Implementation of a Loneliness Screening and Referral Program in Primary Care

Abstract

PURPOSE:

Loneliness screening is recommended as best practice in primary care, yet it is not widely implemented. The purpose of the current study was to assess feasibility of a loneliness screening and referral program (SOCIAL Rx) in a primary care practice.

METHOD:

Loneliness was assessed using the 3-item UCLA Loneliness Scale and curated referrals were provided for those who screened positive. Outcome measures were organized using the RE-AIM framework domains of reach, effectiveness, and adoption. Qualitative interviews were conducted to explore feasibility/acceptability and patient preferences regarding referrals.

RESULTS:

Eighty-one percent of patients were screened for loneliness; 33.3% were somewhat lonely and 17.2% were very lonely. Fifty-two percent of those who were lonely were provided a referral, and 40% of providers referred ≥50% of eligible patients.

CONCLUSION:

Loneliness was prevalent in this population of older adults, highlighting the imperative for screening and intervention.

The U.S. Surgeon General declared loneliness a public health epidemic in the United States (Office of the U.S. Surgeon General, 2023). Loneliness is commonly defined as a negative feeling of dissatisfaction with the quantity or quality of social relationships (Hawkley & Cacioppo, 2010). The negative health effects of loneliness in older adults are well established and include cognitive and functional decline, depression, frailty, Alzheimer’s-type dementia, and mortality (Cacioppo et al., 2010; Davies et al., 2021; Holt-Lunstad et al., 2015; Lara et al., 2019; Wilson et al., 2007). Despite increased understanding of the morbidity and mortality associated with loneliness in older adults, assessment and intervention for loneliness is not commonly implemented in primary care.

The Centers for Disease Control and Prevention, World Health Organization, American Heart Association, and National Academy of Sciences, Engineering, and Medicine (NASEM) are among the national and international organizations calling for action to address social connection. Loneliness screening in primary care is a recommended best practice by these organizations; however, there are barriers to implementation, including limited time with many competing priorities. There are multiple benefits to loneliness screening in primary care, including (a) identification of individuals at risk who require intervention, (b) identification of treatable causes (e.g., hearing loss, mobility impairment, continence issues), and (c) linking individuals to existing community-based resources (NASEM, 2020). Furthermore, loneliness screening is now mandated by the Centers for Medicare & Medicaid Services in the post-acute care setting (e.g., admissions to Certified Home Health Agencies and long-term care). Primary care practices are ideal for loneliness screening and intervention because even lonely and socially isolated older adults who may be hard to reach present for care (NASEM, 2020).

Screening for social determinants of health (SDOH), such as financial strain, environmental safety, and health care access, are becoming more common, and SDOH screening is widely implemented at the Montefiore Health System where the current study took place (Fiori et al., 2020). However, loneliness screening is not included. Loneliness falls under the umbrella term of social connection, which is commonly used to encompass the ways in which human beings interact with others (NASEM, 2020). The three categories of social connection include structural (e.g., marital status, living alone), functional (e.g., social support, social engagement, loneliness), and quality (e.g., marital quality) aspects of social relationships. Loneliness is commonly defined as a subjective, negative feeling of dissatisfaction with the quantity or quality of social relationships. As such, loneliness cannot be determined by objective structural or functional measures, such as marital status, living alone, level of social support, or social engagement (e.g., their spouse regularly brings them to their primary care appointment, or they are a member of a religious institution). Thus, lack of loneliness screening represents a gap in care and requires screening using validated tools to identify individuals in need of intervention.

Social prescribing is a commonly used term to refer to linking patients with non-medical sources of support in the community for social, emotional, or practical needs (e.g., art programs, educational opportunities, exercise activities) (Chatterjee et al., 2018; Galvez-Hernandez et al., 2022). Social prescribing infrastructure varies by communities, especially as these social opportunities are often supported locally rather than nationally. The United Kingdom is an example of a country where social prescribing infrastructure has been leveraged for >30 years, referral networks are widely established, and programs are supported by the national health systems (Chatterjee et al., 2018). Benefits of social prescription include improved physical, psychological, and mental health; improved cognition, self-esteem, and social connection; and reduced primary care service use (Chatterjee et al., 2018). Referrals can be made via clinicians or through a link worker or community health worker who acts as a bridge between primary care and community resources (Chatterjee et al., 2018).

