Participants were asked to complete 4 surveys such as when joining the program (baseline) and then after 30, 60, and 90 days of peer support. All surveys were administered within the app. Each survey was available for participants to complete during a 10-day window. We collected demographics at baseline and used the member’s zip codes to calculate their social vulnerability index according to the Centers for Disease Control and Prevention [44], which accounts for 16 social factors including poverty, lack of vehicle access, and crowded housing. The following measures were included in the surveys.

This validated measure of loneliness [45] includes the following questions: How often do you feel that you lack companionship? How often do you feel left out? How often do you feel isolated from others? For each question, participants were asked to choose among 1=hardly ever, 2=some of the time, and 3=often. A total score of 4 or greater is considered positive for loneliness, and a score of 7 to 9 is considered severely lonely.

The Patient Health Questionnaire 2-Item Scale (PHQ-2) is a validated measure [46] that asks about the frequency of depressed mood and anhedonia over the past 2 weeks. Items are scored from 0 (not at all) to 3 (nearly every day). A total score of 3 or greater indicates that a major depressive disorder is likely.

A single item from the validated General Anxiety Disorder 7-Item Scale (GAD-7) [47] was used: “Over the past 2 weeks, how often have you been bothered with feeling nervous, anxious, or on edge?” This item was scored from 0 (not at all) to 3 (nearly every day).

This validated measure captures information on the physical and mental health status of individuals and on the impact of health status on quality of life [48]. It includes 2 questions: “Thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?” and “Thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” The 2-item measure assessing physically and mentally unhealthy days is highly correlated with traditional measures of health, including morbidity, mortality, and health care costs. Each additional unhealthy day is associated with an incremental cost increase of US $15.64 per member per month [49].

Participants were asked whether they would recommend the program to others, whether they believe that the program should be made available to everyone in Colorado, and to provide subjective comments.

The Wisdo intervention is a peer support and social health platform designed to enable the 4 pillars of social health such as emotional support, reassurance of worth, a sense of belonging, and reliable alliance (Figure 1) [34,50-53]. The platform is available as a native app on iOS and Android devices and as a web app on personal computers and is HIPAA (Health Insurance Portability and Accountability Act) and System and Organization Controls (SOC) 2 + HITRUST (Health Information Trust) compliant. Members can be anonymous on the platform as they are not asked to use their real names. Since its launch in 2018, over 500,000 adults aged 17 to 80 years have participated in the platform’s peer support communities. There are 5 key aspects to the peer support intervention such as self-mapping, connecting with peers, group coaching, referrals to other behavioral health resources, and the ability to become a peer helper.

New members are onboarded to the platform with a self-mapping exercise that is critical for assigning the participant to a curated social support community comprised of helpful peers with shared lived experiences. The self-mapping element includes 4 steps (Figures 2A-2E). First, members select the community they want to join based on a topic they are concerned about and that makes them feel most lonely. Members choose from over 50 communities covering areas such as behavioral health (eg, anxiety, depression, loneliness, and alcohol use), physical health (eg, diabetes, heart failure, and cancer), family (eg, caregiver and parenthood), workplace (eg, burnout, stress, and working remotely), and self-growth (eg, positive thinking and starting to exercise; Figure 2A). On average, members join 6 communities during their first year on Wisdo.

Second, members are asked to identify as either “been there” or “there now” on a series of 25 to 35 key challenges, obstacles to health, gaps in care, and milestones that are frequently experienced by individuals in their selected community (Figure 2B). On average, members click on 20 such milestones per community they join, providing in-depth insights into their experiences to date. As members click through each step, they can see how many people on the platform have clicked the same step, further promoting 2 important pillars of social health such as a sense of belonging and emotional support.

Third, members choose up to 3 personal goals from a list of 10 (eg, making new friends, better managing stressful events, and improving sleep) relevant to their community (Figure 2C). As members click on a goal, they can see how many other members on the platform have already accomplished it. Members can track their progress toward achieving each goal.

Finally, members answer brief validated measures (Figure 2D). These responses establish a baseline for members’ social and mental health and enable them to track changes over time.

As members continue to join additional communities, establish goals, and answer follow-up surveys, the self-mapping information shared by members during onboarding is used by the platform’s artificial intelligence engine to create curated connections with helpful peers, identify members to join group coaching sessions to develop social health skills, and connect members with covered clinical programs and services to support social barriers to care.

The peer support platform continuously suggests connections with helpful peers based on the number of shared steps (ie, “been there” and “there now”) and how helpful these peers have been to similar participants. Once connected, participants can chat on the app one-on-one and in group settings and receive emotional, nonjudgmental support and encouragement (Figure 2E). They can also share “empathy-charged” reactions to posts by other members that build up emotional support and a sense of worth, such as “Helpful,” “Love,” and “Been There.” The community is moderated by trained staff augmented by its artificial intelligence engine, which monitors over 600 keywords and phrases that could indicate a potential risk. Members who fail to follow the program’s code of conduct are warned and may be banned from the platform.

