Institute For Clinical Evaluative Sciences Research Articles

Risk of complications among diabetics self-reporting oral health status in Canada: A population-based cohort study.

BackgroundPeriodontitis has been associated with diabetes and poor health. While clear associations have been identified for the diabetes–oral health link, less is known about the implications of poor oral health status for incident complications of diabetes. This study investigated the risk of diabetes complications associated with self-reported “poor to fair” and “good to excellent” oral health among diabetics living in Ontario, Canada.MethodsThis was a cohort study of diabetics who took part in the Canadian Community Health Survey (2003 and 2007–08). Self-reported oral health was linked to electronic health records held at the Institute for Clinical Evaluative Sciences. Participants aged 40 years and over, who self-reported oral health status in linked databases were included (N = 5,183). Cox proportional hazard models were constructed to determine the risk of diabetes complications. Participants who did not experience any complications were censored. Models were adjusted for age and sex, followed by social characteristics and behavioural factors. The population attributable risk of diabetes complications was calculated using fully adjusted hazard ratios.ResultsDiabetes complications differed by self-reported oral health; 35% of the total sample experienced a complication and 34% of those reporting “good to excellent” oral health (n = 4090) experienced a complication in comparison to 38% of those with “fair to poor” oral health (n = 1093). For those reporting “poor to fair” oral health, the hazard of a diabetes complication was 30% greater (HR 1.29; 95% CI: 1.03, 1.61) than those reporting “good to excellent” oral health. The population level risk of complications attributable to oral health was 5.2% (95% CI: 0.67, 8.74).ConclusionsOur findings indicate that reporting “poor to fair” oral health status may be attributed to health complications among diabetics, after adjusting for a wide range of confounders. This has important public health implications for diabetics in Ontario, Canada.

UTI Management: It’s All in the Culture

UTI Management: It’s All in the Culture

Fractionation in adjuvant radiotherapy for invasive breast cancer and ductal carcinoma in situ in Ontario, Canada from 2009 to 2015.

The use of hypofractionated radiotherapy (HFRT) in patients with breast cancer and ductal carcinoma in situ (DCIS) in Ontario, Canada, from 2009 to 2015 was reported. A retrospective cohort study was conducted using data from the Institute for Clinical Evaluative Sciences (ICES). Patients with a breast cancer or DCIS diagnosis between 2009 and 2015 who received adjuvant breast or chest wall radiation were included. Trends in HFRT use (≤16 fractions) and factors associated with HFRT use in a multivariable logistic regression model with physician-level random effect were reported. The approximate number of hours that could be saved if all patients were to receive HFRT was calculated. A total of 42072 patients were included. All included characteristics were significantly associated with HFRT use. Hypofractionated radiotherapy use in patients with breast cancer and DCIS increased to around 75% in 2015. In stage I/II patients with mastectomy and chest wall radiation, HFRT use increased to 40% in 2015. Hypofractionated radiotherapy use in patients with regional nodal radiation or reconstruction has increased but remains under 20%. For breast cancer patients with breast-conserving surgery (BCS) and breast radiation, 56265 visits corresponding to 7200hours of treatment or 3500 additional HFRT courses could have been saved. In conclusion, HFRT use in Ontario has increased in all patient populations but is nonuniform among physicians and institutions. Use of HFRT in chest wall and regional nodal radiation remains relatively lower than in breast cancer and DCIS patients with BCS.

A population-based study of prescribing trends in a potentially vulnerable paediatric population from 1999 to 2012.

There is a limited understanding of paediatric medication prescribing trends and patterns, thus poorly positioning decision-makers to identify quality and safety concerns related to medication use. The objective of this study was to determine overall medication prescribing trends and patterns among children receiving Ontario Drug Benefits over a thirteen-year period in the province of Ontario, Canada. Administrative health databases housed within the Institute for Clinical Evaluative Sciences (ICES), Ontario, Canada, were used to identify outpatient prescriptions dispensed from 1999 to 2012 through a publicly funded programme to children ≤18 years of age. Medications were classified according to the American Hospital Formulary Service Pharmacologic-Therapeutic Classification system. Descriptive statistics were used to summarize prescribing patterns. This study identified 457 037 children who were dispensed a new prescription between 1999 and 2012. About 56% received their first prescription before 6.5 years of age, and 85% of the children in this study were from families who received social assistance. The most commonly prescribed drugs were antiinfectives (56.1%). Prescriptions for several central nervous system agents, including antipsychotics and agents for attention-deficit/hyperactivity disorder, increased across the study period. Changes in prescribing patterns within opioids, hormones and autonomic agents were noted. The results suggest that historically, prescribing trends have shifted with public policy, pharmaceutical marketing and diagnostic patterns, thus identifying them as a possible tool to measure the impact of policydriven practice changes. Anti-infective prescribing increased markedly with the global H1N1 pandemic. Pharmaceutical marketing, formulary decisions and diagnostic trends may affect the prescribing of ADHD medications globally. The prescribing of codeine-containing products and medroxyprogesterone appeared to fluctuate in response to important publications in the medical literature, and the use of epinephrine syringes increased after public policy changes in the province of Ontario. The steady rise in the use of medications whose long-term effects in children are unknown, such as antipsychotics and proton pump inhibitors, identifies areas in need of future research. This study presents the first overview of Canadian prescribing trends for children, the majority of which are of low socioeconomic status and represent a potentially vulnerable population. Our analysis suggests that future research is required to determine whether prescribing trends could be used as indicators of policy effectiveness, pharmacovigilance and diagnostic trends.

