Across 148 studies that followed more than 300,000 people for an average of seven and a half years, people with stronger social relationships had 50% higher odds of still being alive at the end of the study period. The effect held across age groups, gender, initial health status and cause of death.
The United Kingdom’s response to this evidence was to appoint a Minister for Loneliness in January 2018, backed by a one-off £20 million funding package announced that June. The United States Surgeon General’s response, in May 2023, was an advisory calling on “individuals and families, schools and workplaces, health care and public health systems, technology companies, governments, faith organizations, and communities” to act. No national funding program followed, and no legislation passed.
Measured against a mortality risk of this magnitude, the typical public health investment has been a strategy document and an advisory. To see why, follow a risk of this kind into health systems that have no ready category for it.
The numbers, and what they mean
That opening figure comes from the psychologist Julianne Holt-Lunstad and two colleagues, who pooled prospective studies from populations on several continents to test whether the quality and quantity of people’s social relationships predicted who survived. They expressed the result as an odds ratio of 1.50: the odds of survival for people with adequate social relationships were half again as high as for those with poor ones.
The authors set that figure beside the best-known mortality risk factors and found it comparable to the benefit of quitting smoking and larger than the effects of physical inactivity or obesity, risks that doctors already routinely screen for.
That figure is sometimes paired with the claim that lacking social connection is as dangerous as “smoking up to 15 cigarettes a day.” Vivek Murthy, then the US Surgeon General, put this comparison front and center in his May 2023 advisory, and it made headlines. It should not be repeated uncritically. In 2023, Holt-Lunstad herself and two co-authors argued that the analogy “often oversimplifies the evidence” and risks framing what is structurally produced as an individual-level problem with an individual-level solution. Their objection was to the metaphor, not the finding. The finding has a limit of its own: the studies were observational, so illness that cuts people off from others could account for part of the link. But they followed people forward in time, and the association was much the same whatever participants’ health when the studies began.
Holt-Lunstad’s team later separated the risk into three distinct constructs. Social isolation, an objectively measured lack of social contact, carried an odds ratio of 1.29. Loneliness, the subjective experience of feeling disconnected however much contact one has, carried 1.26. Living alone carried 1.32. A person can be socially isolated without feeling lonely, and someone surrounded by family can still feel an acute disconnection. A program that fixes one need not touch the other.
Three different risks
Social isolation means an objective lack of contact with others: small social networks, infrequent interaction. It is, in principle, countable. Loneliness is the subjective experience of that disconnect, the felt gap between the social contact someone has and the contact they want. The two correlate but are not equivalent. Living alone is a separate structural variable that partly captures both and captures neither fully — yet in the 2015 meta-analysis it carried the largest mortality association of the three. An intervention designed to get isolated people into rooms together will not resolve subjective loneliness if those rooms are full of ambiguous social signals the lonely person is primed to read as rejection.What isolation does to the body
Naomi Eisenberger, Matthew Lieberman and Kipling Williams put people in a brain scanner and had them play Cyberball, a simple computerized ball-tossing game in which the other apparent players stopped throwing them the ball mid-game. Though the exclusion was trivial, brain imaging showed that being shut out activated the anterior cingulate cortex, a region central to the distressing component of physical pain, and the more active it was, the more upset participants felt. Later, finer-grained imaging showed that rejection and physical pain are encoded as different patterns within these regions, so social pain is not literally physical pain. But the two run on overlapping hardware.
The usual evolutionary account of the overlap is that, for a social species, exclusion from the group was a survival threat, so a social alarm system may have come to share hardware with the physical pain system because both dealt with matters of life and death. It is a plausible story rather than a demonstrated one.
The harder question is what happens when that alarm system runs chronically. The late John Cacioppo and his colleagues described lonely people as living in a state of chronic hypervigilance for social threat.
When the team recorded electrical brain activity, the brains of lonely participants told social threats apart from non-social threats about 116 milliseconds after an image appeared. Non-lonely participants took about 252 milliseconds. The sample was small, only 19 people, but the direction fits the model: lonely brains flag social danger in roughly half the time. On Cacioppo’s account, that speed makes evolutionary sense: someone on the social perimeter needs to spot threats faster than those safely inside the group.
In its modern, chronic form this hypervigilance doesn’t switch off. It is thought to keep the hypothalamic-pituitary-adrenal (HPA) axis, the body’s primary stress response system, working overtime, and the likely consequences reach several of the systems most closely tied to early death.
