Dynamics of magnetization at infinite temperature in a Heisenberg spin chain

Understanding universal aspects of quantum dynamics is an unresolved problem in statistical mechanics. In particular, the spin dynamics of the one-dimensional Heisenberg model were conjectured as to belong to the Kardar-Parisi-Zhang (KPZ) universality class based on the scaling of the infinite-temperature spin-spin correlation function. In a chain of 46 superconducting qubits, we studied the probability distribution of the magnetization transferred across the chain's center, [Formula: see text]. The first two moments of [Formula: see text] show superdiffusive behavior, a hallmark of KPZ universality. However, the third and fourth moments ruled out the KPZ conjecture and allow for evaluating other theories. Our results highlight the importance of studying higher moments in determining dynamic universality classes and provide insights into universal behavior in quantum systems.

Early palliative care and quality of dying and death in patients with advanced cancer

OBJECTIVE

Early palliative care (EPC) in the outpatient setting improves quality of life for patients with advanced cancer, but its impact on quality of dying and death (QODD) and on quality of life at the end of life (QOL-EOL) has not been examined. Our study investigated the impact of EPC on patients' QODD and QOL-EOL and the moderating role of receiving inpatient or home palliative care.

METHOD

Bereaved family caregivers who had provided care for patients participating in a cluster-randomised trial of EPC completed a validated QODD scale and indicated whether patients had received additional home palliative care or care in an inpatient palliative care unit (PCU). We examined the effects of EPC, inpatient or home palliative care, and their interactions on the QODD total score and on QOL-EOL (last 7 days of life).

RESULTS

A total of 157 caregivers participated. Receipt of EPC showed no association with QODD total score. However, when additional palliative care was included in the model, intervention patients demonstrated better QOL-EOL than controls (p=0.02). Further, the intervention by PCU interaction was significant (p=0.02): those receiving both EPC and palliative care in a PCU had better QOL-EOL than those receiving only palliative care in a PCU (mean difference=27.10, p=0.002) or only EPC (mean difference=20.59, p=0.02).

CONCLUSION

Although there was no association with QODD, EPC was associated with improved QOL-EOL, particularly for those who also received inpatient care in a PCU. This suggests a long-term benefit from early interdisciplinary palliative care on care throughout the illness.

TRIAL REGISTRATION NUMBER

ClinicalTrials.gov Registry (#NCT01248624).

Real-World Experience of Intensity Modulated Radiation Therapy and Concurrent Chemotherapy for Anal Cancer with Long-Term Follow up and Clinical Outcomes

PURPOSE/OBJECTIVE(S)

The standard treatment for epidermoid anal cancer (AC) is concurrent chemoradiation (CRT). Here we present real world evidence of the safety and outcomes of AC patients managed by IMRT and concurrent chemotherapy at a single academic cancer center.

MATERIALS/METHODS

We retrospectively reviewed the outcomes of 180 AC patients treated with definitive CRT between 2011 and 2018. Patients were managed according to a prospectively designed protocol of IMRT with radiation dose escalated according to tumor stage: 50.4, 55.8 and 63 Gy for T1, T2 and T3/T4 disease respectively, and 36 Gy for elective nodal RT. Involved nodes were given the same dose based on T category. Concurrent chemotherapy consisted of two cycles of mitomycin C (MMC, 12 mg/m2) and 5-fluorouracil (5FU, 1000 mg/m2/day x 4 days) given on week 1 and 5. There was no planned treatment break. Univariate and multivariate analysis for outcomes were performed using Cox proportional hazard method and likelihood ratio statistics. Overall survival (OS) disease free (DFS), colostomy-free survival (CFS) and local failure rates (LFR) were described by Kaplan-Meier methods.

RESULTS

There were 128 female and 52 male patients with a median age of 64 (IQR 55-74). The median size of the primary was 4.0 cm (0.6-11.0 cm). There were 18 T1, 91 T2, 38 T3 and 33 T4 lesions; 50.6% (91/180) of the patients had N0 disease. Thirteen (7.2%) did not receive concurrent chemotherapy, and 16 (8.9%) failed to complete treatment as planned. Forty-three (23.9%) patients had a treatment gap >5 (6-33) days. Eighteen of 147 (12%) with T1-3 disease failed locally, LF was observed in 13/33 (39%) T4 lesions (P = 0.0002). The 5-year OS, DFS, CFS and LFR were 85.1%, 75.6%, 87.6% and 15.5% respectively. On multivariate analysis, increasing age and N+ disease were significant for worse OS, and increasing size of the primary tumor was the only significant factor for worse DFS, CFS and LFR. Grade ≥3 acute toxicities were observed in 42.8% of patients, with grade ≥3 neutropenia and febrile neutropenia observed in 18.9% and 13.9% of patients respectively. Six patients (3.3%) died of acute toxicities. Thirteen (7%) patients experienced grade ≥3 late toxicities.

CONCLUSION

Size of the primary appears to be the most important determinant of outcome following standard CRT using IMRT for AC. Despite IMRT, almost 1 in 4 patients required a treatment break, and over 40% experienced grade ≥3 acute toxicities including neutropenia and febrile neutropenia. Future studies with RT dose escalation or de-escalation, stratifying patients based on tumor size, HPV status and molecular markers are necessary to improve outcomes and decrease treatment related toxicity.

Hospital encounters and associated costs of prostate evaluation for clinically important disease MRI vs. standard evaluation procedures (PRECISE) study from a provincial-payer perspective

INTRODUCTION

Systematic transrectal ultrasonography (TRUS) biopsy has been the standard diagnostic tool for prostate cancer (PCa) but is subject to limitations, such as a high false-negative rate of cancer detection. Multiparametric magnetic resonance imaging (mpMRI) prior to biopsy is emerging as an alternative diagnostic procedure for PCa. The PRECISE study found that MRI followed by a targeted biopsy was more accurately able to identify clinically significant cancer than TRUS biopsy.

METHODS

PRECISE study patients recruited in Ontario between January 2017 and November 2019 were linked to various Ontario provincial administrative databases available at the Institute for Clinical and Evaluative Sciences (ICES ) to determine health resources used, associated costs, and hospitalizations in the 14 days after biopsy. Costs are presented in 2021 CAD.

RESULTS

A total of 281 males were included in this study, with 48.4% of the patients in the TRUS biopsy group, 28.1% in the MRI+, and 23.5% in the MRI- group. Twenty-one patients (15%) from the TRUS biopsy group were seen at a hospital in the 14 days after their biopsy compared to fewer than five patients (6%) from the MRI+ group. The mean per person per year (PPPY) costs for the TRUS and all MRI groups (MRI- and MRI+) were $7828 and $8525, respectively.

CONCLUSIONS

Patients in the TRUS biopsy group experienced more hospital encounters compared to patients who received an MRI prior to their biopsy. This economic analysis suggests that MRI imaging prior to biopsy is not associated with a significant increase in costs.

A 20-Year Review of the Ontario Radiation Therapy Access to Care Crisis: Lessons Applied to the COVID-19 Pandemic

Purpose/Objective(s)

Cancer systems across Canada are struggling with an acute access to care crisis resulting from the COVID-19 pandemic. Policies are being considered and applied across the country to address the backlog of patients needing access to cancer services, including radiation therapy (RT). The purpose of this research was to assess the impact of central (provincial health ministry and cancer agency) and regional (cancer center) policies on access to RT in Ontario between 1997 and 2017, and consider their relevance to today's challenges.

Materials/Methods

The research design was a case study with multiple embedded units. The case was Cancer Care Ontario. The embedded units were four diverse regional cancer centers representing the 14 different Ontario cancer centers. Methods included a document review, longitudinal quantitative data collection, and 43 key informant interviews. The theoretical underpinning was an extension of Kingdon's Multiple Streams Framework, to examine the ‘problem’, ‘policy’ solutions and ‘politics’ surrounding the crisis.

Results

The access to RT problem started as a wait time issue in the 1990s caused by inadequate RT facility capacity and a shortage of RT specialized staff, and evolved into a shortfall in RT utilization. Thirty-seven policies were identified and categorized as: improving existing RT capacity (n=5), system planning (n=7), performance management (n=6), human resources (HR, n=12), and building new RT capacities (n=7). Ten of the HR policies implemented to address recruitment and retention had mixed success because of implementation and political context issues. Many of these same policies are now being applied across Canada to address access to cancer services during the COVID-19 pandemic.

Conclusion

A 20-year case study of the Ontario RT access crisis in the 1990s, and the post-crisis periods, offer many useful learnings that can be applied to current policy challenges in access to care due to the ongoing pandemic.

