Using Big Data to Conduct Innovative Cardiovascular Clinical Trials

Using Big Data to Conduct Innovative Cardiovascular Clinical Trials: The Community Heart Outcomes Improvement and Cholesterol Education Study (CHOICES)

Knowledge Translation Program of St. Michael's Hospital, Heart and Stroke Foundation of Ontario, Heart & Stroke Richard Lewar Centres of Excellence in Cardiovascular Research, CorHealth Ontario, The Ontario Spor Support Unit

Traditional randomized clinical trials (RCTs) have provided extremely valuable information on medical therapies and procedures that have changed the way heart diseases are treated. However, despite these contributions, traditional RCTs are costly, the findings may not be applicable to patients unlike those in the study, and the use of trial findings may be infrequent. These limitations may be addressed by incorporating 'big data' in RCTs, which is the emerging field using electronic information that is routinely collected in various large administrative health databases. The Community Heart Outcomes Improvement and Cholesterol Education Study (CHOICES) will test the potential of using 'big data' in a 'real-world' clinical trial to measure outcomes using routinely collected health information. CHOICES aims to increase the use of cholesterol-lowering statin drugs to prevent heart attack and stroke in high-risk health regions across Ontario using a 'toolbox' of interventions. The 'toolbox' of interventions are informational strategies targeted for both patients and family physicians to help improve cholesterol management and contribute to shared decision making for heart healthy goals.

An estimated 19,500 cardiac events could be prevented each year in Canada by use of statin therapy as recommended in the Canadian Cardiovascular Society's Lipid Guidelines. Despite substantial evidence supporting statin use, several studies suggest dyslipidemia management in Canada remains suboptimal. In Ontario, prior work using the 2008 Cardiovascular Health in Ambulatory Care Research Team (CANHEART) 'big data' registry of almost the entire Ontario population of 9.8 million adults created through linkage of 17+ population health databases at ICES, has documented an approximate 2-fold variation across the province in cardiovascular events that is associated with performance of key cardiovascular preventive measures, particularly lipid screening and statin prescribing. This work noted that the variation did not have a clear association with traditional clinical risk factors or socioeconomic conditions. This observation suggests that heterogeneity in this care process may be modifiable with an intervention geared to improving adherence to national guidelines. In this pragmatic, cluster randomized registry trial, 'big data' is used to test the 'real world' effectiveness of a tailored, multicomponent intervention strategy aimed at improving lipid management (screening, risk assessment, statin initiation, statin adherence) amongst a primary prevention cohort of 40 to 75 year olds individuals living in 14 (of 28) communities in Ontario with higher than average rates of cardiovascular events. A multicomponent intervention strategy will include a 'toolbox' of lipid management resources for both patient and physicians in the intervention (high-risk) communities of the province. The intervention strategy will include tools to enable patients and physicians to make informed and shared decisions about statin therapy and will be implemented in the intervention communities using targeted local and social media strategies. Patient characteristics for those aged 40 to 75 and clinical outcomes in this study will be measured without primary data collection using the 2016 CANHEART 'big data' registry, with the exception of stain use and adherence data available only in adults 66 to 75 years old.

This is a two-arm parallel assignment trial, cluster randomized at a community level. The 14 communities in each arm of the study will receive either the usual standard of care or the intervention in parallel.

The participants and care providers will be masked to what communities are enrolled in the trial, but due to the large scale study promotion that is required for dissemination, they may be aware that the resources they are receiving are part of an ongoing clinical trial. The principal investigator and a statistician will be blinded from which communities are assigned to which arm.

The 14 communities that are in the control arm of the trial will receive usual standard of care. The usual standard of care will follow clinical daily practice patterns provided by family physicians in Ontario for CVD prevention. This follows the periodic standard of care provided by Canadian cholesterol, hypertension, and diabetes best practice guideline recommendations utilized based on each physician's clinical judgement, physical assessment, and discretion. Patients also typically have access to existing cardiovascular prevention materials offered online through publicly available websites.