Cost disparities with age in the treatment of advanced non-small cell lung cancer (NSCLC) in Ontario, Canada.

6629 Background: Previous studies noted an association between age and cost of care in NSCLC. The drivers of these cost disparities have not yet been fully examined. We conducted a cost analysis study examining the differences in, and drivers of, costs of NSCLC care across age groups in Ontario, Canada. Methods: We conducted a retrospective cohort study of patients diagnosed in Ontario from Apr 1, 2007 to Mar 30, 2014, who received palliative chemotherapy for stage IV NSCLC. Variables of interest were extracted from registry data linked by the Institute for Clinical Evaluative Sciences (ICES). The mean total cost of care including systemic therapy and supportive care, was calculated in 2015 CAD dollars by fiscal year of diagnosis. Results: Of all NSCLC cases diagnosed in Ontario (n = 37,786), 17,203 (45.5%) were de novo stage IV, of which 29.7% of patients received any chemotherapy for their disease (n = 5,113), and 281 patients are presumed alive. In this population, median age was 65 to 69 years, 51.9% were male, 43.5% were adenocarcinomas, and 25.1% received second line chemotherapy. After adjusting for comorbidities, income, gender, year of diagnosis, and rural versus urban living, the average lifetime costs per patient remains significantly inversely related to age (p < 0.001). Belonging to the highest income quintile (p = 0.006) and being diagnosed in more recent years (p < 0.001) contributes significantly to increasing overall healthcare costs. Elderly patients (80+) cost less (71%) and have shorter survival time (HR of death 1.28, 95% C.I. 1.10 to 1.50) compared to younger patients (≤45 years old). Accounting for longer survival in younger patients, the youngest group still incur a higher cost per day alive than other age groups ($471/day in ≤45 group, $301/day in > 85 group). Hospitalization accounts for ~30% of total cost in both age groups. Chemotherapy accounts for 1% of total health care costs amongst the elderly (80+) age group and 10% of costs in ≤45 group. Conclusions: Our study shows that, despite everyone receiving systemic therapy in this patient population, younger patients incur significantly higher costs than elderly patients with advanced NSCLC, both before and after adjusting for survival. While hospitalization accounts for the biggest component of total costs, patients with high income and more recent years of diagnosis drive the higher costs of care, and chemotherapy remains a driver of higher costs amongst younger patients.

Neighborhood Material Deprivation Is Associated with Childhood Asthma Development: Analysis of Prospective Administrative Data.

Rationale Material deprivation has been proposed as a more comprehensive measure of socioeconomic status than parental income. Stronger associations between childhood emergency department visits for asthma and air pollution have been demonstrated among children living in neighborhoods with high levels of deprivation, but the associations with asthma development and ongoing asthma are not known. Objectives We determined the associations between neighborhood material deprivation and the development of new and ongoing childhood asthma. Methods Prospectively collected administrative data housed at the Institute for Clinical Evaluative Sciences were examined for Toronto children born from 1997 to 2003. Neighborhood material deprivation, comprising no high school graduation, lone parent families, government transfers, unemployment, low income, and homes needing major repairs, was reported in the Ontario Marginalization Index. Incident asthma was defined by the time of entry into the Ontario Asthma Surveillance Information System (OASIS) database. We measured the risk of incident asthma using Cox proportional hazards models and the associations between ongoing asthma visits and deprivation by year of life with generalized linear mixed models. Results OASIS asthma criteria were met for 21% of the 326,383 children. After adjustment for characteristics strongly associated with asthma, including male sex, prematurity, obesity, and atopic conditions other than asthma, children with high birth neighborhood deprivation were at increased risk of incident asthma (HR 1.11; 95% CI, 1.09–1.13). High deprivation in a given year of life was associated with increased odds of ongoing asthma during that year of life (OR 1.03; 95% CI, 1.02–1.05). Conclusions Children living in high-deprivation neighborhoods are at increased risk of incident and ongoing asthma. This study suggests that neighborhood material deprivation may represent a helpful tool for evaluating the effects of disparities in health and social advantages on the likelihood of developing and continuing to need healthcare visits for ongoing childhood asthma.

Abstract P3-12-05: Fractionation patterns in adjuvant breast radiotherapy in Ontario, Canada from 2009 to 2015