Factors associated with acute kidney injury among patients with cancer treated with immune checkpoint inhibitor therapy: A population-based study

2584

Background: Cancer immune checkpoint inhibitor (ICI) therapy may be associated with kidney immune-related adverse events (IRAEs) and other causes of acute kidney injury (AKI). In clinical trials, the frequency of AKI events was uncommon, however, further real-world study is warranted. Methods: We evaluated the proportion of AKI events among patients with advanced cancer (bladder, head and neck, lung, kidney and malignant melanoma) treated with ICI therapy in Ontario, Canada from 2012 - 2018. AKI was defined by a rise in the concentration of serum creatinine as per Kidney Disease: Improving Global Outcomes (KDIGO) criteria. A multivariable regression model was used to identify predictors of AKI while accounting for the competing risk of death. Results: A total of 4,380 patients received ICI therapy. In follow-up, 1,283 (29%) had recorded AKI event (any stage AKI) and 289 (7%) had a severe AKI event (≥ stage 2). Median time to AKI was 6 months (Interquartile Range 2-16 months) and ≤ 1 % of patients received dialysis therapy. Within 30 days of any observed AKI event, 853 (58%) discontinued ICI therapy, 372 (29%) were hospitalized and 266 (21%) died. Mortality was significantly higher among patients who experiencing a severe AKI event (≥ stage 2) as compared to patients with a less severe AKI event (stage 1) or no observed AKI event. Among patients alive at 30 days following an AKI event, 14% received an outpatient corticosteroid or immunosuppressive therapy prescription, 7% had a visit with a nephrologist. Characteristics associated with a higher risk of AKI included female sex, bladder or kidney cancer (reference malignant melanoma), history of hypertension or diabetes, higher Charlson comorbidity score, a baseline estimated glomerular filtration rate less than 30 mL/min/1.73 m2, or outpatient prescription for either a proton pump inhibitor or non-steroidal anti-inflammatory drug. Among patients with an AKI event and treatment discontinuation, re-challenge of ICI therapy was infrequent (16%) with a significant risk of a recurrent AKI event (57%). Conclusions: In a population-based study among patients with cancer receiving ICI therapy, the rate of AKI was common (29%) but severe AKI was less frequent (7%). Rates of ICI discontinuation, hospitalization and death are substantial following an AKI event. Kidney function should be monitored carefully among patients undergoing ICI therapy who have common risk factors for developing renal disease. Nephrology consultation may be optimized among patients who develop a severe AKI event, especially among individuals who are considered for ICI therapy re-challenge.

Phase-specific costs of care in diffuse large B-cell lymphoma in Ontario, Canada

e18840

Background: This Canadian study analyzes real-world data to estimate phase-specific costs and health care resource utilization for patients with diffuse large B-cell lymphoma (DLBCL). Methods: A cohort of adult patients diagnosed with DLBCL were identified between 2003-2014 from the Ontario Cancer Registry (OCR) and linked to treatment data from Cancer Care Ontario and other provincial administrative health care databases. Health system costs and resource utilization were determined for four defined phases of care in which index date was defined as the OCR diagnosis date, including: pre-diagnosis (90 days prior to index date), initial treatment (index date to 6 months), follow-up (end of treatment phase to beginning of end-of-life or completion of cohort follow-up) and end-of-life (last 6 months of life for patients who died). Costs (Canadian dollars in 2014) and resource utilization were normalized to 30-days. Results: There were 5,392 individuals (53% males; median age 64 years (IQR 53-74)) diagnosed with DLBCL. The median follow-up was 1,903 days (IQR 1,194-2,882) from index date. At follow-up completion, 4,015 (74.5%) individuals were alive. Total mean 30-day cost was $1,175 (±2,267) in pre-diagnosis, $9,166 (±5,581) during initial treatment, $1,462 (±2,757) during follow-up, and $7,965 (±7,104) during end-of-life. The total 30-day cost for inpatient care was $3,864 (±5,013) during treatment (n = 3,362 hospitalized during treatment) and $5,045 (±5,908) for end-of-life (n = 1,367 hospitalized during end-of-life). For home care, the mean 30-day costs were $528 (±512) during initial treatment (n = 3,508, 65% of cohort) and $725 (±884) during end-of-life (n = 1,211, 23% of cohort). Mean 30-day cost of cancer medication was $80 (±455) in pre-diagnosis, $3,115 (±1,236) during initial treatment, $232 (±712) during follow-up, and $305 (±694) during end-of-life. A subset of 239 individuals had autologous stem cell transplantation (ASCT) for relapsed/refractory DLBCL. For the ASCT subgroup, the mean total 30-day cost was $11,653 (±4632) in pre-ASCT phase, $1,542 (±2961) in post-ASCT phase, and $14,094 (±12,845) in end-of-life phase for those that subsequently died (n = 118). A further subset of 52 individuals had relapsed and proceeded to third-line therapy. The total mean total 30-day cost was $8,288 (±8,943) in the first 6-month follow-up, $2,584 (±4,000) in the post 6-month follow-up, and $14,999 (±9,443) in end-of-life phase for those that died (n = 29). Conclusions: Total mean 30-day cost was highest in initial treatment phase and end-of-life phases following a U-shaped pattern. Inpatient care was the cost driver across all phases. Individuals who required ASCT had significantly increased costs. Home care was less frequently accessed during end-of-life phase compared to initial treatment phase (23% vs. 65%). These findings can help allocate appropriate resources throughout the different phases of cancer care.

Real-world cost-effectiveness of first-line gemcitabine + nab-paclitaxel versus FOLFIRINOX in patients with advanced pancreatic cancer: A population-based retrospective cohort study in Ontario, Canada

529

Background: Currently, there are no direct randomized control trials (RCTs) comparing gemcitabine and nab-paclitaxel (Gem-Nab) and FOLFIRINOX for advanced pancreatic cancer (APC). Thus, previous model-based cost-effectiveness analyses were based on indirect comparisons of RCT data. While it is well known that RCT-based efficacy does not often translate to real-world effectiveness, there is limited literature investigating the comparative cost-effectiveness of Gem-Nab versus FOLFIRINOX for APC in the real-world. The objective of this study is to examine the real-world cost-effectiveness of Gem-Nab versus FOLFIRINOX in patients with APC in Ontario, Canada. Methods: This population-based retrospective cohort study compared all patients treated with first-line Gem-Nab or FOLFIRINOX for APC with ECOG performance status 0-1 in Ontario from April 2015 to March 2019. Patients were linked to administrative databases to identify key characteristics and costing data. Using propensity scores and a stabilizing weights method, an inverse probability of treatment weighted cohort was developed. Mean survival and total costs were calculated over a 5-year time horizon, adjusted for censoring and discounted at 1.5% (per Canadian guidelines). Incremental cost-effectiveness ratio and net monetary benefit were computed (measured in life-years and quality-adjusted life years) to estimate cost-effectiveness from the public healthcare payer’s perspective. A sensitivity analysis was conducted using the propensity score matching method. Results: 1,988 patients were identified (Gem-Nab: 928, FOLFIRINOX: 1,060). Mean survival was lower for patients in the Gem-Nab group than the FOLFIRINOX group (0.98 versus 1.26 life-years, incremental -0.28 (95% confidence interval -0.47, -0.13)). Patients in the Gem-Nab group also incurred greater mean 5-year total costs (Gem-Nab: $103,884, FOLFIRINOX: $101,518). Key cost contributors include ambulatory cancer care, acute in-patient hospitalization, and systemic therapy drug acquisition. Gem-Nab was dominated by FOLFIRINOX, as it is less effective and more costly. Results from the sensitivity analysis were similar. Conclusions: In routinely treated unselected patients, Gem-Nab is likely more costly and less effective than FOLFIRINOX and therefore, not considered cost-effective at any commonly accepted willingness-to-pay threshold.

Benefits of High-Volume Medical Oncology Care for Noncurable Pancreatic Adenocarcinoma: A Population-Based Analysis

BACKGROUND

Although pancreatic adenocarcinoma (PA) surgery performed by high-volume (HV) providers yields better outcomes, volume-outcome relationships are unknown for medical oncologists. This study examined variation in practice and outcomes in noncurative management of PA based on medical oncology provider volume.

METHODS

This population-based cohort study linked administrative healthcare datasets and included nonresected PA from 2005 through 2016. The volume of PA consultations per medical oncology provider per year was divided into quintiles, with HV providers (≥16 patients/year) constituting the fifth quintile and low-volume (LV) providers the first to fourth quintiles. Outcomes were receipt of chemotherapy and overall survival (OS). The Brown-Forsythe-Levene (BFL) test for equality of variances was performed to assess outcome variability between provider-volume quintiles. Multivariate regression models were used to examine the association between management by HV provider and outcomes.

RESULTS

A total of 7,062 patients with noncurable PA consulted with medical oncology providers. Variability was seen in receipt of chemotherapy and median survival based on provider volume (BFL, P<.001 for both), with superior 1-year OS for HV providers (30.1%; 95% CI, 27.7%-32.4%) compared with LV providers (19.7%; 95% CI, 18.5%-20.6%) (P<.001). After adjustment for age at diagnosis, sex, comorbidity burden, rural residence, income, and diagnosis period, HV provider care was independently associated with higher odds of receiving chemotherapy (odds ratio, 1.19; 95% CI, 1.05-1.34) and with superior OS (hazard ratio, 0.79; 95% CI, 0.74-0.84).

CONCLUSIONS

Significant variation was seen in noncurative management and outcomes of PA based on provider volume, with management by an HV provider being independently associated with superior OS and higher odds of receiving chemotherapy. This information is important to inform disease care pathways and care organization. Cancer care systems could consider increasing the number of HV providers to reduce variation and improve outcomes.

Impact of Palliative Care Involvement on End-of-Life Care Patterns Among Adolescents and Young Adults With Cancer: A Population-Based Cohort Study

PURPOSE

Evidence suggests that adolescents and young adults (AYAs) with cancer (defined as age 15-39 years) receive high-intensity (HI) medical care at the end-of-life (EOL). Previous population-level studies are limited and lack information on the impact of palliative care (PC) provision. We evaluated prevalence and predictors of HI-EOL care in AYAs with cancer in Ontario, Canada. A secondary aim was to evaluate the impact of PC physicians on the intensity of EOL care in AYAs.

