Developing an Ethical and Legal Interoperability Assessment Process for Retrospective Studies

Abstract

The past decade has witnessed the creation of major international research consortia, aiming to facilitate the sharing of data from different studies to maximize health benefits. However, combining heterogeneous data across existing studies requires addressing issues related to both data harmonization and ethical and legal interoperability. This article proposes a rigorous interoperability assessment process to assess whether different data sets are sufficiently ethically and legally interoperable to allow for a given proposed research use. The methodology used to develop this process is based on a comprehensive analysis of the international ethical and legal framework governing the use of retrospective data in research, and includes the following steps: (I) finding existing processes; (II) comparing processes to identify similarities and differences and determining the limits of the “consistent whole”; (III) establishing common principles and procedures; and, (IV) changing or removing processes that do not contribute to the consistent whole. Each of these four steps were examined using step-specific methodologies, including (a) literature and policy reviews; (b) consultations with international ethical, legal and social implications (ELSI) experts; and (c) a case study piloting the proposed framework in an actual international research consortium. This assessment process takes into account key legal and ethical components such as consent, recontact, and waiver of consent. As a result, this analysis allows the development of a comprehensive filter used to verify the legal and ethical restrictions pertaining to a data set. This in turns helps in determining whether the given data set can to be used for a proposed research project, or is ethically and legally interoperable for use in research collaborations. By integrating this filter to the regular data access processes used by cohorts, not only will researchers be able to create virtual “mega-cohorts” of research participants, but it will also ensure that these cohorts respect basic legal and ethical precepts.

Introduction

The past decade has witnessed the creation of major international research consortia aiming to facilitate the sharing of data from different studiesⁱ across the world¹ to maximize health benefits. Indeed, although most existing studies are stand-alone infrastructures, many are increasingly forming partnerships with others to obtain the number of participants required to explore the interaction of genetic, lifestyle, environmental, and social factors.¹ Such consortia include, for example, the European Network for Genetic and Genomic Epidemiology (ENGAGE)² and the Biobank Standardisation and Harmonisation for Research Excellence in the European Union (BioSHaRE-EU).³ However, combining heterogeneous data across existing studies requires addressing issues related to both data harmonization and ethical, legal and social implications (ELSI) interoperability.⁴

Data harmonization can help in achieving or improving the comparability of similar measures collected from different individuals, by different studies.⁵ A number of resources have been developed to ensure optimal harmonization, integration, and coanalysis of data by research partnerships. For instance, Maelstrom Research and its partners⁶ have developed methods and software to foster a generic, but systematic, approach to retrospective data harmonization and to guide investigators in achieving harmonization and integration of data.^7,8

However, data harmonization, by itself, is not sufficient to ensure that data sets from different studies are interoperable.⁹ Understanding the ethical and legal “availability” of individual participants' data collected by studies is a critical prerequisite to evaluate whether these data can be shared (or not) and how.¹⁰ Evaluation of the data availability is therefore key to establishing research collaborations across studies. Indeed, it is essential to ensure that the study-specific legal and ethical elements adequately address the specific scientific objectives of the research project. Although this exercise can be done for each research question, it is a time- and resource-consuming task. Currently, no formal tool or method exists to provide a rigorous and user-friendly assessment of the ethical and legal interoperability of data across studies.

This article proposes a rigorous ELSI interoperability assessment process and the development of a comprehensive ethical and legal filter. The methodology used is based on a comprehensive analysis of the international ethical and legal framework governing the use of retrospective data in research. The resulting filter can be used either (1) to verify whether the legal and ethical framework governing a data set allows for the proposed research use or (2) to assess whether the different data sets are sufficiently ethically and legally interoperable to allow for a collaborative research use (e.g., use of data sets from different studies for the same research use). In addition to proposing a list of criteria for ELSI assessment of data interoperability, the results of this analysis can also help in assessing models of data sharing that research networks or consortium may want to explore (e.g., federated access process, centralized meta-analysis, etc.), given the identified legal and ethical limits to data access and use. Although this model was initially developed to facilitate the assessment of the ethical and legal “availability” of data from individual participants, it could also be used to assess the research uses of biospecimens.

