Opt-in and opt-out consent procedures for the reuse of routinely recorded health data in scientific research and their consequences for consent rate and consent bias– A systematic review

doi:10.21203/rs.3.rs-1434893/v1

Download PDF

Research Article

Opt-in and opt-out consent procedures for the reuse of routinely recorded health data in scientific research and their consequences for consent rate and consent bias– A systematic review

https://doi.org/10.21203/rs.3.rs-1434893/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 24 Aug, 2022

Read the published version in Journal of Medical Internet Research →

Version 1

posted

You are reading this latest preprint version

Background

Scientific researchers who wish to reuse health data pertaining to individuals can obtain consent through an opt-in procedure or opt-out procedure. The choice of procedure may have consequences for the consent rate and representativeness of the study sample, and the quality of the research, but these consequences are not well known. Therefore, this review aims to provide insight into the consequences for the consent rate and consent bias of the study sample of opt-in procedures versus opt-out procedures for the reuse of routinely recorded health data for scientific research purposes.

Method

A systematic review was performed, based on searches in PubMed, Embase, Cinahl, PsycInfo, Web of Science Core Collection, and the Cochrane Library. Two reviewers independently included studies based on predefined eligibility criteria and assessed whether the statistical methods used in the reviewed literature were appropriate for describing the differences between consenters and non-consenters. Statistical pooling and a narrative synthesis were conducted to report results.

Results

Fifteen studies were included. Of these, 13 implemented an opt-in procedure, one implemented an opt-out procedure, and one implemented both. The average weighted consent rates were 84.0% among the studies that used opt-in and 96.8% in the single study that used opt-out. In the single study that described both procedures, the consent rate was 21.0% in the opt-in group and 95.6% in the opt-out group. Opt-in procedures resulted in more consent bias compared to opt-out procedures. In studies with opt-in, consenting individuals were more likely to be men, had a higher level of education, a higher income, and a higher socioeconomic status.

Conclusion

Consent rates are generally lower when using an opt-in procedure compared to using an opt-out procedure. Furthermore, in studies with opt-in, participants are less representative of the study population. However, both the study populations and the way in which opt-in or opt-out was organized, varied widely between the studies, which makes it difficult to draw general conclusions regarding the desired balance between patient control over data and learning from patient data. Such further use may be hampered by administrative burdens and the risk of bias.

Real-world data

secondary data use

electronic health records

routine health data

consent procedure

opt-in

opt-out

consent rate

representativeness

consent bias

Routinely recorded health data of individuals, e.g., in electronic health records, are increasingly being reused in scientific research. A healthcare system in which the use and reuse of health data is an integral part of the healthcare delivery process is an important aspect of what is called a ‘learning health system.’ Enabling these data to be used for research can be considered to be a prerequisite for a learning health system [1, 2]. The development of learning health systems is important [3, 4], as the secondary use of routinely recorded health data in scientific research allows us to learn more about individuals’ healthcare use and the quality of care without increasing the administrative burden for healthcare providers. Such research can shape policy and practice, resulting in greater quality, accuracy, accessibility, and sustainability of health care systems [2, 5].

The increasing importance of reusing routine health data must, however, be balanced against a patient’s right to data protection [6, 7]. In European Union countries, the General Data Protection Regulation (GDPR) provides the general legal framework for collecting and processing personal health data. The GDPR generally prohibits the processing of sensitive data, except, for instance, in cases in which the data subject has granted consent. Consent in the sense of the GDPR must meet a range of requirements: it must be specific, freely given, informed, and unambiguous[8]. This means that the individual has voluntarily given an express statement of consent or a clear affirmative action that leaves no room for interpretation and that consent is in the context of specific purpose(s). This is also referred to as informed consent [9]. In short, researchers need individuals to actively opt in before the researchers can access personal health data and reuse them in research [8].

There is an option in the GDPR for an exemption from the requirement to obtain informed consent when data are used for research purposes. This research exemption provided for in the GDPR has been implemented differently in the various European Union member states [10]. In the Netherlands, for instance, the legislation states that consent is not required if asking for consent would make it impossible or unfeasible to perform the research. These conditions have recently been refined in a Code of Conduct [11]. The Code of Conduct serves as an important normative framework and expresses the current consensus about the relevant European and Dutch legislation. It states that consent (acquired by the healthcare provider) is in principle necessary to release pseudonymized patient data for research. This consent is not required if: a) requesting consent from an individual is not reasonably possible or would impose too great a burden to the patient, for instance, because the patient is terminally ill; b) the request for permission cannot reasonably be expected from the healthcare provider, for instance, due to the large size of the study group; or c) if asking for permission would lead to a low and/or selective response or participation rate that cannot be corrected for (selection or consent bias). This can threaten the representativeness of the study sample [12, 13] and ultimately render the research unreliable and therefore impossible. If the exemption from having to obtain informed consent applies, further conditions must be met, such as that the research must be in the general interest. Also, researchers must ensure that adequate technical and organizational safeguards, e.g., pseudonymization, are in place when using data pertaining to patients [10]. All of these legal restrictions can be regarded as one of several possible sources of bias in using routine electronic health records for research purposes [10, 14, 15].

As mentioned, under the GDPR, researchers need to obtain study-specific informed consent. Still, national legislation can allow a broad consent procedure. Broad consent is a form of consent in which one consent for multiple potential future research projects in a certain scope is obtained [16]. Specific consent, rather than a one-off broad consent, may increase the risk of low and biased participation rates [9]. Broad consent can thus be a viable alternative for projects intending to use health data for multiple studies.

Generally speaking, in healthcare research, asking for informed consent through an opt-in methodology is more likely to lead to low study samples and consequently consent bias, as a selective, non-random group might opt in. For instance, this would be the case if healthier individuals are more likely to consent. Thus, in healthcare research, an opt-out procedure might be a good alternative. In an opt-out procedure, an individual is presumed to consent if they do not actively refuse the reuse of health data. Opt-out procedures minimize the efforts in approaching patients and reduce the administrative burden for healthcare providers. Also, if an opt-out procedure is used and provided that patients are well informed, patients still have considerable control over the data. They can explicitly request that the data should not be reused for research purposes, which would not be the case if there was no consent procedure at all. An opt-out procedure can also be a good alternative if researchers expect a low and selective response when asking for consent. But, even in opt-out procedures, a selective, non-random group of individuals might actively object to the reuse of the data, resulting in consent bias. In sum, both opt-in and opt-out procedures have their drawbacks for research, as consent rates and consent bias can become problematic [17–20]. However, the extent to which low consent rates and consent bias occur and whether opt-in and opt-out procedures differ in these respects is unknown. Also, the consequences of a study-specific consent or a broad consent on consent rates and consent bias are unknown.

Even though asking for informed consent in itself might lead to invalid study results, in research, there is no consensus on the specific situations in which the exemption for obtaining consent, as described in the Code of Conduct, applies. Therefore, assessing this is not straightforward, and considerations and alternatives must be carefully weighed in each situation. An example of such a specific situation where the exemption for informed consent might apply is in studies in which the involvement of legal representatives plays a major role. The involvement of such a representative can be the case for certain groups of incompetent patients, such as patients with advanced dementia and patients with serious mental health problems. A legal representative might be more apprehensive about providing opt-in consent on the individual’s behalf, which can lead to low study samples. However, science lacks clear insight into the effect of legal representative consent on consent rates and consent bias and whether this differs between opt-in and opt-out procedures.

Naturally, both opt-in and opt-out procedures require individuals to be adequately informed about the research, data protection, and data governance [21]. Research has shown that the extent to which patients understand the substance of the consent request determines how likely they are to provide consent [22, 23]. This indicates that the method of providing information and obtaining consent could affect consent rates. For example, there may be substantial differences in consent rates depending on whether the request for consent is made face-to-face or via post or e-mail. The effect of that method on consent rates may differ between opt-in and opt-out procedures.

Therefore, our review paper aims to provide more insight into the consequences of opt-in versus opt-out procedures for the secondary use of routinely recorded health data of individuals for scientific research purposes. We were particularly interested in the consequences of opt-in versus opt-out for (a) the consent rate and (b) the representativeness of the individuals who gave consent for the study population.

