Our team of researchers and knowledge users (patients, clinicians) conducted 2 population-based cross-sectional surveys using Leger’s consumer panel. Having patients and clinicians as equal partners on the team was done to ensure the study yielded relevant findings [21]. Patients were on the executive research team coleading the study (author CL who is immunocompromised), and author MS stimulated the research topic. Both were engaged in all study aspects. Together, we identified the research objectives, outcomes, procedures, and deliverables for the funded proposal. Next, they were involved in guiding the study. We used the CHERRIES (Checklist for Reporting Results of Internet E-Surveys) guideline [22].

The survey was conducted in Canada during the third wave of COVID-19 [23]. During the first wave of COVID-19 in January-June 2020, 80% of the COVID-19–related deaths occurred in long-term-care homes, and governments issued stay-at-home orders and travel restrictions [24]. Masks became mandated for use in indoor spaces in July 2020. The second wave of COVID-19 started in November 2020 and the third wave in March 2021. When the hospitals became overwhelmed during the third wave, with a high impact on intensive care units, governments reissued a stay-at-home order [23]. COVID-19 vaccines were initially approved by Health Canada in December 2020, with priority for health care workers, Indigenous peoples, the elderly living in group settings, and others at higher risk (eg, people with cancer, with organ transplants, or undergoing dialysis). Vaccination gradually opened to adults in spring 2021 and to children (ages 5-12 years) in fall 2021 [25]. Although the AstraZeneca vaccine was approved for use on February 26, 2021, the rate of vaccine-induced immune thrombotic thrombocytopenia was of concern (1 in 60,000; April 2021). Governments continued to recommend the AstraZeneca vaccine until June 2021 because Canada was in the third wave of infections and there was insufficient supply of messenger RNA (mRNA) vaccines [23].

We recruited 2 groups of participants: (1) adults aged 18 years or older who received or were receiving any health services in the past 12 months for themselves (labeled “adults”) and (2) adults aged 18 years or older who were responsible for children or seniors aged >65 years with cognitive impairment who received or were receiving any health services in the past 12 months (labeled “parents/caregivers”). Leger recruited participants through its consumer panel titled “Leger Opinion (LEO) Panel,” which has about 400,000 Canadians across 10 provinces and 3 territories. The panel includes data on age, gender, and region that can be used for sampling and quota management. At the onset, specific quotas by age, gender, and region were set based on the Canadian population data published by Statistics Canada [26]. Leger monitored throughout the recruitment phase, and sampling was adjusted to ensure the data collected were representative. For example, if a certain cohort was underrepresented, sampling was adjusted to recruit more respondents falling within that cohort. There was no weighting applied to the data after being collected. Respondents received a personalized email containing a unique URL link to the survey. The email invite said, “LEO wants to hear from you!” and it did not state the exact survey topic.

Adapted from previous surveys in Canada [13,14,27], questions were based on the Ottawa Decision Support Framework [11]. Questions from instruments with good reliability and validity included the Decisional Conflict Scale (DCS) [12], the Decision Regret Scale (DRS) [28], and Strull’s roles in decision-making [13]. The survey enquired about health-related decisions adults and parents/caregivers faced during the past 12 months, factors influencing decision-making, trusted information sources, and sociodemographics according to the PROGRESS (place of residence, race/ethnicity/culture/language, occupation, gender/sex, religion, education, socioeconomic status, and social capita) framework [29,30]. To collect data on a broader range of participants’ characteristics that may stratify health opportunities and outcomes due to discrimination [29,30], we also asked respondents to self-identify if they had lived experience as a member of a marginalized group defined as disabled or caregiver of a person with a disability, gender diverse (eg, agender, nonbinary, transgender, cisgender), intersex, LGBTQ+ (lesbian, gay, bisexual, pansexual, transgender, queer, two-spirited, questioning), Indigenous, racialized (eg, person of color), neurodivergent (eg, attention deficit hyperactivity disorder [ADHD], autism, dyslexia), or a marginalized group not listed. Respondents were initially asked to list all the decisions made in the past year and then pick 1 difficult decision for the remaining questions. A difficult decision was defined as having more than 1 option and no clear best option. The online survey, available in English or French, screened eligibility and included 32 questions (Multimedia Appendix 1). Two random test questions were used to ensure respondents were not trying to speed through the survey, and the survey stopped for those who clicked a wrong response on both.

Respondents received 1 question per screen and pressed “continue” to proceed. They could not return to previous screens to change responses. To avoid missing responses, no questions could be skipped. Up to 3 reminders were sent to those who started but did not complete the survey. To avoid multiple entries, respondents were assigned a unique identifier linked to their LEO account; if they tried to complete the survey again, they received an error message.

