Patient consent preferences on sharing personal health information during the COVID-19 pandemic: “the more informed we are, the more likely we are to help”

Rapid ethical access to personal health information (PHI) to support research is extremely important during pandemics, yet little is known regarding patient preferences for consent during such crises. This follow-up study sought to ascertain whether there were differences in consent preferences between pre-pandemic times compared to during Wave 1 of the COVID-19 global pandemic, and to better understand the reasons behind these preferences. A total of 183 patients in the pandemic cohort completed the survey via email, and responses were compared to the distinct pre-pandemic cohort (n = 222); all were patients of a large Canadian cancer center. The survey covered (a) broad versus study-specific consent; (b) opt-in versus opt-out contact approach; (c) levels of comfort sharing with different recipients; (d) perceptions of commercialization; and (e) options to track use of information and be notified of results. Four focus groups (n = 12) were subsequently conducted to elucidate reasons motivating dominant preferences. Patients in the pandemic cohort were significantly more comfortable with sharing all information and biological samples (90% vs. 79%, p = 0.009), sharing information with the health care institution (97% vs. 83%, p < 0.001), sharing information with researchers at other hospitals (85% vs. 70%, p < 0.001), sharing PHI provincially (69% vs. 53%, p < 0.002), nationally (65% vs. 53%, p = 0.022) and internationally (48% vs. 39%, p = 0.024) compared to the pre-pandemic cohort. Discomfort with sharing information with commercial companies remained unchanged between the two cohorts (50% vs. 51% uncomfortable, p = 0.58). Significantly more pandemic cohort patients expressed a wish to track use of PHI (75% vs. 61%, p = 0.007), and to be notified of results (83% vs. 70%, p = 0.012). Thematic analysis uncovered that transparency was strongly desired on outside PHI use, particularly when commercialization was involved. In pandemic times, patients were more comfortable sharing information with all parties, except with commercial entities, where levels of discomfort (~ 50%) remained unchanged. Focus groups identified that the ability to track and receive results of studies using one’s PHI is an important way to reduce discomfort and increase trust. These findings meaningfully inform wider discussions on the use of personal health information for research during global crises.


Introduction
Rapid access to patient health information (PHI) by medical researchers during pandemics is of paramount importance to advance the clinical understanding of novel pathogens and enable the swift development of lifesaving treatments and vaccines. Health-related organizations must be able to develop patient consent policies and processes that are sufficiently nimble to facilitate timely access while adhering to ethical standards defined by governing bodies (e.g. Canada's Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans, TCPS2, 2018 [1]; European Union's General Data Protection Regulation, GDPR, 2018 [2]; United States' National Institute of Health Data Sharing Policy, 2020 [3]), as well as institutional Research Ethics Boards (REBs). While REBs may be able to provide exemptions from requiring consent for some studies during pandemics and other exceptional circumstances (TCPS2, 2018; Article 3.7A) [1], investigating the consent needs and preferences of patients during a pandemic is pivotal to ensuring transparent patient-centered policies that maximize rapid research whilst meeting patient needs.
Pandemics pose unique challenges to traditional informed consent processes, which are ordinarily conducted by in-person, face-to-face conversations with clinicians or research coordinators, with direct opportunities to ask questions and sign physical consent forms. While pandemic policies and approaches will differ across international contexts, in our Canadian context, clinical activities were reduced and entry into hospitals was restricted to protect patients and health care workers. Unfortunately, this also impeded opportunities to engage in the dialogue necessary for valid informed consent. Strict pandemic visitor policies limiting who might be considered a patient's essential care partner also had the potential to reduce a patient's level of comfort with consenting to treatments or research studies. Solutions proposed and enacted to overcome these challenges included utilizing digital tools, documenting verbal consent, using electronic informed consent, and the employment of digital HIPAA-compliant tools such as e-mail surveys or telehealth assessments [4]. Leveraging digital platforms for telehealth and consent processes may have presented additional benefits to research by enabling the inclusion of a larger number of eligible patients who might otherwise be excluded due to geographic, mobility, or language barriers. However, there is little understanding of how these changes may affect the process of informed consent, and whether patients' overall comfort with participation and technology-enabled consent options was increased or decreased under conditions where the vulnerability associated with the experience of critical illness was coupled with pandemic psychosocial stressors.
There is a paucity of literature on patient preferences for research participation during a pandemic. In fact, Gobat et al. (2015 [5]; 2017 [6]; 2019 [7]) have repeatedly highlighted that while patients are primary contributors and beneficiaries of pandemic-related clinical research, systematic studies regarding their views on research participation during a pandemic is lacking. Their preliminary explorations of mostly European patients indicated that while 74.8% thought "special rules" should apply to pandemic-relevant research, most (58.4%) still preferred standard enrollment procedures, such as prospective written informed consent, with only 38.6% opining that simplified procedures would be acceptable. It was also observed that the patients' level of trust in both health professionals and the government was predictive of their willingness to participate in pandemic-relevant research [7]. Furthermore, while use of routinely collected data and clinical samples for pandemic-relevant research without explicit prior consent was supported in principle, it was found less acceptable when a profit motive was perceived [6]. Therefore, it is clear that further empirical research in this area is warranted.
In a prior study by our group [8], we sought to acquire insight to the contemporary and specific consent needs of cancer patients at a large Canadian academic hospital to inform institutional consent policies. In that study (conducted in radiation oncology clinics prior to the COVID-19 pandemic), the majority of the 222 participants (83%) were willing to share PHI with researchers at our own institution, though many preferred a more transparent and reciprocal consent process [8]. Prior to the COVID-19 pandemic, the majority of surveyed patients (63%) desired to be asked for permission before being entered into a research contact pool; 38% preferred study-specific consent (i.e., be given information on each study and decide each time vs. 56% preferring one-time to track and receive results of studies using one's PHI is an important way to reduce discomfort and increase trust. These findings meaningfully inform wider discussions on the use of personal health information for research during global crises. Keywords: Consent during pandemics, COVID-19, Patient consent preferences, Data sharing during pandemics, Research ethics broad consent for all studies; 6% would not share at all), and approximately half of patients were uncomfortable sharing PHI with commercial enterprises [8]. Most patients desired the option to track PHI usage (61%), with the highest proportion reported by the youngest group of patients (≤ 49 years: 71%); and the majority wished to be notified regarding study results (70%). The level of comfort in sharing PHI was highest for those within one's own health organization (83%), followed by researchers at other academic institutions (70%), and followed by not-for-profit companies such as health charities (57%). By a substantial margin, the lowest level of comfort was reported for sharing with for-profit commercial companies (27%). To build on both these findings and the aforementioned preliminary explorations of consent preferences during pandemics, the objectives of the present study were to: (a) ascertain the preferences of patients during the COVID-19 global pandemic to uncover potential differences between pre-pandemic and pandemic cohorts, and (b) acquire a deeper understanding of the reasons underlying these consent preferences.

