By Deborah Borfitz

May 23, 2019 | Electronic medical record (EMR) systems contain a wealth of longitudinal real-world data (RWD) but, as researchers are aware, were never purposefully built for running trials or finding patients to populate them. This secondary use of EMRs has nonetheless grown in popularity due to their ubiquity and sheer volume of data being collected, and the availability of compatible clinical research software.

But data are imperfect proxies for events such as diagnoses and treatment, and what constitutes a service can be highly site dependent. Information is often incomplete or inaccurate, if only because patients obtain care from multiple unconnected locations. The challenges are universal but, in some respects, more easily surmounted in countries like Canada with a publicly funded healthcare system that tightly regulates EMR vendors and supports interdisciplinary healthcare management.

The entire country is served by only a handful of EMR systems versus around 600 in the US, according to Karri Venn, president of research at LMC Manna Research, a Canadian-wide site network whose head office is based in Toronto. But while Canada may have a more integrated health information system, clinical care and clinical research remain largely segregated.

LMC Manna Research is an exception to the rule and, ironically, was born of the chief disadvantage of Canada's healthcare model—restricted access to funding by healthcare professionals in the community. Two decades ago, it was founded by a physician intent on building a community-based "medical home" to better manage the chronic diseases of patients, says Venn. "He realized that research was a creative way to support his model of care."

Automating Searches

The operations of LMC and Manna merged in 2015 to form LMC Manna Research, comprising 20 sites. Twelve of the sites are endocrinology practices with embedded clinical trials and the remainder are a mix of standalone and embedded research sites affiliated with healthcare practices, says Venn. At any one time, the organization has over 100 trials underway in various therapeutic areas. It uses a single clinical trial management system, self-built before such applications were commercially available, and the country's most common EMR (Accuro) is used across the group.

Query functionality is absent or inadequate in most EMR systems today, notes Venn. So LMC Manna Research created its own query system, which to end users appears as a dropdown menu of diagnoses, with the ability to drill down further based on a growing list of variables that includes age, lab value ranges as well as active medications. Automating the task of finding patients potentially eligible for studies saves weeks of staff time per search that can be better spent on educating and otherwise improving the patient experience, she says.

Established companies generating revenues doing this sort of work include Deep 6 AI, TriNetX and IQVIA.

The query functionality has been a "game changer," says Venn, in part because it helps LMC Manna Research know which protocols to say no to from the outset. When there are clearly enough potential participants, the patient recruitment department is ready to jump into action because more than 90% of the network's patient base has already pre-consented to being contacted about study opportunities.

A lesson learned years ago is that study coordinators generally don't favor making recruitment calls, says Venn, so LMC Manna Research instead hired dedicated personnel who enjoy working in a call center-like environment. "Because we're on one EMR in a central location, we can book centrally and make warm transfers to coordinators on-site."

Enrollment Success

Part of the paperwork all new patients fill out at their first patient clinic office visit is a page where they give or withhold consent to be called about study opportunities and to have their de-identified data used for real-world evidence (RWE) studies. The second question was added in the past year after a tiny dip in the number of people saying yes to the first question, Venn says. "We want to see if we're missing people who would be fine with us using their big data but don't necessarily want to participate in a traditional clinical trial."

Pre-consenting patients for future data review for pre-qualification purposes is not currently standard practice but it should be, says Venn. "Accessing even de-identified patient data may be a violation of [patient] privacy so healthcare practices sharing patient data should consider updating their current practices."

The big pool of willing patients can quickly get whittled down by protocol inclusion and exclusion criteria, or because the study is too time-consuming or ill-timed for people, says Venn. But often, LMC Manna Research is among the top enrollers. In one recent global study, three of its network sites were the best performers when it came to randomizing patients—eight times more efficient than other sites in the study, by Venn's calculation.

LMC Manna Research was also one of the first clinical research sites to be part of Sanofi's global investigator network as a center of excellence, Venn says. That allows it to have input on protocol designs and join the "Sanofi Makes Investigators Lives Easier" (aka SMILE) initiative endeavoring to raise awareness of clinical trials and develop new approaches to patient recruitment. Sites in Sanofi's network randomize two to nine times more patients into trials than non-network sites, she adds.

"We don't have to rely on advertisements and hope we can find patients internally," says Venn, "because we've used big data to pre-identify interested patients within our clinic practice environment." Patient retention rates are also impressive. A single LMC Manna Research site randomized 161 elderly patients into one recently completed, six-year trial and maintained a 100% retention rate.

Emerging companies, such as Elligo Health Research, are embedding themselves in healthcare practices and essentially emulating the LMC Manna Research model, she notes. In this age of targeted drug discovery, access to large patient databases is fast becoming a necessity.

Data Collection

A key component of the "research-grade EMR" at LMC Manna Research is a set of standard operating procedures (SOPs) that control how data get entered into the system and what questions get asked, explains Venn. The SOPs limit the use of free-text notes and mandate the collection of demographic data such as ethnicity and education data. To avoid physician workarounds, medical assistants have been hired for medical information intake with clinic patients prior to the physician visit, she says.

Including social determinants of health data in the EMR reflects both the diversity of patients served by LMC Manna Research and the interests of the physicians it employs, says Venn. A published study on refractory patients with type 2 diabetes has found that addressing psychological, socioeconomic and access barriers can significantly improve glycemic control. The information also provides "valuable insights" to the marketing team as it seeks to increase diversity in studies, as is being encouraged by the FDA.

LMC Manna Research has also validated a Skills, Confidence & Preparedness Index (SCPI) it developed to assess self-management in patients with diabetes, which can help determine their barriers to improved care, says Venn. The questionnaire is now publicly available for all healthcare professionals to use, she adds.

Patient-reported outcomes data gets collected via an app, says Venn, although LMC Manna Research will be switching to a new tool soon. Last fall, the FDA released its MyStudiesApp, giving researchers a free, pre-approved tool for inputting RWD from mobile and wearables devices that "takes the complexity out of our hands… and gets data to the FDA in standardized fashion." Open source code has been released for both Apple and Android devices.

Analytics and Pragmatic Trials

LMC Manna Research has a team of data scientists on staff to oversee EMR data extraction and queries, says Venn, must-haves for analytics on a database that includes upwards of 38,500 active type 2 diabetes patients. The data scientists also produce report cards on how well physicians follow standardized diabetes care guidelines—e.g., achieving target LDL cholesterol and A1C levels and prescribing cardiovascular protection medications—using peer-based practices to encourage better patient care management. On-staff epidemiologists analyze questions that require answers, write study protocols and finalize clinical study reports on internally collected data.

Although LMC Manna Research has a 12-bed phase 1 unit, it predominantly conducts phase 2-3 trials. A growing proportion of requests are related to pragmatic clinical trials using RWE in lieu of a traditional phase 4 trial, says Venn. The integrated healthcare model is expected to be well suited to these types of trials for purposes of randomizing standard-of-care patients for pre- and post-marketing study analysis.

The challenge currently is that EMRs are typically not FDA 21 CFR Part 11 compliant. Research will profoundly change once EMRs become more like electronic data capture (EDC) systems, Venn says. "Technically, your EMR would then become your [data] source and EDC all at once." It may happen sooner than some people think, she adds, noting Roche's recent purchase of oncology-specific electronic health record vendor Flatiron Health. Telecommunications company Telus has also been actively acquiring EMRs in Canada.

The traditionally fragmented and slow-to-change research industry will need to become "more open to change," says Venn, in response to the accelerating pace of data collection expected over the next decade.