Assessing patients’ social risks: what can England learn from emerging evidence in the US?

INTRODUCTION

Recognition that social factors, such as income and housing, influence health outcomes has led healthcare systems to invest in improving people’s social conditions. In England, a range of national policies encourage NHS organisations to act on social factors that shape health and inequalities.¹ Since 2020, GP contracts have included funding for ‘social prescribing link workers’ in local practices — staff to connect patients with community activities and services that could help address unmet social needs, such as food assistance or benefits advice.²

Identifying patients’ social needs is an important part of social prescribing, but guidance on how to assess for social needs in general practice is limited, and there are no national guidelines for practices to follow. Variations in practice and lack of robust evaluation of social prescribing schemes in England mean we are not certain that existing interventions are reaching patients with the greatest social needs, or effectively helping them.^3,4

One option to help target social prescribing interventions is to develop more systematic approaches to assessing patients’ social needs — for instance, by standardising social risk assessments and screening all patients in a practice. This would mean introducing some kind of ‘social risk assessment’ in which patients are asked questions about select social issues, such as social isolation or food insecurity, and responses are recorded in medical notes. This is an approach already being taken in the US, where multiple major healthcare organisations have developed standardised social risk assessments and several federal policies incentivise using those instruments for more universal screening.^5,6

The idea is that more standardised and widespread screening for social risks reduces potential bias in who gets asked, as well as stigma for patients, and enables more information to be collected to support person-centred care. But there are also potential downsides, including adding to general practice workload and medicalising social issues that GPs lack tools to address. There are also question marks over whether systematic screening is feasible, or acceptable for patients.

Social prescribing is already being delivered across England but high-quality research to guide its implementation is lacking. Understanding evidence from countries testing different approaches could help inform the development of social prescribing in England. This paper summarises the state of the evidence on social risk screening in the US, drawing on a recent systematic review from the US-based Social Interventions Research and Evaluation Network (SIREN) research group,⁷ and considers implications for policy and practice for the NHS in England.

US EVIDENCE ON SOCIAL RISK SCREENING

Tools for social risk screening

Social screening tools are sets of standardised questions used to assess patients’ social risks. Identification of social risks can then be used to inform treatment decisions and target interventions (such as food bank referrals). Although some tools assess a single domain, such as food insecurity, we focus on tools used in clinical settings that include assessment of two or more social domains. There is no consensus on which domains should be included, but many common tools include domains related to housing, utilities, and food security.⁸ For example, in 2018 the Centers for Medicare and Medicaid Services developed a 10-item screening tool with questions on housing instability, food insecurity, transportation access, utility insecurity, and interpersonal violence.⁹

Validity of screening tools

Understanding whether existing screening tools reliably and accurately capture relevant information about patients’ social context is important. It helps us to identify which tools to use when, how to interpret results, to understand their limitations, and to plan further research. A systematic review found that, although generally favourable pragmatic evidence is available for 21 screening tools, a lack of psychometric testing made it impossible to fully evaluate tool validity.¹⁰ Importantly, no information was provided regarding if or how measures of validity varied by race, ethnicity, or language. An update to this research found that no screening tools had undergone ‘gold standard’ validation, and that only one (an evaluation of the SINCERE tool) attempted to explore whether validity varied by race or ethnicity (it didn’t).⁷ The authors conclude that limited testing of psychometric and pragmatic validity of different screening tools leaves important evidence gaps in guiding which tools to use.

Acceptability of screening to patients and caregivers

Social risk screening can only be widely adopted if patients and caregivers find it acceptable. In the past decade, 18 US studies have explored a total of 3933 patient or caregiver perspectives on multi-domain social screening.⁷ Collectively there are some methodological limitations to these studies: many have small sample sizes (range: 7–1161 participants), and few use experimental study designs. While social risk screening was generally felt to be acceptable, perceptions of acceptability were influenced by a range of factors:

• Setting: patients recruited from settings where most people had public insurance (often used as a proxy for lower income) were most likely to find social risk screening acceptable;

• Trust: patients’ degree of trust in their usual provider influenced the perceived acceptability of social risk screening;

• Content: most people felt that screening for food insecurity, housing (stability and quality), social isolation, transportation, financial strains, employment status, and immigration are acceptable; and

• Demographics: the acceptability of screening did not consistently differ by the race, ethnicity, or gender of responders.

Only two of the 18 studies reported that most people did not find social risk screening acceptable. Both studies were small (one had only 10 patients), and only one explored why. Concerns focused on the potential for stigmatisation (particularly if not everyone gets screened) and on data documentation storage and privacy.

Acceptability of screening to healthcare providers

Significantly more studies have looked at provider perceptions of screening acceptability than patient perspectives on screening (73 versus 18). SIREN found that these studies mostly occurred in primary care settings and were largely of poor quality (with small sample sizes a particular concern).⁷ Participating providers (which included physicians, nurses, allied health professionals, and clinic managers) generally reported feeling that social risk screening in healthcare settings is acceptable, appropriate, and has the potential to improve health outcomes. However, opinions differ around who in the healthcare workforce should conduct social risk screening (for example, physicians, nurses, or support staff). There are also no studies comparing different approaches to frequency of screening (although annual screening is the most common approach).

