A flexible approach for causal inference with multiple treatments and clustered survival outcomes.

When drawing causal inferences about the effects of multiple treatments on clustered survival outcomes using observational data, we need to address implications of the multilevel data structure, multiple treatments, censoring, and unmeasured confounding for causal analyses. Few off-the-shelf causal inference tools are available to simultaneously tackle these issues. We develop a flexible random-intercept accelerated failure time model, in which we use Bayesian additive regression trees to capture arbitrarily complex relationships between censored survival times and pre-treatment covariates and use the random intercepts to capture cluster-specific main effects. We develop an efficient Markov chain Monte Carlo algorithm to draw posterior inferences about the population survival effects of multiple treatments and examine the variability in cluster-level effects. We further propose an interpretable sensitivity analysis approach to evaluate the sensitivity of drawn causal inferences about treatment effect to the potential magnitude of departure from the causal assumption of no unmeasured confounding. Expansive simulations empirically validate and demonstrate good practical operating characteristics of our proposed methods. Applying the proposed methods to a dataset on older high-risk localized prostate cancer patients drawn from the National Cancer Database, we evaluate the comparative effects of three treatment approaches on patient survival, and assess the ramifications of potential unmeasured confounding. The methods developed in this work are readily available in the R $$ \mathsf{R}\kern.15em $$ package riAFTBART $$ \mathsf{riAFTBART} $$ .

Radiation Oncology Alternative Payment Model and Large Urban Academic Centers: Future Implications for Patients and Providers.

The radiation oncology alternative payment model (RO-APM) was developed by the Center for Medicare and Medicaid Innovation, a part of the Centers for Medicare & Medicaid Services, as a vehicle to optimize value for patients undergoing radiation therapy. By shifting reimbursement away from fee-for-service and toward a prospective bundled payment system, the RO-APM is intended to bend the cost curve in radiation oncology while preserving or even enhancing outcomes. As with prior large-scale policy initiatives, the nature and magnitude of the RO-APM's impact on care delivery will vary substantially depending on a host of local factors, including practice setting. Urban academic centers play a key role in radiation oncology by spearheading innovation, managing the most complicated cases, training the next generation of radiation oncologists, and often caring for vulnerable patient populations. Thus, to protect patients' access to this high-quality cancer care, it will be crucial to characterize the RO-APM's projected impact on large urban academic institutions before its implementation, including possible unintended adverse consequences. Here, we provide an overview of this seismic change in radiation oncology reimbursement and discuss its unique potential implications for large urban academic institutions as a means to facilitate necessary preparations and inform future revisions to the model.

Drivers of Medicare Spending: A 15-Year Review of Radiation Oncology Charges Allowed by the Medicare Physician/Supplier Fee-for-Service Program Compared With Other Specialties.

PURPOSE: In 2019, the Centers for Medicare and Medicaid Services proposed a new radiation oncology alternative payment model aimed at reducing expenditures. We examined changes in aggregate physician Medicare charges allowed per specialty to provide contemporary context to proposed changes and hypothesize that radiation oncology charges remained stable through 2017. METHODS AND MATERIALS: Medicare physician/supplier utilization, program payments, and balance billing for original Medicare beneficiaries, by physician specialty, were analyzed from 2002 to 2017. Total allowed charges under the physician/supplier fee-for-service program, inflation-adjusted charges, and percent of total charges billed per specialty were examined. We adjusted for inflation using the consumer price index for medical care from the US Bureau of Labor Statistics. RESULTS: Total allowed charges increased from $83 billion in 2002 to $138 billion in 2017. The specialties accounting for the most charges billed to Medicare were internal medicine and ophthalmology. Radiation oncology charges accounted for 1.2%, 1.6%, and 1.4% of total charges allowed by Medicare in 2002, 2012, and 2017, respectively. Radiation oncology charges allowed increased 44% from 2002 to 2012 ($987.6 million to $1.42 billion) but decreased by 19% from 2012 to 2017 ($1.15 billion), adjusted for inflation. Total charges allowed by internal medicine decreased 2% from 2002 to 2012 ($8.53 to $8.36 billion), adjusted for inflation, and decreased 16% from 2012 to 2017 ($7.05 billion). When adjusting for inflation, ophthalmology charges increased 18% from 2002 to 2012 ($4.53 to $5.36 billion) and increased 3% from 2012 to 2017 ($5.5 billion). CONCLUSIONS: Radiation oncology physician charges represent a small fraction of total Medicare expenses and are not a driver for Medicare spending. Aggregate inflation-adjusted charges by radiation oncology have dramatically declined in the past 5 years and represent a stable fraction of total Medicare charges. The need to target radiation oncology with cost-cutting measures may be overstated.

