Approximately 150 pharmaceutical industry professionals, payers and other subject matter experts came together late last month for the eyeforpharma 5th Annual Real World Evidence Europe conference (25–26 April 2017, Amsterdam, The Netherlands). With real world evidence (RWE) a topic of great interest to various key stakeholders in drug development, the meeting’s agenda encompassed current challenges, latest innovations and future trends in RWE. Day 2’s key themes included RWE methodologies, real world data (RWD) sources, clinical effectiveness measured via RWE vs efficacy measured in randomised controlled trials (RCT), electronic health record (EHR)-based clinical trials and integrated EHRs. The day’s talks are summarised below. A summary of Day 1 is available here.
Chair’s opening address
Day 2’s Chair, Alastair MacDonald (INC Research), started the day with a look at stakeholder perceptions of RWD. He emphasised the importance of selecting the correct methodology for generating credible RWE and highlighted that the day’s agenda would build on Day 1’s discussions by focussing on this issue in further detail.
From hypothesis to RWE – innovative use of RWD from Clinical Practice Research Datalink
Janet Valentine (Clinical Practice Research Datalink [CPRD]) got the day’s presentations underway by considering how digitisation offers new opportunities in the use of RWE. She illustrated this with an overview of the way in which CPRD (a UK-based government service) uses anonymised EHR data. These data are representative of the general UK population, provided by primary care practices, and linked to a range of other health databases. Janet described how EHR data could be used to conduct feasibility studies and for patient recruitment to clinical trials. She gave the recent example of the COPE study in chronic obstructive pulmonary disease (COPD), which used algorithms developed by the study’s researchers to recruit eligible patients in appropriate geographical locations, via their EHR records on CPRD. EHR data were then collected from trial participants via CPRD during the course of the trial. Janet proposed that EHR data, although collected primarily for use in clinical practice, could still provide high-quality data for research. Indeed, CPRD data have been used in 1700 peer-reviewed publications, as well as by the National Institute for Health and Care Excellence (NICE).
Tools and techniques for real-world data science and machine learning
Robin Murray (QuintilesIMS) took delegates on a whistle-stop tour of data science and machine learning. He emphasised the ever-increasing importance of these fields, and the need for greater capacity with respect to infrastructure and staffing — as both the volume of RWD generated and stakeholder demand for RWE continue to grow. Robin described data science techniques, such as phenotype vector-based analysis, as a cornerstone for interpreting RWD and converting these into RWE. He identified the process of converting patient data into vectors as a resource-intensive, potentially rate-limiting step and proposed that if this process could be refined and speeded up, more RWE could be generated.
Draw further insight from your RWD with interactive visualization techniques
Johan Liwing (Janssen) vividly illustrated the application of RWD, using an interactive platform developed to visualise and evaluate data on prostate cancer from the Stockholm-O cohort patient registry. Johan explained that using traditional techniques it can take weeks to analyse RWD from large patient registries to create a single data table. The system developed by Johan and his colleagues at the Karolinska Institute auto-generates data tables as the user views the database. The team plan to refine the platform further and look to apply the system to other therapy areas.
The GSK Salford Lung Study: lessons learnt
David Leather (GSK) and Jørgen Vestbo (University of Manchester) gave delegates an insight into the Salford Lung Study, an RWE-based clinical study considered pioneering in the field, and discussed their learnings from the project.
Jørgen began by outlining the burden of COPD and explaining that the study was set up to evaluate the effectiveness and safety of maintenance drug treatment vs ‘usual care’. Patients were evaluated in a real world setting and data were collected via an integrated EHR system. After an initial randomisation visit, patients continued with care in their usual clinical setting for one year. No research clinic visits or study interventions were conducted during this period, and RWD was collected throughout. Jørgen explained that, due to its real world setting, the patient population recruited into the Salford Lung Study was in some respects quite different to that commonly evaluated in clinical trials, and could be considered closer to the target patient population. Jørgen also highlighted that a larger number of adverse events and higher mortality rate are generally observed in real world studies compared with RCTs, as ‘real world patients’ tend to be in poorer health than those eligible for RCTs. Jørgen concluded with a look at some of the advantages and challenges surrounding the design of the Salford study. The trial enabled assessment of the effectiveness of care – an ‘endpoint’ of greater interest to health care professionals (HCPs) and patients than efficacy – and permitted assessment of patients in the primary care setting (where COPD is primarily managed). It also allowed collection of a large amount of data; an endeavour that does, however, require good infrastructure and can be costly. Jørgen proposed that the study, the first of its kind, sets a precedent in maintaining the scientific rigour of an RCT in a setting as close as possible to routine clinical practice.