The specific review questions addressed are:

1. What are the consequences of opt-in versus opt-out consent procedures for the reuse of health data in scientific research, regarding (a) the consent rate and (b) the extent to which the sample recruited is representative of the study population?

And within opt-in and opt-out procedures, to what extent:

2. Are the consequences for (a) the consent rate and (b) representativeness dependent on whether the consent was study-specific or broad?

3. Are the consequences for (a) the consent rate and (b) representativeness dependent on whether the individual in question or a legal representative provided consent?

4. Are the consequences for (a) the consent rate and (b) representativeness dependent on the method of informing individuals and obtaining consent?

Design

This paper describes the results of a systematic review. The reporting was guided by the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [24].

Search strategy and study selection

The search was performed in August 2021. The databases searched were PubMed, Embase, Cinahl, PsycInfo, Web of Science Core Collection, and the Cochrane Library. The databases were searched with a pre-determined strategy as detailed in Appendix A.

Two reviewers independently selected the studies that satisfied the eligibility criteria (see below). The first selection was made based on the title and abstract (YdM & AF), and in a second step, the full texts were checked (YdM & YJ/MOV). Discrepancies in reviewer selections were resolved by discussion and/or consulting another reviewer/co-author. The selection process is described using a PRISMA flow diagram [24]; see Figure 1.

Eligibility criteria

Inclusion criteria

Studies were included in the review if they met the following eligibility criteria:

Regarding types of studies:

All types of empirical studies were eligible for inclusion, including grey literature.
Full texts had to be available.

Regarding types of research subjects:

Persons of any age who are directed in consent procedures for the reuse of individual, routinely recorded health data. These persons can be patients, citizens in general, or legal representatives of the persons to whom the health data pertain.

Regarding research topics:

Must concern:

The reuse for scientific research of routinely recorded health data pertaining to individuals AND
Health data routinely recorded in relation to good clinical practice involving individuals AND
The involvement of individuals or (legal) representatives in the consent procedure, as opposed to only the healthcare provider providing consent.

Must mention:

The type of consent procedure: opt-in, opt-out, or a combination of opt-in and opt-out.

Regarding variables:

Consent rate AND/OR
Representativeness, in relation to the study population, of the study participants who give consent (whether through an opt-in or opt-out procedure) for secondary use of health data.

Exclusion criteria

Studies were excluded if:

The opt-in or opt-out procedures for secondary use of health data followed earlier consent for study participation, for instance, for participation in a survey or intervention study.
The opt-in or opt-out procedure for secondary use of health data was an integral part of broader consent for study participation, for instance, for participation in a survey or intervention study.
Consent was asked in a hypothetical setting; e.g., the participants were asked in which situations they would be willing to provide consent for the reuse of health data containing information on their treatment.

Editorials, essays, literature reviews, and other non-empirical studies were also excluded. However, reference lists of relevant literature reviews were checked to identify potentially relevant empirical research. The number of excluded studies (including reasons for exclusion for studies excluded following review of the full text) was recorded at each selection stage in the Prisma Flow Diagram.

There were no restrictions regarding publication date or language.

Data extraction and synthesis

For each included study, the information relevant to our review questions was extracted using a pre-structured format. This information included: authors and country, research questions for which the routinely recorded health data were used, the setting, type of consent procedure (opt-in, opt-out, or a combination of both), whether the consent was study-specific or broad, the method of obtaining consent, the person who provided consent, and the characteristics of consenters and non-consenters.

To answer the review questions regarding consent rates, we also extracted the numbers of individuals approached and numbers who gave their consent. We pooled the extracted data regarding consent rates, stratified by consent procedure (opt-in or opt-out) and the outcome of interest (broad vs. specific consent, representative vs. individual consent, and different methods of informing individuals). In the case of missing data regarding the numbers of individuals approached, individuals who responded, consenters or non-consenters, where possible, we deduced the numbers and percentages based on the numbers that were provided by the study concerned. For instance, for opt-in procedures, if the numbers of individuals who were approached and who gave consent were known, we could calculate the non-consenters by subtracting the consenters from the individuals approached.

To answer the review questions regarding the representativeness of the study sample, we extracted all data describing statistical differences in characteristics between the consenters and the non-consenters or differences between the consenters and the total study population. In other words, we extracted information on the consent bias that arose from the type of consent procedure used. We were particularly interested in comparisons for age, sex, ethnicity, education, income, socioeconomic status (SES), and health status. A narrative synthesis to describe these differences was conducted.

Assessment of methodological quality

Two researchers (YdM and BT) independently assessed the statistical quality of the included studies providing data for our review questions regarding the representativeness of the study population (questions 1b, 2b, 3b, and 4b). The statistical quality of the descriptive data on consent rates (for questions 1a, 2a, 3a, and 4a) was not assessed, as descriptive statistics are straightforward and require no further statistical analyses. The quality was assessed based on one question of the Joanna Briggs Institute (JBI) checklist for cross-sectional studies [25]. The JBI checklist comprises eight questions to assess the methodological quality of a research paper. However, the first seven questions concern issues of confounding, measurements of conditions, exposures, and outcomes, and are therefore not applicable to the data extracted for our review questions. We thus assessed the methodological quality using Question 8 of the JBI checklist: “Was appropriate statistical analysis used?” The answer was scored as ‘yes,’ ‘no,’ or ‘unclear.’ A ‘yes’ was scored if a) the study tested for differences between either consenters and non-consenters or consenters/non-consenters and the study population, and b) if appropriate statistical methods were used in the reviewed literature for testing the differences between consenters and non-consenters or consenters and the study population concerning age, sex, ethnicity, education, income, socioeconomic status (SES) or health status. We judged studies to be of ‘good quality’ if both reviewers assigned a ‘yes.’ Any discrepancies in the scoring were resolved via a consensus meeting. If ultimately a study was not assigned a ‘yes’ by both reviewers, the extracted data regarding consent bias were excluded from the results.

Results of the selection procedure

The searches identified 2,367 study references. After removing duplicates, 1,344 studies remained. After the first selection, performed by reading titles and abstracts, 94 studies were selected for full-text screening. Eleven of these 94 studies were deemed eligible for inclusion in the review. An additional 17 potentially relevant studies were identified from the reference lists of the 11 included studies and from relevant literature reviews that were excluded in the screening phase. These 17 studies were also assessed by two reviewers who read the full texts. Four of these 17 were found eligible for inclusion in our review. This resulted in a total of 15 included studies. See Fig. 1 for the PRISMA diagram describing the study screening.

Characteristics of the literature

Countries

Studies were performed in the United States (N = 5), the UK (N = 4), the Netherlands (N = 3), Canada, New Zealand, and Australia (all N = 1) (see Table 1).

Main aim/scope

Ten of the 15 studies had a primary focus on consent rates or consent bias [26, 27, 29, 30, 32–34, 36, 37, 39]. The five remaining studies were not primarily interested in consent rates or consent bias but did report consent rates as a side issue [28, 31, 35, 38, 40].

Study populations

The study populations of the studies reviewed varied considerably. Although most studies focused on adults, four of the 15 studies concerned health data from individuals younger than 18 [26, 33, 35, 38].

Below, the results regarding the review questions are described separately for parts a (consent rates) and b (consent bias), starting with questions 1a, 2a, 3a, and 4a and then continuing with 1b, 2b, 3b, and 4b.

