Advertisement

Trials embedded in cohorts, registries, and health care databases are gaining ground

      Traditionally, randomized trials are organized and funded on a case-by-case basis for each potential new piece of the intervention knowledge mosaic. However, this may not be a sustainable model, given the increasing awareness of important topics to be investigated worldwide in combination with limitations in research capacity, resources, and participating patients.
      It is therefore likely that the system for organizing trials will increasingly embrace alternative approaches, such as trials embedded in cohorts (see Couwenberg et al. 10040 in this issue, highlighted later in this editorial), and trials based on registers and health care data sources [
      • Metsemakers J.F.
      • Höppener P.
      • Knottnerus J.A.
      • Kocken R.J.
      • Limonard C.B.
      Computerized health information in The Netherlands: a registration network of family practices.
      ,
      • Sacristán J.A.
      • Soto J.
      • Galende I.
      • Hylan T.R.
      Randomized database studies: a new method to assess drugs' effectiveness?.
      ,
      • Glynn R.J.
      • Brookhart A.M.
      • Stedman M.
      • Avorn J.
      • Solomon D.H.
      Design of cluster-randomized trials of quality improvement interventions aimed at medical care providers.
      ,
      • Lauer M.
      • Gordon D.
      • Wei G.
      • Pearson G.
      Efficient design of clinical trials and epidemiological research: is it possible?.
      ]. In addition, intensive and creative use of quasi-experimental approaches embedded in routine care is an important development [
      • Bärnighausen T.
      • Røttingen J.A.
      • Rockers P.
      • Shemilt I.
      • Tugwell P.
      Quasi-experimental study designs series-paper 1: introduction: two historical lineages.
      ].
      These approaches are promising with regard to a number of aspects. The efficiency of intervention research will probably increase if already available sampling frames, databases and infrastructures can be used. Opportunities for long-term follow-up, external validity and generalizibility can profit from trials being based on well defined pre-existing population denominators, healthcare-based registers or GPs’ electronic health records [
      • Metsemakers J.F.
      • Höppener P.
      • Knottnerus J.A.
      • Kocken R.J.
      • Limonard C.B.
      Computerized health information in The Netherlands: a registration network of family practices.
      ,
      • Mathes T.
      • Buehn S.
      • Prengel P.
      • Pieper D.
      Registry-based randomized controlled trials merged the strength of randomized controlled trials and observational studies and give rise to more pragmatic trials.
      ]. This can also promote the implementability of the results in clinical practice and improve the conditions for practice-oriented N of 1 trials [
      • Guyatt G.H.
      • Keller J.L.
      • Jaeschke R.
      • Rosenbloom D.
      • Adachi J.D.
      • Newhouse M.T.
      The n-of-1 randomized controlled trial: clinical usefulness. Our three-year experience.
      ]. In addition, if not only the design and practical aspects of trials but also their hypotheses are elaborated in a broader clinical context, the relevance of the research agenda might better match essential health care challenges, as compared with research imposed by academic competitiveness and productivity. Moreover, embedding intervention studies in existing health care data infrastructures can increase the resources for practice-relevant trials by facilitating financial collaborations between research institutions, funding agencies, health insurers, and healthcare organisations. This may also make research groups less dependent on commercial research agendas.
      At the same time, a number of methodological challenges must be effectively tackled [
      • Li G.
      • Sajobi T.T.
      • Menon B.K.
      • Korngut L.
      • Lowerison M.
      • James M.
      • et al.
      2016 Symposium on Registry-Based Randomized Controlled Trials in Calgary. Registry-based randomized controlled trials- what are the advantages, challenges, and areas for future research?.
      ]. Valid baseline and outcome measurements and adequate trial follow-up procedures must be safeguarded in the context of pre-existing cohort- and healthcare-related data bases. Also the need and possibilities for blinding must be considered. When employing quasi-experimental designs, solutions for confounding by indication need special attention.
      Major ethical issues must be considered such as: data protection and privacy; and the question whether screening patients for eligibility for clinical intervention research is allowed in cohorts and healthcare-based registers that were designed for other purposes, without a priori consent for yet unknown research questions. How far could a more general, broad a priori consent reach to serve the interest of future patients? And what could be the ethical implications of the consideration that clinical research and practice both serve the patients’ interests and might increasingly converge and integrate as a living laboratory for continuous learning and healthcare progress?
      Furthermore, the conflict of interest profile in embedded trials needs more attention. While there may be less dependence on specific commercial interests, the influence of healthcare professionals’ interests and beliefs may be more prominent. In addition, a need to financially maintain cohort or registry data infrastructures could influence the research agenda. For considering the various types of conflict of interest and how these could be and are being managed, the analyses in the papers by Grundy 10038 and Lund 10029 that are published in this issue (see also below), present useful new insights.
      Addressing these challenges will often require tailored approaches. For example, in the case of scientifically truly innovative interventions mainly MRC-like funding will be necessary, while for critical evaluation of already accepted interventions public-private co-financing in the interplay of clinical research and health care should work. Specific single patient care-research could even be covered by health insurance only.
      The aforementioned paper by Couwenberg et al. 10040, underlines the advantages of embedded trials. In a clinical oncology setting, they studied whether the ‘trials within cohorts (TwiCs) design’ indeed facilitates efficiency, more specifically regarding patient enrollment, and generalizability. Interestingly, while previous TwiCs studies used batch randomization in longitudinal large cohorts, their study randomized patients directly after cohort enrollment. For this purpose, they used their experience in conducting the first TwiCs, based on the infrastructure of two radiotherapy centers. They asked patients with rectal cancer to participate in a prospective cohort study and to provide ‘broad consent’ for randomization and patient-reported outcomes (PROs). Those who consented and met the trial's inclusion criteria were randomized directly after cohort enrollment in order to investigate the added value of a radiotherapy boost. The acceptance rate and its impact on sample size were evaluated, and the clinical characteristics of the patients in the trial and of the patients included in the Dutch national cancer registry were compared. The authors found that the TwiCs design is feasible, resulted in efficient patient recruitment, allowed enrollment of a high proportion of randomizable patients, improved generalizability, and allowed easy sample size adaptation. They suggest that the TwiCs design be considered more often for pragmatic trials in a clinical oncology setting, to improve generalizability of the results and shorten trial duration.
      Regarding the various types of conflict of interest and how to deal with these, Grundy c.s. 10038 investigated the range of issues labeled as ‘‘non-financial conflicts of interest’’ in biomedicine, the associated concerns, and the implications of defining these issues as conflicts of interest. In a qualitative study they triangulated data from three sources - literature, policies, and interviews - with participation of the corresponding authors of the sampled literature. A wide range of personal, social, professional, intellectual attributes and relationships, and even financial interests were labeled as ‘nonfinancial conflicts of interest’. It was found that, although there is no consensus regarding the nature of the problem, many ‘non-financial’ interests are currently subject to policy action. Interestingly, according to the authors labeling ‘non-financial’ interests as conflicts of interest may obscure underlying problems related to lack of diversity, representation, and power dynamics within evidence-led processes. This may have unintended consequences, such as exclusion of diverse perspectives. They recommend: defining concerns characterized as non-financial conflicts of interest in terms of what they are rather than what they are not, preventing financial conflicts of interest, and ensuring inclusive and equitable representation within evidence-based processes. In addition, they emphasize the need for new conceptual tools to address ‘nonfinancial’ interests and a renewed focus on managing conflicts of interest.
      For the evaluation of conflicts of interest, methods to appraise them and the way they are handled are also important. Lund and co-authors 10029 identified and summarized appraisal tools and other guides that address conflicts of interest in medical research and they reviewed top journals policies on managing conflicts of interest. After an extensive search in bibliographic databases, other sources, and websites of 30 top medical journals, they included 27 appraisal tools. These were not specifically designed for addressing conflicts of interest and included only one or two short items on conflicts of interest. Most of them addressed study funding and authors’ conflicts of interest. Less than half of the tools addressed availability of conflicts of interest information (9), reported conflicts of interest (13), and influence from conflicts of interest (5). Accordingly, the authors concluded that the identified appraisal tools address conflicts of interest typically superficially, and rarely address how conflicts of interest may influence studies. They also found that a minority of journals had explicit conflicts of interest managing policies beyond standard disclosure practices. The investigators stress the need for more comprehensive evidence-based guidance for addressing conflicts of interest in research.
      In conclusion, the work of the groups of Couwenberg and Lund makes it clear that there is much room for improvement in defining, appraising, and managing conflict of interest in biomedicine.