Abstract Purpose: To report the patterns of use of hypofractionated radiotherapy (HFRT) (≤16 fractions)in breast cancer and ductal carcinoma in situ (DCIS) patients in Ontario, Canada from 2009 to 2015 and identify factors related to HFRT use. Methods: A retrospective cohort study of Ontario women diagnosed with breast cancer or DCIS followed by adjuvant breast or chest wall radiation (RT) from 2009 to 2015 was conducted using data from the Institute for Clinical Evaluative Sciences (ICES). Logistic regression models were used to identify factors associated with HFRT use. Physician was included as a random effect. To calculate the potential amount of time that could be saved if all patients were to receive 16 fractions of HFRT, the number of extra RT treatments after the 16th visit was multiplied by the median amount of time spent to treat one patient at our cancer centre (8.76 minutes). This time was found with Sunnybrook Health Sciences Centre data collected from October 2017 to March 2018 (n=523) and represents the amount of time from when a patient enters the RT unit for treatment setup, until the RT beam is turned off. Results: A total of 42,072 patients were included. Most patients were aged between 50 and 69 years old (56.7%) with stage I or II breast cancer (74.6%) and had breast conserving surgery (BCS) (76.9%). Half previously received chemotherapy and 7.9% were DCIS patients. Use of sequential boost, simultaneous boost, and regional nodal RT was 17.2%, 3.1%, and 31.7% respectively. Institutional variation in HFRT use ranged from 27.5% to 71.6%. HFRT use has increased in all patient populations over the study period. HFRT use was more common in breast cancer and DCIS patients with BCS than in chest wall and nodal RT. Simultaneous boost (OR=0.09), nodal RT (OR=0.08), previous chemotherapy (OR=0.7), stage 0, II, and III breast cancer (OR=0.06 relative to stage I), were correlated with less HFRT use. Older age, later year of diagnosis, sequential boost (OR=0.09), BCS and no surgery (OR=1.2, 1.5 relative to mastectomy) were correlated with higher HFRT use. Institution was significantly correlated to HFRT use. The variance estimate for the physician random effect was 0.33 (p<0.0001). For breast cancer patients with BCS and breast RT specifically, 62,396 extra visits occurred from 2009 to 2015, corresponding to ˜9100 hours of treatment if all patients received HFRT. For the entire patient population, a total of 190,726 extra visits occurred or ˜27,900 hours. Trends in the use of HFRT (%) from 2009 to 2015 for DCIS and breast cancer patients in Ontario.YearStage I-IV + BCS + Breast RTStage I-IV + BCS + Breast + Nodal RTStage I-IV + Mastectomy + Breast RTStage I-IV + Mastectomy + Breast + Nodal RTDCIS + BCS200962.610.329.09.545.2201062.06.323.76.951.2201162.24.121.34.752.7201263.79.621.711.452.0201365.112.629.315.057.3201469.37.835.415.164.5201573.711.839.217.573.4BCS: breast conserving surgery, RT: radiation treatment Conclusions: HFRT use in Ontario has increased over time for all patient populations, and reflects the current evidence supporting HFRT in different patient populations, with lower HFRT use seen in chest wall and nodal RT and higher HFRT use in early-stage breast cancer patients with BCS. Citation Format: Chow E, Chan S, Yao Z, Sutradhar R, Lee J, Wan BA, Paszat L. Fractionation patterns in adjuvant breast radiotherapy in Ontario, Canada from 2009 to 2015 [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-12-05.

Incidence of cirrhosis in young birth cohorts in Canada from 1997 to 2016: a retrospective population-based study

Recent data show that the prevalence of chronic liver disease and cirrhosis is increasing in adolescents and young adults in the USA. We aimed to describe the epidemiology of cirrhosis using an age-period-cohort approach to define birth-cohort effects on the incidence of cirrhosis in Ontario, Canada. We did a retrospective population-based cohort study in Ontario, Canada, using linked administrative health data from the databases of ICES, formerly the Institute for Clinical Evaluative Sciences. Patients aged at least 18 years with cirrhosis were identified by use of a validated case definition (defined as at least one inpatient or outpatient visit with a diagnosis of cirrhosis or oesophageal varices without bleeding). We calculated annual standardised incidence and prevalence in the general population. We used an age-period-cohort approach to assess the independent association between birth cohort and incidence of cirrhosis in men and women. Between Jan 1, 1997, and Dec 31, 2016, 165 979 individuals with cirrhosis were identified. The age-standardised incidence increased over the study (from 70·6 per 100 000 person-years in 1997 to 89·6 per 100 000 person-years in 2016) as did the prevalence (from 0·42% in 1997 to 0·84% in 2016). Using age-period-cohort modelling and the median birth year as the reference, the incidence of cirrhosis was higher in participants born in 1980 (incidence rate ratio 1·55, 95% CI 1·50-1·59, p<0·0001); and in participants born in 1990 (2·16, 95% CI 2·06-2·27, p<0·0001) compared with a person of the same age born in 1951. The increase in incidence of cirrhosis was greater in women than in men (eg, women born in 1990: 2·60, 95% CI 2·41-2·79; men born in 1990: 1·98, 1·85-2·12). The incidence of cirrhosis has increased over the past two decades, and more so in younger birth cohorts and in women. Future studies to define the cause and natural history of cirrhosis in these groups are essential to develop strategies that could reverse these trends for future generations. Southeastern Ontario Academic Medical Association New Clinician Scientist Award; American Association for the Study of Liver Disease (AASLD) Foundation Clinical, Translational and Outcomes Research Award in Liver Disease (JAF).