Table 1

Characteristics of included studies
Author(s) year & country	General aim/scope as described in the study	Purpose for which routine health data were obtained and analyzed	Study population	Sources and types of routine health data
Berry 2012, AUS [26]	To determine which approach (opt-in or opt-out) for gaining parental consent for linkage of vaccination data with hospital data provides the highest consent rate for a program of childhood vaccine safety surveillance.	To examine adverse events following immunization.	Children of mothers aged 18 years and over, who resided in South Australia, and had given birth at the Women’s and Children’s Hospital between July 27, 2009, and October 25, 2009.	Sources: Vaccination and hospital medical records. Type: Vaccinations and hospital visits.
Jacobsen 1999, USA [27]	To analyze the influence on consent rate of changes to Minnesota statutes for use of medical records for research.	To achieve study aim.	Sample of patients who received hospital medical care during the three years before January 1, 1997, aged 20+, living in the US.	Source: Hospital medical records; inpatient and outpatient. Type: Demographics, diagnoses, and care utilization.
Barnes 2005, UK [28]	1) To examine use, effectiveness, and tolerability of the drug montelukast for treating asthma; 2) To explore prognostic factors that could predict a favorable response to the drug.	To achieve study aim.	Patients (with physician-diagnosed asthma) who had been prescribed montelukast at any time between February 1998 and June 2000.	Source: Hospital and family-doctor medical records. Type: Demographics and diagnoses.
Damery 2011, UK [29]	To establish the level of consent bias that may occur should individual patient consent be sought.	To improve the understanding of the reasons for anemia.	Adult patients, ≥ 40 years of age, who received a hematological or clinical diagnosis of iron deficiency anemia between 2001 and 2006.	Source: family-doctor medical records. Type: Demographics and diagnoses.
Elwood 2019, NZ [30]	To give a measure of consent bias resulting from consent for inclusion.	To achieve study aim.	All patients in New Zealand, diagnosed with a first primary invasive breast cancer or DCIS from 2000 to 2012.	Source: Hospital medical records. Type: Demographics, diagnoses, treatment, and mortality data.
Evenhuis 2004, NL [31]	To identify unanticipated obstacles for population-based epidemiological research on visual and hearing impairment and reasons for non-participation.	To determine the cause and degree of IDs, and assess the reliability of reported audiometric and ophthalmological data.	Adults with IDs (26% of whom were diagnosed with Down syndrome) making use of ID care services.	Source: Hospital medical records Type: Demographic and ophthalmological and audiometric data.
Jackson 2008, UK [32]	To assess potential consent biases arising from an opt-in procedure.	To achieve study aim.	All patients with a stroke or TIA admitted to the hospital or seen in outpatient clinics from October 2002 to March 2004.	Source: Hospital medical records (inpatient and outpatient). Type: Demographics, data on process of care, and clinical variables.
Knoester 2005, NL [33]	To evaluate whether the consent procedure induces consent bias.	To conduct a retrospective chart review study of the effectiveness of lamotrigine.	Patients with a first lamotrigine prescription between August 1, 1997, and December 31, 2000, aged 18+.	Source: Pharmacy medical charts. Type: Medication.
Kramer 2017, CA [34]	1) To determine success in obtaining consent from parents to allow review of their child’s personal health information for emergency research screening; 2) to examine the variables associated with successful consent.	To assess future research eligibility.	Parents/guardians of children under 18 years of age who presented to the study institution’s ED from July 27, 2015 to January 24, 2016, between the hours of 08:00 to 23:00.	Source: Hospital ED medical records. Type: Demographics, diagnoses, and treatments.
Marrie 2007, USA [35]	1) To compare self-reported diagnoses of MS to physician-reported diagnoses; 2) to compare physician-reported diagnoses with diagnoses based on expert review of medical records; 3) to establish markers that will identify registry participants who have a high probability of not having MS.	To achieve study aim.	Patients with MS included in the North American Research Committee on Multiple Sclerosis Registry.	Source: Hospital medical records. Type: Demographics and treatment.
McCarthy 1999, USA [36]	To determine the effects of state legislation requiring the patient’s informed consent prior to medical record abstraction by external researchers for a study improving the potential safety of pain medication.	To conduct a pharmaco-epidemiologic study of seizures associated with the use of pain medication.	Users and non-users of oral analgesics enrolled in the Minnesota independent practice association health plan between November 1997 and April 1998.	Source: Health-plan administrative claims consisting of longitudinal pharmaceutical, medical, and enrollment files; and hospital medical records. Type: -
Nijhof 2017, NL [37]	To test potential differences between youths who refuse permission and youths who permit the use of their clinical data.	To use data for benchmarking and scientific and policy research.	Youths from six Dutch secure residential care institutions.	Source: Routine outcome measurements from Dutch secure residential youth care. Type: Demographics, diagnoses, and treatments.
Noble 2009, UK [38]	To evaluate the effectiveness and cost of obtaining consent for a review of medical records in a study.	To achieve study aim.	Men aged 50–69 with incident prostate cancer; invited between 2001 and 2008 for prostate-specific antigen testing.	Source: family-doctor medical records. Type: Demographics and mortality data.
Yawn 1998, USA [39]	To gather information on the number and characteristics of patients who refused authorization.	To achieve study aim.	All patients seen at the hospital for their first visit, during January or February 1997.	Source: Hospital medical records (outpatient, emergency department. and inpatient). Type: Demographics and diagnoses.
Zell 2000, USA [40]	To describe the sample, the survey design, and the data collection procedures for the National Immunization Survey.	To compare self-reported vaccinations with provider records.	Children 19–35 months of age included in the National Immunization Survey.	Sources: Vaccination and hospital medical records. Type: Vaccinations and hospital visits.
-=not reported; ~=approximately; DCIS = ductal carcinoma in situ; ED = emergency department; ID = intellectual disability; MS = multiple sclerosis; TIA = transient ischemic attack

Question 1a: What Are The Consequences For Consent Rates Of Opt-in Versus Opt-out Consent Procedures?

All 15 studies reported on the type of consent procedure (opt-in or opt-out) and consent rates (Table 2).

One study comparing an opt-in procedure with an opt-out procedure

Berry et al. investigated whether consent rates differ depending on whether an opt-in or opt-out procedure is used for the same research goal. It was the only study providing results of both procedures [26]. In this study, in which health data regarding children’s vaccinations and hospital visits were linked, 564 mothers were randomly assigned to the opt-in group and 565 to the opt-out group. The sociodemographic characteristics were comparable in the two groups. The consent rate was 21% in the opt-in group and 95.6% in the opt-out group. The differences in consent rates were statistically significant [26].

One study describing an opt-out procedure

The participants in this study by Jacobsen et al. were actively asked for their consent for a review of hospital medical records but, according to the Minnesota law at the time of the study, non-responders were classified as passive consenters [27]. Therefore, we classified this study as a study using an opt-out procedure and reported the data accordingly. In total, 2,463 individuals were approached, of whom 3.2% actively refused, resulting in a consent rate of 96.8% [27].

Thirteen studies describing an opt-in procedure

In 13 studies, participants provided active consent to health data reuse [28–40]. In all of these studies, non-responders were classified as non-consenters. Consent rates ranged from 18.6–99%. Overall, of the 72,418 individuals approached in the 13 studies, 84.0% actively gave their consent for researchers to reuse the routinely recorded health data for research purposes.

To address the review questions 2a, 3a, and 4a, we have pooled the results of the 13 studies with an opt-in procedure [28–40], together with the results of the opt-in group from the study by Berry et al. in which opt-in and opt-out consent rates are compared [26]. We also grouped the results of the one study with an opt-out procedure [27] with the results of the opt-out group from the study by Berry et al. [26].

Question 2a: What are the consequences for consent rates of whether the consent procedure is study-specific or broad?

Table 2 describes for each study whether the consent was study-specific or broad.

Studies with opt-out

In the study by Berry et al., the consent was study-specific [26], while in the study by Jacobsen et al., the procedure was a broad opt-out for the authorization to use medical records for research purposes in general [27]. The two studies had similar consent rates (95.6% [26] versus 96.8% [27]).

Studies with opt-in

Four opt-in studies obtained broad consent. A total of 28.685 individuals were approached and 25.850 consented to the reuse of their health data (90.1%) [30, 32, 37, 39]. The average weighted consent rate was higher than the average weighted consent rate of the ten opt-in studies acquiring a study-specific informed consent, in which a total of 44.297 individuals were approached and 35.070 individuals consented (79.2%) [26, 28, 29, 31, 33–36, 38, 40].

Question 3a: What are the consequences for consent rates of whether the individual in question or their legal representative provided consent?

All 15 studies reported who provided consent and consent rates (Table 2).