Overall, the process aims to help in leveraging collaborations and establishing data-sharing infrastructures. It is an essential step in the goal to streamline data sharing; however, it should be noted that, in most cases, access to data will require that study-specific procedures be followed to complete the access process.

Materials and Methods

A rigorous methodology was used to develop the legal and ethical assessment process. To do so, the following steps were undertaken¹⁰:

I. Finding existing processes.

II. Comparing processes to identify similarities and differences, and determining the limits of the “consistent whole” (as defined hereunder).

III. Establishing common principles and procedures.

IV. Changing or removing processes that do not contribute to the “consistent whole.”¹⁰

As no fixed method currently exists for ethical and legal interoperability assessment, the methodology used in this article is based on the aforementioned methodology, developed based on a recognized legal harmonization method.¹⁰ A few necessary adjustments were made to adapt it to a legal and ethical analysis.

Results

In the following sections, each of these four steps is examined in the context of legal and ethical interoperability, using step-specific methodologies, including (a) literature and policy reviews (steps I, II, and III); (b) consultations with international ELSI experts to identify key legal and ethical components (step II); and (c) a case study piloting the proposed framework in an actual international research consortium (step IV). The literature and policy reviews constitute a preliminary step toward the development of a comprehensive ethical and legal assessment process. However, due to the magnitude of the task, these reviews have been divided into three parts and published as stand-alone publications.¹¹

Finding existing processes

The first step toward the development of an ethical and legal interoperability assessment procedure consists of finding existing processes, which, in turn, entails identifying the common standards applicable to the analysis. In this case, it requires the identification of legal and ethical elements of importance with regard to the use of study-specific data.

To coanalyze data in the context of a research collaboration, the framework for the secondary use of data (i.e., using data in a way that differs from the original purpose it was collected for) must be examined for each study, to determine the extent to which data can be transferred to an external server and if access to individual data is allowed. This is done to determine if, and how, ethically and legally harmonized data sets can be generated across studies. The legal and ethical processes related to the secondary use of data in research were examined in two prior policy and literature reviews.

The first article reviewed the legal and ethical frameworks governing the use of deceased individuals' data for research. It highlighted that despite national legal and ethical specificities, the general process allowing for the determination of the ELSI availability of data for research is similar in all the jurisdictions studied.¹² According to the results of this comparative legal analysis, it was determined that “[i]nternational and national legal frameworks provide guidance to promote a wider (albeit limited) secondary use of data, while protecting research participants' rights and interests. Among the proposed solutions, authors note: (1) reviewing the initial consent form; (2) re-contacting and re-consenting research participants, when legally and ethically acceptable; and (3) obtaining a waiver of consent from a competent research ethics board […]”¹³

Based on these prior analyses, the general process to assess the ELSI “availability” of data for secondary use in research is determined to be based on the following three key components:

A. Content evaluation of the consent form used for the initial collection of data.

B. If the initial consent form does not allow for the broad/new research uses (secondary uses), assess whether it is possible to recontact participant to obtain a new consent (recontact/reconsent).

C. If obtaining a new consent is impossible or not feasible, assess whether it is possible to obtain a waiver of consent.

These three components provide a portrait of all the practices adopted by various international organizations and across jurisdictions, and allow for a thorough overview of the existing processes.

Consent

Consent constitutes a focal point for the ethical and legal assessment process, given its central role in research ethics. Indeed, consent is often seen as the “corner stone” of biobanking research,¹⁴ and is an essential element of all applicable laws, ethical norms, guidelines,¹³ and policies¹¹ pertaining to human biomedical research.

Therefore, a comparative analysis of the consent forms used by the different cohorts participating in a collaborative endeavor is the first step toward the evaluation of the ethical and legal availability of data for research. However, the content of consent forms varies from one study to another, and although most of the information found in the consent form has no implication on the availability of data for research, a few key elements are crucial in determining whether data can be used for specific secondary research or not.¹¹

Recontact and reconsent

The review of international and national norms demonstrates that the identification of processes for recontact and reconsent of participants is not only prescribed by norms or regulations,¹³ but also by provisions contained in the initial consent form.¹¹ Indeed, the consent form may (or may not) allow for recontact of participants for reconsent purposes. In all cases, the wishes of research participants not to be recontacted must be respected. Alternatively, the consent form may be silent on the issue.¹¹