METHODS

A retrospective decedent cohort of AYAs with cancer who died between 2000 and 2017 in Ontario, Canada, was assembled using a provincial registry and linked to population-based health care data. On the basis of previous studies, the primary composite measure HI-EOL care included any of the following: intravenous chemotherapy < 14 days from death, more than one emergency department visit, and more than one hospitalization or intensive care unit admission < 30 days from death. Secondary measures included the most invasive (MI) EOL care (eg, mechanical ventilation < 14 days from death) and PC physician involvement. We determined predictors of outcomes using appropriate regression models.

RESULTS

Of 7,122 AYAs, 43.8% experienced HI-EOL care. PC physician involvement (odds ratio [OR], 0.57; 95% CI, 0.51 to 0.63) and older age at death (OR, 0.60; 95% CI, 0.48 to 0.74) were associated with a lower risk of HI-EOL care. AYAs with hematologic malignancies were at highest risk for HI and MI-EOL care. PC physician involvement substantially reduced the odds of mechanical ventilation at EOL (OR, 0.36; 95% CI, 0.30 to 0.43).

CONCLUSION

A large proportion of AYAs with cancer experience HI-EOL care. Our study provides strong evidence that PC physician involvement can help mitigate the risk of HI and MI-EOL care in AYAs with cancer.

Real-world cost-effectiveness of pertuzumab (P) with trastuzumab + chemo (T+Chemo) in patients (pts) with metastatic breast cancer (MBC): A population-based retrospective cohort study by the Canadian Real-world Evidence for Value in Cancer Drugs (CanREValue) collaboration

1048

Background: Addition of P to T+chemo for MBC pts has been shown to improve overall survival (OS) in a pivotal randomized trial (hazard ratio [HR] = 0.66, 95% CI: 0.52, 0.84) (Baselga et al., NEJM 2012). In Canada, the manufacturer submission to the health technology assessment agency estimated that P produced 0.64 life years gained (LYG) with an incremental cost-effectiveness ratio (ICER) of $187,376/LYG over 10 years (CADTH-pCODR, 2013). This retrospective cohort analysis aims to determine the comparative real-world population-based effectiveness and cost-effectiveness of P among MBC pts in Ontario, Canada. Methods: MBC pts were identified from the Ontario Cancer Registry and linked to the New Drug Funding Program database to identify receipt of treatment between 1/1/2008 and 3/31/2018. Cases received P-T-chemo after universal public funding of P (Nov 2013) and controls received T-chemo before. Demographic (age, socioeconomic, rurality) and clinical (comorbidities, prior adjuvant treatments, prior breast cancer surgery, prior radiation, stage at diagnosis, ER/PR status) characteristics were identified from linked admin databases balanced between cases and controls using propensity score matching. Kaplan-Meier methods and Cox regressions accounting for matched pairs were used to estimate median OS and HR. 5-year mean total costs from the public health system perspective were estimated from admin claims databases using established direct statistical methods and adjusted for censoring of both cost and effectiveness using inverse probability weighting. ICERs and 95% bootstrapped CIs were calculated, along with incremental net benefit (INB) at various willingness-to-pay values using net benefit regression. Results: We identified 1,823 MBC pts with 912 cases and 911 controls (mean age = 55 years), of which 579 pairs were matched. Cases had improved OS (HR = 0.66; 95% CI: 0.57, 0.78), with median 3.4 years, compared to controls median OS of 2.1. P provided an additional 0.63 (95% CI: 0.48 – 0.84) LYG at an incremental cost of $196,622 (95% CI: $180,774, $219,172), with a mean ICER = $312,147/LYG (95% CI: $260,752, $375,492). At threshold of $100,000/LYG, the INB was -$133,632 (95% CI: -$151,525, -$115,739) with < 1% probability of being cost-effective. Key drivers of incremental cost increase between groups included drug and cancer clinic costs. Conclusions: The addition of P to T-chemo for MBC increased survival but at significant costs. The ICER based on direct real-world data was higher than the initial economic model due to higher total costs for pts receiving P. This study demonstrated feasibility to derive ICER from person-level real-world data to inform cancer drug life-cycle health technology reassessment.

The PROVIEW+ tool: Developing and validating a tool to predict risk of poor performance status and severe symptoms in cancer patients over time

12095

Background: There are numerous predictive cancer tools that focus on survival. However, no tools predict risk of low performance status or severe symptoms, which are important for patient decision-making and early integration of palliative care. The aim of this study was to develop and validate a model for all cancer types that predicts the risk for having low performance status and severe symptoms. Methods: A retrospective, population-based, predictive study using linked administrative data from cancer patients from 2008-2015 in Ontario, Canada. Patients were randomly selected for model derivation (60%) and validation (40%). The derivation cohort was used to develop a multivariable logistic regression model to predict the risk of having the reported outcomes in the subsequent 6 months. Model performance was assessed using discrimination and calibration plots. The main outcome was low performance status using the Palliative Performance Scale. Secondary outcomes included severe pain, dyspnea, well-being, and depression using the Edmonton Symptom Assessment System. Outcomes were recalculated after each of 4 annual survivor marks. Results: We identified 255,494 cancer patients (57% female; median age of 64; common cancers were breast (24%) and lung (13%)). At diagnosis, the risk of having low performance status, severe pain, well-being, dyspnea, and depression in 6-months is 1%, 3%, 6%, 13% and 4%, respectively for the reference case (i.e. male, lung cancer, stage I, no symptoms). Generally these covariates increased the outcome risk by > 10% across all models: obstructive lung disease, dementia, diabetes; radiation treatment; hospital admission; high pain; depression; Palliative Performance Scale score of 60-10; issues with appetite; or homecare. Model discrimination was high across all models. Conclusions: The model accurately predicted changing cancer risk for low performance status and severe symptoms over time. Providing accurate predictions of future performance status and symptom severity can support decision-making and earlier initiation of palliative care, even alongside disease modifying therapies.

An evaluation of administrative data linkage for measurement of real-world outcomes of large clinical panel sequencing for advanced solid tumors

283

Background: There is limited real-world evidence of impact of large clinical panel sequencing on treatment-matching for patients with advanced solid tumors. The province of Ontario has a single payer, publicly funded health care system. We linked genomic testing results from a prospective province-wide trial, OCTANE (Ontario-Wide Cancer TArgeted Nucleic Acid Evaluation), to administrative data to determine the feasibility of this approach for evaluating survival and the impact of sequencing on treatment matching. Methods: We linked all Ontario patients from Princess Margaret (PM) with panel testing results (tumor-only 555-gene panel) to province-wide administrative data on treatments and outcomes. Patients were recruited from August 2016 to August 2018. Only clinically actionable variants based upon OncoKB annotation (Level 1 and 2) were assessed for genotype-informed treatment matching. Results: All 888 eligible patients were successfully linked to administrative data. Mean age was 58 (±13) years, 635 (71.5%) were female. Most common disease sites were ovary (26.4%), uterus (14.0%), colorectal (11.8%) and breast (9.5%). Administrative data vital status was more complete than trial collected data with 262 of 476 deaths only recorded in administrative data. Median survival was 1.70 years (95% confidence interval 1.50-1.91). 247 (27.8%) had actionable mutations, most commonly PIK3CA (54.7%), BRCA1 (15.8%), BRCA2 (15.0%) and BRAF (8.9%). 37 (15.0%) and 42 (17.0%) patients with actionable mutations received targeted therapy within 6 and 12 months of test report date, respectively. Conclusions: This is the first known feasibility study of linked administrative data to measure outcomes of large clinical panel sequencing for patients with advanced solid tumors. Vital status was more complete with administrative data compared to trial-collected data, and treatment data was successfully linked. About one in twenty-one enrolled patients received genome-informed treatments within 12 months, or about one in six of all patients with actionable mutations. This may be due to short interval follow up, trial and drug access, successful standard of care treatments, early patient deterioration or limited alterations covered by the panel, among other causes.

A method to audit and score implementation of knowledge translation (KT) interventions in large health regions - an observational pilot study using rectal cancer surgery in Ontario

Background

Across Ontario, since the year 2006 various knowledge translation (KT) interventions designed to improve the quality of rectal cancer surgery have been implemented by the provincial cancer agency or by individual researchers. Ontario is divided administratively into 14 health regions. We piloted a method to audit and score for each region of the province the KT interventions implemented to improve the quality of rectal cancer surgery.

Methods

We interviewed stakeholders to audit KT interventions used in respective regions over years 2006 to 2014. Results were summarized into narrative and visual forms. Using a modified Delphi approach, KT experts reviewed these data and then, for each region, scored implementation of KT interventions using a 20-item KT Signature Assessment Tool. Scores could range from 20 to 100 with higher scores commensurate with greater KT intervention implementation.

Results

There were thirty interviews. KT experts produced scores for each region that were bimodally distributed, with an average score for 2 regions of 78 (range 73–83) and for 12 regions of 30.5 (range 22–38).

Conclusion

Our methods efficiently identified two groups with similar KT Signature scores. Two regions had relatively high scores reflecting numerous KT interventions and the use of sustained iterative approaches in addition to those encouraged by the provincial cancer agency, while 12 regions had relatively low scores reflecting minimal activities outside of those encouraged by the provincial cancer agency. These groupings will be used for future comparative quantitative analyses to help determine if higher KT signature scores correlate with improved measures for quality of rectal cancer surgery.