Studies with opt-out

In the study by Berry et al., parents had to opt out to refuse consent for the children’s vaccination data to be linked [26], while in the study by Jacobsen et al., the participants themselves had to opt out of the health data reuse [27]. However, the two studies had similar consent rates (95.6% [26] versus 96.8% [27]).

Studies with opt-in

In the four opt-in studies in which a representative provided the consent, 32.074 of the 39.120 approached individuals consented, resulting in a weighted average consent rate of 82.0%) [26, 31, 34, 40]. For the ten studies in which the participant themselves provided consent, 28.846 of the 33.862 approached individuals consented to the reuse of their health data, resulting in an average weighted consent rate of 85.2% [28–30, 32, 33, 35–39].

Question 4a: What are the consequences for consent rates of the method of informing and obtaining consent?

All 15 studies reported on the method of informing and obtaining consent and on consent rates (Table 2).

Studies with opt-out

In the two studies with an opt-out procedure, the study information and an explanation of how to refuse study participation were provided by post [26, 27]. The two studies had similar consent rates (95.6% [26] versus 96.8% [27]). The authors of the study of Jackson et al. sent reminders [27]. They did this because individuals were asked to either grant or refuse consent and return the form in both cases. Sending reminders increased the refusal rate from 2.7–3.2% [27].

Studies with opt-in: In the 14 studies with opt-in procedures, four different methods of informing and asking consent were used, namely: by post [26, 28, 29, 31, 33, 35, 36, 38]; verbally, mainly at the start of treatment or during a patient’s first hospital visit or stay [30, 34, 40]; in writing, as part of the registration or intake procedure [37, 39]; or by a combination of postal and verbal information (providing an information leaflet and consent form and asking for verbal consent, or if more time was needed, the form could be returned via post) [32]. For the studies in which opt-in consent was requested verbally, 46.736 individuals were approached and 39.973 consented, resulting in an average weighted consent rate of 85.5%. When consent was obtained as part of the intake procedure, 15.121 of the 16.738 approached individuals consented. The average weighted consent rate was 90.3%. Obtaining consent via post resulted in the lowest weighted average consent rate, namely 56.5% (8.447 approached and 4.776 consented). However, in the studies in which reminders were sent [29, 31, 35, 36, 38], the average weighted consent rate increased to 75.7% compared to 52.9% in the studies in which no reminders were sent [26, 28, 33].

Table 2

Consent and non-consent rates by consent procedure
	Study	Content of consent	Response and consent rates				Broad or specific consent	Method and setting for obtaining consent	Sent reminders	Consent provided by
Consent procedure			Approached, (N)	Response, N (%)	Consent, N (% of approached)	Non-consent, N (% of approached)
Opt-in and opt-out	Berry 2012 [26]	To include their medical record data in the vaccination registry (data-linkage).	Opt-in: 564	Opt-in: 120 (21.3%)	Opt-in: 120 (21.3%)	Opt-in: 444 (78.7%)	Specific	Via post, signed by a pediatrician asking parents to consent/refuse via a reply form, telephone, or e-mail.	-	Parents of the infants
Opt-in and opt-out	Berry 2012 [26]		Opt-out: 565	Opt-out: 25 (4.4%)	Opt-out: 540 (95.6%)	Opt-out: 25 (4.4%)				Parents of the infants
Opt-out	Jacobsen 1999 [27]	To review medical records for research purposes.	2463	2023 (82.1%)	~ 2384 − 2380^* (96.8%)^†	~ 53–79^* (3.2%)^†	Broad	Via post with post-paid return envelope.	Reminders sent after 4, 6, and 12 weeks.	Participant
Opt-in	Barnes 2005 [28]	To review medical records for research purposes.	~ 2500	1429 (57.2%^*)	1400 (56.0%)	1071 (42.8%)^*	Specific	Via post.	-	Participant
	Damery 2011 [29]	To review medical records for research purposes.	592	425 (71.8%)	371 (62.7%)	221 (37.3%)^*	Specific	Via post with an enclosed reply slip.	Reminder sent after 2 weeks.	Participant
	Elwood 2019 [30]	To include their medical record data in the registry (data-linkage).	Invasive cancer: 9244 DCIS: 1642	Invasive cancer: 9244 (100%) DCIS: 1642 (100%)	Invasive cancer: 8282 (89.6%) DCIS: 1397 (85.1%)	Invasive cancer: 962 (10.4%) DCIS: 245 (14.9%)^‡	Broad	Verbally by a clinician during the patients' first hospital visit.	-	Participant
	Evenhuis 2004 [31]	To review and screen medical records	2706	2656 (98.2%)	1660 (61.3%^*)	1046 (38.7%)^*	Specific	Via post through contact persons of intellectual disabilities services (usually a physician and a speech and hearing therapist or medical secretary)	Reminders sent after one month.	Legal representative. Clients who were able to communicate verbally were asked for written or verbal consent.
	Jackson 2008 [32]	To review medical records for research purposes.	1061	1061 (100%)	1050 (99.0%)	11 (1.0%)	Broad	Inpatients: Verbally after an information leaflet with a consent form was provided Outpatients: Via post or during consultations after an information leaflet with a consent form was provided	-	Participant (94%), or relative in case of patients with incapacity (6%)
	Knoester 2005 [33]	To review medical records for research purposes.	1636	1069 (65.3%)	968 (59.2%)	668 (40.8%)	Specific	Via a recruitment letter through community pharmacists.	-	Participant
	Kramer 2017 [34]	To review medical records for research purposes.	2506	2506 (100%)^*	1852 (73.9%)	654 (26.1%)^*	Specific	Verbal consent was obtained during a visit to the emergency department by a volunteer delegate.	-	Parents/guardians
	Marrie 2007 [35]	To review medical records for research purposes.	109	81 (74.3%)	52 (47.7%)	57 (52.3%)^*	Specific	Via post.	Telephone call.	Participant
	McCarthy 1999 [36]	To review medical records for research purposes.	140	73 (52.1%)	26 (18.6%)	114 (81.4%)^*	Specific	Via a letter from the health plan's medical director.	Second mailing after six weeks and a telephone call after an additional month	Participant
	Nijhof 2017 [37]	To use their medical records for research purposes and benchmarking.	949	887 (93.5%)	628 (66.2%)	316 (33.3%) (and 5 unaccounted for by the researchers) ^*	Broad	Via a questionnaire at the start of their treatment.	-	Participant. However, for youths under the age of 16, the parent/legal guardian also had to consent to the data use.
	Noble 2009 [38]	To review medical records for research purposes.	193	184 (95.3%)	179 (92.7%)	14 (7.3%)^*	Specific	Via consent packages sent out by family doctor practices.	Reminder pack sent after 3 weeks.	Participant
	Yawn 1998 [39]	To review medical records for research purposes.	15789	15069^* (95.4%)	14493 (91.8%)	1296 (8.2%)^*	Broad	Via a written consent form, as part of the hospital registration procedure with a receptionist or hospital registration clerk.	-	Participant. However, if a patient had died, or was aged 16 or younger, a legal representative was asked to sign.
	Zell 2000 [40]	To contact providers who have administered vaccinations, who would then provide medical records for research purposes.	33344 children	33305 children (99.9%)	28442 children (85.3%)	4902 (14.7%)^*	Specific	Verbal consent was obtained at the end of a telephone interview regarding vaccination questions.	-	Parents/ guardians
- = not reported; ~=approximately; DCIS = ductal carcinoma in situ; ED = Emergency department
* Self-calculated values based on the numbers provided by the study.
^† This study only reported weighted percentages. The number of participants actively consenting and refusing are therefore estimates and do not add up to the total who responded.
^‡ Patients who did not give consent may have declined it, or may not have been offered the relevant information and consent forms; the data does not distinguish between these categories.

Question 1b: What are the consequences for consent bias of opt-in versus opt-out consent procedures?

Eight of the 15 studies using an opt-in and/or opt-out procedure reported information on the representativeness of the study sample [26, 27, 29–31, 33, 34, 37].

