This secondary analysis followed a prespecified descriptive analytic plan focused on characterizing engagement patterns at the clinic level. Primary outcomes included email delivery, open rates, click-through rates, device type, and conversion rates. No hypothesis testing was conducted, as the study objective was exploratory and formative. This paper presents data from reports and indicators gathered during the Patient-Reported Indicator Survey (PaRIS) national survey in Quebec, Canada. The PaRIS is an international survey led by the Organisation for Economic Co-operation and Development (OECD) to develop, standardize, and implement a new generation of indicators that measure the outcomes (patient-reported outcome measures [PROMs]) and experience of health care (patient-reported experience measure [PREMs]) most relevant to primary care patients worldwide. The PaRIS initiative has been described in detail elsewhere [14,15]. The OECD provided guidance on patient recruitment across participating countries, enabling each country to tailor its recruitment methods to its chosen approach, whether by phone, mail, email, or in-person at clinics. Quebec opted to recruit patients through a large-scale email campaign. The sample consisted of patients attending primary care clinics participating in the Canadian PaRIS-OECD national survey in the province of Quebec. The inclusion criteria for patients were as follows: (1) being 45 years or older, (2) having consulted their primary care clinician within the last 6 months, and (3) providing an email address. In accordance with the OECD’s guidelines for the PaRIS study, the targeted sample size was established at 75 patients per clinic, although clinics could recruit more if feasible [15].

Our team, comprising a registered nurse and researcher, a research coordinator, a patient partner, and a scientific graphic designer, cocreated an invitation template for patient recruitment. STW and MEP are coleads for PaRIS-OECD in Canada. First, we identified the essential items to include in the email invitation. We then drafted the initial content, refining it before the scientific graphic designer further refined it to enhance its visual clarity and usability. As Canada has 2 official languages, French and English, we created bilingual email recruitment materials.

Building on the Canadian Institutes for Health Research’s strategy for patient-oriented research [16] and our previous findings, which demonstrated that recruitment facilitated by a patient partner results in a higher participation rate, we incorporated a patient partner video clip in the email invitation. This 1-minute 50-second French video, featuring English subtitles, showcased a patient partner explaining the project’s objectives and emphasizing the importance of participation. The scientific graphic was produced using Adobe Premiere Pro (version 25.4). The email template (Multimedia Appendix 1) was designed using the online marketing platform CyberImpact 2025, which provides delivery and engagement metrics and contains the following 5 sections:

We also created a reminder template with adjusted wording to include the reminder element and remove links to video clips discussing PREMs and PROMs, thereby streamlining the invitation template. Finally, we created support materials for participating clinics, including a PDF illustrating the steps for importing emails into the marketing platform and how to schedule email campaigns and several examples of email subjects such as “PaRIS Study-Survey on your care experience at [name of the clinic],” “[name of the clinic]: PaRIS research project,” and “We need your opinion, please help us!” Each clinic was instructed to program 1 initial invitation and 3 reminders, all scheduled to be sent at 6 AM.

Participating primary care clinics obtained data from the statistics provided by CyberImpact and shared anonymized reports with the research team, as approved by the ethics committee. The following categories of data were used: mailing performance, opening statistics, and statistics on embedded links. Briefly, email open and click rates were measured using hidden tracking pixels. Specifically, the pixel was activated only when the images embedded in the email were downloaded, which helped reduce false openings generated by automated systems that scanned messages. A certain margin of error remained, for example, due to image blocking or user privacy settings. However, CyberImpact considers the reliability of these measures to be very high, estimating data accuracy at approximately 95%. We used the analytic data from YouTube to track metrics (number of views, average view duration, and average percentage of video viewed) for the patient partner introduction video.

The open rate was calculated as the percentage of patients who opened the email out of the total number of patients who received the email. The click-through rate refers to the percentage of people who click a link in an email out of those who open it. The conversion rate is the number of recipients who take the desired action compared with the number of recipients who received the email. The unsubscribe rate refers to the number of patients who unsubscribe out of the number of patients who received the email [17]. Quantitative data were analyzed using descriptive statistics, including measures of central tendency (mean) and variability (SD) using a 95% CI, as well as the minimum and maximum values of the variables’ means across clinics, using SPSS (IBM Corp) software. We did not use a reporting guideline. Instead, we emphasized transparency in describing the data sources, metrics examined, unit of analysis, and descriptive analytic approach, consistent with the exploratory objectives of the study and the reporting standards commonly used for formative and implementation-focused research.