Prior studies additionally provide evidence for screening and referral support to increase utilization of services and improve health outcomes (Berkowitz et al., 2016; Fiori et al., 2020; Smith et al., 2021; Yan et al., 2022). A wide range of community-based interventions exist in the community to address loneliness, including intergenerational visits, friendly calls, volunteer and work opportunities, and older adult centers with a broad array of programming (e.g., shared meals, activities, group exercise activities), similar to those studied in other social prescribing schemes (Chatterjee et al., 2018). The variety of interventions provides options that can be tailored according to patient-specific preferences. If patients are screened for loneliness, they can be appropriately referred to community-based services to match their needs. The purpose of the current study was to (1) assess whether implementation strategies lead to increased screening, and (2) whether screening leads to increased referrals and better use of available resources.

METHOD

Design

This was a mixed-methods implementation pilot of a loneliness screening and referral program called SOCIAL Rx. As such, we focused on programmatic outcomes to assess feasibility of implementation in a primary care practice. We used the RE-AIM evaluation model (Glasgow et al., 1999) to organize outcome measures including quantitative outcomes (e.g., reach, effectiveness, adoption) and qualitative outcomes (e.g., acceptability, feasibility). The strength of a mixed methods design is that it includes quantitative data regarding program outcomes accompanied by exploration of acceptability of the program to the end-user (i.e., patient). Findings are presented using CONSORT guidelines for reporting non-randomized feasibility studies (Eldridge et al., 2016).

Study Population and Setting

The current research was conducted at a geriatric ambulatory clinic embedded within a large, urban health system in Bronx County, New York. The geriatrics practice is an academic primary care and consultative practice with eight geriatricians, one social worker, and one geriatrics fellow. Approximately 910 patients aged ≥65 years are served at the practice with an additional 503 patients seen annually for specialty consultations; 75.3% of patients at the practice are female, 32.6% identify as Black/African American, 37.5% identify as Hispanic/Latino, and 18.4% identify as White/Caucasian. Most patients are English-speaking; 74.2% speak English as their primary language and 23.5% speak Spanish as their primary language. Most (96.2%) patients at the practice are covered by Medicare (i.e., health insurance provided by the federal government for adults aged ≥65 years and those with disabilities) and 54.4% are covered by Medicaid (i.e., health insurance provided by the federal government for low-income adults and children). Primary care practice follow-up visits are scheduled for 30-minute blocks; new patient visits and specialty consultations are scheduled for 1-hour blocks. The study was approved by the Albert Einstein College of Medicine Institutional Review Board.

SOCIAL Rx Pilot Program Description

The SOCIAL Rx program was developed to address a gap within established social needs screening programs and leverage the trusted provider–patient relationship to intervene. Given loneliness screening in primary care is an established best practice, we used an implementation science approach. The SOCIAL Rx program comprised several implementation strategies designed to promote adoption of loneliness screening at the practice (Proctor et al., 2013).

Implementation Strategies: Educational Sessions/Materials.

Prior to implementation, we conducted a 1-hour focus group with geriatric providers during the weekly departmental meeting to discuss how providers were addressing social needs and barriers and facilitators of screening at the practice and used the information learned to design the study. In addition, we conducted a 10-minute educational session 2 weeks prior to implementation to review (a) why loneliness screening matters, (b) how the screening process will work, (c) how the loneliness screens will be scored, (d) what to do with a positive screen, and (e) how to document a positive screen in the electronic medical record (EMR). We compiled a curated list of online and in-person referral resources available in the community to be provided to patients identified as lonely. We posted educational flyers to inform patients that a new questionnaire would be implemented practice-wide with instructions on how to fill out the loneliness screen (Supplement A, available in the online version of this article)

Implementation Strategies: Implement New Workflow.

A screening workflow protocol was developed and implemented using an iterative approach with administrative and clinical team feedback. Development of the workflow primarily involved identifying who to screen, who would conduct the screening and when, and who would be responsible for reviewing the results and providing a referral/intervention. To streamline screening processes, we screened all patients presenting to the geriatric practice for any visit type (e.g., new patient, sick visit, consult). Patients aged <65 years who presented to the clinic for comprehensive cognitive testing were excluded. The loneliness questionnaire was provided to the patient by the research team after they checked in for their visit as a new social questionnaire that was being offered to all patients. Once completed, the questionnaire was brought to the evaluation room by the patient for discussion with the provider.