Results of the methodological appraisal

all eight studies clearly described the study methodology. The study by Evenhuis et al. only provided descriptive statistics when comparing the consenters in their study to the base population [31]. No statistical testing regarding consent bias was performed. Therefore, this study was not scored positively with regard to the statistical quality by the reviewers. The other seven studies used various statistical analyses to assess the differences between consenters and non-consenters, such as chi-square tests and logistic regression and reported results including P-values, relative risks (RR), odds ratios (OR), or hazard ratios (HR). The overall consensus of the reviewers was that the statistical analyses regarding the assessment of potential consent bias were appropriate for these seven studies.

Results regarding representativeness: The seven studies reporting on representativeness and of good statistical quality are: the study by Jacobsen et al. in which an opt-out procedure is described [27], the study by Berry et al. in which both an opt-in procedure and an opt-out procedure are described [26], and five studies in which an opt-in procedure is described [29, 30, 33, 34, 37]. These seven studies reported varying consent rates, ranging from 21.3–97.8%, and reported comparisons between consenters versus non-consenters. The studies reported age, sex, ethnicity, education, income, socioeconomic status (SES), and health status comparisons. Even though Berry et al. compared opt-in and opt-out parental consent rates for childhood vaccine safety surveillance using data linkage and tested for significance, they did not test whether the two procedures (opt-in and opt-out) differed significantly with regard to consent bias. Therefore, we could only report differences between consenters and non-consenters for the opt-in group and the opt-out group separately. Below, we describe the comparisons for the six opt-in procedures [26, 29, 30, 33, 34, 37] and the two opt-out procedures [26, 27] separately. Table 3 shows the extracted data regarding these outcomes.

Studies with opt-out

Jacobsen et al. and Berry et al. reported characteristics of consenters and non-consenters by age [26, 27]. Jacobsen et al. found that non-consenters were more likely to be older than consenters [27]. Berry et al. found no age-related differences between consenters and non-consenters [26].

Both studies also reported the characteristics of consenters versus non-consenters by sex. Berry et al. found that fewer men gave their consent (RR: 0.9 CI:0.81–0.99) [26]. Jacobsen et al. reported no significant differences in sex [27].

Only Berry et al. compared consenters with non-consenters by educational level, SES, and income. They found no significant differences related to these variables [26].

Jacobsen et al. compared consenters with non-consenters by health status; they found no significant differences for this variable [27]. Neither study reported information regarding ethnicity [26, 27].

Studies with opt-in

All six studies with an opt-in procedure reported analyses by age. In the study by Berry et al., consenters were older [26]. In two other studies (Damery 2011 and Elwood 2019), consenters were found to be younger, or non-consenters were found to be older [29, 30]. Knoester et al. found no differences in age [33], and Kramer et al. and Nijhof et al. similarly reported no significant differences between consenters and non-consenters by age [34, 37].

Four of the studies with opt-in, that recruited both men and women, reported on sex [26, 29, 33, 37]. In the study by Damery et al., women were less likely to provide consent than men [29]. The other three studies reported no differences in women’s odds of consenting compared to men [26, 33, 37].

Elwood et al. and Nijhof et al. reported analyses of ethnicity [30, 37]. In the New Zealand clinical breast cancer registry study by Elwood et al., the non-consenting proportion was similar in Maori women (9.8%) but significantly higher in Pacific women (14.4%) compared to European New Zealand women (9.9%) [30]. In the Dutch youth residential study by Nijhof et al., Caucasian youths were more likely to consent to health data reuse [37].

Berry et al. and Nijhof et al. reported differences by educational level [26, 37]. Both found that consenters had a higher level of education than non-consenters.

Berry et al. and Knoester et al. compared consenters and non-consenters by income [26, 33]. Berry et al. showed that consenters were more likely to be in the highest annual household income bracket than non-consenters (RR 2.04 CI: 1.08–3.87) [26]. Knoester et al. found no significant differences in income [33].

Berry et al. and Damery et al. reported data on socioeconomic status (SES) [26, 29]. Damery et al. reported that patients living in more deprived areas were significantly more likely not to give their consent for access to medical records than patients living in more affluent areas [29]. Berry et al. found no significant differences over the socioeconomic quintiles for consenters versus non-consenters [26].

Of the four studies with opt-in that reported on health status [29, 30, 33, 37], Damery et al. showed that colorectal cancer status was not significantly associated with consent in patients with anemia [29]. Nijhof et al. showed that consenting youths in residential care institutions were more likely to have a longer treatment duration [37]. However, in this study, consent was obtained during intake and assessed retrospectively. In the two remaining studies (Elwood, 2019 and Knoester, 2005), poorer health status was reported in non-consenters [30, 33]. Elwood et al. showed that, compared to consenters, non-consenters had a poorer prognosis and were more likely to have a metastatic disease [30]. In addition, non-consenting patients less frequently chemotherapy, radiotherapy, hormonal therapy, or biological therapy (all p < 0.001), in part because more non-consenting patients declined these treatments. In the second study, non-consenters were more likely to have a chronic disease score (CDS) above 6 [33] compared to consenters. CDS is a measure of chronic disease status derived from population-based automated pharmacy data. A higher CDS is associated with poorer health status [41]. Furthermore, previous use of two or more antiepileptic drugs was more likely for non-consenters than for consenters. Finally, the use of antidepressants and anti-migraine drugs was significantly associated with non-consent.

Questions 2b, 3b, and 4b: What are the consequences for consent bias of whether the consent was study-specific or broad, whether the representative or the individual in question provided consent, and the methodology for obtaining consent?

There were no associations reported between the degree of consent bias and whether a study-specific or broad consent was obtained (review question 2b), or whether the individual to whom the data were related or the legal representative gave the consent (review question 3b). Also, there were no associations reported in the reviewed literature between the degree of consent bias and the method of informing people about the reuse of health data and of obtaining consent (review question 4b).