Evaluation of Definitions for Oral Anticoagulant–Associated Major Bleeding: A Population-Based Cohort Study

Background: Major bleeding is the most serious complication of oral anticoagulation. Consensus criteria to define major bleeding have been established by the International Society for Thrombosis and Haemostasis (ISTH), Bleeding Academic Research Consortium (BARC) and Thrombolysis in Myocardial Infarction (TIMI). Significant variability exists across these definitions, and their agreement for identifying oral anticoagulant (OAC) related major bleeding is unknown. Furthermore, the association between each definition and mortality in cases of OAC bleeding has not been evaluated. We therefore, sought to evaluate the agreement of cases identified as major bleeding by the ISTH, BARC and TIMI definitions, and to assess associated in-hospital (emergency department or inpatient) and 30-day mortality of cases identified by these criteria.Methods: We used an existing dataset of individuals ≥66 years in Ontario, Canada, who presented to one of five tertiary care institutions with OAC related bleeding across three cities from 2010-2015. Detailed clinical data on consecutive episodes of OAC-associated bleeding were linked to population-based databases held at the Institute for Clinical Evaluative Sciences. We calculated Cohen's κ for agreement between the three major bleeding definitions, and used Pearson's χ2 to determine any differences in in-hospital and 30-day mortality for cases defined as major bleeding by ISTH, BARC and TIMI criteria.Results: We included 2,002 cases of OAC related bleeding in the analysis (460 on direct oral anticoagulants, 1,542 on warfarin). ISTH, BARC and TIMI major bleeding definitions were met in 75%, 77% and 29% of cases, respectively. 18% of cases did not meet criteria for major bleeding by any definition. Age, sex, CHA2DS2-VASc and HAS-BLED score, as well as proportion of chronic kidney disease were similar across ISTH-, BARC- and TIMI-defined cases. Over 9 in 10 cases of TIMI-defined major bleeding events involved an intracranial hemorrhage (94.4%), compared to 37% and 36% of cases identified by ISTH or BARC definitions respectively (p<0.001 across three groups). Agreement in case identification between ISTH and BARC was substantial (agreement 89%; Cohen's κ=0.69). On the other hand, agreement between TIMI and both ISTH (agreement 54%; Cohen's κ =0.24) and BARC (agreement 52%; Cohen's κ=0.21) were poor. The association between in-hospital mortality and TIMI-defined major bleeding was higher (29%) than that for ISTH and BARC (17% for both; p<0.001 for TIMI vs. ISTH and TIMI vs. BARC). The association with 30-day mortality showed a similar trend (30%, 18% and 18% for TIMI-, ISTH- and BARC- defined major bleeding events respectively; p<0.001 for TIMI vs. ISTH and TIMI vs. BARC). 6% of cases that were not categorized as major bleeding by ISTH or BARC definitions died within 30 days of hospital presentation, and this was 10% for cases not meeting criteria for TIMI major bleeding (10%, p=0.036 by Pearson's χ2).Conclusions: Among patients with OAC-associated bleeding, major bleeding events identified by ISTH and BARC criteria showed good agreement and similar prognostic utility. Meanwhile, TIMI criteria identified patients with higher clinical risk and subsequent mortality. Patients presenting with OAC-associated bleeding who did not fulfill ISTH or BARC major bleeding criteria had considerable risk of 30-day mortality and was even higher among those not meeting the TIMI criteria. Our findings suggest the need to refine current major bleeding definitions to identify additional patients at risk of death. DisclosuresWells:Bayer: Honoraria; BMS: Honoraria, Research Funding; Sanofi: Honoraria; Janssen: Honoraria. Sholzberg:Amgen: Research Funding; CSL Behring: Research Funding; Octapharma: Research Funding; Shire: Research Funding.

Pilot study of the ability to probabilistically link clinical trial patients to administrative data and determine long-term outcomes.

Clinical trials are important but extremely costly. Utilization of routinely collected administrative data may simplify and enhance clinical trial data collection. The aim of this study was to test the feasibility of use of administrative databases in Ontario, Canada, for long-term clinical trial follow-up, specifically (a) to determine whether limited patient identifiers held by the Canadian Cancer Trials Group can be used to probabilistically link with individuals in the Institute for Clinical Evaluative Sciences databases and if so, (b) the level of concordance between the two data sets. This retrospective study was conducted through collaboration of established health service (Institute for Clinical Evaluative Sciences) and clinical trial (Canadian Cancer Trials Group) research groups in the province of Ontario, Canada, where healthcare is predominantly funded by the government. Adults with pre-treated metastatic colorectal cancer previously enrolled in the Canadian Cancer Trials Group CO.17 and CO.20 randomized phase III trials were included, limited to those in Ontario. The main outcomes were rate of successful probabilistic linkage and concordance of survival data, stated a priori. Probabilistic linkage was successful in 266/293 (90.8%) participants. In those patients for whom linkage was successful, the Canadian Cancer Trials Group (trial) and the Institute for Clinical Evaluative Sciences (administrative) data sets were concordant with regard to the occurrence of death during the period of clinical trial follow-up in 206/209 (98.6%). Death was recorded in the Institute for Clinical Evaluative Sciences, but not the Canadian Cancer Trials Group, for 57 cases, where the event occurred after the clinical trial cut-off dates. The recorded date of death matched closely between both databases. During the period of clinical trial conduct, administrative databases contained details of hospitalizations and emergency room visits not captured in the clinical trial electronic database. Prospective use of administrative data could enhance clinical trial data collection, both for long-term follow-up and resource utilization for economic analyses and do so less expensively than current primary data collection. Recording a unique identifier (e.g. health insurance number) in trial databases would allow deterministic linkage for all participants.

The Association Between Continuity of Primary Care and Preventive Cancer Screening in Women With Intellectual Disability.