Implementation Strategies: Audit and Feedback.

Audit and feedback processes comprised informal check-ins with providers and nurses at the practice to iterate the screening process and adjust study procedures to improve screening.

Loneliness Screen

We used the 3-item UCLA Loneliness Scale (Russell, 1996). The UCLA Loneliness Scale is a valid and reliable tool (Cronbach’s alpha = 0.89 to 0.94, r = 0.73; Russell, 1996) that does not require trained staff to administer. The scale includes three questions: (1) How often do you feel you lack companionship?; (2) How often do you feel left out?; and (3) How often do you feel isolated from others? Responses were scored on a Likert scale where 1 = hardly ever, 2 = some of the time, or 3 = often. Scores ≥4 were considered lonely (Perissinotto et al., 2012). The screening was conducted on paper in English or Spanish and included an additional question to inquire whether the patient agreed to contact following their visit for follow up.

Referrals

Referral sources were compiled based on knowledge of existing community-based organizations as well as national and local sources for support and connection in coordination with the practice social worker. Referral sources were free or low cost and included virtual and in-person opportunities for connection ranging from a prescription to call a family member or friend, friendly call center, friendly visiting, a social club where individuals can share life stories, volunteer opportunities, exercise, transportation, technology services, and local older adult centers. Providers were asked to refer patients to services based on their own knowledge of patient needs, preferences, and abilities and engage in shared decision-making with the patient.

Theoretical Framework

We used an implementation science framework for the current pilot study as we were seeking to implement an intervention (loneliness screening) that is widely recommended yet not systematically implemented. We adapted the RE-AIM (reach, effectiveness, adoption, implementation, maintenance) framework to provide a structure for organizing and evaluating our outcomes. The RE-AIM framework was developed to evaluate the impact of public health and community-based interventions, and the impact of an intervention is understood to be due to its combined effects on these five evaluative dimensions (Glasgow et al., 1999). Given that this was a pilot study, we adapted the RE-AIM framework to reflect the type of data and results we expected to collect during a trial run of the program (e.g., no maintenance measure as the program was short term and no patient-facing efficacy measures, such as reduction in loneliness, as we were focused on implementation of the program). Reach refers to the proportion of individuals who receive care or are affected by the program; adoption refers to the proportion and representativeness of settings that adopt a given program; and effectiveness of a program is determined by whether the program was delivered as intended (Glasgow et al., 1999)..

Outcome Measures

Primary outcome measures for implementation of SOCIAL Rx were organized based on the RE-AIM framework (Glasgow et al., 1999). Reach of the SOCIAL Rx pilot was defined as the proportion of patients who presented to the practice for any visit type over the duration of the 2-week pilot who were screened for loneliness. Effectiveness of the SOCIAL Rx pilot was defined as the proportion of patients who screened positive for loneliness who were provided a referral for community resources. Adoption of SOCIAL Rx was measured by the proportion of providers who referred >50% of patients who scored positive for loneliness. Qualitative outcome measures, including feasibility and acceptability of the SOCIAL Rx pilot, were assessed via semi-structured interviews.

Follow-Up Calls and Qualitative Interviews

All patients who agreed to a follow-up call were contacted approximately 4 weeks after their visit to inquire whether they followed up on their referral (if applicable) and gauge interest in participating in qualitative interviews for the study. We attempted to contact all patients who agreed to contact via telephone on three separate occasions.

Semi-structured interviews were conducted to (a) examine acceptability of responding to a loneliness screen at a primary care visit, (b) examine acceptability of referral sources provided (if applicable), and (c) assess barriers and facilitators to social connection. All patients who agreed to a follow-up call were outreached for recruitment and those who were reached were invited to participate in the qualitative study. Only one participant preferred to be interviewed in person; 13 participants preferred to be interviewed via telephone. Verbal informed consent was obtained by the interviewer before conducting the interview and participants received a $25 gift card upon completion of the interview. Data were derived from in-depth interviews using a semi-structured interview guide that was iteratively developed using feedback from gerontological experts outside the research team as well as feedback from potential participants (adults aged ≥65 years living in the community). Interviews lasted 30 to 60 minutes and were conducted by two authors (C.P., J.W.). Interviews were digitally recorded and transcribed verbatim.