An evaluation of administrative data linkage for measurement of real-world outcomes of large clinical panel sequencing for advanced solid tumors

e19303

Background: There is limited real-world evidence of impact of large clinical panel sequencing on treatment-matching for patients with advanced solid tumors. The province of Ontario has a single payer, publicly funded health care system. We linked genomic testing results from a prospective province-wide trial, OCTANE (Ontario-Wide Cancer TArgeted Nucleic Acid Evaluation), to administrative data to determine the feasibility of this approach for evaluating survival and the impact of sequencing on treatment matching. Methods: We linked all Ontario patients from Princess Margaret (PM) with panel testing results (tumor-only 555-gene panel) to province-wide administrative data on treatments and outcomes. Patients were recruited from August 2016 to August 2018. Only clinically actionable variants based upon OncoKB annotation (Level 1 and 2) were assessed for genotype-informed treatment matching. Results: All 888 eligible patients were successfully linked to administrative data. Mean age was 58 (±13) years, 635 (71.5%) were female. Most common disease sites were ovary (26.4%), uterus (14.0%), colorectal (11.8%) and breast (9.5%). Administrative data vital status was more complete than trial collected data with 262 of 476 deaths only recorded in administrative data. Median survival was 1.70 years (95% confidence interval 1.50-1.91). 247 (27.8%) had actionable mutations, most commonly PIK3CA (54.7%), BRCA1 (15.8%), BRCA2 (15.0%) and BRAF (8.9%). 37 (15.0%) and 42 (17.0%) patients with actionable mutations received targeted therapy within 6 and 12 months of test report date, respectively. Conclusions: This is the first known feasibility study of linked administrative data to measure outcomes of large clinical panel sequencing for patients with advanced solid tumors. Vital status was more complete with administrative data compared to trial-collected data, and treatment data was successfully linked. About one in twenty-one enrolled patients received genome-informed treatments within 12 months, or about one in six of all patients with actionable mutations. This may be due to short interval follow up, trial and drug access, successful standard of care treatments, early patient deterioration or limited alterations covered by the panel, among other causes.

Dementia risk among post-menopausal women treated with endocrine therapy for early-stage breast cancer in Ontario, Canada

521

Background: The association between anti-estrogen therapy and risk of dementia remains controversial. We performed a population-based real-world study investigating the association between endocrine therapy use and dementia. Methods: We used health administrative data collected from post-menopausal women (aged ≥66 years) who were diagnosed with breast cancer and started on adjuvant endocrine therapy from 2005-2012. Patients were classified by use of either an aromatase inhibitor or tamoxifen and followed to measure an unadjusted cumulative incidence of developing dementia. A multivariable analysis adjusting for age, income quintile, medical comorbidities, and duration of endocrine therapy was completed using a Cox-proportional hazards model. Results: We identified 12,077 patients of whom 73% were treated with an aromatase inhibitor and 27% with tamoxifen. The median age was 73 years (IQR 69-78), 64% of patients were treated with lumpectomy, and 19% received adjuvant chemotherapy. The unadjusted event rate for developing dementia was Hazard Ratio (HR)= 0.70 (95% confidence interval (CI)=0.63-0.78, p-value<0.0001) among patients receiving an aromatase inhibitor versus tamoxifen and the 5-year dementia incidence rate was 7.4% versus 9.2% respectively. Our multivariable analysis showed a significant decrease in the rate of dementia in patients treated with an aromatase inhibitor compared to tamoxifen (HR=0.88, 95% CI 0.78-0.98, p-value=0.02) with a median of 5.9 years of follow-up. Factors associated with the development of dementia included older age, previous history of ischemic heart disease, diabetes, hypertension and stroke. Duration of endocrine therapy and previous use of adjuvant chemotherapy were not associated with dementia in our study. Conclusions: This investigation indicates that use of aromatase inhibitors compared to tamoxifen is associated with a lower risk of developing dementia among post-menopausal breast cancer patients. Further prospective studies investigating the neurocognitive effects of endocrine therapy are warranted.

Development and validation of a risk prediction model for poor performance status and severe symptoms among cancer patients

12097

Background: Existing cancer predictive tools focus on survival, but few incorporate patient-reported outcomes to predict quality-of-life domains, such as symptoms and performance status. The objective was to develop and validate a predictive cancer model (called PROVIEW) for poor performance status and severe symptoms over time. Methods: We used a retrospective, population-based, cohort study of patients, with a cancer diagnosis, in Ontario, Canada between 2008-2015. We randomly selected 60% of patients for model derivation and 40% for validation. Using the derivation cohort, we developed multivariable logistic regression models with baseline characteristics, using a backward stepwise variable selection process. The primary outcome was odds of having poor performance status six months from index date, as measured by a score < = 30 out of 100 on the Palliative Performance Scale. The index date for each model was diagnosis (Year 0), which was then re-calculated at each of 4 annual survivor marks after diagnosis (up to Year 4). Secondary outcomes included having severe pain, dyspnea, well-being, or depression, as measured by a score of > = 7 out of 10 on the Edmonton Symptom Assessment System. Covariates included demographics, clinical information, current symptoms and performance status, and healthcare utilization. Model performance was assessed by AUC statistics and calibration plots. Results: Our population-based cohort identified 125,479 cancer patients for the performance status model in Year 0. The median diagnosis age was 64 years, 57% were female, and the most common cancers were breast (24%), lung (13%), and prostate (9%). 32% had Stage 3 or 4 disease. In Year 0 after backwards selection, the odds of having a poor performance status in 6 months was increased by more than 10% when the patient had: COPD, dementia, diabetes; radiation treatment; a hospital admission in the prior 3 months; high pain or depression; a current performance status < = 30; any issues with appetite; or received end-of-life homecare. Generally, these variables were also associated with a > 10% increased odds in other years and for the secondary outcomes. The average AUC across all 25 models is 0.7676 which indicates high model discrimination. Conclusions: The PROVIEW model accurately predicts risk of having a poor performance status or severe symptoms over time among cancer patients. It has the potential to be a useful online tool for patients to integrate earlier supportive and palliative care.

Development and Validation of a Prognostic Survival Model With Patient-Reported Outcomes for Patients With Cancer

IMPORTANCE

Existing prognostic cancer tools include biological and laboratory variables. However, patients often do not know this information, preventing them from using the tools and understanding their prognosis.

OBJECTIVE

To develop and validate a prognostic survival model for all cancer types that incorporates information on symptoms and performance status over time.

DESIGN, SETTING, AND PARTICIPANTS

This is a retrospective, population-based, prognostic study of data from patients diagnosed with cancer from January 1, 2008, to December 31, 2015, in Ontario, Canada. Patients were randomly selected for model derivation (60%) and validation (40%). The derivation cohort was used to develop a multivariable Cox proportional hazards regression model with baseline characteristics under a backward stepwise variable selection process to predict the risk of mortality as a function of time. Covariates included demographic characteristics, clinical information, symptoms and performance status, and health care use. Model performance was assessed on the validation cohort by C statistics and calibration plots. Data analysis was performed from February 6, 2018, to November 6, 2019.

MAIN OUTCOMES AND MEASURES

Time to death from diagnosis (year 0) recalculated at each of 4 annual survivor marks after diagnosis (up to year 4).

RESULTS

A total of 255 494 patients diagnosed with cancer were identified (135 699 [53.1%] female; median age, 65 years [interquartile range, 55-73 years]). The cohort decreased to 217 055, 184 822, 143 649, and 109 569 patients for each of the 4 years after diagnosis. In the derivation cohort year 0, and the most common cancers were breast (30 855 [20.1%]), lung (19 111 [12.5%]), and prostate (18 404 [12.0%]). A total of 47 614 (31.1%) had stage III or IV disease. The mean (SD) time to death in year 0 was 567 (715) days. After backward stepwise selection in year 0, the following factors were associated with increased risk of death by more than 10%: being hospitalized; having congestive heart failure, chronic obstructive pulmonary disease, or dementia; having moderate to high pain; having worse well-being; having functional status in the transitional or end-of-life phase; having any problems with appetite; receiving end-of-life home care; and living in a nursing home. Model discrimination was high for all models (C statistic: 0.902 [year 0], 0.912 [year 1], 0.912 [year 2], 0.909 [year 3], and 0.908 [year 4]).

CONCLUSIONS AND RELEVANCE

The model accurately predicted changing cancer survival risk over time using clinical, symptom, and performance status data and appears to have the potential to be a useful prognostic tool that can be completed by patients. This knowledge may support earlier integration of palliative care.