Women with intellectual disability have low screening rates for breast and cervical cancer. This population-based cohort study examined the association between the level of primary care continuity and breast and cervical cancer screening rates in women with intellectual disability. Data were obtained from the Institute for Clinical Evaluative Sciences and the Ontario Ministry of Community and Social Services. Neither high (adjusted OR [aOR] = 1.06; 95% CI: 0.88-1.29) nor moderate (aOR = 1.11; 95% CI: 0.91-1.36) continuity of care were associated with mammography screening. Women were less likely to receive a Pap test with high (aOR = 0.70; 95% CI: 0.64-0.77) and moderate (aOR = 0.81, 95% CI 0.74-0.89) versus low continuity of care. Improving continuity of care may not be sufficient for increasing preventive screening rates.

Impact of novel chronic lymphocytic leukemia drugs on public spending.

103 Title: Impact Of Novel Chronic Lymphocytic Leukemia Drugs On Public Spending Background: Chronic lymphocytic leukemia (CLL) is a common hematologic malignancy that mainly affects the elderly. Over the past five years, the treatment pathway for CLL has dramatically changed with the emergence of multiple novel agents. In Ontario, Canada, the Ontario Drug Benefit (ODB) program and the New Drug Funding Program (NDFP) primarily provide public coverage for CLL drugs. Given advances in treatment, we evaluated utilization trends for publicly-funded CLL drugs over a five year period (fiscal years 2012/13 to 2016/17). Methods: Drugs with primary CLL indications funded under the two public funding programs (i.e., ODB and NDFP) by 16/17 were included (n = 6). Claims and costs data were obtained from the Institute for Clinical Evaluative Sciences and Cancer Care Ontario‘s datasets. Results: Over five years, expenditures on CLL drugs have increased approximately ten-fold (1000 percent), reaching nearly CAD 43 million (i.e., USD 32.8 million) in 16/17. In the front-line setting, spending on rituximab remained consistent over the five years. Spending on single agent bendamustine decreased with the introduction of obinutuzumab which became the main cost driver by 16/17. In the previously-treated CLL population, ibrutinib has dominated expenditures since it was funded in July 2015. By 16/17, ibrutinib accounted for approximately ninety eight percent of spending in the previously-treated population. Conclusions: In the past five years, public spending on CLL drugs increased rapidly and substantially with the introduction of four novel agents, and may continue to evolve as Canadian provinces consider funding additional indications or newer agents. These findings warrant exploring whether the incremental spending is providing survival benefit or improvement in patients’ quality of life in a real-world setting. Declaration of funding: None

Our health counts: population-based measures of urban Inuit health determinants, health status, and health care access

ObjectiveHealth determinants and outcomes are not well described for the growing population of Inuit living in southern urban areas of Canada despite known and striking health disparities for Inuit living in the north. The objective of this study was to work in partnership with Tungasuvvingat Inuit (TI) to develop population prevalence estimates for key indicators of health, including health determinants, health status outcomes, and health services access for Inuit in Ottawa, Canada.MethodsWe employed community-based respondent driven sampling (RDS) and a comprehensive health assessment survey to collect primary data regarding health determinants, status, and service access. We then linked with datasets held by the Institute for Clinical Evaluative Sciences (ICES), including hospitalization, emergency room, and health screening records. Adjusted population-based prevalence estimates and rates were calculated using custom RDS software.ResultsWe recruited 341 Inuit adults living in Ottawa. The number of Inuit living, working or accessing health and social services in the City of Ottawa was estimated to be 3361 (95% CI 2309–4959). This population experiences high rates of poverty, unemployment, household crowding, and food insecurity. Prevalence of hypertension (25%; 95% CI 18.1–33.9), chronic obstructive pulmonary disease (6.7%; 95% CI 3.1–10.6), cancer (6.8%; 95% CI 2.7–11.9), and rates of emergency room access were elevated for Inuit in Ottawa compared to the general population. Access to health services was rated fair or poor by 43%. Multiple barriers to health care access were identified.ConclusionsUrban Inuit experience a heavy burden of adverse health determinants and poor health status outcomes. According to urban Inuit in Ottawa, health services available to Inuit at the time of the study were inadequate.

P2.15-31 The Evolution of Costs in the Treatment of Advanced Non-Small Cell Lung Cancer (NSCLC) in Ontario, Canada Between 1999 to 2014

Treatment options for patients with advanced non-small cell lung cancer (NSCLC) have evolved substantially since the 1990s. Newer agents are more expensive to acquire and administer and the perception is the cost of lung cancer care has increased considerably in the last two decades. We conducted a cost analysis study to evaluate changes in the total cost of NSCLC care over time. We conducted a retrospective cohort study of all NSCLC patients diagnosed in Ontario from Apr 1, 1999 to Mar 30, 2014, who received palliative chemotherapy for advanced disease. Variables of interest were extracted from provincial registry data electronically linked by the Institute for Clinical Evaluative Sciences (ICES). The use of oral systemic therapy is not universally captured in these databases. The mean total cost of care including, systemic therapy, and supportive care (hospitalizations, physician billings, lab tests, out-patient visits, emergency visits, home care, and most prescription medications), was calculated in 2015 CAD dollars by fiscal year of diagnosis. Regression analysis was used to project costs in years with missing supportive care costs. Of all NSCLC cases diagnosed in Ontario (n=89,936), 21.6% (n=19,447) received any chemotherapy for advanced disease. In this population, median age ranged from 65 to 69 years, 54.9% were male, 15.4% resided rurally, 45.0% were adenocarcinoma, 58.6% with de novo stage IV disease, and 25.8% received second line chemotherapy. At the time of the data analysis, 6.8% of patients are presumed to be alive. The average cost of care per patient treated rose from $85,339 to $102,026. The cost of systemic therapy rose from $3,856 to $12,554. During the same time period, average supportive care costs changed from $84,732 to $93,495. The proportion of total costs of systemic therapy changed from 4.5% to 12.3%. Utilization of palliative chemotherapy in only those with de novo stage IV disease remained stable over time (29.8%). Patients who received second line chemotherapy cost 5.9 times that of patients who did not ($11,342 and $1,924 respectively). Costs in the final year of life increased from 61.9% to 68.0% of total lifetime costs from 2003 to 2014. The cost of systemic therapy for lung cancer patients is rising disproportionately to that of supportive care. Ongoing analyses are assessing the main drivers of cost of care and model the impact of oral targeted therapies and immunotherapies on the cost of lung cancer care.