Data Collection and Analysis

All data used for the current study were sourced from the EMR (age, gender, daily practice schedule) and the SOCIAL Rx program data for loneliness screening/scoring and referral data, date of visit, service provider, and contact information.

Quantitative outcomes were analyzed using descriptive statistics. Results are presented as means and proportions. For qualitative analysis, transcripts were analyzed by two authors independently (C.P., J.W.) using inductive analysis by deriving codes borne out of the data (Bingham, 2023). This analysis resulted in codes that reflected topics from the interview guide (Supplement A). Once the codebook was finalized, the same two authors organized participant responses by the corresponding codes, including the most representative supporting quotes. Disagreements regarding coding were resolved by discussion. Common themes were developed by synthesizing participant responses across codes regarding loneliness screening and referral in primary care, desirable referral sources, and barriers/facilitators of connection.

RESULTS

Quantitative Outcomes

A summary of patients screened during the pilot is presented in Table 1. A total of 102 patients were screened over the 2 weeks of the study. Mean patient age was 82.2 years, and the majority (76.7%) were female. Most patients completed the screen in English (83.5%). Fifty-one (50%) screened positive for loneliness; 33.3% were categorized as somewhat lonely, and 17.2% were categorized as very lonely. These data reflected the general composition of patients at the practice.

Table 1.

Summary Characteristics of Screened Patients (N = 102)

Characteristic	n (%)
Sex
Female	82 (76.7)
Male	20 (23.3)
Lonelinessa
Not lonely (0 to 3)	49 (49.5)
Somewhat lonely (4 to 6)	33 (33.3)
Very lonely (7 to 9)	17 (17.2)
Language spoken
English	86 (83.5)
Spanish	16 (15.5)
Other	1 (1)
	Mean (SD) (Range)
Age (years)	82.2 (7.3)
Loneliness scorea	4.6 (2.04) (3 to 9)

^aMeasured using the 3-item UCLA Loneliness scale, where scores ≥4 indicated loneliness.

Implementation outcomes are summarized in Table 2. Of 126 patient visits, 102 (81%) patients were screened for loneliness. Reasons patients were not screened included patients presenting for repeat visits during the 2 weeks of the study (n = 3), inability to participate due to cognitive impairment (n = 8), declined to participate (n = 2), aged <65 years (n = 2), and missed by study team (n = 8). Of the 51 (50%) patients who screened positive for loneliness, 26 (52%) were provided referrals for community resources. Three participants were unable to receive a referral due to dementia, one was unable to receive a referral related to lack of a telephone and being homebound, and two declined. Of nine attending physicians and one geriatric fellow seeing patients at the practice, four (40%) provided referrals to ≥50% of eligible patients.

Table 2.

Implementation of Loneliness Screening and Referral Pilot

Variable	n (%)
Total in-person visits in 2 weeks	126
Implementation Outcomes
Reach/penetration (patients screened)	102 (812)
Reasons patients were not screened
Repeat patient	3 (2.9)
Unable to participate	8 (7.8)
Declined to participate	3 (2.9)
Aged <65 years	2 92)
Missed	8 (7.8)
Patients who screened positive	51 (50)
Service Outcomes
Effectiveness
Total positive patients referred	26 (52)
Declined	2 (3.9)
Adoption
Proportion of providers who referred >50% of eligible patients	4 (40)

Qualitative Outcomes

Of 102 patients screened, 57 (55.9%) agreed to be contacted after their visit and 14 (24.6%) agreed to participate in qualitative interviews. Twenty-seven (47.4%) patients were unable to be reached via telephone; 10 (17.5%) declined to participate; three (5.2%) were unable to participate due to dementia, aphasia, or hearing impairment; and three (5.2%) had missing or invalid contact information. Summary characteristics of participants who agreed to be interviewed are displayed in Table 3. Participants included 11 women and three men with a mean age of 79.1 years (range = 71 to 88 years). Most interviews were conducted in English (n = 12) and two were conducted in Spanish. Of 14 participants, six (46.2%) were not lonely, five (35.7%) were somewhat lonely, and three (23.1%) were very lonely. Themes and supporting quotes are described in Table 4.