      References

        • Metsemakers J.F.
        • Höppener P.
        • Knottnerus J.A.
        • Kocken R.J.
        • Limonard C.B.
        Computerized health information in The Netherlands: a registration network of family practices.
        Br J Gen Pract. 1992; 42: 102-106
        • Sacristán J.A.
        • Soto J.
        • Galende I.
        • Hylan T.R.
        Randomized database studies: a new method to assess drugs' effectiveness?.
        J Clin Epidemiol. 1998; 51: 713-715
        • Glynn R.J.
        • Brookhart A.M.
        • Stedman M.
        • Avorn J.
        • Solomon D.H.
        Design of cluster-randomized trials of quality improvement interventions aimed at medical care providers.
        Med Care. 2007; 45: S38-S43
        • Lauer M.
        • Gordon D.
        • Wei G.
        • Pearson G.
        Efficient design of clinical trials and epidemiological research: is it possible?.
        Nat Rev Cardiol. 2017; 14: 493-501
        • Bärnighausen T.
        • Røttingen J.A.
        • Rockers P.
        • Shemilt I.
        • Tugwell P.
        Quasi-experimental study designs series-paper 1: introduction: two historical lineages.
        J Clin Epidemiol. 2017; 89: 4-11https://doi.org/10.1016/j.jclinepi.2017.02.020
        • Mathes T.
        • Buehn S.
        • Prengel P.
        • Pieper D.
        Registry-based randomized controlled trials merged the strength of randomized controlled trials and observational studies and give rise to more pragmatic trials.
        J Clin Epidemiol. 2018; 93: 120-127https://doi.org/10.1016/j.jclinepi.2017.09.017
        • Guyatt G.H.
        • Keller J.L.
        • Jaeschke R.
        • Rosenbloom D.
        • Adachi J.D.
        • Newhouse M.T.
        The n-of-1 randomized controlled trial: clinical usefulness. Our three-year experience.
        Ann Intern Med. 1990; 112: 293-299
        • Li G.
        • Sajobi T.T.
        • Menon B.K.
        • Korngut L.
        • Lowerison M.
        • James M.
        • et al.
        2016 Symposium on Registry-Based Randomized Controlled Trials in Calgary. Registry-based randomized controlled trials- what are the advantages, challenges, and areas for future research?.
        J Clin Epidemiol. 2016; 80: 16-24https://doi.org/10.1016/j.jclinepi.2016.08.003