Cortisol is the most visible marker. In one group of middle-aged working adults, loneliness was associated with a larger cortisol awakening response, meaning the body’s first hormonal surge of the day arrived already amplified. Lonely participants also showed bigger diastolic blood pressure reactions to acute mental stress (in women) and larger stress-induced rises in fibrinogen, a clotting and inflammatory protein linked to cardiovascular disease.
Inflammation may be the more consequential pathway. Chronic loneliness is linked to higher levels of interleukin-6 (IL-6), though the link is inconsistent across studies and shows up mainly once other factors are accounted for, and in lonely people the immune system’s gene expression tilts toward inflammation while antiviral defenses are turned down. That matters because IL-6 is implicated in cardiovascular disease, type 2 diabetes, cognitive decline and all-cause mortality. Inflammation is a leading candidate for how chronic psychological stress turns into physical illness.
Then there is sleep. Lonely people do not sleep less, but they sleep worse, with lower sleep efficiency and more time awake after first dropping off, a pattern seen both in the lab and at home. The night, which should reset the stress system, no longer fully does its job, and that sets up a second reinforcing loop.
Cardiovascular aging compounds these effects over time. Among middle-aged and older Chicagoans followed for five years, loneliness predicted steadily larger rises in systolic blood pressure, independent of other risk factors.
The trap that tightens
Cacioppo's hypervigilance model has a further consequence that explains why chronic loneliness is so resistant to resolution. People who are chronically lonely want connection, yet repeated activation of the threat-detection system has left them with implicit behavioral tendencies that make sustained connection harder to achieve: wariness of ambiguous social signals, anticipation of rejection, scanning behavior that reads neutral interactions as subtly hostile. In the lab, lonely brains told social threats from other threats in about a tenth of a second. But that vigilance produces behaviors that others read as cool, guarded or withdrawn. Those reactions make rejection more likely, and each rejection confirms the threat model and tightens the hypervigilance, so the people most entrenched in loneliness are the hardest for the usual interventions to help.The trend
The standard narrative runs like this: loneliness has been rising steadily since the 1970s, accelerated by smartphones, and is now epidemic. Some of this is accurate, and the inaccurate parts determine what, if anything, you would do about it.
The “rising since the 1970s” claim is not well supported. Long-running loneliness surveys from wealthy countries show no consistent aggregate increase in loneliness before COVID. Among Americans over 50, those born between 1920 and 1947 and those born between 1948 and 1965 report similar levels of loneliness.
Late midlife is where the data show a change. By around age 60, adults in the United States, England and Mediterranean Europe born after 1945 report higher loneliness than earlier-born cohorts did at the same age. The US gap is substantial: middle-aged Americans score 0.3 to 0.8 standard deviations higher on loneliness than their peers in every European region studied. Nordic countries with similar levels of economic development, by contrast, show no historical increases at all.
The Nordic exception is a warning against single-cause stories. The Nordic countries are among the most digitally connected societies on earth, yet their middle-aged adults show no rise in loneliness across generations. Whatever drove the American increase, technology alone cannot be the whole explanation.
Among teenagers, the picture shifted measurably from around 2012 onward: in international school surveys of 15- and 16-year-olds, school loneliness rose between 2012 and 2018 in 36 of 37 countries, in step with the spread of smartphones and heavy internet use. For adults, the most rigorous recent evidence points the same way. Among Dutch adults followed for nine years, both passive social media use (scrolling without engaging) and active use (posting and messaging) predicted rising loneliness over time. The relationship ran both ways: lonelier people went on to use social media more, and heavier use was followed by more loneliness. The frequently offered advice to “engage more actively online” is not supported by this evidence.
The Nordic countries that escaped the rise have shorter working hours, more generous parental leave, lower economic inequality, and social infrastructure such as parks, libraries and community spaces that creates incidental contact. These differences predate smartphones by decades and may help explain why smartphone adoption landed differently in populations that already had more non-mediated social time built into the architecture of daily life.
Cacioppo’s hypervigilance model suggests one reason screens may not help. Digital interaction is characteristically low in nonverbal cues, ambiguous in tone, and stripped of the multisensory context of proximity, touch, timing and breath that the social pain system was calibrated to interpret. If the threat-detection circuitry activated by loneliness evolved for embodied, face-to-face contact, digital interaction may multiply the ambiguity instead of resolving it.