Table 3.

Summary of Qualitative Interview Process and Participants

Variable	n (%)
Patients screened	102 (80)
Patients who agreed to contact	57 (55.9)
Patients who agreed to participate in qualitative interviews	14 (24.6)
Reasons for not participating
Unable to reach	27 (47.4)
Declined	10 (17.5)
Unable to participate	3 (5.2)
Missing/invalid contact info	3 (5.2)
Characteristics of interviewed participants
Mean age (years)	79.1
Female	10 (76.9)
Language spoken
English	11 (84.6)
Spanish	2 (15.4)

Table 4.

Patient Perspectives on the SOCIAL Rx Program: Themes and Representative Quotes

Themes/Subthemes	Quotes
Feasibility/Acceptability of SOCIAL Rx
Showing care and concern	“I feel he’s concerned like with my grieving he was very supportive and very understanding I find nothing wrong with that.”
Opportunity to open up	“I think it is important for the doctor to ask. You may need help. It was in privacy that the questions were asked so you can get good answers.” “I think it’s important for the doctor to ask what the person is going through because you know normally the doctors don’t ask these questions and sometimes people are reluctant to say anything about it.”
Acceptability/Preferences of Referral Sources
Referrals should be personalized	“The thing is that some of this is depending on a person’s ability to get out to do things...I’ve had some major health issues this past year so that I was not able to get out. I still use a walker. So, some of the questions regarding interests what you would like to do may be limited by the physical ability to do them...I really hedge my wants and my anticipation of what I can do.” “I think part of the solution is really looking into people individually as to asking them what would make them feel better. And not just saying hey, go to this, do that.”
Direct and repeated outreach would be most effective	“I get emails, it used to be letters but now it’s emails you know you’re welcome to offer this offer that and... maybe if they get like communication, like almost a barrage like,‘Oh, if you’re interested we have this service’ because then I said you know, ‘OMG...let me try this, they have this, too,’you know kind of thing.”
Feeling generative	“When I’m home I’m not doing nothing, and I’ve always been active you know I’m not doing nothing. I would go out and find someone to help.” “Going to the technology class that I’m going to now that makes me feel good because I’m learning something...that we’re eventually going to have to learn—that I like.”
Barriers/Facilitators of Social Connection
Technology as a facilitator	“I have my iPad I can reach out to you know, Zoom, different news sources to people. If someone doesn’t have any ability to reach out to a person, to a book, anything, that could be very debilitating.” “The phone rings all the time. I’m on my phone a lot.”
Community spaces and institutions as opportunities for connection	“I like going to the senior center, you know meeting new people during my exercise, and I like going to the mall even though I don’t shop just looking around.” “Like when I went to the senior center, I just sat down and started talking to this lady and told her what kind of things do they do at the senior center and she mentioned about the therapist. It was so little time that we talked I never met this lady. I told her a little about the loss that I had, and she told me that there’s hope.” “I have friends at church. Some of them are my country people. We get along. As a matter of fact, July 6 it’s going to be international day so I’m going to see a lot of people there. We have an international day where you wear...your African attire, you cook the food and you come, you invite friends.”
Environment can be a barrier or facilitator	“Every other weekend...I would invite people to my home, and I would cook, and we would sit down and eat and have a good time. I was always inviting people. But now some of them are in their seventies and they’re older and they can’t do the steps...” “I live on a 4th floor walkup. Sometimes I go up and down these stairs two and three times a day and eventually it’s going to get to me and I don’t have the means to move.” “I’m very...lucky because across the street I got the post office, the pharmacy, the supermarket, and the bodega...and I can go across with somebody that goes with me because I can’t go by myself.”
Needing a little more support	“I’ll tell you one thing that’s been a great help for me is my grandson. Because he comes over. And he fixes my computer. He updates me... So you know it’s I’m very fortunate in that he’s so supportive so helpful. If I had to do all this by myself, I wouldn’t have been able to do it.” “I need somebody in here because I’m always so lonely and I can’t go outside by myself because sometimes I’m dizzy. And I can’t walk, I could fall. I don’t go out by myself.”
Temporary relief	“When I get to the class of course I start interacting with the person who’s teaching me, the volunteer, students, young adults, and they’re very, very smart so they’ll help me, and I find in that hour I feel very relaxed and happy because I’m talking but then when I leave the technology, boom, I’m back to reality.” “I have a problem because I am alone. I have a nurse with me until 5. She is with me from 9 o’clock until 5 o’clock. After that then I am alone.”
Financial barriers	“I have six grandkids and six great-grandkids, and they all love me. I want to go down there but I have a lot of fear and phobias so I can’t take the train. And they live so far. So, I can’t really see them. It has to be like by cab and that’s a lot of money or they will come and pick me up. So I haven’t seen them.” “It’s also two ways you have to want to help yourself and the means, too. That’s very important.”
Sources of Connection and Disconnection
Religion/spirituality	“I noticed that everyone can bring their trouble to me and it’s like I’m a problem solver but I can’t call them and say you know what’s happening in my life but if I go on my knees... I say it’s midday, let me go to my prayers. When I get up from my prayers, I’m a brand-new person you wouldn’t believe.” “I’m very Catholic. Before I go to bed I sit on my bed, and I pray to God for everything, and I tell God to bless my kids you know everybody. I always do that. That helps me a lot.”
Loss of important relationships	“There’s one that dropped out of my life who was a joy to talk to because we covered so many subjects...that was a huge blow. It really was. So, there I was without the companionship and the time.” “Missing my soulmate... I miss him a lot. This weekend if he would have been alive, we would have been on the beach with two cans of beer and a sandwich, and you know we would talk and listen to music. He was my best friend.”
Being physically alone	“I really try; I shouldn’t be like this. Living alone. Because I brought my brother from Africa, I brought my son. I really tried to help people so that I wouldn’t be alone. Why at this point in my life I say I shouldn’t be alone.” “Right now, I need somebody to [ac]company me, to help me clean the house.... I do what I can you know.”