Methods
This study employed a mixed methods design with both quantitative (i.e. survey), and qualitative (i.e. focus group) strands, which were intricately linked in that the focus group guide was designed to further explore key findings from our pre-pandemic and pandemic cohort survey responses (e.g. who should decide if health information is shared, contact pool permissions, tracking future use, and sharing for commercial purposes).

Surveys
A study information letter and link to a 10-min online survey (Additional file 1: Appendix 1) were digitally sent to patients in Canada's Princess Margaret (PM) Cancer Centre's Virtual Care database who had consented to email contact (~ 1000 patients in total) during Wave 1 of the COVID-19 pandemic. This survey design was based on existing evaluations of patient consent preferences across Canada and internationally (e.g. [9][10][11][12][13]), and was developed with the input of experts from medicine, bioethics, digital technology, and public health policy. This exact survey was administered in-person to a distinct cohort of 222 patients at the same Canadian cancer center during calendar year 2019 [8], allowing the unique opportunity to glean insights into how patient perceptions for sharing health information for general use outside clinical care might differ during ordinary times versus a global pandemic. Key survey items focused on: (a) broad versus study-specific consent; (b) opt-in versus opt-out approaches for a research contact pool; (c) levels of comfort in sharing with different types of recipients; (d) perceptions related to commercialization; and (e) options to track use of information and notification of study results.
Questionnaire responses, stratified by cohort, were summarized using descriptive statistics. Differences in response distribution between cohorts were analyzed using the Chi-square test. To evaluate whether differences in response distributions remained statistically significant after accounting for patient demographics, multivariable logistic and proportional odds ordinal regression models were fitted to the data. The ordinal regression models were used for questions evaluating comfort level (Q7 to Q17), with comfort level modelled as a three-level ordinal outcome: uncomfortable/ very uncomfortable versus neutral versus comfortable/ very comfortable. Therefore, an odds ratio > 1 denoted increased odds of reporting comfort. Logistic regression models were used for questions not assessing comfort levels (Q1 to Q5, Q18, Q19). The models adjusted for age group (< 49 vs. 50 to 74 vs. 75+), sex, treatment phase and education levels (completed post-secondary versus not completed post-secondary education). The proportional odds assumption was evaluated for the ordinal regression models.