The cost of chemotherapy administration: A systematic review and meta-analysis

810

Background: Cancer treatment is a significant driver of healthcare costs worldwide, however, the economic impact of treating patients with anti-neoplastic agents is poorly elucidated. Hence, we conducted a systematic review and meta-analysis to estimate the direct costs associated with administering intravenous chemotherapy in an outpatient setting. Methods: We systematically searched four databases from 2010 to present and extracted hourly administration costs and the respective components of each estimate. Separate analyses were conducted of Canadian and United States (US) studies, respectively, to address a priori hypotheses regarding heterogeneity amongst administration cost estimates. The Drummond checklist was used to assess risk-of-bias. Data were summarized using medians with interquartile ranges and five outliers were identified; costs were presented in 2019 USD. Results: A total of 44 studies were analyzed, including sub-analyses of 19 US and seven Canadian studies. 26/44 studies were of moderate or high quality. When components of administration cost were evaluated, physician costs were reported most frequently (24 studies), followed by lab tests (13) and overhead costs (9). The median cost estimate when outliers were excluded was $142/hour (IQR = $103-166). Sensitivity analyses determined the median administration cost in the US was $149/hour (IQR = $118-158), and was $128/hour (IQR = $102-137) in Canada. Conclusions: There is currently a paucity of literature addressing the costs of chemotherapy administration, and existing studies utilize a patchwork of reporting methodologies which renders direct comparison challenging. Our results demonstrate that the cost of administering chemotherapy is approximately $125-150/hour, globally. This value is dependent upon the region of analysis, inclusiveness of cost subcomponents as well as the methodology used to estimate unit prices, as described here.

Economic impacts of care by high-volume providers for noncurative esophagogastric cancer: A population-based analysis

339

Background: Esophagogastric cancer (EGC) is one of the deadliest and costliest malignancies to treat. Care by high-volume providers can provide better outcomes for patients with EGC. Cost implications of volume-based cancer care are unclear. We examined the cost-effectiveness of care by high-volume medical oncology providers for non-curative management of EGC. Methods: We conducted a population-based cohort study of non-curative EGC over 2005-2017 by linking administrative healthcare datasets. High-volume was defined as >11 patients/provider/year. Healthcare costs ($USD/patient/month-survived) were computed from diagnosis to death or end of follow-up from the perspective of the healthcare system using validated costing algorithms. Multivariable quantile regression examined the association between care by high-volume providers and costs. Sensitivity analyses were conducted by varying costing horizons and high-volume definitions. Results: Among 7,011 non-curative EGC patients, median overall survival was superior with care by high-volume providers with 7.0 (IQR: 3.3-13.3) compared to 5.9 (IQR: 2.6-12.1) months (p < 0.001) for low-volume providers. Median costs/patient/month-lived were lower for high-volume providers ($5,518 vs. $5,911; p < 0.001), owing to lower inpatient acute care costs, despite higher medication-associated and radiotherapy costs. Care by high-volume providers was independently associated with a reduction of $599 per patient/month-lived (95% confidence interval: -966 to -331) compared to low-volume providers. The incremental cost-effectiveness ratio was -393. Care by high-volume providers remained the dominant strategy when varying the high-volume definition and the costing time horizon. Conclusions: Care by high-volume providers for non-curative EGC is associated with superior survival and lower healthcare costs, indicating a dominant strategy that may provide an opportunity to improve cost-effectiveness of care delivery.

Real-world outcomes of FOLFIRINOX vs gemcitabine and nab-paclitaxel in advanced pancreatic cancer: A population-based propensity score-weighted analysis

BACKGROUND

In Ontario, FOLFIRINOX (FFX) and gemcitabine + nab-paclitaxel (GnP) have been publicly funded for first-line unresectable locally advanced pancreatic cancer (uLAPC) or metastatic pancreatic cancer (mPC) since April 2015. We examined the real-world effectiveness and safety of FFX vs GnP for advanced pancreatic cancer, and in uLAPC and mPC.

METHODS

Patients receiving first-line FFX or GnP from April 2015 to March 2017 were identified in the New Drug Funding Program database. Baseline characteristics and outcomes were obtained through the Ontario Cancer Registry and other population-based databases. Overall survival (OS) was assessed using Kaplan-Meier and weighted Cox proportional hazard models, weighted by the inverse propensity score adjusting for baseline characteristics. Weighted odds ratio (OR) for hospitalization and emergency department visits (EDV) were estimated from weighted logistic regression models.

RESULTS

For 1130 patients (632 FFX, 498 GnP), crude median OS was 9.6 and 6.1 months for FFX and GnP, respectively. Weighted OS was improved for FFX vs GnP (HR = 0.77, 0.70-0.85). Less frequent EDV and hospitalization were observed in FFX (EDV: 67.8%; Hospitalization: 49.2%) than GnP (EDV: 77.7%; Hospitalization: 59.3%). More frequent febrile neutropenia-related hospitalization was observed in FFX (5.8%) than GnP (3.3%). Risk of EDV and hospitalization were significantly lower for FFX vs GnP (EDV: OR = 0.68, P = .0001; Hospitalization: OR = 0.76, P = .002), whereas the risk of febrile neutropenia-related hospitalization was significantly higher (OR = 2.12, P = .001). Outcomes for uLAPC and mPC were similar.

CONCLUSION

In the real world, FFX had longer OS, less frequent all-cause EDV and all-cause hospitalization, but more febrile neutropenia-related hospitalization compared to GnP.

Costs after incident breast cancer diagnosis among high-deductible health plan members

120

Background: High-deductible health plans (HDHP) are associated with breast cancer treatment delays of up to 10 months, but their impact on health outcomes is unknown. We hypothesized that, compared with women in generous plans, HDHP members would present with more advanced disease and thus experience higher total costs of early care. Methods: We studied 2004-2014 claims data from a large US health insurer. We included women aged 25-64 who were in traditional low-deductible (≤$500) health plans for 1 baseline year then experienced either an employer-mandated switch to HDHPs (≥$1000) for up to 4 or years or an employer-mandated continuation in low deductible plans. We defined the HDHP switch date as the index date. We then restricted to women who developed incident breast cancer after the index date. Using baseline characteristics, we closely matched HDHP members with incident breast cancer to contemporaneous women with incident breast cancer who remained in low-deductible plans. We measured total costs of all health care services in the 60 days after incident breast cancer diagnosis as a proxy for the intensity of incident breast cancer care. We used negative binomial regression adjusted for baseline characteristics to compare total 60-day costs among HDHP and control members. We also subset analyses to low-income women. Results: We included 1514 HDHP members and 9283 matched controls. 60-day costs after incident breast cancer diagnosis were $24,151 (95% CI: $22,766, $25,535) among HDHP members and $22,474 ($21,952, $22,996) among controls, an absolute difference of $1677 ($197, $3156) and a relative difference of 7.5% (8.1%, 14.1%). Low-income HDHP members had corresponding absolute and relative differences of $2653 ($368, $4939) and 12.5% (1.5%, 23.5). Conclusions: HDHP members with incident breast cancer had 7.5% higher health care costs in the 60 days after incident breast cancer than women with more generous coverage, a finding driven 12.5% higher costs among low-income HDHP members. Results raise concerns that delays in breast cancer care among HDHP members are associated with more advanced disease and adverse outcomes.

Underuse of ECG monitoring in oncology patients receiving QT-interval prolonging drugs

OBJECTIVE

We examined use of ECG monitoring in oncology patients prescribed QT-prolonging drugs.

METHODS

Patients ≥66 years diagnosed with cancer between 2005 and 2011 were identified through the Ontario Cancer Registry and linked to multiple population-based administrative databases to ascertain demographics, comorbidities, prescription drug use, systemic therapy and ECG. QT-prolonging drugs were identified as per drug lists developed by the Arizona Center for Education and Research on Therapeutics. Univariable and multivariable analyses were used to examine factors associated with ECG use in patients on first-line systemic therapy.

RESULTS

A total of 48 236 patients (median age 74; 49% women) received one or more drugs associated with a risk of QT-interval prolongation but only 27% of patients had an ECG performed. Factors associated with more ECG use on multivariable analysis included recent cancer diagnosis (p for trend <0.001 between 2005 and 2011), use of concurrent QT-prolonging drugs (OR=1.15 per each additional QT-prolonging drug, 95% CI 1.12 to 1.17) and the presence of coronary artery disease (OR 1.31; 95% CI 1.25 to 1.38) and heart failure (OR 1.25; 95% CI 1.17 to 1.35). Use of anticancer (OR 0.74; 95% CI 0.70 to 0.79) and antiemetic (OR 0.93; 95% CI 0.88 to 0.99) QT-prolonging drugs was paradoxically associated with less ECG use.

CONCLUSIONS

Our study highlights common use of QT-prolonging drugs and underuse of ECG in oncology patients. Since ECG is an inexpensive, non-invasive and widely available test, it may be readily incorporated in the monitoring of patients for toxicities in routine clinical practice.

The impact of routine ESAS use on overall survival: Results of a population-based retrospective matched cohort analysis

6509

Background: The study objective was to examine the impact of routine Edmonton Symptom Assessment System (ESAS) use on overall survival among adult cancer patients. We hypothesized that patients exposed to ESAS would have better overall survival rates than those who didn’t have ESAS. Methods: The effect of ESAS screening on survival was evaluated in a retrospective matched cohort study. The cohort included all Ontario patients aged 18 or older who were diagnosed with cancer between 2007 and 2015. Patients completing at least one ESAS assessment during the study were considered exposed. The index date was the day of their first ESAS assessment. Follow up time for each patient was segmented into one of three phases: initial, continuing, or palliative care. Exposed and unexposed patients were matched 1:1 using hard (birth year ± 2 years, cancer diagnosis date ± 1 year, cancer type and sex) and propensity-score matching (14 measures including cancer stage, treatments received, and comorbidity). Matched patients were followed until death or the end of study at Dec 31, 2015. Kaplan-Meier curves and multivariable Cox regression were used to evaluate the impact of ESAS on survival. Results: There were 128,893 pairs well matched on all baseline characteristics (standardized difference < 0.1). The probability of survival within the first 5 years was higher among those exposed to ESAS compared to those who were not (73.8% vs. 72.0%, P-value < 0.0001). In the multivariable Cox regression model, ESAS assessment was significantly associated with a decreased mortality risk (HR: 0.49, 95% CI: 0.48-0.49) and this protective effect was seen across all phases. Conclusions: ESAS exposure is associated with improved survival in cancer patients, in all phases of care. To the extent possible, extensive matching methods have mitigated biases inherent to observational data. This provides real world evidence of the impact of routine symptom assessment in cancer care.