Theme 1: Feasibility/Acceptability of SOCIAL Rx.

Implementing a loneliness screen in the primary care setting was feasible and acceptable to patients. They believed the survey indicated care and concern on the part of the provider and presented an opportunity to discuss concerns a patient may not otherwise raise with the provider. As one 67-year-old woman noted, “I think it’s important for the doctor to ask what the person is going through because you know normally the doctors don’t ask these questions and sometimes people are reluctant to say anything about it.”

Theme 2: Acceptability/Preferences of Referral Sources.

Participants who received referral prescriptions from their provider generally did not recall receiving any referral and did not follow through on recommendations. Participants preferred social prescriptions that were personalized according to their preferences and abilities, rather than blanket recommendations. They also preferred to receive information regarding available resources over multiple occasions and were encouraged to take advantage of resources that were directly recommended to them during health care visits. Resources that offered the opportunity to feel generative and learn something new were preferable.

Theme 3: Barriers/Facilitators of Social Connection.

Several participants took advantage of activities offered in community spaces and institutions, such as dances, exercise classes, and communal meals at older adult centers and religious institutions. Technology also offered an opportunity to connect with others, particularly for those for whom it was difficult to leave the home. The built environment can be a barrier or facilitator for social connection and lack of accessible spaces hindered connection by making it difficult to leave the home or invite people over. Importantly, the availability of a bare minimum of support, such as somebody to go out with or somebody to assist with technology, facilitated social connection. Still, opportunities for connection were only a temporary relief from loneliness after which patients were alone again or no longer distracted. Finances were another barrier to connection, as they reduced access to other people who lived at a distance as well as the accessibility of resources.

Theme 4: Sources of Connection and Disconnection.

Religion/spirituality were a great source of comfort and when they had no one to talk to, participants would talk to God. Loss of important relationships, including widowhood or loss of friends, were a great source of disconnection. In addition, being physically alone or not having others around led to feelings of disconnection.