The burden is real, and the evidence for it has been building since at least 2010. In other domains, that has been long enough to redirect public health infrastructure.
The policy record
In January 2018, Theresa May’s government appointed Tracey Crouch as the world’s first Minister for Loneliness. The appointment followed the recommendations of the Jo Cox Commission on Loneliness, set up by the MP Jo Cox and carried on in her name after her murder in June 2016. In October 2018, the UK published its first national loneliness reduction strategy. General practitioners were enabled to refer patients to community activities through “social prescribing” schemes, and link workers were employed to connect patients with local resources.
The money, though, was modest: a £20 million package announced in June 2018, including an £11.5 million Building Connections Fund for charities and community groups — spread across an entire country. Social prescribing ran through the existing NHS. The loneliness brief never had a minister to itself; it was added to the portfolio of the Minister for Sport and Civil Society. And the strategy deliberately set no quantitative target for reducing loneliness. Later funding rounds followed; by early 2024 ministers put the amount invested by the government and its partners since 2018 at more than £80 million.
In May 2023, US Surgeon General Vivek Murthy released “Our Epidemic of Loneliness and Isolation,” an advisory with six pillars for action. It asked for contributions from effectively every institutional actor in American society, from families and schools to technology companies and faith organizations, but brought no money or law of its own. Murthy left office with the change of administration in January 2025, signing off with a “parting prescription” that urged Americans to “choose community.” He also co-chaired the World Health Organization’s Commission on Social Connection, whose 2025 report took the issue global. In Washington, the advisory remained the high-water mark.
Compare this to the response to tobacco. In fiscal 2024 the CDC sent about $85 million to the states and the District of Columbia for tobacco control programs and quitlines, with further grants to territories and tribal organizations. Even that machinery has proved fragile: the CDC’s Office on Smoking and Health was eliminated in the April 2025 federal health layoffs. Britain, meanwhile, passed the Tobacco and Vapes Act in April 2026; from 1 January 2027 it will be illegal ever to sell tobacco to anyone born on or after 1 January 2009. Both countries have spent decades layering on taxation, advertising bans, regulation and pharmaceutical support for quitting.
By Holt-Lunstad’s benchmark, weak social ties carry a mortality risk comparable to the gain from quitting smoking. Yet nothing resembling that infrastructure exists for social connection.
Part of the reason is classification. Loneliness does not appear on death certificates. There is no pathogen to target, no product to regulate or tax, no pharmaceutical approval pathway, and no body count in the official statistics — the World Health Organization’s 2025 estimate of more than 871,000 deaths a year linked to loneliness is a modeled figure, not a tally from death certificates. It fits none of the usual categories of public health emergency. The mechanisms Western health systems have built for responding to population-level mortality mostly need either a biological agent or a commercial actor to act on, and loneliness has neither.
The other part is political economy. Each structural cause of loneliness, from overlong working hours and hostile urban design to declining access to shared community space and digital architectures optimized for engagement over connection, has powerful commercial constituencies that benefit from the status quo. Addressing these causes requires interventions in time, space, and the social conditions of daily life. Few of these interventions offer a return that a private actor can readily capture. There is no loneliness industry that can fund the political infrastructure of reform the way pharmaceutical companies fund cardiovascular research.
What the evidence supports
Ray Oldenburg first used the term “third places” in 1982, writing with Dennis Brissett, and made it famous in his 1989 book The Great Good Place. He meant spaces that are neither home nor work, where informal social contact happens by proximity rather than by appointment: bars, barbershops, parks, community centers, libraries, diners. Their design is what produces that contact. Older residents of three Stockholm neighborhoods described accessible local third places, where they could walk, eat and meet people, as vital resources against loneliness. This is perceptual evidence from a small qualitative sample of 30 people, not a measured outcome. But the logic is strong: incidental contact with weak-tie acquaintances (familiar faces and people known by sight, not close friends) is the connective tissue of everyday belonging.
Meanwhile the time Americans spent with friends fell from an hour a day in 2003 to 34 minutes by 2019, with the decline steepening markedly after 2013. Third-place closure may contribute to that contraction, though the evidence cannot yet say how much.