In all cases, a Research Ethics Board (REB) must approve the recontacting and reconsenting processes. However, even when consented to, the recontact and reconsent raises major ethical issues, and could be either impossible or unfeasible.¹⁵

Waiver of consent

The processes related to the waiver of consent in the context of secondary research were identified by undertaking a legal analysis of the international and selected national legal frameworks (Australian, American, Canadian, British, and French laws) governing the retrospective secondary use of research data, and are presented as part of a separate publication.¹³ Although the possibility of obtaining a waiver of consent was found to be addressed in all jurisdictions examined, there is a high level of variation between the national and local requirements with respect to obtaining a waiver of consent.¹³

Comparing processes to identify similarities and differences and determining the limits of the “consistent whole”

After having identified existing processes, these need to be compared to identify similarities and differences between them. Removal of incompatible differences will then reveal the “consistent whole,” which can be defined as a coherent set of principles or procedures, allowing for the development of a common framework. This comparison entails: (A) comparing the elements of consent identified in the literature and norms to establish a list of common principles; (B) identifying situations where reconsent is required, and, finally; (C) comparing processes to determine the situations where waiving consent could be an option.

Consent

The consent form constitutes one of the main sources of information governing the right to use (or not) the participants' data in international research. Indeed, not only should the particularities of national legal and ethical frameworks be taken into consideration, but the specificities of each cohort's consent form must also be considered. However, as previously mentioned, the assessment of the ethical and legal “availability” of data by comparing the content of consent forms is a tremendous task, especially in the context of international research consortia. For example, the assessment of the 52 ENGAGE cohorts' consent forms identified 13 distinct ELSI-related themes, which were further divided into 41 ELSI-specific questions to be used to explore content of the study-specific consent forms. This analysis yields more than 3*10E666 potential and theoretical combinations of these various consent elements.¹⁰

In addition, only some of the already mentioned consent elements are of importance when considering the ethical and legal availability of data for research. Indeed, it is possible to identify a “consistent whole” composed of a subset of consent elements that must be interoperable to allow for the new use of retrospective data. The exercise of refining these criteria was undertaken in collaboration with members of the ELSI Stream of the BioSHaRE-EU project. Membership to the ELSI Stream is composed of nine international legal and ethics experts, from five European and North American countries.

In collaboration with these ELSI Stream members, a list of consent items to be considered when evaluating the interoperability potential of retrospective cohorts was developed. This list retained four items of major importance to determine whether consent, as already provided by research participants, allows for the proposed secondary use, namely:

• Nature and objectives of the study

• Genetic research

• Data sharing/linkage

• Commercial/industrial use

These items constitute the main components of consent, as a “consistent whole,” based on rounds of consultation with the group of experts in the domain which led to their selection, as compared with the other processes.

Recontact and reconsent

The many mechanisms used to allow an optimal use of data in research are addressed in a separate publication.¹¹ To develop that article, the consent forms from the 52 cohorts of the ENGAGE consortium were analyzed and compared, and results of this analysis demonstrated that cohorts have developed a wide range of mechanisms to facilitate the use of their data. This includes, for example, foreseeing participants' recontact and reconsent for further use of their data in research (although only a minority of cohorts typically include this option).¹¹

Given that these mechanisms are legally and ethically compliant with the international and national frameworks described previously, they could allow for the secondary use of data when the initial consent form used by a cohort does not permit that data it has collected be used for new proposed research initiatives. However, recontact and reconsent clauses are rather rare, and are present, for example, in only 36% of the ENGAGE cohorts' consent forms.¹¹

Waiver of consent

A comparative study analyzing the international and national (five countries) legal and ethical requirements for the use and secondary (retrospective) use of data for research purposes¹¹ demonstrated that “despite the international consensus on the requirement for a prior, free and informed consent from the data subjects for the research use of their data, all jurisdictions recognize that a waiver of consent could be ethically justified,”¹³ if the following criteria are met:

• Obtaining new consent must be impossible or impracticable.

• The waiver of consent must be granted by a REB or another authorized committee.

• Access to identifiable personal data must be essential for the research.

• If access to identifiable data is not essential for research, data must be either coded or deidentified.

• The proposed research use must correspond to an important public interest.