DISCUSSION

Our main findings were that loneliness was highly prevalent in our sample and screening and referrals for loneliness among older adults was integrated into a primary care practice setting during implementation of the SOCIAL Rx pilot study. Several barriers and facilitators of the program were identified during implementation. The program was acceptable as demonstrated by high participation as well as qualitative feedback, and screening was feasible in the setting of a designated screener. Several themes were identified related to barriers and facilitators of social connection in this population.

Prevalence of Loneliness

Our findings that 50.5% of participants were lonely at least some of time and 17.2% were very lonely is comparative to other national samples that reported loneliness prevalence of 43% in older adults (Perissinotto et al., 2012).

Implementation of SOCIAL Rx

Because few studies have addressed loneliness screening in primary care (Gunn et al., 2023), we considered reports on loneliness screening alongside several studies that reported on the implementation of social needs screening that focused on SDOH to contextualize findings from the current study (Fiori et al., 2020; Meyer et al., 2020; Yan et al., 2022; Zhang & Fornili, 2023). Most patients consented to social needs screening, which is similar to another study that implemented social needs screening in the primary care setting where >96% of patients were screened (Zhang & Fornili, 2023). The same study reported social needs screening was acceptable to patients and that patients appreciated referrals for support (Zhang & Fornili, 2023). We similarly found that the SOCIAL Rx screening and referral program was well received by patients and patients appreciated the care and concern demonstrated by the implementation of the screening.

Barriers and Facilitators of Implementation

The practice was an ideal site to pilot the SOCIAL Rx program, as providers at the practice are geriatricians who are aware of the influence of social connection and health outcomes in older adults and already engage in informal discussions of social needs with their patients. Therefore, the study provided a formalized process for screening, documentation, and reporting what they were already doing informally. The focus group and educational session prior to implementation was an important entry point to introduce the program and address questions/concerns for end-users who would be charged with implementation.

Lack of time for screening/referrals is often cited as a primary factor for low screening in primary care (Galvez-Hernandez et al., 2022). In the current study, time constraints were addressed by assigning a research assistant to screen patients and score the tool to remove the burden of screening from the provider. Patients completed the questionnaire prior to the start of the visit to minimize the time it took for the provider to address the screen with the patient. Another challenge was who will be conducting the screening and where the screening will be placed in the workflow. Given time constraints in a busy primary care setting, we designated research staff to assist with screening, similarly to other studies that used volunteers (Meyer et al., 2020) or community health workers (Fiori et al., 2020). The most challenging point to address was who would conduct the screening, even in a practice where buy-in for social screening was already high, related to time constraints and concerns regarding workload for social work or nursing staff. Studies on the implementation of SDOH screening reported similar challenges to integration of social needs screenings in practice workflows and variability in uptake across sites (Fiori et al., 2020). Studies reported integration of loneliness screening with routine clinical screening (Meyer et al., 2020) and the EMR (Fiori et al., 2020; Gunn et al., 2023; Meyer et al., 2020; Zhang & Fornili, 2023) may improve screening rates. Our long-term goal is to implement loneliness screening as part of the Medicare annual health assessment benefit, which already has an established workflow at the practice via nursing that would ensure high levels of screening related to mandated assessments. Paper-based screening lays the groundwork for integration into the EMR as it establishes feasibility and proof of concept for such screening, which was one of the goals of the current work.

Another challenge was lack of follow through with referral sources on the part of the patient. The purpose of the provider as the key implementer of the intervention was to leverage the authority of the provider as a trusted confidante and source of health information to stress the importance of loneliness intervention for health upon the patient. However, none of the patients followed through on provider recommendations. It is possible that because the program was new, program processes had not yet become routine. In addition, qualitative findings revealed that personalized referrals tailored to the individual’s physical abilities and preferences are ideal and that frequent presentation and outreaches regarding social connection opportunities and programs might incentivize or encourage individuals to take advantage of available resources. It is possible that given the opportunity to revisit referral recommendations at later visits by the provider and patient, better follow through could be achieved. Furthermore, shared decision-making with the patient in terms of recommendations may also improve uptake. Indeed, prior reviews on social prescribing in primary care agreed patient-informed content yielded increased engagement (e.g., asking individuals what they would like to do, as our qualitative results show) and that the referral should be provided by a knowledgeable provider (Galvez-Hernandez et al., 2022). It is possible that dedicated staff for referrals may improve participation and use of resources, as dedicated staff may have time to explore patient-specific preferences and provide personalized instructions and referrals. We could not assess reasons for lack of follow through as patients who received referrals did not recall receiving one. It is possible that the number of referral sources offered was too overwhelming and fewer, more specific referral sources might be better received. Finally, in the current study, lonely participants with advanced dementia were not provided with a referral given their condition, highlighting a need for available resources appropriate for this population that bears consideration in future studies. Similarly, individuals who are homebound and/or lack technology require alternative forms of social connection and also deserve further study.