This study was in accordance with the ethical standards of the responsible committee on human experimentation and with the World Medical Association Declaration of Helsinki. Ethics approval was obtained from the research ethics board of the Ministry of Health and Social Services of the province of Quebec (CCER-22-23). Patients gave their consent to participate in the PaRIS-OECD national study, and this consent covered secondary analysis. Patients participating in the PaRIS study in Quebec were eligible for a draw to win one of three US $25 grocery gift cards among respondents from their clinic. Demographic characteristics are presented in aggregate, and no identifiable participant information is included in this study.

Overall, this formative secondary analysis provides practical insights into how primary care patients engage with email-based research invitations in real-world clinical contexts. Our findings suggest that email invitations are an efficient and effective method for reaching a large number of primary care patients. High delivery and open rates suggest that emails sent on behalf of trusted primary care clinics are effective at reaching patients. Moreover, conversion rates indicate that a meaningful proportion of recipients proceeded to survey participation. These findings support the feasibility of email recruitment as a low-burden approach for clinics and research teams. These results lead us to the following observations: (1) email is a highly feasible recruitment method in primary care settings; (2) email makes it possible to reach older patients, but equity issues persist; and (3) commitment varies greatly depending on the clinic, which indicates that the organizational and relational context matters.

We found that a large-scale email campaign using a marketing platform effectively reached a significant number of primary care patients. The high delivery rate of 96.6% (14,758/15,277) is consistent with previous studies [9,18]. Only a few email addresses were incorrect, indicating that patients generally provided valid addresses to their clinic. As expected, reminders were effective for primary care patients recruited by their clinic [10]. Unlike paper questionnaires sent by traditional mail, which remain physically present and serve as a constant reminder, digital questionnaires do not provide a lasting prompt [19]. A high proportion of primary care patients (10,828/14,758, 73.4%) opened the invitation, which is high compared to 35% to 51% rates reported in other studies [20,21]. The average open rate in email marketing campaigns related to health care services sent by the industry is 23% [22]. Within the email marketing industry, open rates above 50% are considered outstanding, typically reflecting the presence of a highly loyal or well-targeted niche audience [23]. This may be attributed to the sender’s email address being from a reputable authority (ie, a primary care practice), which could lead patients to perceive it as more relevant [24] and reduce the likelihood of them mistaking our invitation for spam or a fraudulent solicitation. Indeed, participants may hesitate to open emails from unknown or suspicious addresses [9,25]. Additionally, the clinic used a catchy email subject to capture patients’ attention, which may have contributed to the high open rate [8]. Targeted email campaigns can also enhance recruitment effectiveness, as all invited patients were directly concerned with the study topic [26]. This confirms the organizational feasibility of large-scale recruitment campaigns while imposing minimal burden on clinics.

Our email invitations achieved high engagement indicators, successfully engaging more than 1000 primary care patients. The conversion rate of our email recruitment strategy was 9.7% (1430/14,758), which is higher than the rates typically reported in similar studies (2.9%-4%) [9,24,25,27-29]. General email marketing campaigns in Canada average a conversion rate of less than 3% [8]. This may also be attributed to the clear division of materials into sections, allowing patients to quickly assess the relevance and interest of the information provided [18]. Including multimedia elements, such as illustrations and video clips, likely contributed to higher performance [18]. A total of 1512 clicks were recorded on the patient partner video link, which suggests that this content seems to be of interest to patients. Moreover, simply knowing a patient partner was involved in the study and made a video may have reassured some patients and added external validity [30]. However, the study design does not allow conclusions regarding its causal impact on participation. These findings suggest that while cocreation may enhance acceptability and trust, its contribution should be understood as contextual and supportive rather than as a measurable driver of engagement. This strategy, along with the contribution of patient partners in research and recruitment, is highly recommended as best practice guidelines for patient-oriented research [16,31].