• The waiver of consent will not adversely affect the rights and welfare of the participants (balance of risks vs. benefits).

• The research design must involve no more than minimal risk.

• Privacy of research participants and confidentiality of data must be protected.

• The waiver must be consistent with international and domestic law.¹³

However, it is important to note that the last item of the list refers to domestic law, which can generate uncertainty regarding the possibility to implement waivers of consent, nationally or regionally.

Establishing common principles and procedures

The results of this analysis lead us to the next step, which is to establish the common principles and procedures stemming from the comparison of processes detailed under step II. The limits of the “recontact/reconsent” and “waiver of consent” processes are dictated by the international and national legal and ethical framework, as described previously, and are generally well defined. However, the limits of the “consent” process need to be refined to include only elements of relevance for the assessment of cohorts' ethical and legal “availability” in the context of a retrospective secondary use of data.

For instance, consent can be either broad, allowing for a wide range of research uses, or it can be specific to certain types of research uses, for example, limitations related to a certain type of disease.¹³ Although broad consent offers great potential for future unforeseen uses of the data, restricted consent often requires a deeper analysis of the conditions for the use of data to assess the potential availability of such data sets. This assessment requires a rigorous comparison of the content of the consent form as well as the specific research questions addressed by the potential data user. In the context of collaborative research, this process would have to be repeated for each proposed research question. However, this time-consuming task can be facilitated. For instance, continuing with the example of disease-type consent restrictions, the number of possible variations is limited to the number of existing diseases and conditions. For instance, the original consent form's proposed use could be matched with an internationally accepted list of diseases, such as the World Health's Organization's International Classification of Diseases (ICD).¹⁶ This would ensure that the proposed research uses and the original consent language are interoperable.

As a result of the analysis of common processes and procedures, a three-step assessment (for consent, reconsent, and waiver) was developed to assess the ethical and legal availability of data for unforeseen, secondary research uses. This filter is summarized in Figure 1.

FIG. 1.

Ethical and legal interoperability assessment filter. REB, Research Ethics Board.

Changing or removing processes that do not contribute to the “consistent whole”

Finally, the last step in producing an ethical and legal interoperability filter is to change or remove processes that do not contribute to the “consistent whole.” This requires streamlining the selected processes to ensure that they suit the concrete needs of actual cohorts and proposed data use projects. To test this final methodological step, the BioSHaRE-EU project was used as a case study.

More specifically, the ethical and legal interoperability assessment method, described in Figure 1, was used to evaluate the collaboration potential between nine BioSHaRE-EU cohorts wanting to participate in the healthy obese project (HOP). The HOP aims to characterize genetic and metabolic profiles of metabolically healthy obese individuals and to better understand genetic and metabolic differences between healthy versus metabolically obese individuals.

Consent

The consent forms and information brochures/pamphlets of the nine targeted cohorts were analyzed and compared to assess the ethical and legal interoperability of data for inclusion in the HOP (Table 1. Targeted cohorts include LifeGene, UK Biobank, LifeLines, HUNT 3, KORA 4, CHRIS Study, NCDS, Estonian Biobank, and MICROS). To do so, the following elements were examined for each cohort: (1) the nature and objectives of the study, (2) whether genetic research is allowed, (3) data sharing/linkage provisions, and (4) commercial and industrial uses. Given the objectives of the HOP, these elements were selected as determining factors in the inclusion, or not, of the BioSHaRE cohorts in the HOP.

Table 1.

Comparative Table of BioSHaRE-EU Consent Forms for Healthy Obese Project Study

	LifeGene	UK Biobank	LifeLines	HUNT 3	KORA 4	CHRIS Study	NCDS	Estonian Biobank	MICROS
Country	Sweden	United Kingdom	Netherlands	Norway	Ger-many	Italy	United Kingdom	Estonia	Italy
Nature and objectives of study—does the nature and objectives of the study allow for inclusion of collected data in the HOP?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Genetic research—is genetic research allowed?	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Data sharing/linkage—is data sharing/linkage allowed?	Yes	Yes	Yes	Not specified	Yes	Yes	Not specified	Yes	Yes
Commercial/industrial use—is commercial/industrial use allowed?	Yes	Yes	Yes	Yes, if reconsented	Yes	Not specified	No	Yes	Not specified

BioSHaRE-EU, Biobank Standardisation and Harmonisation for Research Excellence in the European Union; HOP, Healthy Obese Project.