Focus groups
All participants who completed the virtual survey during the COVID-19 pandemic were invited by a clinical research coordinator email to participate in one virtual 60-90-min focus group. The focus group guide was pilot tested with a preliminary focus group of three patient partners to ensure that the topics were presented in a clear and balanced fashion. Four focus groups were subsequently conducted with 12 participants (response rate 6%) by the study's lead author (ST; a Scientific Associate in the Radiation Medicine Program with no clinical relationship to the participants). For each focus group, participants were read an introductory script informing them of the study purpose and indicating that their participation constituted consent to participate in the study. The focus group guide (Additional file 2: Appendix 2) was built around four key questions: (1) Who should decide if PHI is shared?; (2) How should we contact patients to participate in research?; (3) What information do you need about future studies that propose to use your PHI?; and (4) How do you feel about commercial companies using your PHI?
Interviews were audio-recorded and transcribed verbatim, with all names and identifying features removed. Thematic analysis was employed to analyze the data; transcripts were read fully to capture holism, and the text was grouped into meaningful pieces of information known as meaning units (MUs), which were grouped based on similar features to create prominent themes and sub-themes [14]. Three members of the research team reviewed the transcripts and codes to ensure inter-rater reliability; and discussion ensued until consensus was reached to resolve any discrepancies. Constant comparison method [15] was used to ensure consistent classification of MUs. The ethnographic methodology of grounded theory [16] served as the framework for analysis as we aimed to construct theory from the data and allow themes to emerge iteratively, rather than being classified based on existing theories or pre-established categories. The Research Ethics Board (REB) of the University Health Network (UHN) provided full ethics approval of this study, and all methods were carried out in accordance with relevant guidelines and regulations. All study participants were provided with a detailed Study Information Letter to ensure informed consent, and were given the opportunity to ask any questions prior to consenting to participate.

Cohort characteristics
A total of 183 and 222 patients in the pandemic and pre-pandemic cohort respectively, completed the survey ( Table 1). The pandemic cohort contained more patients at the treatment stage (57% vs. 39%, p < 0.001), had fewer patients in the oldest age group of 75+ years (13% vs. 22%, p < 0.001), and was more likely to have completed a college or university degree (76% vs. 63%, p < 0.001). The sexes were similarly distributed across the two cohorts, with 50% and 49% of the pandemic and pre-pandemic cohorts identifying as female.

Multivariable regression models
Results from the multivariable regression models are reported in Tables 3 and 4. Table 3 includes results from the ordinal regression models used for questions assessing comfort level (Q7-Q17).

Focus groups Theme 1 mixed preferences and a diverse array of reasons justifying them emerged with respect to who should make PHI-sharing decisions, opt-in versus opt-out approaches for a research contact pool, and broad versus specific consent strategies
Some participants noted that hospital committees would be necessary to render informed and ethically defensible decisions because expert knowledge is needed, and this will expedite research. However, individual consent was felt to be important to respect individual preferences, and there remained a notable lack of trust surrounding privacy and uses of information.
The nature of some studies that could be done you The bold values are statistically significant Fig. 1 Percentage of patients reporting comfort sharing PHI as a function of recipient comparing pre-pandemic versus pandemic cohorts. Inverted pyramid chart showing the progressive decline in percentage of participants reporting comfort sharing with recipients outside the circle of care from within one's own health organization to for-profit commercial companies, and that during a pandemic, participants felt more comfortable sharing with all recipients except for-profit commercial companies In a similar sense, opt-out approaches to assemble a contact pool were acknowledged by participants to be important/valuable for researchers to access the largest participant group, because some participants maintained that the communal benefit of research should outweigh individual privacy risks.

You might get a lot of no's, and eliminate the number of people that the researchers can select from, and they might need that. FG2, P3 I do think the benefits of a larger contact pool would far outweigh personal freedom in this matter. FG2, P3
However, due to both legal and ethical/pragmatic concerns, opt-in (i.e. being asked for permission before research contact pool entry) was often still preferred, as opt-out (i.e., automatic entry with a chance to later be removed) was perceived to be a breach of privacy and was associated with a loss of control.  The outcome column outlines how the outcome was coded for each question Finally, the benefits of broad consent were recognized for practical reasons with many acknowledging logistical struggles with providing information on a study-specific basis.

I think that if we are all automatically entered, there might be a breach of privacy…not all patients are
In an ideal world, it would be nice to be asked. But, again, we go back to the question of practicality. FG2, P1 It would be a nightmare for researchers to call everybody, every time they want to submit their information. FG2, P3 However, many participants still expressed preferences for study-specific consent due to ethical concerns about potential research uses, and so they indicated a preference to be informed, educated, and given a choice. Reasons behind participant discomfort often centered on companies, which respondents suggested may not be held to the same legal and ethical norms as academic institutions or hospitals, with profits being their primary interest.
The problem with companies is that they're not obliged to the same legal and ethical laws and norms as academic and non-profit organizations like academic centers, institutions and hospitals.