Beyond third places, the physical architecture of the neighborhood matters. Walkable, mixed-use neighborhoods with accessible shared public space are associated with lower social isolation. In Galway, Ireland, residents of walkable neighborhoods were significantly more likely to know their neighbors, participate socially, and report higher levels of trust than those in car-dependent suburbs. The cross-sectional evidence can’t fully rule out self-selection; people who prioritize social connection may prefer walkable environments. But each link in the chain is plausible: accessible shared space enables incidental contact, incidental contact builds weak-tie networks, weak-tie networks reduce felt isolation. Planning decisions about whether streets are walkable, whether public space is accessible and free, and whether neighborhoods have gathering places are made routinely, at every level of government, without reference to their consequences for population loneliness.
Work hours set the limit on all of this. Social connection requires discretionary time, and at the extreme a seventy-hour week with a long commute at each end leaves almost none. In 2025 the average American worker put in about 1,800 hours, several hundred more than a German or Dutch worker. National statistics count hours in different ways, so the gap is approximate, but it comes to something like nine to twelve additional 40-hour work weeks. The transatlantic loneliness gap is consistent with that structural difference: the countries with the shortest working hours and strongest leave entitlements are the ones showing no historical increase in loneliness. The causal evidence is indirect, since countries cannot be randomized to different labor regimes, but longer American working hours leave less of the unscheduled time in which connection happens.
Against these structural changes, the digital fixes look thin. In controlled trials, digital tools that only connect people or leave them to self-guided activities do little for loneliness; the digital programs that work are structured psychological interventions (especially those with a group component), group-based activities and, for older people in care homes, robotic pets. Video calling can supplement in-person contact, but if the hypervigilance model is right it is unlikely to replace it, because touch, shared space and other cues the social pain system may use to resolve ambiguity are missing.
Social prescribing and its limits
In the UK's social prescribing model, GPs refer patients to community activities, link workers help them navigate options, and outcomes are tracked. It is the most coherent institutional response to loneliness that any major health system has attempted at scale. Evaluations mostly report good news, with patients describing feeling less isolated and some schemes reporting fewer GP visits, but the UK studies in an early systematic review, which covered work up to 2016, were mostly small, mostly without control groups and all at high risk of bias, too weak to judge success or value for money. The model also has a structural weakness. Social prescribing depends on places to send patients (third places, activities, accessible venues), and that infrastructure is declining at the same time. Prescribing people to spaces that no longer exist, or that require transport they don't have, or that cost money they can't spend, leaves the problem in place.Classified as misfortune
Social connection predicts survival with an effect size that rivals quitting smoking and outstrips physical inactivity and obesity. How much of that difference any policy could win back is still unknown, and finding out is the job that funded programs and large trials exist to do. For smoking, governments built them.
For loneliness, the likeliest levers are familiar enough: working hours short enough that social life is possible, accessible shared space that is free at the point of use, and urban design that generates incidental contact rather than eliminating it. These are ordinary features of ordinary life in several high-income countries that have avoided the loneliness trajectories of the US and UK.
These features are the product of political decisions. The data has been available for more than fifteen years, and the main pathways are well mapped.
The response so far has been advisory documents and ministers with token budgets, and loneliness is still treated as a personal misfortune. Over the same years, Britain passed a law making it illegal ever to sell tobacco to anyone born on or after 1 January 2009.
Gen AI Disclaimer
Some contents of this page were generated and/or edited with the help of a Generative AI.