Table 3

Representativeness of the study samples regarding age, ethnicity, education, income, socioeconomic status, and health status/comorbidity.
Consent procedure	Study	Age	Sex	Ethnicity/migration background	Education	Income	SES	Health status/comorbidity
Opt-in and opt-out	Berry 2012 [26]	Opt-in: Consenters are older Ref:18–24 years 25–29: RR = 1.81 (0.77–4.25) P-value: 0.174 30-34: RR = 3.02 (1.38–6.61) P-value: 0.006 35-39: RR = 3.36 (1.52–7.44) P-value: 0.003 40+: RR = 4.16 (1.77–9.81) P-value: 0.001	Opt-in: NS	-	Opt-in: Consenters are more highly educated Ref: up to year 10 (~ 16 years old) Up to year 12: RR = 1.68 (0.64–4.36) P-value: 0.291 Trade or certificate: RR = 1.90 (0.79–4.58) P-value: 0.155 University or higher: RR = 3.37 (1.46–7.81) P-value: 0.005	Opt-in: Consenters have higher income Ref: <$20.800 $20 800 − 41 599: RR = 0.87 (0.40–1.85) P-value: 0.708 $41600–83199: RR = 1.49 (0.78–2.82) P-value: 0.226 >$83.200: RR = 2.04 (1.08–3.87) P-value: 0.029	Opt-in: NS	-
Opt-in and opt-out	Berry 2012 [26]	Opt-out: NS	Opt-out: Fewer men participate RR = 0.9 (CI:0.81–0.99) P-value:0.035	-	Opt-out: NS	Opt-out: NS	Opt-out: NS	-
Opt-out	Jacobsen 1999 [27]	Non-consenters are older (60+): P-value: <0.001	NS	-	-	-	-	NS
	Damery 2011 [29]	Non-consenters are older Ref: 40–60 years old. 60+: OR = 2.84 (2.01–4.02) P-value: <0.0001	Non-consenters more likely to be women OR = 1.62 (1.13–2.34) P-value:0.008	-	-	-	Non-consenters more likely to be deprived Ref: affluent. Deprived: OR = 1.61 (1.15–2.26) P-value = 0.005	Colorectal cancer status NS
	Elwood 2019 [30]	Consenters are younger: P-value=-	-	Varied over the strata Adjusted for age, Compared to NZ European women (9.9% non-consented) the non-consenting proportion was similar in Maori women (9.8%) but significantly higher in Pacific women (14.4%). Consented; non-consented - (% non-consented and 95%CI): NZ European: 5498; 633 (9.9%; CI: 9.2–10.7) Maori: 629; 47 (9.8%; CI: 6.2–13.4) Pacific: 616; 90 (14.4%; CI: 11.1–17.6) Asian: 708; 97 (12.9%; CI: 9.9–15.9) Other: 331; 39 (11.0%; CI: 7.4–14.7)	-	-	-	Non-consenting patients had a poorer prognosis, and 13% had metastatic disease compared with 3% of consenting patients (P < 0.0001). Non-consenting patients more often had no primary surgery (26.2% compared with 4.0% of consenting patients) and less frequently had chemotherapy (P < 0.0001), radiotherapy (P < 0.0001), hormonal therapy (P = 0.0004), or biological therapy ( p < 0.0001), in part because more non-consenting patients declined these treatments.
	Knoester 2005 [33]	NS	NS	-	-	NS	-	Non-consenters have higher chronic disease score (CDS): Ref: CDS 0-2 CDS > 6: HR = 1.24 (CI: 1.01–1.53) More previous use of antiepileptic drugs (AEDs) Ref: 1 AED 2 + AEDs: HR = 1.33 (CI: 1.12–1.57) More co-medication Ref: absence Antidepressants: HR = 2.01 (CI: 1.64–2.63) Antimigraine drugs: HR: 1.74 (CI: 1.16–2.61)
	Kramer 2017 [34]	Child’s age: NS	-	-	-	-	-	-
	Nijhof 2017 [37]	NS	NS	Consenters more likely to be Caucasian: P-value: <0.05 Consent vs refuse: Non-Western foreign: 27.1% vs 35.9% Western foreign: 8.1% vs 7.7% Caucasian: 48.4% vs 33.6% Unknown: 16.4% vs 22.8%	Consenters are more highly educated: P-value: <0.01 Consent vs refuse: (Special) education: 5.7% vs 5.4% HAVO: 2.4% vs 1.9% MBO: 7.8% vs 6.9% Practical education: 5.9% vs 6.9% Special higher education: 22.0% vs 10.8% VMBO: 37.4% vs 20.5% VWO: 0.8% vs 0.4% Unknown: 18.0% vs 47.1%	-	-	Consenters have a longer treatment duration: Consent vs refuse 215 vs 140 days P-value: <0.01
- = not reported; AUD = Australian dollar; HAVO = school of higher general secondary education; ID = intellectual disability; MBO = senior secondary vocational education; NS = not significant; VMBO = preparatory secondary vocational education; VWO = pre-university education;

This review paper addresses the consequences of opt-in and opt-out procedures for the consent rate and the representativeness of the study sample when reusing routinely recorded health data in scientific research. We reviewed 13 studies using an opt-in procedure, one study using an opt-out procedure, and one study using both opt-in and opt-out. Overall, consent rates were lower in the 14 studies using an opt-in procedure compared to the two studies using an opt-out procedure. In addition, participants in studies with an opt-in procedure were more likely not to be representative of the study population compared to studies with an opt-out procedure. In three studies with opt-in, no statistically significant differences were reported between consenters and non-consenters regarding age, sex, or income. However, in the studies that did report significant differences, an opt-in procedure resulted in a study sample with a higher proportion of men, a higher level of education, a higher income, and socioeconomic status, and better health status than the study population, thereby introducing a risk of biased and invalid study results. No statistically significant differences were reported in the two studies with an opt-out procedure for education, income, and health status. And while significant differences in age and sex were found between consenters and non-consenters in one opt-out study, no statistically significant differences were reported in the other.

This review paper also aimed to examine whether differences in consent rates and representativeness for studies with opt-in versus studies with opt-out were affected by: 1) the implementation of a study-specific versus broad consent procedure; 2) different methods for informing individuals and obtaining consent; 3) whether the individual to whom the routinely recorded health data pertained or a legal representative provided consent for the reuse of the data in scientific research.

First, our results indicate that in opt-in studies, consent rates were higher in studies that implemented a broad consent procedure compared to studies that obtained study-specific consent. Approaching individuals repeatedly for consent, as would be the case when a study-specific consent is required, has the disadvantage that it can potentially result in ‘consent fatigue’. Even though more research is needed regarding this phenomenon in the context of reusing health data for scientific research, consent fatigue may lead to participants taking less time to understand the consent request [42], endangering the concept of informed consent. Obtaining consent repeatedly can also potentially lower an individual’s willingness to provide consent, leading to low study samples and an increased risk of consent bias. Therefore, broad consent, would be a good alternative to study-specific consent, provided it would be allowed by national legislation.

Second, our review indicates that the method for informing individuals and asking for their consent affects consent rates for opt-in procedures. Participants can be informed through different methods, such as by post or e-mail, during their visit to the healthcare provider, or as part of the registration or intake procedure. For studies using an opt-in procedure, informing individuals via post results in the lowest response and consent rates. Nevertheless, informing participants by post or e-mail has a considerable potential advantage in terms of a low workload for healthcare professionals, as it can be done by secretariats. Also, consent rates can be further increased by sending reminders by post or e-mail. However, even sending reminders does not increase the consent rate to the rate seen for opt-in procedures in which participants are asked verbally for their consent. Moreover, obtaining consent verbally in opt-in procedures still does not deliver the consent rates seen in the studies with opt-out procedures in which individuals were recruited by post. We argue that this could be because each element of extra effort required from participants to provide consent in an opt-in or opt-out procedure could be expected to reduce consent rates and thereby increase consent bias. This holds especially for opt-in procedures, as this procedure requires the participant to take action to give their consent, e.g., to return a consent form by post to the researcher.

So far, the research literature has mainly explored the effects of specific methods for asking individuals to participate in a survey or RCT, while the literature on methods to increase consent rates in opt-in or opt-out procedures to reuse routinely recorded health data is limited. Edwards et al., for instance, report that frequent contact between the individuals requesting consent and the research subjects increases response rates in survey research [43]. Even though the main aim of the studies in our review was not to describe the effect of specific methods for obtaining consent (e.g., by post or verbally) on response rates, we argue that a high response rate is a prerequisite for achieving a high consent rate. Therefore, this finding is in line with our review finding that consent rates in opt-in procedures in which participants are recruited by post could be increased by sending reminders. On the other hand, Caldwell et al. found in their systematic review that how, when, and by whom information was provided when recruiting patients for an RCT was not associated with response rates. Instead, the content of the information was more important: the individuals’ awareness of the scope of the research and the potential health benefits does increase consent rates [44].

Third, in studies with an opt-in procedure, consent rates were slightly lower when a legal representative was requested to provide consent rather than the individual in question. In contrast, no differences in consent rates were found for the two studies using an opt-out procedure. This indicates that opt-in studies involving a legal representative might be more prone to low study samples. Based on the two opt-out studies, an opt-out procedure could be a good alternative; however, more research is needed to confirm this conclusion. In our review, legal representatives provided consent in the case of children’s health data or health data pertaining to patients with intellectual disabilities. However, a representative will also be involved in some other groups. In groups such as persons with advanced dementia, frail nursing home residents, and people with serious mental or intellectual disorders [23], individuals will often be unable to make informed choices [45]. So far, little is known about the decision-making process of legal representatives in opt-in or opt-out procedures for the reuse of routinely recorded health data. An explanation for the lower consent rates of legal representatives can be that they are more afraid or cautious regarding potential risks — such as violation of privacy — when deciding whether to give consent involves someone else, especially when this person is vulnerable. Thus, specific attention needs to be paid to groups in which legal representatives, rather than the individuals themselves, are involved in opt-in or opt-out procedures. In the reviewed studies of Berry et al., Evenhuis et al., Kramer et al., and Zell et al., we have seen already that when a legal representative is asked for their consent, this can lead to lower consent rates [26, 31, 34, 40]. This could indicate that an opt-out procedure might be more suitable for these specific populations to prevent low study samples and consent bias.