Comparison of outcomes in a population-based cohort of women with metastatic breast cancer receiving anti-HER2 therapy with clinical trial outcomes

1037

Background: Little data exist for comparing cardiac safety and survival outcomes of anti-HER2 therapy with concurrent trastuzumab (T) and pertuzumab (P) or ado-T emtansine (TDM1) in metastatic breast cancer (MBC) patients enrolled in randomized clinical trial (RCT) vs those in the real world. Furthermore, whether older patients have worse outcomes is unknown. Methods: This was a retrospective population-based cohort of all women with MBC treated with concurrent T with P or TDM1 in Ontario (between 2012 and 2017), identified from New Drug Funding Program and linked to Ontario Cancer Registry and other administrative datasets. Outcomes were incident heart failure (HF, defined as hospitalization or emergency room visit for HF) and overall survival (OS). RCT data were obtained from digitizing survival curves as per established methods and compared with cohort OS data using log-rank test. Age based comparison of outcomes was conducted for women ≥ 65 years old vs younger. Results: Our cohort composed of 833 (28% > 64 years old), and 397 (28% > 64 years old) women treated with P and TDM1, respectively, of which 46 and 30 had baseline HF, respectively. 49% and 99.5% of women received T prior to P and TDM1, respectively. Incident HF following P or TDM1 initiation was low (P 26 women, TDM1 8 women; Table). HF events was not more in women ≥ 65 years old compared to women < 65 treated with P (16 vs. 10, p = 0.23). Unadjusted OS was significantly worse than RCT OS (Table; P HR 1.67, 95% CI 1.37-2.03, p < 0.0001; TDM1 HR 2.80, 95% CI 2.27-3.44, p < 0.0001). Older women had worse OS than younger women for P (HR 1.54, 95% CI 1.22-1.96, p = 0.0003), but not for TDM1 (HR 1.08, 95% CI 0.81-1.43, p = 0.62). Conclusions: HF incidence during P or TDM1 therapy in this real world cohort was relatively low. Survival in this cohort was significantly worse compared to RCT, particularly for older women, suggesting importance of evaluating effectiveness in an unselected patient population to facilitate informed decision-making based on real-world risks and survival outcomes.[Table: see text]

A weighted criterion-based approach to value assessment of oncology drugs

6627

Background: The rising cost of anti-cancer therapy has motivated recent efforts to quantify the overall value of new cancer treatments. Multi-criteria decision analysis offers a novel approach to establish an explicit framework to evaluate new cancer treatments. Methods: We recruited a diverse multi-stakeholder group who identified and weighted key criteria to establish the Drug Assessment Framework (DAF). Strength of evidence (SOE) modifiers deducted points for lower quality evidence. Through one-on-one meetings with stakeholders, face and content validity of the DAF were established in an iterative process. Construct validity assessed the degree to which DAF scores were associated with the pan-Canadian oncology drug review (pCODR) funding decisions and European Society for Medical Oncology Magnitude of Clinical Benefit score (ESMO-MCBS, version 1.1). Sensitivity analyses were performed on the final results. Results: The final validated DAF includes ten criteria: overall survival, progression-free survival, response rate, quality-of-life, toxicity, unmet need, equity, feasibility, disease severity and caregiver well-being. The first five clinical benefit criteria represent 64% of the total weight. DAF scores range from 0 to 300, reflecting both the expected impact of the drug and the quality of the supporting evidence. When the DAF was retrospectively applied to the last 60 drugs (in blinded fashion) reviewed by pCODR (2015-2018), the mean total DAF score was 94 (range, 18-179). Drugs with positive pCODR funding recommendation had higher DAF scores than drugs not recommended for reimbursement (103 vs. 63, t-test p = 0.0007). Funded drugs had fewer SOE points deducted than those that were not funded (median 0 vs. 24 points deducted, Wilcoxon p = 0.03). The correlation coefficient for DAF and ESMO-MCBS was 0.37 (95% CI, 0.10 to 0.59). Sensitivity analyses that varied the subjective criteria either positively or negatively did not change the results. Conclusions: Using a structured and explicit approach, a criterion-based valuation framework was designed and validated. The DAF can provide a transparent and consistent method to value and prioritize cancer drugs, in order to facilitate the delivery of affordable cancer care.

Moving toward the elimination of cervical cancer: modelling the health and economic benefits of increasing uptake of human papillomavirus vaccines

Background

The human papillomavirus (hpv) is a common sexually transmitted infection and a primary cause of cervical cancer. The Government of Canada has set a target of reaching 90% hpv vaccine coverage among adolescents by 2025. Here, we examine hpv vaccine uptake in school-based immunization programs across Canada and explore how achieving the 90% target could affect the future incidence of cervical cancer, mortality, and health system expenditures in a cohort of Canadian women.

Methods

Data for hpv vaccine uptake in the most recent reported school year available in each jurisdiction were provided in 2017 by jurisdictional school-based immunization programs and were used to estimate a national weighted average of 67%. The OncoSim microsimulation model (version 2.5) was used to compare 3 different levels of hpv vaccine uptake (0%, 67%, 90%) on health and economic outcomes for a hypothetical cohort of all 5- to 10-year-old girls in Canada in 2015.

Results

Vaccine uptake for girls in school-based programs varied from 55.0% to 92.0% in the jurisdictions reviewed. The OncoSim model projects that increasing uptake to 90% from 67% would result in a 23% reduction in cervical cancer incidence rates (to 3.1 cases from 4.0 cases per 100,000, averaged across the lifetime of the cohort) and a 23% decline in the average annual mortality rate (to 1.0 deaths from 1.3 deaths per 100,000). Finally, the model projects that the health system will incur a cost of $9 million (1% increase) during the lifetime of the cohort if uptake is increased to 90% from 67%. Costs are discounted (1.5%) and expressed in 2016 Canadian dollars. Costs reflect the payer perspective.

Conclusions

Our model shows that increasing hpv vaccine uptake to 90% from current levels for girls in school-based immunization programs could result in substantial reductions in the future incidence and mortality rates for cervical cancer in Canada.

Vulnerable And Less Vulnerable Women In High-Deductible Health Plans Experienced Delayed Breast Cancer Care

The effects of high-deductible health plans (HDHPs) on breast cancer diagnosis and treatment among vulnerable populations are unknown. We examined time to first breast cancer diagnostic testing, diagnosis, and chemotherapy among a group of women whose employers switched their insurance coverage from health plans with low deductibles ($500 or less) to plans with high deductibles ($1,000 or more) between 2004 and 2014. Primary subgroups of interest comprised 54,403 low-income and 76,776 high-income women continuously enrolled in low-deductible plans for a year and then up to four years in HDHPs. Matched controls had contemporaneous low-deductible enrollment. Low-income women in HDHPs experienced relative delays of 1.6 months to first breast imaging, 2.7 months to first biopsy, 6.6 months to incident early-stage breast cancer diagnosis, and 8.7 months to first chemotherapy. High-income HDHP members had shorter delays that did not differ significantly from those of their low-income counterparts. HDHP members living in metropolitan, nonmetropolitan, predominantly white, and predominantly nonwhite areas also experienced delayed breast cancer care. Policies may be needed to reduce out-of-pocket spending obligations for breast cancer care.

Survival and cost associated with chemotherapy and chemoradiotherapy among resected pancreas cancer patients

351

Background: Pancreas cancer is expensive to treat, and the effectiveness of adjuvant chemotherapy (CT) and chemoradiation (CRT) following resection is debated. We compared both survival and healthcare costs by adjuvant therapy after curative-intent pancreaticoduodenectomy (PD) for pancreas adenocarcinoma (PC). Methods: All patients with resected PC in Ontario, Canada diagnosed 2004 to 2014 were identified and linked to administrative healthcare databases. Stratified Kaplan—Meier survival curves and log-rank test compared survival across treatment groups. Costs were assessed from the perspective of Ontario’s single-payer healthcare system and compared between CT and CRT. A one-year time horizon was used from the date of surgery. Results: 677 PC patients met all inclusion/exclusion criteria and underwent curative-intent PD with 77% receiving CT and 23% CRT. Median survival after resection was 21.7 and 18.9 months for CT and CRT groups, respectively. Patients receiving CRT were less likely to have high comorbidity burden (ADG ≥ 10), but were similar across other demographics. CRT patients were more likely to have margin positive disease. In a subgroup of 489 patients with margin negative disease, median survival in the node negative patients (n = 156) was 28.0 months for CRT and 24.7 months for CT (p = 0.8297, logrank). Median survival in the node positive patients (n = 333) was 20.6 months and 21.8 months for the CRT and CT patients, respectively (p = 0.9856, logrank). The median total one-year cost for CT was $52,575 (USD); CRT was $68,216 (Table 1). Conclusions: Patients who underwent adjuvant CT and CRT after PD for PC had similar overall survival, but healthcare expenditures were significantly higher in the CRT group. [Table: see text]

