket costs of treatments. They also overestimate the benefits of treatments that sometimes extend life by only weeks or months or not at all. Oncologists are generally aware of this conundrum but uncertain about whether and how the cost of care should affect their recommendations.14 Although raising awareness of costs and providing tools to assess value may help to manage costs while maintaining high-quality care, some oncologists see this as being in conflict with their duty to individual patients.15 Recent American Society of Clinical Oncology Efforts Motivated by our responsibility to help oncologists deliver the highest-quality care to patients everywhere, the American Society of Clinical Oncology (ASCO) formed the Task Force on the Cost of Cancer Care in 2007. Its mission includes educating oncologists about the importance of discussing costs associated with recommended treatments, empowering patients to ask questions pertaining to the anticipated costs of their treatment options, identifying the drivers of the rising costs of cancer care, and ultimately developing policy positions that will help Americans move toward more equal access to the highest-quality care at the lowest cost.16 In 2012, through the work of the Task Force, ASCO responded to the Choosing Wisely Campaign of the American Board of Internal Medicine Foundation and identified specific instances of overuse in the delivery of cancer care. ASCO used a deliberative consensus process to identify five common clinical practices that are not supported by high-level evidence. A second list of five was developed using the same process and submitted to the Choosing Wisely Campaign in 2013. ASCO amplified the evidence basis for both top-five lists in two publications17,18 and is now developing measures to evaluate the use of these practices as part of its Quality Oncology Practice Initiative. These exercises have provided opportunities to develop a rigorous but flexible approach to assessing efficacy across diagnostic and treatment domains.

LBA1 Background: Racial disparities in cancer outcomes remain a societal challenge. The ACA sought to improve equity in healthcare access and outcomes by permitting states to expand Medicaid and providing subsidies for purchase of private insurance. We assessed the impact of Medicaid expansions on racial disparities in time to treatment among patients (pts) with advanced cancer. Methods: We selected pts ages 18-64 years with advanced or metastatic cancer (NSCLC, breast, urothelial, gastric, colorectal, renal cell, prostate, and melanoma), diagnosed between Jan 1, 2011 and Dec 31, 2018, from the nationwide Flatiron Health electronic health record-derived database. We assigned expansion status based on whether the pts’ state of residence had expanded Medicaid as of the diagnosis date. We estimated Medicaid expansion-related changes in the rate of “timely treatment,” an outcome defined as first-line treatment start within 30 days of advanced or metastatic diagnosis. Regression model covariates included race (White, African American, Asian, and Other race), age, sex, practice type, cancer type, stage, and unemployment rate, using time and state fixed-effects. Regression results present predictive margins. Results: The study included 34,067 pts (median age 57 years; 12% African American). Racial disparities were observed pre-expansion: African American pts were 4.9 percentage points (%pt) less likely to receive timely treatment (Table). Regardless of race, Medicaid expansion trended toward an increase in timely treatment overall (p = 0.05). Expansion was associated with a differential benefit for African American vs white pts (6.9 %pt and 1.8 %pt). Prior racial disparities were no longer observed after Medicaid expansion. Conclusions: Implementation of Medicaid expansions as part of the ACA differentially improved African American cancer pts’ receipt of timely treatment, reducing racial disparities in access to care.[Table: see text]

6620 Background: The OCM is a voluntary Center for Medicare and Medicaid Innovation alternative payment model pilot program. As of Oct 2017, OCM practices are required to report the biomarker status for NSCLC pts. Our objective was to assess the effect of OCM reporting on quality of care in aNSCLC. Methods: We developed a decision-analytic model to compare the likelihood of receiving biomarker testing and corresponding appropriate therapy. We populated the model using real-world data from pts (n=7,075) at OCM sites (n=45) and non-OCM sites (n=105) in the Flatiron Health electronic health record (EHR)-derived database. The pre-period control included pts diagnosed with aNSCLC from Jan 1, 2011 - Dec 31, 2015; the post-period included pts diagnosed Oct 2017 - Nov 2018. For OCM vs non-OCM sites, we estimated probabilities and unadjusted odds ratio (OR) of biomarker testing (EGFR, ROS1, or ALK) and subsequent delivery of appropriate therapy, defined as use of a biomarker-guided tyrosine kinase inhibitor (TKI) for positive pts, or non-TKI therapy for negative pts. Results: No differences in rates of biomarker testing or delivery of appropriate therapy were detected between OCM and non-OCM practices prior to the reporting requirement. In the post-period, OCM was associated with higher odds of biomarker testing and appropriate therapy (Table). Conclusions: To our knowledge, this is the first study of the association of OCM reporting requirements with downstream quality of care. Our results suggest that OCM documentation and reporting requirements are associated with modestly higher quality of care for pts with aNSCLC. Ongoing sensitivity analyses will determine the relative contribution of provider and practice characteristics to these findings. Careful measurement of the impact of reporting requirements is essential to measure the impact of payment reform interventions and inform policy. [Table: see text]