The English and French versions were pretested by members of our team, including patient partners. On May 18, 2021, we field-tested the surveys with 139 adults and 26 parents/caregivers to monitor completion time (adults: ~10 minutes; parents/caregivers: ~9 minutes) and potential problems with surveys, including predefined skip logic. Given no changes were made to the surveys based on field testing, the collected data were retained.

Based on our previous research [13,31,32], sample sizes of 1500 adults and 500 parents/caregivers were adequate to estimate proportions with significant decisional conflict (SDC) associated with decision delay. SDC is defined as a total score of >37.5 out of 100 on the DCS based on 253 studies [33] using a 2-sided 95% CI with a margin of error of +1.80% or +3.85% when the estimated proportions are 0.15 or 0.26, respectively [34,35]. This is the most severe cutoff; sensitivity analyses were performed with a more relaxed cutoff of 25, also sometimes seen in the literature (data not shown), but we privileged a more severe cutoff. This means it could have been showing more people with SDC. To account for an anticipated 10% response rate, target sample sizes of adults and parents/caregivers were 15,000 and 5000, respectively.

We classified respondents into adults who made a health-related decision for themselves (adults) or made a decision on behalf of a child or a senior with cognitive impairment (parents/caregivers). Next, we used descriptive statistics to identify decisions and described decisional needs using the Ottawa Decision Support Framework for each group (adults and parents/caregivers) [10]. We classified respondents as having SDC [36], and for those who had made the decision, we classified respondents as having significant decision regret based on a cutoff of >25 out of 100 on the DRS [37].

The University of Ottawa research ethics board approved our study (H-03-21-6752). Invitees provided study consent at the start of the online survey and only initiated the survey questions after clicking the link consenting to participate. Respondents voluntarily answered questions and were guaranteed confidentiality. Leger offered an incentive of 2000 points (equivalent to CA $1.60, or US $1.18) to complete the survey.

From May 18 to June 4, 2021, 14,459 adults and 6542 parents/caregivers were invited to participate (Figure 1). The view rate (unique survey visitors/unique survey invitees) was 15.5% (2236/14,459) and 28.3% (1850/6542), respectively. The participation rate (unique visitors who consented to participate/unique survey visitors) was 69.3% (1549/2236) and 28.7% (531/1850). The completion rate (users who completed the survey/users who consented to participate) was 97.3% (1507/1549) and 95.1% (505/531), respectively. Among those who completed the surveys, we removed respondents if they did not identify a difficult health decision. After reading open text describing options, 36 were moved to the adult database and 16 were moved to the parent/caregiver database.

There were 1454 adults and 438 parents/caregivers included in the analysis. The survey was completed in English (1598/1892, 84.5%) or French (294/1892, 15.5%). Respondents represented a range of ages, education levels, civil statuses, ethnicities, and annual household income (Table 1). Of 1892 respondents, 541 (28.6%) self-identified as members of marginalized groups.

Health-related decisions, in order of frequency, were about COVID-19 vaccination, managing a health condition, social COVID-19 decisions (masking, limiting contacts), mental health care or addiction treatment, medication, surgery, pain management, health care for COVID-19 (testing, seeking care for symptoms), stay at home or move to an assisted care facility (eg, nursing home), pregnancy, birth control, staying safe at home (eg, adapt or retrofit home), and end-of-life care (Table 2). Decisions were similar when respondents were asked to pick a single decision described as more difficult, with the exception of health care for COVID-19, which a few rated as difficult. Other single decisions for caregivers were about moving a family member to/from a retirement or nursing home and stopping them from driving a car. Decisions (adults vs parents/caregivers) were within the past month (318/1318, 24.1%, vs 70/366, 19.1%), 1-6 months (524/1318, 39.8%, vs 134/366, 36.6%), 6-12 months (307/1318, 23.3%, vs 97/366, 26.5%), or no response (169/1318, 12.8%, vs 65/366, 17.8%).

Of 1454 adults and 438 parents/caregivers, 323 (22.2%) and 95 (21.7%), respectively, had SDC (Tables 3 and 4). Decisions, in order of frequency, for adults with SDC were mental health care, managing a health condition, taking medications, pain management, and COVID-19 vaccination (Figure 2). Decisions for parents/caregivers with SDC were COVID-19 vaccination, managing a health condition, health care for COVID-19, and mental health care.

Respondents experiencing SDC were more likely to feel uninformed (mean 44, SD 18.6, vs mean 18, SD 14.3, out of 100), have unclear values (mean 45, SD 20.3, vs mean 18, SD 14.6), feel they made an ineffective decision (mean 48, SD 16.1, vs mean 16, SD 13.3), and have decision regret (mean 41, SD 17.5, vs mean 16, SD 15.7) compared to those not experiencing SDC. In addition to feeling worried about choosing the wrong option (193/418, 46.2%), respondents experiencing SDC identified the following factors as making the decisions more difficult: public health restrictions due to COVID-19 (133/418, 31.8%), difficulty separating misinformation from scientific evidence (113/418, 27.0%), information overload (107/418, 25.6%), and no or limited access to information on the decision (97/418, 23.2%); see Table 3 for all respondents. Adults identified difficulty discussing the decision with clinicians (85/323, 26.3%). Parents/caregivers identified difficulty discussing the decision with significant others (eg, family, friends; 22/95, 23.2%).