Interestingly, we only found studies in which the legal representatives (mainly parents) represented children. No eligible studies were found for other groups where a legal representative would be involved, such as the elderly with Alzheimer’s disease or other vulnerable patient groups. Also, the results for the opt-out studies are based on two studies only and must thus be interpreted with caution. Therefore, we recommend future research targeting various patient groups with legal representatives and the effect on consent rates and consent bias in both opt-in and opt-out procedures.

Overall, the results of our systematic review contribute to the debate on individuals’ control of the health data that pertains to them [46, 47]. Laws such as the General Data Protection Regulation (GDPR) in European Union member states, which regulate the processing of sensitive data pertaining to individuals, including routine health data, aim to protect individuals’ privacy and reduce the risk of misuse of the data [10]. The GDPR states that explicit consent must be obtained from individuals, implying that an opt-in procedure is required when reusing routinely recorded health data for research. But, as mentioned in our introduction, based on national law, exemptions apply when, for instance, the representativeness of the study sample is at stake [12, 13]. The results of our systematic review indicate that for scientific research, based on routinely recorded health data, opt-in consent procedures are prone to low study samples and consent bias, therefore reducing the representativeness of the study sample. Specifically, persons with low-income, low-education level, and/or poor health status, who could benefit most from a learning health system, are less represented in studies with opt-in procedures. This could indicate that opt-out procedures might be more appropriate for research with routinely recorded health data to increase consent rates and reduce the risk of consent bias.

A question that remains is at what point the efforts needed to obtain a sufficient consent rate and to create a representative study sample are proportionate given 1) the privacy risks that reusing the data impose on participants; 2) the benefits that data reuse has for society. In the introduction, we mentioned that the conditions under which the implementation of an opt-out procedure applies, are not straightforward. A better understanding of how the benefits, the aforementioned efforts, and the risks can be weighed against one another can help further define the conditions under which an opt-out should be allowed.

Strengths, limitations, and recommendations for further research

To our knowledge, this is the first systematic review assessing the consequences of opt-in procedures versus opt-out procedures for consent rates and consent bias in research where health data routinely recorded by healthcare professionals are reused. Another strength of our review is that we used a broad search strategy in six literature databases and performed a statistical quality assessment of the included studies.

However, regarding our results, four limitations should be kept in mind. First, only two studies were found that implemented an opt-out procedure to obtain authorization for the reuse of routinely recorded health data. Therefore, our results must be interpreted with caution, and more research is needed to further validate the results. An explanation for the low number of studies with an opt-out procedure in our review is that overall, consent rates and consent bias might not be seen as big issues in opt-out procedures. Therefore, research exploring the consequences of opt-out procedures on consent rate and consent bias when reusing routinely recorded health data might be limited. Another reason might be that opt-out procedures are not commonly used when a study reuses routinely recorded health data, as it appears to be far from straightforward to meet the legal conditions for an opt-out procedure. The second limitation is that only one study directly compared an opt-in procedure and opt-out procedure, using an RCT design, and tested for significant differences in consent rates. Extending this methodology is recommended for further research, as this could provide greater insight into whether an opt-in procedure or opt-out procedure is preferred concerning the consent rate and representativeness. Third, the weighted consent rates we calculated for research questions 2 through 4 are unadjusted estimates. Further research accounting for interactions and confounding is necessary to strengthen these indications. Fourth, 11 of the 15 reviewed studies reused data recorded in a hospital setting. We suggest further empirical research to explore whether the results found in our review will be similar in other settings. Also, almost all the studies concerned patients diagnosed with one specific condition. Only four studies looked at a more general population. Patients with a specific diagnosis will probably be more engaged with a study topic related to their illness. This is likely to increase their tendency to consent to the reuse of health data. Whether this is the case cannot be determined from our results; we, therefore, recommend further studies focusing on consent procedures in more general health settings to assess the effect of this on consent rates and consent bias.

Our results indicate that consent rates are lower in studies using opt-in procedures for the secondary use of routinely recorded health data in scientific research compared to studies using opt-out procedures. Also, in studies using opt-in procedures, consenters are more likely not to be representative of the study population. Specifically, consenters are more likely to be male, with a higher level of education, a higher income and socioeconomic status, and better health status. This implies that people with poor health or other vulnerable conditions will be underrepresented in studies using opt-in procedures. In addition, our review indicates that implementing study-specific consent as opposed to broad consent can decrease consent rates in opt-in studies. Also, the method of informing individuals and asking for their consent affects consent rates. In opt-in procedures, a more direct approach and sending reminders lead to higher consent rates. In studies using an opt-in procedure, consent rates were slightly lower when a legal representative was asked to provide consent rather than the individual in question. At the same time, no difference was found for studies using an opt-out procedure. However, both the study populations and the way in which opt-in or opt-out was organized, varied widely between the studies, which makes it difficult to draw general conclusions regarding the desired balance between patient control over data and learning from patient data. Such further use may be hampered by administrative burdens and the risk of bias. More research is needed comparing opt-in and opt-out procedures and more knowledge regarding the consequences (benefits and efforts) of opt-in versus opt-out procedures so that researchers can weigh up all the consequences when they want to reuse routinely recorded health data for scientific research.

CDS

Chronic disease score

GDPR

General Data Protection Regulation

Hazard Ratio

JBI

Joanna Briggs Institute

Odds Ratio

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analysis

Relative Risk

SES

Socio economic Status

Ethics approval and consent to participate: Not applicable.

Consent for publication: Not applicable.

Availability of data and materials: Not applicable.

Competing interests: The authors declare that they have no competing interests.

Funding: Not applicable.

Authors' contributions: AF, YJ, and RV conceptualized and designed this review. YdM, AF, MOV, and YJ were responsible for screening the literature. Data extraction was done by YdM and YJ and checked by KdW, BT, EBvV, and MOV. YdM and BT conducted the methodological quality assessment. EBvV was responsible for the legal aspects of this paper. AF supervised and guided the project throughout the entire process. YdM and YJ were responsible for writing the initial draft of the manuscript. MOV, AF, RV, KdW, BT, and EBvV reviewed and edited the paper in the writing process. All authors approved the final paper.

Acknowledgments: The authors would like to thank Librarian Linda Schoonmade for her contribution to the search strategy and Clare Wilkinson of Tessera Translation Services for the final edit.