Barriers and Facilitators of Social Connection

Our findings indicated free or low-cost opportunities would be crucial to overcome financial barriers to social connection. Accessibility was also a crucial facilitator as participants identified the physical environment as a key factor in their ability to participate in social connection opportunities and that distance was a barrier. Other reviews of loneliness interventions also point to affordable interventions facilitating engagement (Galvez-Hernandez et al., 2022). Accessible and/or free/low-cost transportation could also facilitate social connection and should be included in policy proposals as well as the planning and designing of interventions. In addition, financing and elevating community spaces as opportunities for connection is another policy opportunity to facilitate social connection. Many participants took advantage of activities and programs at the local older adult center and these spaces provided opportunities to connect and network with others.

An important theme we identified was that opportunities for social connection only provided temporary relief from loneliness and that feelings of connection did not endure once the individual returned to their own reality. This finding has important implications in terms of the limitations of some of the social connection interventions that are commonly proposed and highlights an area of opportunity for future research to explore effective interventions for loneliness and dosing of interventions to affect lasting impact. Given most existing loneliness interventions were investigated over the short term and generation of trust is key to participation (Galvez-Hernandez et al., 2022), this highlights an area of opportunity for future research in the field for sustainable loneliness interventions with persistent effects.

STRENGTHS AND LIMITATIONS

Strengths of our study include piloting the implementation of a recommended screening that is not commonly implemented in primary care. The study was implemented in a busy, primary care practice that serves diverse older adults, targeting a vulnerable population at risk for loneliness and associated negative health outcomes. In addition, our findings were enhanced by qualitative insights that have important implications for future research on social connection interventions in the older adult population. Implementation of the current study in a practice where buy-in was already high was an asset to the study and lessons learned regarding implementation strategies and tailoring workflows, strategies, and referrals according to practice and patient-specific needs can be leveraged for future studies on the implementation of loneliness screens and referrals in primary care.

Our study also had several limitations. First, the study was limited by the short study time frame and difficulty integrating the screening process into the practice workflow and findings bear further study. Second, most patients were not able to be reached via telephone for qualitative interviews, which may limit the generalizability of our findings as the sample was small. Third, most interviews were conducted via telephone based on patient preference, which may affect the quality of the interview. Qualitative interviews with providers were not completed for this study and may provide further insight into barriers and facilitators of screening and referrals. Finally, in an effort to adhere to the nature of a pragmatic study, we did not collect information regarding social network size, social support, marital status, living arrangements, functional status, or other information that may be informative to our results and these factors may be important to consider in future research.

FUTURE DIRECTIONS

Future directions include a larger study sample across primary care sites, conducted over a longer period. In addition, an iterative process to integrate screening and referrals into existing workflows would improve implementation and maintenance of screening and referral programs. Given that staffing resources may vary by site, these processes should be tailored according to the needs and preferences of site-specific implementers. Another opportunity for further study includes effectiveness trials of existing social connection interventions, such as use of community resources, to establish the evidence base for these interventions and inform future development of interventions for social connection.

CONCLUSION

Loneliness was highly prevalent in this population of older adults, highlighting the imperative for screening and intervention. Loneliness screening was overwhelmingly acceptable to older adults in the primary care setting and patients believed the screening offered an opportunity to confide in their provider and showed care and concern on the part of the provider. Personalized referrals that are curated according to individual needs and preferences were most preferable and direct and repeated outreaches may increase participation in social connection opportunities and programs. Public health policies and interventions should focus on providing free/low cost, accessible opportunities that leverage community spaces to facilitate social connection. Further study is needed in larger samples, over longer periods to investigate sustainability of screening and referral programs and effectiveness of personalized referrals to community-based programs as an intervention for loneliness.