PCN210 - USING REAL WORLD DATA TO DETERMINE HEALTH SYSTEM COSTS OF A 5-YEAR NON-SMALL CELL LUNG CANCER COHORT

With the recent advances in the treatment of advanced non-small cell lung cancer (NSCLC) and fiscal constraints of publicly funded healthcare systems, understanding the real world economic impact of lung cancer is increasingly important. The objective of this study was to determine the health system costs of a 5-year cohort of NSCLC patients treated in Ontario, Canada. Patients diagnosed between April 1, 2010 and March 31, 2015 were identified in the Ontario Cancer Registry and linked to provincial administrative databases, capturing hospitalizations, cancer clinic visits, physician services and expensive chemotherapies. Activity level reporting of radiation and inexpensive chemotherapies was not reported. Total cohort, annual and mean per patient costs (CAD 2017) were determined using a costing methodology available at the Institute for Clinical Evaluative Sciences called GETCOST. Short-term episodes (e.g., hospital-based encounters) calculate costs by multiplying the encounter’s resource intensity weight by an annual cost per weighted case for total costs, long-term episodes (e.g., complex continuing care) calculate costs by weighted days, and costs of visit-based encounters are determined at utilization. It was assumed that costs were attributable to NSCLC. 24,729 NSCLC patients were identified (median age 70yr; 48.5% female; stage I=20.7%, II=9.3%, III=20.8%, IV=49.2%). The overall total cost was $1.9B (mean cost per patient= $76,816); inpatient hospitalizations ($635.2M), cancer clinic visits ($323.7M) and physician services ($301.4M) were the top three cost contributors while the total cost of expensive chemotherapies was $61.5M. Year 1 after diagnosis had the highest costs ($1.3B) followed by year 2 ($316.3M). Mean cost per patient varied by stage, being highest for stage II ($94,553) and lowest for stage IV ($66,652). Real world costs during the 5-year period (prior to the use of immuno-oncology drugs) were substantial even with poor survival of NSCLC patients and those receiving no active treatment.

PCN221 - STAGE III UNRESECTABLE NON-SMALL CELL LUNG CANCER TREATMENT PATTERNS AND COSTS: AN ONTARIO, CANADA ANALYSIS

With recent changes in the treatment of Stage III, non-small cell lung cancer (NSCLC), understanding the economic impact from a public payers’ perspective is key. A longitudinal, population-level study was carried out to determine baseline treatment patterns and health system costs prior to the introduction of emerging treatments. Patients diagnosed between April 1, 2010 and March 31, 2015 were identified in the Ontario Cancer Registry. Stage III patients were unresectable if no surgery was undertaken within 3 months of diagnosis. Initial treatments included radiation (RT), targeted therapy, chemotherapy, concurrent and sequential chemo+RT (cCRT, sCRT). Costs from provincial administrative databases included inpatient hospitalizations, cancer clinic visits, physician services, and chemotherapies. Activity level reporting of radiation and inexpensive chemotherapies was not reported. Total cohort, annual and mean per patient costs (CAD 2017) were determined using a costing methodology from the Institute for Clinical Evaluative Sciences called GETCOST. Short-term episodes (e.g., hospital-based) calculate costs by multiplying the encounter’s resource intensity weight by an annual cost per weighted case, long-term episodes (e.g., complex continuing care) calculate costs by weighted days, and costs of visit-based encounters are determined at utilization. It was assumed costs were attributable to NSCLC. 4,542 (18.4%) of the 24,729 patients identified were Stage III unresectable (median age 70yr, 54.2% male). Treatment distribution was: cCRT (21.6%), palliative RT (21.3%), curative RT (20.2%), no treatment (19.6%), chemotherapy (11.6%), sCRT (4.9%) and targeted therapy (0.7%). The overall total cost was $396.9M (mean cost per patient= $87,393); 68.4% of total costs ($271.3M) were incurred in the first year. Cost drivers were inpatient hospitalizations ($108.4M), cancer clinic visits ($107.5M), and physician services ($56.9M). cCRT was only administered in 21.6% of Stage III unresectable patients, even though considered the standard of care and chemotherapies were not a cost driver in this cohort.