Members can join weekly group video coaching sessions conducted by certified life coaches via video. For anonymity, participants can choose not to turn on their video cameras. Each group coaching topic allows members to connect around a shared goal and skill (eg, building emotional resilience, developing a sense of purpose, and developing a sense of self-worth) or interest (eg, travel, books, and exercise). Together, this support from peers and groups provides the key elements of a loneliness intervention, namely emotional support, reassurance of worth, a sense of belonging, and reliable alliance.

The program matches participants to covered benefits and community services offered by their employer, health plan, government agencies, or local nonprofits. Once a participant is flagged as a candidate for a covered benefit, the program deploys a combination of several tactics, including in-app messages, push notifications, and an invitation to relevant coaching sessions to inform and motivate the member to engage with the relevant covered benefit.

Highly rated members who are frequently considered helpful by other community members are invited to enroll in a web-based training program to earn a Helper badge. The program includes 5 modules covering topics such as providing emotional support, motivational interviewing skills, and caring for oneself. Each module is followed by a quiz. Users who successfully complete the training are awarded the Helper badge and provided with an explanation of how their performance will be continuously monitored and assessed to ensure safety and quality on the platform. On average, 10%-20% of members joining the peer support platform successfully complete the training and earn the Helper badge. This format promotes 2 important pillars of social health such as reliable alliance and reassurance of worth.

Of the 815 participants, 595 completed the baseline survey, and 130 completed one or more of the subsequent surveys.

The average loneliness score at baseline was 7.41 (SD 1.65), with 98% (n=583) screening positive for loneliness (UCLA-3 Loneliness Scale score of 4-9) and 68.7% (n=409) screening as severely lonely (UCLA-3 Loneliness Scale score of 7-9). After 90 days, 41.5% (n=54) of individuals who screened as severely lonely (UCLA-3 Loneliness Scale score of 7-9) when joining the study had lower loneliness scores and 4.6% (n=6) scored as no longer lonely (UCLA-3 Loneliness Scale score of 3). Overall, 10.4% (n=13) individuals who screened as lonely when joining the study (UCLA-3 Loneliness Scale score of 4-9) scored as not lonely after 90 days (UCLA-3 Loneliness Scale score of 3) and 31.5% (n=41) reported no change in their loneliness level.

There was a statistically significant decrease over time in levels of loneliness (b=–0.01; t₁₉₇=–5.65; P<.001; Figure 3). Overall, loneliness decreased 11.6% (mean 6.46) within the first 30 days and 14.6% (mean 6.47) between the baseline and 90-day assessments. The participants who only reported loneliness at baseline and at 30 days had the greatest improvement in loneliness. The change over time in loneliness did not differ significantly between participants with complete data versus individuals with only partial survey data (t₁₉₆=0.62).

The average depression score at baseline was 3.49 (SD 1.91). In total, 61% (n=363) of participants joined the study with a score of 3 or higher (ie, screened at risk for depression); of these, 65% (n=236) reported below-risk levels of depression (<3 on PHQ-2) within 30 to 90 days. Overall, there was a 33.1% (mean 2.12) decline in depression symptoms by 30 days and a 50.1% (mean 1.89) decline in depression symptoms between baseline and day 90 (Figure 4; b=–0.02; t₁₉₇=–6.48; P<.001). This decline did not vary as a function of whether the participant completed all 4 surveys (t₁₉₆=–0.40).

The average anxiety score at baseline was 1.86 (SD 1.00), with 87.9% (n=523) of participants reporting feeling nervous, anxious, or on the edge for at least several days over the past 2 weeks and 30% (n=179) saying they experience these feelings nearly every day. Participants had a 20.1% (mean 1.35) reduction in anxiety symptoms by day 30 and a 29% (mean 1.42) reduction in anxiety from the baseline to day 90 (Figure 5). There was a statistically significant decline in levels of anxiety over time (b=–0.16; t₁₉₇=–3.42; P<.001). This pattern of results did not appear to vary for participants who provided complete data versus individuals who did not complete all the surveys (t₁₉₆=–0.18).

The average number of unhealthy mental health days at baseline was 20.1 (SD 9.9) days, and the average number of unhealthy physical health days at baseline was 15.9 (SD 11.4) days. At baseline, 88.9% (n=529) reported having 6 or more mentally unhealthy days in the prior month and 52.9% (n=315) reported having 6 or more physically unhealthy days during the same period. There was a statistically significant (13%) improvement in health-related quality of life from the baseline survey (mean 24.63, SD 10.34) to the 90-day survey (mean 21.35, SD 9.13; t₃₉=2.02; P=.05). On average, participants reported 3.28 less monthly unhealthy days after 90 days on the platform. This indicates the potential of the platform to drive an annual reduction of medical costs equaling US $615 per participant (3.28 days×US $15.64 estimated cost reduction per unhealthy day×12 months).

Overall, 88% (n=91) reported that they would recommend the peer support platform to others, and 98% (n=102) of participants said that they believed that the program should be made available to everyone in Colorado. In subjective comments, participants frequently wrote that being able to connect and talk with others who have shared lived experiences in a safe, judgment-free, and supportive environment was the most valuable part of their experience. No participants wrote a negative comment.