Identifying Potentially Avoidable Hospitalizations in Medicare Patients With Prostate Cancer: A Retrospective Analysis.

PURPOSE:: If identifiable, potentially avoidable hospitalizations (PAHs) can serve as an important target for cost containment efforts in oncology. METHODS:: PAHs among a cohort of Medicare patients with prostate cancer were identified using a two-stage consensus-driven review process. In stage 1, two clinicians independently evaluated admissions records using a case review form, which we modified from a previous study to assess for PAHs. In stage 2, any admissions that the reviewers disagreed on or were unsure of were re-examined in a larger group of clinicians to yield a consensus determination regarding avoidability. Univariable and multivariable regression analyses were performed to identify factors predictive of PAH. RESULTS:: There were 160 admissions among this cohort of 210 patients from January 2012 to June 2015, of which 99 were evaluable. Consensus-driven clinical review yielded an overall PAH rate of 28.3%. Factors associated with increased PAH risk were admission for symptoms related to cancer (odds ratio [OR], 7.33; P < .001), presence of a social contributor to admission (OR, 4.40; P = .014), and history of alcohol or drug abuse (OR, 4.93; P = .025). Admission for a noncancer condition was associated with decreased PAH risk (OR, 0.32; P = .011). On multivariable analysis, presence of a social contributor to admission (OR, 9.35; P = .002) and admission as a result of a noncancer condition (OR, 0.16; P = .038) remained predictive of PAH risk. CONCLUSION:: A significant proportion of hospitalizations among patients with prostate cancer are potentially avoidable. Understanding factors predictive of risk for PAH can help inform programs aimed at avoiding such admissions to improve overall care quality and value.

Interpreting Oncology Care Model Data to Drive Value-Based Care: A Prostate Cancer Analysis.

PURPOSE:: The Oncology Care Model (OCM) must be clinically relevant, accurate, and comprehensible to drive value-based care. METHODS:: We studied OCM data detailing observed and expected expenses for 6-month-long episodes of care for patients with prostate cancer. We constructed seven disease state-treatment dyads into which we grouped each episode on the bases of diagnoses, procedures, and medications in OCM claims data. We used this clinical-administrative stratification model to facilitate a comparative cost analysis, and we evaluated emergency department and hospital utilization and drug therapy as potential drivers of cost. RESULTS:: We examined 377 episodes of care, pertaining to 210 patients, that took place within our health system from January 2012 to June 2015. Ninety-six percent of episodes were assigned to clinically meaningful dyads. Excessive expenses were seen in metastatic, castration-resistant dyads containing second-line hormone therapy (ratio of observed to expected expenses [O/E], 2.66), chemotherapy (O/E, 2.09), and radium-223/sipuleucel-T (O/E, 3.01). An OCM update correcting for castration-resistant prostate cancer led to small differences in observed expenses (0% to +2%) but large changes in expected expenses (-17% to -27% for hormone-sensitive dyads and +136% to +141% for castration-resistant dyads). O/E increased up to 38% for hormone-sensitive dyads and decreased up to 58% for castration-resistant dyads. Emergency department and hospital utilization seems to drive cost for castration-resistant dyads but not for hormone-sensitive dyads. In the revised OCM model, overall O/E for all episodes improved by 22%, from 1.48 to 1.15. CONCLUSION:: Our experience with OCM highlights the limitations of administrative claims data within this model and illustrates a method of translating these data into clinically meaningful information to improve value.