Media
Key Sources and References
Holt-Lunstad J, Smith TB, Layton JB. Social relationships and mortality risk: a meta-analytic review. PLoS Medicine. 2010 Jul 27;7(7):e1000316. https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1000316
Holt-Lunstad J, Smith TB, Baker M, Harris T, Stephenson D. Loneliness and social isolation as risk factors for mortality: a meta-analytic review. Perspectives on Psychological Science. 2015;10(2):227-237. DOI: 10.1177/1745691614568352. https://doi.org/10.1177/1745691614568352
Smith RW, Holt-Lunstad J, Kawachi I. Benchmarking social isolation, loneliness, and smoking: challenges and opportunities for public health. American Journal of Epidemiology. 2023;192(8):1238-1242. DOI: 10.1093/aje/kwad121. https://doi.org/10.1093/aje/kwad121
Eisenberger NI, Lieberman MD, Williams KD. Does rejection hurt? An fMRI study of social exclusion. Science. 2003;302(5643):290-292. DOI: 10.1126/science.1089134. https://doi.org/10.1126/science.1089134
Woo CW, Koban L, Kross E, Lindquist MA, Banich MT, Ruzic L, Andrews-Hanna JR, Wager TD. Separate neural representations for physical pain and social rejection. Nature Communications. 2014;5:5380. DOI: 10.1038/ncomms6380. https://doi.org/10.1038/ncomms6380
Cacioppo S, Capitanio JP, Cacioppo JT. Toward a neurology of loneliness. Psychological Bulletin. 2014;140(6):1464-1504. DOI: 10.1037/a0037618. https://doi.org/10.1037/a0037618
Cacioppo S, Bangee M, Balogh S, Cardenas-Iniguez C, Qualter P, Cacioppo JT. Loneliness and implicit attention to social threat: A high-performance electrical neuroimaging study. Cognitive Neuroscience. 2016;7(1-4):138-159. DOI: 10.1080/17588928.2015.1070136. https://doi.org/10.1080/17588928.2015.1070136
Steptoe A, Owen N, Kunz-Ebrecht SR, Brydon L. Loneliness and neuroendocrine, cardiovascular, and inflammatory stress responses in middle-aged men and women. Psychoneuroendocrinology. 2004 Jun;29(5):593-611. DOI: 10.1016/S0306-4530(03)00086-6. https://doi.org/10.1016/S0306-4530(03)00086-6
Smith KJ, Gavey S, Riddell NE, Kontari P, Victor C. The association between loneliness, social isolation and inflammation: a systematic review and meta-analysis. Neuroscience & Biobehavioral Reviews. 2020;112:519-541. DOI: 10.1016/j.neubiorev.2020.02.002. https://doi.org/10.1016/j.neubiorev.2020.02.002
Cole SW, Capitanio JP, Chun K, Arevalo JMG, Ma J, Cacioppo JT. Myeloid differentiation architecture of leukocyte transcriptome dynamics in perceived social isolation. Proceedings of the National Academy of Sciences. 2015;112(49):15142-15147. DOI: 10.1073/pnas.1514249112. https://doi.org/10.1073/pnas.1514249112
Cacioppo JT, Hawkley LC, Berntson GG, Ernst JM, Gibbs AC, Stickgold R, Hobson JA. Do lonely days invade the nights? Potential social modulation of sleep efficiency. Psychological Science. 2002;13(4):384-387. DOI: 10.1111/1467-9280.00469. https://doi.org/10.1111/1467-9280.00469
Hawkley LC, Thisted RA, Masi CM, Cacioppo JT. Loneliness predicts increased blood pressure: 5-year cross-lagged analyses in middle-aged and older adults. Psychology and Aging. 2010;25(1):132-141. DOI: 10.1037/a0017805. https://doi.org/10.1037/a0017805
Ortiz-Ospina E. Is there a loneliness epidemic? Our World in Data, 2019. https://ourworldindata.org/loneliness-epidemic
Infurna FJ, Dey NEY, Gonzalez Avilés T, Grimm KJ, Lachman ME, Gerstorf D. Loneliness in midlife: historical increases and elevated levels in the United States compared with Europe. American Psychologist. 2025;80(5):744-756 (online 2024). DOI: 10.1037/amp0001322. PMC: PMC11758892. https://pmc.ncbi.nlm.nih.gov/articles/PMC11758892/
Twenge JM, Haidt J, Blake AB, McAllister C, Lemon H, Le Roy A. Worldwide increases in adolescent loneliness. Journal of Adolescence. 2021;93:257-269. DOI: 10.1016/j.adolescence.2021.06.006. https://doi.org/10.1016/j.adolescence.2021.06.006
Roberts JA, Young PD, David ME. The epidemic of loneliness: a 9-year longitudinal study of the impact of passive and active social media use on loneliness. Personality and Social Psychology Bulletin. 2026;52(5):1111-1125. DOI: 10.1177/01461672241295870. https://doi.org/10.1177/01461672241295870
