Principal findings
This qualitative analysis provided important insight about high-risk individuals’ perceptions regarding receiving personalized depression risk information. Risk prediction algorithms for mental health are a new topic in health care research, this study is one of the few to investigate the perceptions about how using these algorithms will affect the population. The main motivation for receiving the risk score was out of curiosity and to improve self-awareness of mental health. Receiving the risk information can provide a new perspective to discuss mental health problems openly because many of the participants in the study were open to sharing their score with friends and family. Participants had mixed views about the utility of personalized depression risk information, and the utility largely depends on the context and situation in which individuals were embedded. The optimal delivery of the personalized risk score was with a counselor, either in person or on the telephone. Providing additional resources, more follow-up conversations, and a paper copy are strategies that may enhance depression risk communication.
Comparing to current literature
As part of this RCT, we conducted a qualitative study on participant perspectives 1 month after receiving the personalized depression risk information at baseline [
21]. Consistent with that qualitative study, participants had a positive view about the risk information and a majority of them (over 93%) were interested in knowing their risk. Importantly, the RCT found that providing personalized depression risk information does not have a negative impact on physical or mental health [
21]. A qualitative study conducted by Bellon and colleagues [
29] in the Spanish primary care setting reported the same results. The Spanish study also indicated that primary care patients preferred health care professionals providing resources and helping them understand ways to prevent depression. This is, in part, consistent with our results; participants are often interested in sharing their risk information with their family doctors, but only if their doctor is both aware that they have depression and has been providing treatment for depression. On the other hand, many participants in this study reported that they preferred to share the information with spouses, family members and friends whom they trust and can rely on; this demonstrates the acceptability of the personalized depression risk information by the participants and its potential broader health promotion impact.
Implications
Understanding motivations behind receiving their risk score is an important step in understanding whether or not someone will seek help, what those help seeking behaviors will be and their effect on the individual [
30]. Similarly to what was found in this study, people often seek out health information with the intent to make positive behavior changes to improve their health or prevent an illness [
31].
The results of this qualitative study offered some important insights about the formats of delivery and how personalized depression risk information may be better communicated. Personalized depression risk scores are new to individuals in the community, and many have little knowledge about what constitutes “high risk”. As such, a large proportion of high-risk individuals either over or underestimate their risk of having depression [
32]. To this end, providing a comparative risk (e.g., the average risk of having an MDE in the general population) may help address this issue and improve users’ understanding of their risk. Participants also strongly recommended providing resources and effective strategies for risk reduction. Pertinent research in cardiology and oncology has shown that effective risk communication often include the following components: (1) Individualized risk presented as an absolute risk, as opposed to a relative risk [
33]. The risk information may also list the individual’s risk factors. (2) Appropriate format of presentation (e.g., graphic, visual). The format of presentation can influence the degree to which individuals perceive their risk and will affect behavioural change, and (3) providing clear evidence-based information on available choices. It is clear that experience in other medical disciplines resonates well with the views of the participants in this study.
Participants preferred the idea of receiving depression risk information from a counselor. Moreover, participants also endorsed follow-ups by the research team in order for individuals to have a chance to have questions answered, discuss potential risk factors, receive information about effective self-help strategies, as well as information about mental health resources. Risk communication in oncology and cardiology is delivered either through written materials or the Internet, or with professional guidance (e.g., in-person or telephone education and counselling by trained coaches or health professionals) [
34,
35]. It is possible that depression risk communication guided by counselors or trained coaches may be more effective than un-guided risk communication in motivating behaviour change. Evidence of that was found in the Spanish study by Moreno-Peral and colleagues [
36] where they investigated the effect of personalized risk of depression information guided by health care professionals on anxiety. They found that this guided information decreased the anxiety of the participants. The decision about using guided or un-guided format will have cost implication when it comes to large scale implementation. A cost-effectiveness study was completed in Spain using a guided format in the clinical setting; they found that there was not increase in cost for increased quality of life [
37]. Future studies are needed to compare the cost-effectiveness of guided and un-guided risk communication interventions and their benefits.
This study has several limitations. First, this study was completed during the global COVID-19 pandemic, which may have affected the perceptions of the participant’s mental health. Second, participants were asked to recall information from their past which may be vulnerable to reporting and recall biases. As a means to decrease the chances for recall bias, we only included participants who has recently completed their 12-month assessment in the qualitative interviews. This study was conducted in Canada, among individuals who were at high risk of having depression. Therefore, the results of this qualitative study should be understood as specific to this context. Finally, the results were not validated by the participants because they were not able to read and comment on the results of the study. To address that limitation, the analysis was completed by two different people separately than compared.
With the emerging interests in applying risk predictive analytics and machine learning techniques in the realm of mental health, it is anticipated that more risk predictive models for mental health problems will be developed in the clinical arena or in the general population. Nevertheless, these risk predictive models can only become useful when the baseline risk is communicated to the target population and the information is being acted upon. This qualitative study shed light on user perceptions about personalized depression risk information and how it may be best communicated, offering important information for designing personalized risk communication tools not only for depression and other mental health problems.