The Impact of Economic Recession on the Incidence and Treatment of Cancer.

PURPOSE: The impact of economic recessions on the incidence and treatment of cancer is unknown. We test the hypothesis that cancer incidence and treatment rates decrease during a recession, and that this relationship is more pronounced in cancers that present with mild, more easily ignored symptoms. METHODS AND MATERIALS: Data on incidence and treatment for all cancers, and breast and pancreatic cancers specifically, from 1973-2008, were collected using Surveillance Epidemiology and End RESULTS (SEER). The data was adjusted for race, income, and education. Unemployment rate was used as the measure of economic recession. Data was log-transformed, and multivariate linear mixed regression was used. RESULTS: Adjusting for socioeconomic factors, the data revealed a significant inverse correlation between unemployment and rates of cancer incidence and treatment. Every 1% increase in unemployment was associated with a 2.2% (95% CI: 1.6-2.8%, p<0.001) reduction in cancer incidence, a 2.0% (1.2-2.8%, p=0.0157) decrease in surgery, and a 9.1% (8.2-10.0% p<0.001) decrease in radiation therapy (RT). Breast cancer incidence and treatment had a dramatic inverse relationship - 7.2% (6.3-8.1%), 6.7% (5.7-7.6%), and 19.0% (18.1-19.8%), respectively (p<0.001 for all). The decrease in incidence was only significant for in situ and localized tumors, but not in regional or distant breast cancer. Compared to breast cancer, pancreatic cancer had a weaker relationship between unemployment and incidence: 2.6% (1.8-3.3%, p=0.0005), surgery: 2.4% (2.0-2.7%, p<0.001), and RT: 1.9% (1.5-2.2% p<0.001). CONCLUSIONS: Increasing unemployment rates are associated with a decrease in the incidence and treatment of all cancers. This effect is exaggerated in breast cancer, where symptoms can more easily be ignored and where there are widely used screening tests relative to pancreatic cancer.

Assessment of nodal target definition and dosimetry using three different techniques: implications for re-defining the optimal pelvic field in endometrial cancer.

PURPOSES: 1. To determine the optimal pelvic nodal clinical target volume for post-operative treatment of endometrial cancer. 2. To compare the DVH of different treatment planning techniques applied to this new CTV and the surrounding tissues. METHODS AND MATERIALS: Based on the literature, we selected a methodology to delineate nodal target volume to define a NEW-CTV and NEW-PTV. Conventional 2D fields, 3D fields based on anatomic guidelines per RTOG 0418, 3D fields based on our guidelines, and IMRT based on our guidelines were assessed for coverage of NEW-CTV, NEW-PTV, and surrounding structures. CT scans of 10 patients with gynecologic malignancies after TAH/BSO were used. DVHs were compared. RESULTS: For NEW-PTV, mean V45Gy were 50% and 69% for 2D and RTOG 0418-3DCRT vs. 98% and 97% for NEW-3DCRT and NEW-IMRT (p < 0.0009). Mean V45Gy small bowel were 24% and 20% for 2D and RTOG 0418-3DCRT, increased to 32% with NEW-3DCRT, and decreased to 14% with IMRT (p = 0.005, 0.138, 0.002). Mean V45Gy rectum were 26%, 35%, and 52% for 2D, RTOG 0418-3DCRT, and NEW-3DCRT, and decreased to 26% with NEW-IMRT (p < 0.05). Mean V45Gy bladder were 83%, 51%, and 73% for 2D, RTOG 0418-3DCRT, and NEW-3DCRT, and decreased to 30% with NEW-IMRT (p < 0.002). CONCLUSIONS: Conventional 2D and RTOG 0418-based 3DCRT plans cover only a fraction of our comprehensive PTV. A 3DCRT plan covers this PTV with high doses to normal tissues, whereas IMRT covers the PTV while delivering lower normal tissue doses. Re-consideration of what specifically the pelvic target encompasses is warranted.