UK Department for Digital, Culture, Media & Sport. A connected society: a strategy for tackling loneliness. October 2018. https://assets.publishing.service.gov.uk/media/5fb66cf98fa8f54aafb3c333/6.4882_DCMS_Loneliness_Strategy_web_Update_V2.pdf
UK Parliament. Loneliness. House of Commons debate, Hansard, 11 January 2024. https://hansard.parliament.uk/Commons/2024-01-11/debates/AFB87B09-694F-4E6A-8356-2086F9FBDE6E/Loneliness
US Department of Health and Human Services, Office of the Surgeon General. Our Epidemic of Loneliness and Isolation: The U.S. Surgeon General’s Advisory on the Healing Effects of Social Connection and Community. HHS, 2023. https://www.hhs.gov/sites/default/files/surgeon-general-social-connection-advisory.pdf · https://web.archive.org/web/20250101013549id_/https://www.hhs.gov/sites/default/files/surgeon-general-social-connection-advisory.pdf
The Hill (via Yahoo News). Surgeon general offers ‘parting prescription’ for US. January 7, 2025. https://www.yahoo.com/news/surgeon-general-offers-parting-prescription-222500858.html
World Health Organization. Social connection linked to improved health and reduced risk of early death. News release, June 30, 2025. https://www.who.int/news/item/30-06-2025-social-connection-linked-to-improved-heath-and-reduced-risk-of-early-death
CDC Office on Smoking and Health. National and State Tobacco Control Program: Program Funding. Centers for Disease Control and Prevention, FY2024 (the CDC Office on Smoking and Health was eliminated in April 2025). https://www.cdc.gov/tobacco/php/tobacco-control-programs/program-funding.html
Truth Initiative. What federal health agency cuts mean for tobacco control. Updated April 1, 2026. https://truthinitiative.org/FederalCuts
The New York Times (via The Philadelphia Inquirer). Report on the cancellation of CDC anti-smoking campaigns and the Office on Smoking and Health. July 6, 2026. https://www.inquirer.com/news/nation-world/trump-cdc-antismoking-ads-campaigns-cancel-funds-staff-cut-20260706.html
UK Department of Health and Social Care. Tobacco and Vapes Bill becomes law. GOV.UK, April 29, 2026. https://www.gov.uk/government/news/tobacco-and-vapes-bill-becomes-law
UK Department of Health and Social Care. Smoking and vaping: guidance, policy and regulation (Tobacco and Vapes Act 2026 implementation timeline). GOV.UK. https://www.gov.uk/government/collections/smoking-and-vaping-guidance-policy-and-regulation
Oldenburg R, Brissett D. The third place. Qualitative Sociology. 1982;5(4):265-284. DOI: 10.1007/BF00986754. https://doi.org/10.1007/BF00986754
Oldenburg R. The Great Good Place. Paragon House, 1989.
Jing Jing, Dahlberg L, Canter D, Plater-Zyberk E. The role of third place concerning loneliness in the context of ageing in place: three neighbourhoods in Stockholm. Health and Social Care in the Community. 2024. DOI: 10.1155/2024/4172682. https://doi.org/10.1155/2024/4172682
Kannan VD, Veazie PJ. US trends in social isolation, social engagement, and companionship — nationally and by age, sex, race/ethnicity, family income, and work hours, 2003-2020. SSM — Population Health. 2022;21:101331. https://pmc.ncbi.nlm.nih.gov/articles/PMC9811250/
Leyden KM. Social capital and the built environment: the importance of walkable neighborhoods. American Journal of Public Health. 2003;93(9):1546-1551. DOI: 10.2105/AJPH.93.9.1546. https://doi.org/10.2105/AJPH.93.9.1546
OECD. Average annual hours actually worked per worker (DF_AVG_ANN_HRS_WKD). OECD Employment and Labour Market Statistics, 2025 data. https://sdmx.oecd.org/public/rest/data/OECD.ELS.SAE,DSD_HW@DF_AVG_ANN_HRS_WKD,/USA+DEU+NLD+DNK+NOR+SWE…………?startPeriod=2022
Hansen T, Johansen R, Kirkøen B, Stene-Larsen K, Straiton M, Tornes RA, Reneflot A. Digital bridges to social connection: a systematic review and meta-analysis of digital interventions for loneliness and social isolation. Internet Interventions. 2025;41:100856. DOI: 10.1016/j.invent.2025.100856. https://doi.org/10.1016/j.invent.2025.100856
Bickerdike L, Booth A, Wilson PM, Farley K, Wright K. Social prescribing: less rhetoric and more reality. A systematic review of the evidence. BMJ Open. 2017;7(4):e013384. DOI: 10.1136/bmjopen-2016-013384. https://doi.org/10.1136/bmjopen-2016-013384