The 14 communities that are in the intervention arm of the trial will receive a multicomponent intervention that provides both physicians and patients with access to a 'toolbox' of lipid management resources. The components planned for the 'toolbox' are all evidence-based interventions and chosen after consultations with Canadian family physicians and implementation science experts based on their potential for scalability to the entire population, cost and practicality. Online tools will be used and the trial will leverage pre-existing implementation initiatives (e.g., newsletters, listservs) wherever possible to minimize study costs and increase accessibility.

The intervention toolbox will include: 1) community-level report cards on lipid management (developed using an updated version of the 2016 CANHEART 'big data' registry of ~10.9 million adults created through linkage of 19+ population health databases) to distribute to family physicians, 2) printed and electronic patient education materials on cholesterol screening and management, 3) a new online clinical decision aid to facilitate shared decision-making between patients and their family physicians regarding statin utilization, 3) patient educational videos, and 4) physician educational videos and material.

Proportion of FRS determined intermediate- and high-risk residents (aged 66 to 75) in each community who filled a statin prescription within 100 days, as measured by the CANHEART registry at the completion of the 3 year intervention period.

The number of lipid-related visits to primary care physicians for the primary prevention cohort of 40 to 75 year olds in each community.

Adherence rates to statins in FRS determined intermediate- and high-risk statin users 66 to 75 year olds in each community. Adherence will be measured at 1.2 times the prescription length.

Proportion of 40 to 75 year olds in the primary prevention cohort for each community receiving lipid screening from lab data.

The proportion of AMI, stroke or CVD death in the primary prevention cohort of 40 to 75 year olds in each community, along with the individual components of these incident composite outcomes.

The proportion of AMI, stroke, or CVD death in addition to revascularization procedures in the primary prevention cohort of 40 to 75 year olds in each community, along with the individual components of these incident composite outcomes.

The incident rates of DM in 40 to 75 year olds in the primary prevention cohort of 40 to 75 year olds in each community.

Inclusion Criteria: Community with CVD incidence rates higher than the Ontario provincial average Community with a population size greater than 5,000 40 to 75 year olds Community with at least 1,000 66 to 75 year olds Community with 20 to 130 active and practicing family physicians Exclusion Criteria: Patients with established CVD within each community

ICES; Women's College Hospital; Peter Munk Cardiac Centre, Toronto General Hospital; University of Toronto

James S, Rao SV, Granger CB. Registry-based randomized clinical trials--a new clinical trial paradigm. Nat Rev Cardiol. 2015 May;12(5):312-6. doi: 10.1038/nrcardio.2015.33. Epub 2015 Mar 17.

Anderson TJ, Gregoire J, Pearson GJ, Barry AR, Couture P, Dawes M, Francis GA, Genest J Jr, Grover S, Gupta M, Hegele RA, Lau DC, Leiter LA, Lonn E, Mancini GB, McPherson R, Ngui D, Poirier P, Sievenpiper JL, Stone JA, Thanassoulis G, Ward R. 2016 Canadian Cardiovascular Society Guidelines for the Management of Dyslipidemia for the Prevention of Cardiovascular Disease in the Adult. Can J Cardiol. 2016 Nov;32(11):1263-1282. doi: 10.1016/j.cjca.2016.07.510. Epub 2016 Jul 25.

Tu JV, Chu A, Maclagan L, Austin PC, Johnston S, Ko DT, Cheung I, Atzema CL, Booth GL, Bhatia RS, Lee DS, Jackevicius CA, Kapral MK, Tu K, Wijeysundera HC, Alter DA, Udell JA, Manuel DG, Mondal P, Hogg W; Cardiovascular Health in Ambulatory Care Research Team (CANHEART). Regional variations in ambulatory care and incidence of cardiovascular events. CMAJ. 2017 Apr 3;189(13):E494-E501. doi: 10.1503/cmaj.160823.

Tu JV, Donovan LR, Lee DS, Wang JT, Austin PC, Alter DA, Ko DT. Effectiveness of public report cards for improving the quality of cardiac care: the EFFECT study: a randomized trial. JAMA. 2009 Dec 2;302(21):2330-7. doi: 10.1001/jama.2009.1731. Epub 2009 Nov 18.