MS Risk in Immigrants in the McDonald Era: A Population-Based Study in Ontario, Canada

Background: Migration studies may provide insights into environmental determinants of multiple sclerosis (MS) and whether a critical window for exposure exists. We aimed to determine risk factors for MS in immigrants and compare MS risk in immigrants versus long-term residents in Ontario, Canada. Methods: We applied a validated algorithm to linked, population-based immigration and health claims data to identify incident MS cases in immigrants and long term-residents between 1994 and 2016. We conducted two multivariable Cox proportional hazards regression analyses including one limited to the immigrant cohort assessing potential risk factors for developing MS; and one comparing MS risk between immigrants and matched long-term residents (1:3 match). Results: We identified 2,304,302 immigrants for the immigrant-only analysis, of whom 1,526 (0.066%) developed MS. Younger age at migration, female sex, greater co-morbidity burden, lower neighbourhood income, and time since landing were associated with increased MS risk. Risk was greatest in those <15 years old at landing (Ref: <15 years; 16-30 years: HR 0.73, 95%CI: 0.63-0.85; 31-45 years: HR 0.55, 95%CI: 0.47-0.64). Region/country of origin was also associated with MS risk. Immigrants from the Middle East (HR 1.22, 95%CI: 1.06-1.40) were at greater risk than immigrants from Western nations; all other regions had lower MS risk (p<.0001). The matched analysis included 2,207,751 immigrants and 6,362,169 long-term residents. Immigrants were less likely to develop MS than long-term residents (p<.0001), although this lower risk was attenuated with longer residence in Canada. Interpretation: MS incidence in immigrants to Ontario, Canada varied widely by region of origin, with greatest risk seen in those from the Middle East. Although MS risk declined with increasing age at arrival, considerable risk was sustained with migration after age 15, suggesting that environmental exposures into adulthood contribute to MS risk. Funding Statement: This study was supported by the Institute for Clinical Evaluative Sciences (ICES), which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care (MOHLTC). The opinions, results and conclusions reported in this paper are those of the authors and are independent from the funding sources. No endorsement by ICES or the Ontario MOHLTC is intended or should be inferred. Declaration of Interests: DR has received research support from the Consortium of Multiple Sclerosis Centers and from the Multiple Sclerosis Society of Canada. She has served as a consultant for Sanofi Aventis, Biogen, Novartis, EMD Serono, and Roche. RAM receives research funding from the Canadian Institutes of Health Research, Multiple Sclerosis Society of Canada, National Multiple Sclerosis Society, Research Manitoba, the Consortium of MS Centers, Crohn’s and Colitis Canada, and the Waugh Family Chair in Multiple Sclerosis. CJM has received research funding from the Canadian Institutes of Health Research, P.S.I. Foundation, MS Society of Canada, Consortium of MS Centers, MS Scientific Research Foundation, Ontario Ministry of Health and Long-Term Care-Health System Research Fund, Canadian Frailty Network, Partners for Canadian Consortium on Neurodegeneration in Aging, Network for Aging Research-University of Waterloo. She is supported by a University Research Chair at the University of Waterloo. KT receives a research scholar award from the Department of Family and Community Medicine at the University of Toronto. Ethics Approval Statement: This study was approved by the research ethics boards at St. Michael’s Hospital in Toronto, Ontario, Canada. Informed consent by participants was not required.

Perspectives on Linkage Involving Indigenous data

Topic: Perspectives on Linkage Involving Indigenous dataIndigenous populations across the globe are reaffirming their sovereignty rights in the collection and use of Indigenous data. The Indigenous data sovereignty movement has been widely influential and can be unsettling for those who routinely use population-level linked data that include Indigenous identifiers. Ethical policies that stipulate community engagement for access, interpretation and dissemination of Indigenous data create an enabling environment through the critical process of negotiating and navigating data access in partnership with communities. This session will be designed to create space for leading Indigenous voices to set the tone for the discussion around Indigenous population data linkage.&#x0D; Objectives:&#x0D; &#x0D; To provide participants with an opportunity to build on the themes of Indigenous Data Sovereignty presented in the keynote session as they apply to diverse Indigenous populations.&#x0D; To explore approaches to the linkage of Indigenous-identified population data across four countries, including First Nations in three Canadian regions.&#x0D; To share practical applications of Indigenous data sovereignty on data linkage and analysis and discussion.&#x0D; To center Indigenous-driven data linkage and research.&#x0D; &#x0D; Facilitator:Jennifer Walker. Canada Research Chair in Indigenous Health, Laurentian University and Indigenous Lead, Institute for Clinical Evaluative Sciences.&#x0D; Collaborators:&#x0D; &#x0D; Alberta: Bonnie Healy, Tina Apsassin, Chyloe Healy and William Wadsworth (Alberta First Nations Information Governance Centre)&#x0D; Ontario: Carmen R. Jones (Chiefs of Ontario) and Jennifer Walker (Institute for Clinical Evaluative Sciences)&#x0D; British Columbia: Jeff Reading (Providence Health Centre) and Laurel Lemchuk-Favel (First Nations Health Authority)&#x0D; Australia: Raymond Lovett (Australian National University)&#x0D; Aotearoa / New Zealand: Donna Cormack (University of Otago)&#x0D; United States: Stephanie Rainie and Desi Rodriguez-Lonebear (University of Arizona)&#x0D; &#x0D; Session format: 90 minutesCollaborators will participate in a round-table introduction to the work they are doing. Collaborators will discuss the principles underlying their approaches to Indigenous data linkage as well as practical and concrete solutions to challenges. Questions to guide the discussion will be pre-determined by consensus among the collaborators and the themes will include: data governance, community engagement, Indigenous-led linkage and analysis of data, and decision-making regarding access to linked data. Other participants attending the session will be encouraged to listen and will have an opportunity to engage in the discussion and ask questions.&#x0D; Intended output or outcome:The key outcome of the session will be twofold. First, those actively working with Indigenous linked data will have an opportunity for an in-depth and meaningful dialogue about their work, which will promote international collaboration and sharing of ideas. Second, those with less experience and knowledge of the principles of Indigenous data sovereignty and their practical application will have an opportunity to listen to Indigenous people who are advancing the integration of Indigenous ways of knowing into data linkage and analysis.&#x0D; The output of the session will be a summary paper highlighting both the diversity and commonalities of approaches to Indigenous data linkage internationally. Areas where consensus exists, opportunities for collaboration, and challenges will be highlighted.