The age range of patients who responded to the survey was broad. Notably, patients aged 70 years or older also answered the questionnaire, suggesting that email can be an effective method for recruiting older patients. This observation aligns with the increasing internet use among older adults in Canada over recent years, as the proportion of those aged 75 years or older using the internet rose from 62% in 2020 to 72% in 2022 [32]. The online survey received a higher response rate from women, which is consistent with comparable studies [33-37]. Additionally, because patients were invited if they had an encounter with their health care professional within the previous 6 months, and women tend to consult primary care more frequently than men, this may have led to a higher proportion of women being invited to participate in the survey [28,38]. For ethical and privacy reasons, our team did not have access to the identity of admissible invited patients and, consequently, to their sociodemographic characteristics. We also observed a low number of responses among people of non-White ethnic origin. Due to the unavailability of data on the ethnic origins of the invited patients, it is not feasible to calculate or compare specific response rates by ethnicity. However, the literature suggests that ethnicity may be a source of nonresponse bias [39,40]. Heerman et al [28] observed similar results, finding that technology-enabled recruitment approaches are not the most efficient for enrolling non-White patients. Therefore, it is essential to consider strategies to enhance the participation of these underrepresented groups in health surveys, primarily because it is known that ethnic groups other than White have a less positive primary care experience [41].

Engagement metrics varied substantially across participating primary care clinics, including differences in open rates, click-through rates, and conversion rates. These variations suggest that a strictly standardized recruitment protocol may not be optimal in all primary care contexts. Although a common email template and overall recruitment strategy were used, clinics adapted key implementation elements—such as timing of invitations, number of reminders, and communication practices—to better align with their organizational context and patient population. This adaptability appears necessary to maintain acceptability and feasibility, highlighting that digital recruitment strategies in primary care may benefit from a balance between standardization (to ensure efficiency and comparability) and local tailoring (to reflect clinic-specific realities and patient expectations). These findings emphasize the importance of context-sensitive implementation when deploying digital recruitment approaches in real-world primary care settings.

This study should be interpreted as a formative, implementation-oriented analysis rather than a test of behavioral theory or recruitment effectiveness. The use of descriptive engagement metrics offers actionable information for research teams seeking to monitor, refine, and adapt digital recruitment strategies over time. Future studies could build on this work by incorporating comparative designs, experimental testing of message components, or qualitative inquiry to elucidate mechanisms of engagement further.

This study had a few limitations that must be underscored. Ideally, participation invitations would have been sent to all patients across clinics simultaneously, and the time between reminders would have been consistent across clinics, allowing for a deeper exploration of factors influencing response rates across clinics. Ideally, to assess the effect of the various strategies embedded in the email, different versions of the message should have been sent to separate groups of patients (for instance, a plain or less visually appealing email or the same message without the patient partner video). However, within the PaRIS study context, it was not feasible to send potentially less-engaging messages to patients, as the primary goal was to maximize participant recruitment. Due to the PaRIS study’s design, objectives, and ethical constraints, we could access only the sociodemographic characteristics of patients who responded to the questionnaire, not those who received the email invitation. Sample representativeness is a major concern, given that only 9.7% (1430/14,758) of the invited patients partially or fully completed the questionnaire. Additionally, staff from some clinics discussed the study with some patients, which may have contributed to a higher conversion rate. However, this intervention was not measured, and its effect cannot be assessed. Our study also revealed variation in indicators across clinics, demonstrating that patients do not all respond similarly to email invitations. These differences may be attributed to the type of clinic (private, public, or university-affiliated) and the characteristics of the patients enrolled in the study. However, our study lacks sufficient power to determine these characteristics.

This formative, descriptive secondary analysis documents the functioning of email-based recruitment in real-world primary care research settings. Using routinely collected engagement metrics, we show that emails sent on behalf of trusted primary care clinics are a feasible and low-burden approach for reaching patients and supporting participation in web-based surveys. These findings offer practical, implementation-oriented insights for research teams seeking scalable recruitment approaches. They can inform the design, refinement, and ongoing monitoring of digital recruitment strategies in future primary care studies.