From this analysis, it was noticed that the data held by LifeGene, UK Biobank, LifeLines, KORA, and the Estonian Biobank cohorts could be used for the HOP without further action required. However, in the case of the HUNT 3, NCDS, CHRIS, and MICROS studies, the documents analyzed did not confirm the ethical and legal availability of data, especially in the case where commercial/industrial partners are involved.

This being said, the HOP does not include commercial or industrial partners, which makes the last item “commercial/industrial use” nonrelevant to the evaluation of the inclusion of these cohorts' data in the HOP. Therefore, in the context of this HOP case study, it was concluded that the commercial use category could be removed from the list of criteria assessed, or replaced with another category that better reflects the specific considerations of the HOP project (e.g., the deposit of research data in a public or open database, mandatory return of results, etc.). Indeed, after removing the “commercial/industrial use” criteria, we notice that data from the CHRIS and MICROS cohorts can now be used in the HOP, but further information is still needed from HUNT 3 and NCDS.

Recontact and reconsent

Finally, to complete the assessment process, upon further analysis of both the HUNT 3 and NCDS consent documents, it was found that HUNT 3 participants could be recontacted to obtain a new consent for unforeseen research uses, if required. However, the documents available from the NCDS study did not address this issue. Therefore, although recontacting HUNT 3 participants could be envisaged for their inclusion in the HOP, it would be impossible to use the data held by NCDS, unless obtaining authorization from a competent REB.

In sum, this exercise effectively demonstrates that different elements described in our process will be relevant to the research uses proposed. The proposed assessment method provides an overall general filter applicable to a wide range of data-sharing scenarios. However, it can be adjusted, by adding or removing thematic categories, to reflect the particularities of the proposed initiatives, as illustrated by the HOP case study.

Discussion

This article provides a demonstration that retrospective assessment of the ethical and legal “availability” of data for secondary uses, and of cohorts' interoperability in research consortia, is possible. This assessment process takes into account key legal and ethical elements (consent, recontact, and waiver of consent) as summarized in Figure 1. The developed filter allows us to verify whether the legal and ethical requirements or restrictions pertaining a data set allow for data to be used for a proposed research use. In doing so, this method effectively assesses whether data sets are sufficiently ethically and legally interoperable to allow for a given collaborative research use.

In the context of research consortia, the identification of interoperable cohorts can help to further refine the data access process by targeting only the cohorts that are ethically and legally “available.” In doing so, the process also contributes to protecting research participants' rights and interests. Integrating this filter to the regular data access process, will not only enable researchers to create virtual “mega-cohorts” of research participants, but will also ensure that these cohorts respect basic legal and ethical precepts.

For instance, integration of this process using currently available data harmonization/access IT tools, such as the Maelstrom Research web-based platform,⁶ could even strengthen and accelerate the data access process by preidentifying eligible cohorts from a legal and ethical standpoint.

However, the limits of consent assessment processes are generally well documented. The main limit consists in the jurisdictional-specific requirements regarding consent, recontact/reconsent, and waiver of consent. Indeed, cultural norms and regulatory requirements may vary depending on the country, and, therefore, the method to assess cohort interoperability is limited by the potential divergence between the proposed use of the cohort's data sets and country-specific ethical and legal requirements.¹⁷ In addition, although the filter results from an analysis of general norms, specific regional and local contexts may call for additional ethical requirements. For instance, broad consent may not be acceptable in some jurisdictions, or research with data from deceased individuals or from minor or incompetent individuals may bring about additional considerations in relation to the impact of research on family members,¹² which are not taken into account by this general filter.

Although these limitations advocate for maintaining a decision-making body to govern access rights (e.g., data access and compliance office, REB, etc.), the proposed process may contribute to streamline the access process by setting up the key elements for a prior assessment of the availability of research resources.

The benefits of this framework could also extend beyond the identification of useable data sets. Indeed, this method may serve to determine collaboration models, upstream of networking and collaborative projects. For instance, data-sharing collaborative efforts can take on many forms, including local data sharing, centralized data sharing, or a federated model.¹⁸ Although researchers are free to adopt the model most adapted to their needs, a pre-emptive legal and ethical interoperability assessment may well influence which data-sharing model may present less ethical hurdles to implement.