Provider-volume associated with variable receipt of therapy and outcomes for noncurative pancreas adenocarcinoma: A population-based analysis

352

Background: While high-volume providers for pancreatic adenocarcinoma (PA) surgery yield better outcomes, variation in practice and the role of provider-volume has not been investigated for systemic therapy. We examined variation in practice and outcomes in the management of non-curative PA, based on medical oncology provider-volume. Methods: We conducted a population based retrospective cohort study of non-resected PA over 2005-2016 by linking administrative healthcare datasets. High-volume (HV) medical-oncology providers were defined as the 5th quintile of number of PA seen per provider per year. Outcomes were receipt of chemotherapy and overall survival (OS). Brown Forsythe Levene (BFL) test for equality of variances assessed outcomes variability between provider-volume quintiles (Q1 to 5). Multivariate regressions examined the association between management by HV provider and receipt of systemic therapy and OS. Results: Of 10,881 non-curative PA patients, 7,062 consulted with medical oncology. Among 341 medical oncology providers, 3% were HV, defined as > 16 patients/year. There was variability in receipt of chemotherapy based on provider-volume, with 44% (IQR: 25-54) for Q1 and 47% (IQR: 43-54) for Q5, and in median survival, with 4.1 months (IQR: 2.7-6.2) for Q1 and 7.5 months (IQR: 6.6-8.0) for Q5. Variability between provider-volume quintiles was significant for receipt of chemotherapy and median survival (both BFL p < 0.001). After adjusting for age, sex, comorbidity burden, rurality, income quintile, and diagnosis year, HV provider was independently associated with higher odds of receiving chemotherapy (OR 1.19 [1.05-1.34]), and superior OS (HR 0.79 [0.74-0.84]). Conclusions: There was significant variation in non-curative management and outcomes of PA based on provider-volume. Management by a HV provider was independently associated with higher odds of receiving chemotherapy and superior OS, after adjusting for case-mix. This information is important to inform disease care pathways and care organization. Cancer care systems could consider initiatives to increase the number of HV providers to reduce variation and improve outcomes.

Undertreatment of noncurative pancreatic adenocarcinoma?: A population-based analysis of patterns of care

439

Background: Noncurative pancreatic adenocarcinoma (PA) portends a guarded prognosis. Advancements in systemic therapy have improved this outlook. It is unknown whether patients get access to these therapies. We sought to define patterns of access to care and therapy for noncurative PA. Methods: We conducted a population-based analysis of nonresected PA over 2005-2016 by linking administrative healthcare datasets. Primary outcome was nonreceipt of cancer-directed therapy (radiation/chemotherapy; NRCDT). The first contact and overall consultations with specialized care (surgery, medical, or radiation oncology) were examined. Multivariate models examined factors associated with NRCDT. Results: Of 10,881 patients surviving a mean of 3.3 months (IQR: 1.2-8.5), 62% had NRCDT. More of patients of older age (65% of 71-80 years old, 89% of ≥81 years old), high comorbidity burden (68%), and lower socioeconomic status (69%), had NRCDT. Distance from residence to nearest cancer centre did not differ based on NRCDT. 35% of all patients did not see medical oncology, including 56% of NRCDT patients; 17% had no consultation with specialists. First contact with specialized care was surgery for 55% of all patients, and 50% with NRCDT. Most patients saw palliative care (81%) at median 27 days (IQR: 9-75) after diagnosis. Older age (OR 0.42 [0.37-0.48], and OR 0.14 [0.12-0.16] for 71-80 and ≥81 years old respectively), lowest income quintile (OR 0.62 [0.54-0.71]) and rurality (OR 0.63 [0.56-0.71]) were independently associated with lower odds of seeing medical oncology. First contact with oncology was independently associated with higher odds of receiving therapy (OR 1.48 [1.34-1.62]), compared to surgery. Conclusions: The majority of patients with noncurative PA did not receive cancer-directed therapy. Of those, more than half did not see medical oncology. While some patients may not be eligible to therapy, we identified disparities in receipt of cancer-directed therapy that indicate potential gaps in assessment for therapy and undertreatment, especially for vulnerable populations. This information is important to optimize access to and delivery of evidence-based care, and improve PA outcomes.

Is Performance Status Associated with Symptom Scores?

Background: Symptom scores and performance status are both important measures for patients with cancer. However, since performance status is not often part of routinely collected data, there is interest in exploring whether it can be calculated from symptom scores. Methods: This was a population-based longitudinal study of cancer outpatients in Ontario, Canada in the year following their cancer diagnosis and among the subset of patients during the last year of their lives. Results: In the first year after diagnosis, there was a significant relationship between performance status and fatigue and appetite; fatigue and well-being had a significant association with performance status in the last year of life. In both periods, the associations, although statistically significant, were not large enough to be clinically meaningful. Conclusion: Performance status is an important measurement that cannot be substituted or captured with symptom scores; it is important for healthcare providers to record performance scores on a regular basis.

Initiation of Chemotherapy in Cancer Patients with Poor Performance Status: A Population-Based Analysis

Objective: Practice guidelines indicate that patients who have months to weeks left to live should not be offered chemotherapy. We examined factors associated with clinician-reported poor performance status as determined by the Palliative Performance Scale (PPS) and subsequent initiation of intravenous (IV) chemotherapy in an ambulatory cancer population in Ontario, Canada. Methods: In this retrospective study, patients who had at least one PPS assessment indicating poor performance status (a PPS score of 50 or lower) comprised the study cohort. Using linked administrative databases, we observed the cohort for initiation of IV chemotherapy within 30 days of the first (index) poor PPS assessment. Results: We excluded patients for whom IV or oral chemotherapy was on going or recently completed or whose performance status improved following the index assessment. Of the remaining cohort, 9.3 percent (264/2,842) received IV chemotherapy within 30 days of the index PPS. Conclusion: A small number of cancer patients with poor performance status began IV chemotherapy in the month following assessment.

Objectif: Les directives de pratiques cliniques recommandent que l'on ne propose pas de traitements de chimiothérapie aux patients présentant une espérance de survie de quelques semaines à quelques mois. Nous avons examiné, chez un groupe de patients atteints du cancer en Ontario, au Canada, les éléments associés au faible statut des malades selon les paramètres de l'Έchelle de performance pour soins palliatifs afin d'identifier les facteurs qui ont déterminé l'amorce de la chimiothérapie. Méthode: Cette étude rétrospective comprenait les patients chez qui lors du test d'évaluation de performance on avait noté au moins un élément négatif, soit un score de performance de 50 ou moins. En utilisant plusieurs banques de données administratives interreliées, nous avons étudié cette cohorte de patients devant être traités par chimiothérapie au cours de la période de 30 jours suivant leur évaluation. Résultats: Nous avons exclu les patients pour lesquels le traitement de chimiothérapie orale ou intraveineuse était déjà en cours ou récemment terminé ou ceux dont le statut s'était amélioré selon l'Έchelle de performance. De la partie restante de la cohorte, 9,3 pourcent (264/2 842) ont reçu le traitement par voie intraveineuse à l'intérieur des 30 jours suivant l'indice de l'Έchelle de performance. Conclusion: Un petit nombre de patients ayant un faible statut ont commencé la traitement de chimiothérapie au cours du mois suivant l'évaluation.

Can administrative data improve the performance of clinical trial economic analyses?

2

Background: Economic analyses of trials often rely on trial-collected health resource utilization data, which is expensive and may be incompletely recorded. We investigated whether routinely collected health administrative data (RCD) can be utilized to improve trial economic analysis performance. Methods: Health administrative data was probabilistically linked to Ontario patient data from the Canadian Cancer Trials Group CO.17 trial (n = 572), evaluating cetuximab plus best supportive care (n = 75 linked Ontario patients) versus best supportive care alone (n = 73). Completeness of trial data was compared to RCD. Cost-effectiveness in 2007 Canadian dollars was determined using RCD up to trial date of last contact (DOLC), and up to RCD DOLC. Incremental cost effectiveness ratio (ICER) confidence intervals (CI) were determined using bootstrapping with 5000 iterations. Cost-effectiveness acceptability curves were determined. Sensitivity analyses were performed. Results: Among 148 Ontario patients, up to trial DOLC, RCD vital status was concordant in > 96%. 29 deaths occurred after trial DOLC. Up to trial DOLC there were 34 net additional hospitalizations in RCD, and 28 net additional emergency room visits. Using RCD, total cetuximab group costs were $3,023,034, and $1,191,118 for best supportive care alone up to trial DOLC. Cost difference was driven by cetuximab drug costs ($1,531,370). Using RCD, the ICER was $211,128 per life-year gained (90% CI: $101,396, $694,950) when data was limited to trial DOLC, and $164,378 (90% CI: -$138,260, $644,555) using routinely collected data DOLC. ICER estimates were similar to the original economic analysis using trial-collected data ($199,742 (95% CI $125,973, $652,492)). Estimates were robust in sensitivity analysis. Conclusions: Administrative data were more complete than trial-collected utilization data, even under optimal conditions. There was also longer follow-up. We found that cost differences were robust to varying costing assumptions. Our findings demonstrate the potential of administrative data sources to relieve institutions, sponsors and patients from the burden of collecting key utilization information which requires considerable effort and cost.