Of 1318 (90.6%) adults and 366 (83.6%) parents/caregivers who made a decision, 353 (26.8%) and 125 (34.2%), respectively, had decision regret (Table 4). Those facing mental health care decisions had higher decision regret (mean 32.0%, SD 46.8%). Adults (322/1318, 24.4%) and parents/caregivers (77/366, 21.0%) reported not getting their preferred option.

Most respondents indicated an active role in making the decision either alone (768/1318, 58.3%, adults vs 181/366, 49.5%, parents/caregivers) or after considering the opinion of their clinician (348/1318, 26.4%, adults vs 105/366, 28.7%, parents/caregivers). Few decided together with their clinician (148/1318, 11.2%, adults vs 63/366, 17.2%, parents/caregivers) or deferred to their clinician (54/1318, 4.1%, adults vs 17/366, 4.6%, parents/caregivers). If asked to make the decision again, most preferred an active role (1067/1318, 81.0%, adults vs 264/366, 72.1%, parents/caregivers) or together with their clinician (218/1318, 16.5%, adults vs 83/366, 22.7%, parents/caregivers).

Respondents indicated that trustworthy information sources were health professionals (1190/1454, 81.8%, adults vs 323/438, 73.7%, parents/caregivers), Health Canada (931/1454, 64.0%, adults vs 273/438, 62.3%, parents/caregivers), and provincial health departments (755/1454, 51.9%, adults vs 211/438, 48.2%, parents/caregivers). Fewer respondents trusted information from specific health organizations (583/1454, 40.1%, adults vs 167/438, 38.1%, parents/caregivers), consumer/patient organizations (265/1454, 18.2%, adults vs 79/438, 18.0%, parents/caregivers), companies that produce health information (199/1454, 13.7%, adults vs 49/438, 11.2%, parents/caregivers), or health insurance companies (99/1454, 6.8%, adults vs 43/438, 9.8%, parents/caregivers). Those experiencing SDC and decision regret were less trusting of these sources (Table 5).

In the first year of the pandemic, the most frequent decisions identified by Canadians were about COVID-19 vaccination, managing a health condition, social COVID-19 decisions, mental health care, medication treatments, and caregiver decisions about moving family members to or from residential facilities. One in five respondents had SDC, and a third reported decision regret. Factors making decisions more difficult were public health restrictions due to COVID-19, information overload, difficulty separating misinformation from scientific evidence, and difficulty discussing decisions with clinicians. Demographics of and the types of decisions made by respondents indicating iniquities above a level of 30% differences were ethnicity and mental health, respectively. The most trusted information sources were health care professionals and governmental sources. Our results led to the following observations:

The involvement of knowledge users on the study team was beneficial, and they guided all aspects of this study, including survey design, interpretation of findings, and drafting of results. No negative effects of involving knowledge users were encountered. Our view rate of 15.5% for adults and 28.3% for parents/caregivers was comparable to the typical rate for online surveys of 10%-20% [27,56-58]. In addition, the geographic distribution across provinces and territories was consistent with Canadian distribution [59].

With regard to limitations, our data may be subject to recall bias, given respondents were asked to identify a difficult decision made within the previous year. Second, it would have been easier to compare findings with the 1999 survey had we used “facing a complex health decision” as an eligibility criterion [17]. However, we were interested in all decisions (including those with any level of SDC or decision regret) for those who had interactions within the health care system. Compared to Canadian census data, fewer respondents had high school education or less (33.7% invited, 20.5% participated, 44.8% census) and were from the Canadian territories (0.2%-0.4% invited, 0.5% from Yukon participated, 0.1% census); see Multimedia Appendix 2. Finally, the question about costs influencing the decision preceded the question asking to list options, and some only listed options related to costs.

Our survey of Canadians identified COVID-19–related decisions that emerged during the first year of the pandemic and how the pandemic influenced other health-related decisions. Many adults, parents, and caregivers had unmet decision-making needs, resulting in SDC and decision regret. Factors making decisions more difficult were public health restrictions due to COVID-19, information overload, difficulty separating misinformation from scientific evidence, and difficulty discussing decisions with clinicians. Most Canadians made the decisions on their own, with few sharing the decisions with others. Canadians need their decision-making needs recognized for providing person-centered care to achieve improved health outcomes and to inform future pandemic preparedness.