Approaches to Capacity Building for Machine Learning and Artificial Intelligence Applications in Health

Many health systems and research institutes are interested in supplementing their traditional analyses of linked data with machine learning (ML) and other artificial intelligence (AI) methods and tools. However, the availability of individuals who have the required skills to develop and/or implement ML/AI is a constraint, as there is high demand for ML/AI talent in many sectors. The three organizations presenting are all actively involved in training and capacity building for ML/AI broadly, and each has a focus on, and/or discrete initiatives for, particular trainees.&#x0D; P. Alison Paprica, Vector Institute for artificial intelligence, Institute for Clinical Evaluative Sciences, University of Toronto, Canada. Alison is VP, Health Strategy and Partnerships at Vector, responsible for health strategy and also playing a lead role in “1000AIMs” – a Vector-led initiative in support of the Province of Ontario’s \$30 million investment to increase the number of AI-related master’s program graduates to 1,000 per year within five years.&#x0D; Frank Sullivan, University of St Andrews Scotland. Frank is a family physician and an associate director of HDRUK@Scotland. Health Data Research UK \url{https://hdruk.ac.uk/} has recently provided funding to six sites across the UK to address challenging healthcare issues through use of data science. A 50 PhD student Doctoral Training Scheme in AI has also been announced. Each site works in close partnership with National Health Service bodies and the public to translate research findings into benefits for patients and populations.&#x0D; Yin Aphinyanaphongs – INTREPID NYU clinical training program for incoming clinical fellows. Yin is the Director of the Clinical Informatics Training Program at NYU Langone Health. He is deeply interested in the intersection of computer science and health care and as a physician and a scientist, he has a unique perspective on how to train medical professionals for a data drive world. One version of this teaching process is demonstrated in the INTREPID clinical training program. Yin teaches clinicians to work with large scale data within the R environment and generate hypothesis and insights.&#x0D; The session will begin with three brief presentations followed by a facilitated session where all participants share their insights about the essential skills and competencies required for different kinds of ML/AI application and contributions. Live polling and voting will be used at the end of the session to capture participants’ view on the key learnings and take away points.&#x0D; The intended outputs and outcomes of the session are:&#x0D; &#x0D; Participants will have a better understanding of the skills and competencies required for individuals to contribute to AI applications in health in various ways&#x0D; Participants will gain knowledge about different options for capacity building from targeted enhancement of the skills of clinical fellows, to producing large number of applied master’s graduates, to doctoral-level training&#x0D; &#x0D; After the session, the co-leads will work together to create a resource that summarizes the learnings from the session and make them public (though publication in a peer-reviewed journal and/or through the IPDLN website)

How integration of the federal Indian Register has enhanced First Nations-specific analysis of ICES data

IntroductionIn Ontario, First Nations are increasingly seeking population-level data about the health of their citizens. However, First Nations people are not readily identified in standard health administrative data and indirect strategies, such as the use of on-reserve addresses, are limited in scope and validity.&#x0D; Objectives and ApproachThe Chiefs of Ontario entered into a Data Governance Agreement with the Institute for Clinical Evaluative Sciences (ICES) that enabled the linkage of the federal Indian Register (IR) to data at ICES. This study examined the impact of the IR linkage on First Nations population estimates and location of residence, measured by postal code or residence code. Overall, and for each First Nation community in Ontario, we compared First Nations population estimates from the ICES data with and without the IR linkage to estimates available from Indigenous and Northern Affairs Canada (INAC).&#x0D; ResultsWithout the IR, using only Ontario residence codes or postal codes that were unique to a given community, 62,242 individuals were identified as living in First Nations communities. This is approximately 30% lower than the current INAC on-reserve population estimate of 92,234 for First Nations communities in Ontario. Adding the IR allowed the use of non-unique postal codes as well, resulting in the identification of an additional 15,183 First Nations individuals. It also allowed the identification of over 113,000 First Nations individuals who live outside of First Nations communities, especially in urban areas. Finally, the combination of residence information and the IR permits communities to identify their registered member living within and outside their communities.&#x0D; Conclusion/ImplicationsUsing the IR in combination with geographic residence information, made possible through the Data Governance Agreement signed between Chiefs of Ontario and ICES, will provide First Nations communities with more accurate and complete population estimates, which is key to the production of useful and relevant First Nations-specific health research.