David went on to describe some of the key drivers for performing the study in further detail. He emphasised the need to evaluate the drug in an environment as close to real clinical practice as possible, to determine real world effectiveness and to assess endpoints that may not be feasible in a standard RCT design. He explained that the selection of Salford as a location was also important. Salford is a geographically defined region, has a ‘paperless’ main hospital, and researchers were able to link to databases from a range of clinical settings to gather real-time RWD (including pharmacies, primary care physician practices, hospitals, research nurses and ‘out of hours’ clinics). David described the importance of wide collaboration with the whole local health care community to achieve this. He also spoke to the difficulties of recruiting a large number of patients and primary care physicians, and ensuring that ‘usual care’ could proceed during the one-year period, while still collecting data. He described how these challenges were overcome via an innovative ‘grassroots’ approach that incorporated advertising within the local community and involved local HCPs from the outset. David concluded by outlining some of the next steps for the project. With over 255 million rows of data still to explore, further insights are likely to be available as data analysis continues. An asthma study is also ongoing in Salford, with results expected in the third quarter of this year.
Innovative approaches to real world uses of EHR data
The ability to harness the potential of EHR data was also discussed by Jacqueline Brereton (Ignite Data). She described how an EHR-based system could be used for study planning and feasibility assessments, as well as in identifying trial sites and clinicians, recruiting patients in a resource-efficient manner, direct data collection and monitoring, and data analysis.
The patient perspective: real world evidence = real world patients: Design your real-world strategy around the key stakeholder
In the afternoon session, delegates were given a different perspective of RWE; that of the patient. Alan Thomas (Ataxia and Me) described his experience of living with ataxia and the importance of the patient voice in the drug development process, especially in the case of rare diseases. Reiterating a theme from Day 1 of the meeting that terminology relating to RWD and RWE is not standardised, Alan emphasised the detrimental impact this can have on communication between the pharmaceutical industry, healthcare professionals and patients. He informed delegates that patients gain information and share their experiences via online and social media platforms, rather than in a clinical setting. In many cases, patients may find it easier to enter into a dialogue via these methods. He closed by urging delegates to speak to patients: the ‘real world experts’ on RWE.
Real world evidence: who cares? Investigating how EU payers actually base decisions on RWE
Gavin Outteridge and Korina Papadopoulou (both from Kinapse) outlined some of the challenges that exist regarding payer perceptions of RWE and explored the potential of RWE pre-, peri- and post-health technology assessment (HTA). While most HTA bodies use RWE in post-review reimbursement decisions, few use RWE before or during this review, and overall there is a lack of a clear approach among payers across Europe. Gavin and Korina suggested three principles that pharmaceutical companies should follow to ensure that RWE is impactful: 1) Educate: build a cross-functional consensus on the business and patient benefits of investing in RWE; 2) Integrate: include RWE as standard in evidence-generation decisions, processes and templates from early phases onwards; 3) Communicate: engage proactively with payers and other stakeholders to convince them of the value of RWE for informing decision making.
Increase your understanding of the burden of illness through text mining
In the final presentation of the meeting, Asif Khan also drew delegates’ attention to the patient. Highlighting that patient history can represent up to 75% of a diagnosis, Asif proposed that the patient voice should be integral to the development of RWE. He highlighted that the evolving model of medicine globally – from a paternalistic to partnership-based relationship between the HCP and the patient – and the importance of patient-reported outcomes (PROs) in regulatory and HTA body evaluations add further weight to this. Asif went on to discuss the challenges that can be involved in PRO development, with new instruments taking 2–3 years to develop and validate. He also suggested that the relatively small number of patients from whom data is gathered during development of these measures could lead to important endpoints being overlooked. He suggested that a potential solution could be to gather information via open-entry fields in patient questionnaires and then to use text mining to analyse these for additional RWD. This exercise could be conducted at Phase II and used to inform the design of Phase III trials.
Asif also highlighted the emerging potential of social media as a source of RWD. While the impact of social media on regulatory and HTA body assessments remains to be seen, he believes this could become increasingly important as the field of drug development moves to ‘beyond the pill’ approaches. He concluded by describing some unresolved questions associated with the use of ‘big data’; such as who has access to the data, for what purpose(s), who benefits, who pays and – underlying all of these points – who decides on these issues.
Chair’s concluding remarks
Alastair MacDonald (INC Research) drew the day to a close with a summary of the topics covered over the two-day meeting, reminding delegates of the importance of issues such as RWE credibility, and emerging trends such as the use of ‘big data’. Paul Simms (eyeforpharma) thanked the Chairs, speakers and delegates for their participation in the meeting and in the useful discussions that had arisen. Alastair closed the session with a call to action: to generate credible RWE to assist patients – the ultimate goal of drug development.
Day 1 meeting report available here.
AN ANNOUNCEMENT FROM EYEFORPHARMA:
The American counterpart of this event, Data, Evidence and Access Summit 2017 will be held 13–14 November, Philadelphia, USA. For more information, or if you want to get involved, please contact James Mackintosh at Jmackintosh@eyeforpharma.com