Footnotes

Acknowledgments

The authors would like to thank the following organizations for their financial support: Canadian Institutes of Health Research FRN: R18293; European Commission for the ENGAGE (FP7/2007-2013, grant agreement HEALTH-F4-2007-201413); BioSHaRE-EU (FP7/2010-2015, grant agreement FP7-HEALTH 2010-261433) projects; Quebec Ministère de l'Économie, Innovation et Exportation; and Genome Quebec and Genome Canada. The authors would also like to thank the BioSHaRE-EU ELSI Stream members, Bartha Maria Knoppers (Canada), Jane Kaye (United Kingdom), Madeleine Murtagh (United Kingdom), Susan Wallace (United Kingdom), Anne Cambon-Thomsen (France), Emmanuelle Rial-Sebbag (France), Jasper Bovenberg (the Netherlands), Jennifer Harris (Norway), and Isabelle Budin-Ljøsne (Norway) for their support for the identification of the “consistent whole.”

Author Disclosure Statement

No conflicting financial interests exist.

References

Knoppers

, Zawati

, Kirby

. Sampling population of humans across the world: ELSI issues. Annu Rev Genomics Hum Genet, 2012; 13:395–413.

European Network of Genomic and Genetic Epidemiology (ENGAGE). Available from: www.euengage.org/ Accessed April 18, 2016.

Biobank Standardisation and Harmonisation for Research Excellence in the European Union (BioSHaRE-EU). Available from: www.BioSHaRE-EU/ Accessed April 18, 2016.

Zika

, Paci

, Schulte

, et al. Biobanks in Europe: Prospects for Harmonisation and Networking. JRC-IPTS Working Papers JRC57831. Institute for Prospective and Technological Studies, Joint Research Centre; 2010.

Wolfson

, Wallace

, Masca

, et al. DataSHIELD: Resolving a conflict in contemporary bioscience—performing a pooled analysis of individual-level data without sharing the data. Int J Epidemiol, 2010; 39:1372–1382.

Maelstrom Research. Available from: www.maelstrom-research.org/ Accessed April 18, 2016.

Fortier

, Burton

, Robson

, et al. Quality, quantity and harmony: The DataSHaPER approach to integrating data across bioclinical studies. Int J Epidemiol, 2010; 39:1383–1393.

Doiron

, Raina

, Ferretti

, et al. Facilitating collaborative research: Implementing a platform supporting data harmonization and pooling. Norw J Epidemiol, 2012; 21:221–224.

Budin-Ljøsne

, Isaeva

, Knoppers

, et al. Data sharing in large research consortia: Experiences and recommendations from ENGAGE. Eur J Hum Genet, 2014; 22:317–321.

10.

Tassé

. From ICH to IBH in biobanking? A legal perspective on harmonization, standardization and unification. Stud Ethics Law Technol, 2013; 7:1–15.

11.

Tassé

, Budin-Ljøsne

, Knoppers

et al.. Retrospective access to data: The ENGAGE consent experience. Eur J Hum Genet, 2010; 18:741–745.

12.

Tassé

. Biobanking and deceased persons. Hum Genet, 2011; 130:415–423.

13.

Tassé

. A comparative analysis of the legal and bioethical frameworks governing the secondary use of data for research purposes. (in press.) Biopreserv Biobank, 2016; 14.

14.

Hoeyer

. The ethics of research biobanking: A critical review of the literature. Biotechnol Genet Eng, 2008; 25:429.

15.

Steinsbekk

. Biobanks—when is re-consent necessary?. Public Health Ethics, 2011; 4:236–250.

16.

World Health Organization.. International Classification of Diseases (ICD). ICD-10. 2015. Available from: http://apps.who.int/classifications/icd10/browse/2015/en Accessed April 18, 2016.

17.

Wallace

, Knoppers

. Harmonised consent in international research consortia: An impossible dream?. Genomics Soc Policy, 2011; 7:35–46.

18.

Maelstrom Research. Environment for Co-Analysis. Available from: www.maelstrom-research.org/about-harmonization/environment-for-co-analysis/ Accessed April 18, 2016.