Olsen L, Aisner D, McGinnis JM: Institute of Medicine. The learning healthcare system: workshop summary. Washington DC: The National Academies Press; 2007.
Friedman CP, Wong AK, Blumenthal D: Achieving a nationwide learning health system. Science translational medicine. 2010;2(57):57cm29-57cm29.
Verheij R, Witvliet C, Jansen T, Hooiveld M, Hilten OV: Hergebruik van routine zorgdata voor beleid en wetenschap: hoe het nog beter kan. 2019.
Buntin MB, Burke MF, Hoaglin MC, Blumenthal D: The benefits of health information technology: a review of the recent literature shows predominantly positive results. Health affairs. 2011;30(3):464–471.
Chaudhry B, Wang J, Wu S, Maglione M, Mojica W, Roth E, Morton SC, Shekelle PG: Systematic review: impact of health information technology on quality, efficiency, and costs of medical care. Annals of internal medicine. 2006;144(10):742–752.
van Veen EB: Observational health research in Europe: understanding the General Data Protection Regulation and underlying debate. European Journal of Cancer. 2018;104:70–80.
Taylor MJ, Whitton T: Public interest, health research and data protection law: Establishing a legitimate trade-off between individual control and research access to health data. Laws. 2020;9(1):6.
Voigt P, Von dem Bussche A. The EU general data protection regulation (GDPR). A Practical Guide, 1st Ed, Cham: Springer International Publishing. 2017;10:3152676.
European Data Protection Board. Guidelines 05/2020 on Consent Under Regulation 2016/679. EDPB Brussels; 2020.
Hansen J, Wilson P, Verhoeven E, Kroneman M, Kirwan M, Verheij R, et al. Assessment of the EU Member States’ rules on health data in the light of GDPR. 2021.
Coreon. Gedragscode gezondheidszorg onderzoek. Verantwoord omgaan met (persoons)gegevens en lichaamsmateriaal bij gezondheidsonderzoek 2022.
Dutch Medical Treatment Contracts Act (WGBO), Stat. artikel 7:446, lid 1 BW Afdeling 5.
Kamerstukken II. 1993–1994, 21 561, nr. 20, p. 3.
Verheij RA, Curcin V, Delaney BC, McGilchrist MM. Possible sources of bias in primary care electronic health record data use and reuse. Journal of medical Internet research. 2018;20(5):e9134.
Kuchinke W, Ohmann C, Verheij RA, van Veen E-B, Arvanitis TN, Taweel A, et al. A standardised graphic method for describing data privacy frameworks in primary care research using a flexible zone model. International journal of medical informatics. 2014;83(12):941–57.
Maloy JW, Bass PF. Understanding broad consent. Ochsner Journal. 2020;20(1):81–6.
Young AF, Dobson AJ, Byles JE. Health services research using linked records: who consents and what is the gain? Australian and New Zealand journal of public health. 2001;25(5):417–20.
Woolf SH, Rothemich SF, Johnson RE, Marsland DW. Selection bias from requiring patients to give consent to examine data for health services research. Archives of family medicine. 2000;9(10):1111.
Kho ME, Duffett M, Willison DJ, Cook DJ, Brouwers MC. Written informed consent and selection bias in observational studies using medical records: systematic review. Bmj. 2009;338.
Emam KE, Jonker E, Moher E, Arbuckle L. A review of evidence on consent bias in research. The American Journal of Bioethics. 2013;13(4):42–4.
Lensink MA, Boers SN, Jongsma KR, Bredenoord AL. Understanding (in) consent for governance. The American Journal of Bioethics. 2019;19(5):43–5.
Jäckle A, Beninger K, Burton J, Couper MP. Understanding data linkage consent in longitudinal surveys. Advances in Longitudinal Survey Methodology. 2021:122 – 50.
Sugarman J, McCrory DC, Hubal RC. Getting meaningful informed consent from older adults: a structured literature review of empirical research. Journal of the American Geriatrics Society. 1998;46(4):517–24.
Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJM, Gavaghan DJ, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Controlled clinical trials. 1996;17(1):1–12.
Joanna Briggs Institute. Critical appraisal checklist for analytical cross sectional studies. Adelaide: The Joanna Briggs Institute Critical Appraisal Tools for Use in JBI Systematic Reviews. 2017.
Berry JG, Ryan P, Gold MS, Braunack-Mayer AJ, Duszynski KM, Group VAULDW. A randomised controlled trial to compare opt-in and opt-out parental consent for childhood vaccine safety surveillance using data linkage. Journal of medical ethics. 2012;38(10):619–25.
Jacobsen SJ, Xia Z, Campion ME, Darby CH, Plevak MF, Seltman KD, et al. Potential effect of authorization bias on medical record research. Mayo Clinic Proceedings; 1999: Elsevier.
Barnes N, Thomas M, Price D, Tate H. The national montelukast survey. Journal of allergy and clinical immunology. 2005;115(1):47–54.
Damery S, Ryan R, McManus R, Warmington S, Draper H, Wilson S, et al. The effect of seeking consent on the representativeness of patient cohorts: iron-deficiency anaemia and colorectal cancer. Colorectal Disease. 2011;13(11):e366-e73.
Elwood JM, Marshall RJ, Tin ST, Barrios ME, Harvey VJ. Bias in survival estimates created by a requirement for consent to enter a clinical breast cancer registry. Cancer epidemiology. 2019;58:178–83.
Evenhuis H, Van Splunder J, Vink M, Weerdenburg C, Van Zanten B, Stilma J. Obstacles in large-scale epidemiological assessment of sensory impairments in a Dutch population with intellectual disabilities. Journal of Intellectual Disability Research. 2004;48(8):708–18.
Jackson C, Crossland L, Dennis M, Wardlaw J, Sudlow C. Assessing the impact of the requirement for explicit consent in a hospital-based stroke study. QJM: An International Journal of Medicine. 2008;101(4):281–9.
Knoester P, Belitser S, Deckers C, Keyser A, Renier W, Egberts A, et al. Recruitment of a cohort of lamotrigine users through community pharmacists: differences between patients who gave informed consent and those who did not. Pharmacoepidemiology and drug safety. 2005;14(2):107–12.
Kramer AI, Stephenson E, Betel A, Crudden J, Boutis K. Success of University Student Volunteers in Obtaining Consent for Reviewing Private Health Information for Emergency Research. Accountability in research. 2017;24(6):329–43.
Marrie R, Cutter G, Tyry T, Campagnolo D, Vollmer T. Validation of the NARCOMS registry: diagnosis. Multiple Sclerosis Journal. 2007;13(6):770–5.
McCarthy DB, Shatin D, Drinkard CR, Kleinman JH, Gardner JS. Medical records and privacy: empirical effects of legislation. Health services research. 1999;34(1 Pt 2):417.
Nijhof KS, van Domburgh L, Strijbosch E, Labun A, Delsing M, Scholte RH, et al. Who wants to participate in routine outcome monitoring? Differences between youth in secured residential care who consent or refuse the use of their clinical data for research and policy making. Residential Treatment for Children & Youth. 2017;34(3–4):244–65.
Noble S, Donovan J, Turner E, Metcalfe C, Lane A, Rowlands M-A, et al. Feasibility and cost of obtaining informed consent for essential review of medical records in large-scale health services research. Journal of Health Services Research & Policy. 2009;14(2):77–81.
Yawn BP, Yawn RA, Geier GR, Xia Z, Jacobsen SJ. The impact of requiring patient authorization for use of data in medical records research. Journal of Family Practice. 1998;47:361–5.
Zell ER, Ezzati-Rice TM, Battaglia MP, Wright RA. National Immunization Survey: the methodology of a vaccination surveillance system. Public health reports. 2000;115(1):65.
Von Korff M, Wagner EH, Saunders K. A chronic disease score from automated pharmacy data. Journal of clinical epidemiology. 1992;45(2):197–203.
Schermer BW, Custers B, Van der Hof S. The crisis of consent: How stronger legal protection may lead to weaker consent in data protection. Ethics and Information Technology. 2014;16(2):171–82.
Edwards PJ, Roberts I, Clarke MJ, DiGuiseppi C, Wentz R, Kwan I, et al. Methods to increase response to postal and electronic questionnaires. Cochrane database of systematic reviews. 2009(3).
Caldwell PH, Hamilton S, Tan A, Craig JC. Strategies for increasing recruitment to randomised controlled trials: systematic review. PLoS medicine. 2010;7(11):e1000368.
Manti S, Licari A. How to obtain informed consent for research. Breathe. 2018;14(2):145–52.
Ploug T. In Defence of informed consent for health record research-why arguments from ‘easy rescue’,‘no harm’and ‘consent bias’ fail. BMC Medical Ethics. 2020;21(1):1–13.
Jones KH, Laurie G, Stevens L, Dobbs C, Ford DV, Lea N. The other side of the coin: Harm due to the non-use of health-related data. International journal of medical informatics. 2017;97:43–51.

Appendix A is not available with this version.

No competing interests reported.

Download PDF

Journal Publication

published 24 Aug, 2022

Read the published version in Journal of Medical Internet Research →

Version 1

posted

You are reading this latest preprint version

Opt-in and opt-out consent procedures for the reuse of routinely recorded health data in scientific research and their consequences for consent rate and consent bias– A systematic review

Status:

Journal Publication

Version 1

Abstract

Background

Method

Results

Conclusion

Figures

Introduction

Methods

Results

Results of the selection procedure

Characteristics of the literature

Question 1a: What Are The Consequences For Consent Rates Of Opt-in Versus Opt-out Consent Procedures?

Question 1b: What are the consequences for consent bias of opt-in versus opt-out consent procedures?

Discussion

Strengths, limitations, and recommendations for further research

Conclusions

List Of Abbreviations

Declarations

References

Appendix

Additional Declarations

Status:

Journal Publication

Version 1