Theme 3 transparency through the provision of information and multiple consent options was identified as essential to building trust
The ability to be informed and to receive study results was strongly preferred by nearly all participants.

Discussion
This study's findings show that cancer patients at a large Canadian hospital were substantially more comfortable sharing PHI during a pandemic compared to pre-pandemic times. This overall finding reflects an altruistic desire of some patients to contribute to the advancement of medical science during a pandemic and supports the notion that provisions to modify/revisit consent policies during times of crisis may be justified. Our findings however, also underscore the continuing need for organizations to exercise caution around the continual concerns over sharing data for commercial uses even during a pandemic. Commercial entities were the only potential recipients of PHI with whom participants did not feel comfortable (~ 50% discomfort levels before and during the pandemic). Previous research has documented that patients are reluctant to share PHI for commercial purposes [10-12, 17, 18]. Our focus group findings demonstrate that this heightened discomfort is owing to the perception that commercial companies are driven mainly by profit, and less by humanitarian motives (i.e., as opposed to academic hospitals). Our focus group participants opined that companies are not held to the same legal and ethical standards as health care institutions, and substantial concerns were raised regarding inequity in access to care or research-related opportunities or options. It was noted that while PHI is obtained from all patients to develop novel therapies, not everyone however, could afford them when needed. While some of these issues may seem insurmountable, innovative solutions emerged around how trust needs to be built, and this trust can serve as a framework for healthcare organizations to adapt their consent policies and processes. Participants underscored that transparency was key to building trust, in that they desired more information, wanted to be asked for permission, and indicated a desire to be able to change preferences over time. Digital dynamic consent platforms (e.g., [19][20][21]) are clearly essential to meeting the needs of patient-centred research-related decision-making, and should be adopted on a broad scale.
Our results regarding mixed preferences around individual versus committee consent decisions, opt-in versus opt-out approaches, and broad versus specific consent strategies echo similar reports that preferences vary with the circumstances of the purpose, user, and the controls placed on use throughout the literature (e.g., [13,22]); further underscoring the complex and heterogeneous nature of these issues. It is evident that clarity and transparency are of paramount importance to patients, as is the provision of information upon entry into the healthcare system, the option to choose whether to share PHI outside the circle of care, and the ability to track and change preferences over time. Attention to these considerations will help institutions in devising consent policies that will meet the complex and diverse needs of our patients, and move towards the implementation of consent processes that are dynamic, and reflecting individual patient values and preferences. Indeed, governments around the world are revising laws around data usage; the EU's GDPR [2] is recognized as amongst the most stringent privacy and security law in the world.
We acknowledge that this study was not without limitations, most notably the 6% response rate. While we acknowledge that this is a low rate, this was not unexpected during a time of global crisis; many people were adjusting to significant challenges related to employment, childcare, and lockdown restrictions. Despite this, we were still able to reach theme saturation (i.e. no new themes emerged from the data as verified by three independent coders; Malterud, 2015) [23], and indeed, it has been reported that 3-6 focus groups are sufficient to reach saturation in similar qualitative studies in the health care field [24]. It is also important to note that the participants were a specific sample (i.e., cancer patients receiving care from a research-intensive hospital who have consented to being contacted for research purposes), and may not be representative of a broader sample of patients with different diagnoses or international locations. These perceptions were also gleaned during Wave 1 of the pandemic, and it is possible that preferences may shift over the course of multiple pandemic waves. These areas represent important avenues for future inquiry.
In conclusion, the Canadian Institute for Health Research (CIHR) Tri-Council Policy Statement for Ethical Conduct of Research with Humans [1] articulates three pillars of consent: informed, voluntary, and ongoing; this is precisely what our participants desired regarding the use of the PHI outside the circle of care. Organizations have the responsibility to meet these needs with respect to sharing PHI by adopting digital dynamic consent platforms. If this were implemented on a broad scale, patients' discomfort in sharing PHI with commercial entities (particularly when their healthcare institution is not involved in the work) might be assuaged. The COVID-19 pandemic has been an unpreceded time, revealing the power of medical science to save lives and to advance the common good. At the same time, it has highlighted profound social inequities and the high costs associated with mistrust in healthcare experts and researchers. Opportunities to demonstrate respect for patients' values and preferences (e.g. dynamic consent approaches) must be undertaken if our communities are to form teams united by a shared vision to advance science and research for the benefit of our patients and society.