e16003 Background: There are limited data comparing first-line (1L) C versus IO in cisplatin-ineligible mUC patients. The primary evidence guiding treatment decisions was a May 2018 Food and Drug Administration (FDA) safety alert based on early review of two ongoing phase III trials, reporting shorter survival in PD-L1 negative patients receiving IO relative to chemotherapy. Final results from these trials are unknown and may not be applicable to a real-world cohort. Methods: We conducted a retrospective cohort study of mUC patients receiving 1L C or IO from January 1, 2011 to May 18, 2018 using the Flatiron Health electronic health record-derived database. The primary outcome was overall survival (OS) from start of 1L treatment to date of death. We compared 12- and 36-month OS, and hazard ratios before and after 12 months. Propensity score-based inverse probability of treatment weighting (IPTW) was used to address confounding in Kaplan-Meier (KM) and Cox multivariable regression model estimates of comparative-effectiveness. Results: Of 2,243 patients, 562 received IO and 1,681 received C. Baseline characteristics were balanced between groups, with few exceptions (Table). In the first 12 months, IO was associated with an increased hazard of death compared to C (HR 1.38, 95% CI 1.16-1.66). Among patients who survived one year, survival was improved with IO compared to C (HR 0.50, 95% CI 0.27-0.92). Conclusions: In routine clinical practice, 1L IO was associated with inferior 12-month OS relative to C, but superior OS beyond 12 months. Clinicians and patients should carefully consider how to balance the early benefit of C against the late benefit of IO. Currently pending trial results will contribute additional evidence to inform treatment decision making. [Table: see text]

6540 Background: RCTs are the gold standard for understanding the efficacy of new treatments, however, patients (pts) in RCTs often differ from those treated in the real-world. Further, selecting a standard of care (SOC) arm is challenging as treatment options may evolve during the course of a RCT. Our objective was to assess the generalizability and relevance of RCTs supporting recent FDA approvals of anticancer therapies. Methods: RCTs were identified that supported FDA approvals of anticancer therapies (1/1/2016 - 4/30/2018). Relevant pts were selected from the Flatiron Health longitudinal, EHR-derived database, where available. Two metrics were calculated: 1) a trial’s pt generalizability score (% of real-world pts receiving treatment consistent with the control arm therapy for the relevant indication who actually met the trial's eligibility criteria) and 2) a trial’s SOC relevance score (% of real-world pts with the relevant indication and meeting the trial's eligibility criteria who actually received treatment consistent with the control arm therapy). All analyses excluded real-world pts treated after the relevant trial’s enrollment ended. Results: 14 RCTs across 5 cancer types (metastatic breast, advanced non-small cell lung cancer, metastatic renal cell carcinoma, multiple myeloma, and advanced urothelial) were included. There was wide variation in the SOC relevance and pt generalizability scores. The median pt generalizability score was 63% (range 35% - 88%), indicating that most real-world pts would have met the RCT eligibility criteria. The median SOC relevance score was 37% (range 15% - 74%), indicating that most RCT control arms did not reflect the way trial-eligible real-world pts in the US were actually treated. Conclusions: There is great variability across recent RCTs in terms of pt generalizability and relevance of SOC arms. Real-world data can be used to inform selection of control arms, predict impact of inclusion/exclusion criteria, and also assess the generalizability of the results of completed trials. Incorporating real-world data in planning and interpretation of prospective clinical trials could improve accrual and enhance relevance of RCT outcomes.

e20643 Background: Recent data suggest that modulating the microbiome may affect response to ICI. Using electronic health record (EHR) data, we examined whether use of ABX influences OS in pts with aNSCLC and aMel treated with ICI. Methods: We performed two retrospective cohort studies (aNSCLC, aMel) of pts who received ICI +/- chemotherapy as first-line treatment in the nationwide Flatiron Health EHR-derived database. The impact of ABX exposure on OS was assessed using Cox regression, adjusting for age, sex, body mass index (BMI), performance status (ECOG PS), Elixhauser Comorbidity Index, glucocorticoid (GC) use, and smoking. In the primary analysis, ABX exposure was defined as use from -6 to +4 weeks around start of ICI. In secondary analyses, we separately evaluated ABX use within 6 weeks pre- and 4 weeks post-initiation of ICI. Results: Baseline characteristics are displayed (Table). In the primary analysis, there was no association between ABX use and OS in aNSCLC (HR 1.16, 95% CI 0.54-2.47, p = 0.7) or in aMel (HR 0.70, 95% CI 0.17-2.81, p = 0.6). The only secondary analysis to reveal a negative impact of ABX on OS was the time window of 0 to +4 weeks after ICI start in aNSCLC (HR 3.41, 95% CI 1.38-8.39, p = 0.008). Conclusions: In this large, real-world dataset, ABX exposure around the start of front-line ICI did not influence OS in pts with aNSCLC or aMel. However, ABX exposure was associated with inferior OS in the subgroup of aNSCLC pts who received ABX within 4 weeks following start of ICI. Further characterization of completeness of ABX exposure data in the oncology EHR will assist with interpretation of these results. [Table: see text]