Active identification of patients appropriate for palliative care: Impact on use of palliative care and home care resources

101

Background: This research evaluates whether active identification of patients who may benefit from a palliative approach to care changes the use of palliative care and home care services. Methods: Between 2014 and 2017, Cancer Care Ontario implemented the INTEGRATE project at 4 cancer centres and 4 primary care teams. Physicians in participating sites were encouraged to systematically identify patients who were likely to die within 1 year and would benefit from a palliative approach to care. Patients in the INTEGRATE intervention group were 1:1 matched to non-intervention controls selected from provincial healthcare administrative data based on a publicly funded health system using the propensity score-matching. Palliative care and home care services utilization was evaluated within 1 year after the date of identification (index date), censoring on death, or March 31, 2017, the study end date. Cumulative incidence function was used to estimate the probability of having used care services, with death as a competing event. Rate of service use per 360 patient-days was calculated. Analyses were done separately for palliative care and home care. Results: Of the 1,187 patients in the INTEGRATE project, 1,185 were matched to a control. The intervention and the control groups were well-balanced on demographics, diagnosis, comorbidities, and death status. The probability of using palliative services in the intervention group was 80.3%, which was significantly higher than that in the control group (62.4%) with more palliative care visits in the intervention group [29.7 (95%CI: 29.4 to 30.1] per 360 patient-days) than in the control group [19.6 (95%CI: 19.3 to 19.9) per 360 patient-days]. The intervention group had a greater probability of receiving home care (81.4%) than the control group (55.2%) with more homecare visits per 360 patient-days [64.7 (95%CI: 64.2 to 65.3) vs. 35.3 (95%CI: 34.9 to 35.7)] The intervention group also had higher physician home visits as compared to the control group (36.5% vs. 23.7%). Conclusions: Physicians actively identifying patients that would benefit from palliative care resulted in increased use of palliative care and home care services.

The impact of routine ESAS use on receiving palliative care services: Results of a population-based retrospective matched cohort analysis

191

Background: In 2007 Cancer Care Ontario began standardized symptom assessment as part of routine clinical care using the Edmonton Symptom Assessment System (ESAS). The purpose of this project was to evaluate the impact of this program on referrals to palliative care. We hypothesized that patients exposed to ESAS would be more likely to be referred. Methods: A retrospective matched cohort study was conducted to examine the impact of ESAS screening on the initiation of palliative care services provided by physician or homecare nurse among newly diagnosed cancer patients in Ontario, Canada. The study included all adult patients who were diagnosed with cancer between 2007 and 2015. Exposure was defined as completing ≥1 ESAS during the study period. Using four hard matched variables and propensity-score matching with 14 variables, cancer patients exposed to ESAS were matched 1:1 to those who were not. Matched patients were followed from first ESAS until initiation of palliative care, death or the end of study at Mar 31, 2017. Results: The final cohort consisted of 204,688 matched patients with no prior palliative care consult. The pairs were well matched. The probability of receiving palliative care within the first 5 years was higher among those exposed to ESAS compared to those who were not (20.6% vs. 15.2%, p < .0001). The risk of death without receipt of palliative care within the same period was low in both groups. In the adjusted cause-specific Cox proportional hazards model, ESAS assessment was associated with a 6% increase in palliative care services (HR: 1.06, 95% CI: 1.04-1.08). Conclusions: Cancer patients who completed ESAS were more likely to initiate palliative care services than those who didn’t. ESAS screening may help identify patients who would benefit from a palliative approach to care earlier in their clinical course.

Impact of Follow-Up Colonoscopy Quality on Canadian Colorectal Cancer Outcomes and Costs

Background: Most colorectal cancer (CRC) cases develop from precancerous polyps. Screening using fecal testing for occult blood, with follow-up diagnostic colonoscopy to remove polyps, can prevent invasive cancer from occurring. However, there is variation in the quality of colonoscopy, which may result in nonoptimal health outcomes. Aim: We evaluated the impact of follow-up colonoscopy quality on health outcomes, resource utilization and costs using the OncoSim-CRC microsimulation model (version 2.5). Methods: OncoSim is a microsimulation model led by the Canadian Partnership Against Cancer with model development by Statistics Canada. We compared results of high quality follow-up colonoscopy after positive fecal immunochemical testing (FIT) (colonoscopy sensitivity for cancer detection= 95%; compliance to follow-up colonoscopy = 85%) with that of reduced quality colonoscopy. Variations in colonoscopy performance were simulated through plausible overall effectiveness reduction (ER) and incomplete colonoscopy (IC). Screening system/patient follow-up deficiencies were simulated through poor compliance to diagnostic colonoscopy (PC). Modeling assumptions included: Biennial FIT screening of average-risk people aged 50-74; positive FIT followed by diagnostic colonoscopy; ER = 20% reduction in overall sensitivity; IC = zero sensitivity in proximal colon; PC = compliance reduction by 50%. Overall cost was calculated for 2017-2036 in undiscounted 2016 CAD, and included screening, treatment and end-of-life costs. Results: Compared with high quality colonoscopy follow-up, incomplete colonoscopy with poor compliance over 20 years led to as many as 12% new cases of CRC; 23% more CRC deaths; 89% more interval cancers; and 6% increased costs to the health care system, annually. Conclusion: Reduced colonoscopy quality can lead to considerable declines in the predicted effectiveness of screening and to increased costs to the healthcare system. Efforts to increase and maintain colonoscopy performance is a necessary component of CRC control planning.

Geographic Variation in Colorectal Cancer Incidence and the Disparities in the Prevalence of Modifiable Risk Factors Across Canada

Background: Colorectal cancer is the third most common cancer worldwide. There is wide geographic variation in incidence with rates varying ten-fold between high- and low-income countries. This heavy burden can be mitigated given previous research has estimated that nearly half of all colorectal cancer cases could have been prevented through healthier diets and physically active lifestyles. In Canada, there is considerable geographic variation in age-adjusted incidence rates for colorectal cancer between jurisdictions, greater than that seen for many other cancers. These wide variations likely reflect differences in the prevalence of risk factors across provinces and territories. Aim: To describe the extent of the variation in colorectal cancer incidence rates across Canada and the disparities in the prevalence of modifiable risk factors across jurisdictions known to contribute to this burden. Methods: Colorectal cancer incident cases were obtained from the Canadian Cancer Registry; 2014 was used for provinces (except Quebec where 2010 was the most recent year available) and years 2012 to 2014 were combined to achieve more stable rates for the territories, which are much smaller in population. Data on four known modifiable risk factors for colorectal cancer (excess weight, physical inactivity, alcohol intake and low fruit and vegetable consumption) were obtained from the 2015-16 combined Canadian Community Health Survey. Results: Findings suggest that there is a north-south and east-west gradient in colorectal cancer modifiable risk factors in Canada. For instance, the percentage of adults with excess body weight ranged from 56.8% in British Columbia (west) to 73.1% in New Brunswick (east) and the percentage of adults not meeting physical activity guidelines ranged from 31.8% in Yukon (north) to 50.3% in New Brunswick (east). Generally, this pattern also reflects colorectal cancer incidence rates. The highest prevalence of modifiable risk factors and rates of colorectal cancer are typically in the northern (territories) and eastern provinces of Canada. Conclusion: The global burden of colorectal cancer is expected to increase by nearly 60% by 2030; therefore, targeted interventions are needed to ensure there is not a widening gap in colorectal cancer burden worldwide. Based on current knowledge, the most effective approaches to reduce the burden of colorectal cancer include: 1) adopting public policies that make healthy choices easier and create healthier environments where people live, work and play, and 2) continuing emphasis on screening and early detection. Strategic approaches to addressing modifiable risk factors, as well as mechanisms for detecting colorectal cancer before it develops, have the potential to translate into positive effects on population health and less people developing and dying from cancer.

International meta-analysis of 684,660 men with vasectomies: a study utilising the International Population Data Linkage Network

IntroductionEvidence on the effect of vasectomy and vasectomy reversal on risk of prostate cancer is conflicting, with the issue of detection bias a key criticism. In this study we examined the effect of vasectomy reversal on prostate cancer risk in a cohort of vasectomised men. Objectives and ApproachA proof of concept study involving the International Population Data Linkage Network which pooled aggregated result data from participating centres in Australia, Canada and the United Kingdom. De-identified linked data extractions took place at each centre. Each participating centre locally conducted Cox proportional hazards regression analysis compared the risk of prostate cancer in those with/without vasectomy reversal in a cohort of vasectomised men. These results were then combined in a meta-analysis. Evidence of a protective effect of vasectomy reversal would suggest the harmful effect of vasectomy on prostate cancer risk, while nullifying detection bias. ResultsData were received from Australia (the states of Western Australia and New South Wales), Canada (the province of Ontario), Wales and Scotland. In total, there were 9,754 men with vasectomy reversals, and 684,660 men with a vasectomy. The combined analysis showed no protective effect of vasectomy reversal on incidence of prostate cancer when compared to those who had vasectomy alone (HR, 95%CI: 0.92, 0.70-1.21). As such, the results align with previous studies which found little or no evidence of a link between vasectomy and prostate cancer. Conclusion/ImplicationsThe study, originally conceived at the first IPDLN meeting in London, found no obvious protective effect of vasectomy reversal on prostate cancer in vasectomised men. The project demonstrated the utility and feasibility of collaborative studies fostered through the IPDLN, despite methodological challenges faced when aggregating international data.