Provider concerns about screening fall into four categories: concerns about time; provider discomfort with screening; worries that topics raised might cause patients to feel stigma, damaging patient–provider relationships; and concerns that providers lack knowledge or resources to address identified risks. But there is some evidence to suggest that provider concerns about social risk screening reduce after training and education about screening or after participation in a screening programme.

Screening implementation

Integration of social risk screening tools into clinic processes is a key determinant of how useful they are. But few studies report on factors shaping the successful implementation of screening — for example, studies might report that adoption of screening is higher in one population over another without exploring reasons why. SIREN found that implementation research on social risk screening is not robust: studies are typically small and descriptive. Of 30 studies describing factors affecting implementation of social risk screening, common facilitators to screening included: clear communication with patients and caregivers about screening rationale and process; regular communication with the healthcare team about screening progress and processes; and training healthcare teams on empathic enquiry and trauma-informed care.⁷

Common barriers reported by providers included staffing and time, but few studies have assessed how long screening takes. Only five studies examined this question, finding significant variability (1–9 minutes) in the time required for screening. We also know very little about whether it matters who conducts the screening. Only one study compared social risk disclosure rates across different health professional groups, finding that people were more likely to disclose social risks to community health workers than to nurses.

DISCUSSION

Policymakers in the US and UK are promoting social prescribing as a way to identify and address patients’ social risks in clinical settings. Healthcare systems in the US and UK vary widely in structure, funding, and context. Bringing information about patients’ social risks into a market-based system such as the US’s has particular risks.¹¹ But assessing patients’ social risks — via a structured approach or ad hoc questions from providers — is a core component of any social prescribing intervention. Evidence from the US suggests that social risk screening is generally feasible and acceptable but the quality of evidence on screening is not rigorous and several evidence gaps remain. So, what can we learn from the existing evidence on social risk screening to inform policy and practice on social prescribing in the English NHS?

The answer isn’t clear cut. The NHS already employs around 2600 social prescribing link workers in England.¹² This is a significant investment, but the number of link workers per head of population remains low, and these resources are not targeted at the most deprived areas.¹³ Even within areas, it is not clear how patients are prioritised for social prescribing — and there are risks of inequities if some patients are asked about social risks and others aren’t. Introducing universal social risk screening could be one way of targeting limited resources more effectively and ensuring that social interventions reach those with the greatest needs.

Better data on patients’ social needs could also help to improve medical decisions by giving clinicians a better understanding of patients’ social context. For example, knowing that a patient with diabetes has no access to a fridge may nudge doctors to choose medication that can be stored at room temperature — increasing the chance of adherence with recommended treatment.

But introducing social risk screening across general practice is a big ask. General practice in England is already overloaded and the service is at crisis point. Social risk screening doesn’t need to be carried out by GPs, but it would likely add to overall practice workload. One way to limit this would be to screen only certain population groups (for example, those with multiple medical conditions). This might undermine potential benefits of universal screening, but would at least ensure that patients are asked a common set of questions linked to defined social needs.

Policymakers must consider whether developing and implementing social risk screening through general practice represents the best use of limited resources. Politicians are asking general practices to focus on improving access to appointments, and there is already an extensive list of evidence-based interventions that GPs and their teams struggle to deliver.¹⁴ Would the time and resources needed to deliver universal social risk screening be better spent on other priorities? US evidence suggests that, although systematic screening is possible, there is no strong evidence that screening leads to use of the acquired data (in the form of interventions), or that interventions lead to improved health.¹⁵

Investing in social risk screening and referrals also rests on the assumption that resources to address people’s social needs are available, just poorly connected to people who need them.¹⁶ Yet there have been major cuts in spending on local government and other social services in the UK over the last decade.¹⁷ GPs in England already report asking patients about a mix of social and economic needs — particularly housing issues, social isolation, and domestic violence — but identify inadequate staffing and lack of follow-up from social and community services as barriers to coordinating support.¹⁸ Screening in clinical settings also risks medicalising patients’ social needs by framing structural drivers of health as individual problems that can be diagnosed and treated.

Another approach is to use area-level data on social and economic needs as a substitute for person-level screening. This would mean using area-level data to prioritise investments in social prescribing — for instance, where to hire link workers. But a recent study in the US found that area-level data missed around 40% of self-reported social deprivation.¹⁹ This is unsurprising, given that area-based deprivation indices have limited ability to identify individuals who are income or employment deprived.²⁰

WHERE NEXT?

Social prescribing has been implemented quickly across England, but policy enthusiasm has outpaced evidence on impact. More systematic assessment of patients’ social risks might help to target social prescribing interventions. But there are potential risks and opportunity costs to introducing social risk screening in the English NHS, and the evidence base is weak. A sensible starting point for meaningful UK research would be to identify, test, and evaluate social risk assessment tools in the NHS. Policymakers should avoid rushing to a(nother) national roll-out before establishing a clear evidence base.

Funding

Dr R Fisher is a Harkness Fellow, supported by the Commonwealth Fund. The views presented here are those of the authors and should not be attributed to the Commonwealth Fund or its directors, officers, or staff.

Provenance

Freely submitted; externally peer reviewed.

Competing interests

Professor LM Gottlieb is founding co-director of the Social Interventions Research & Evaluation Network where Dr E De Marchis works as a researcher. Dr R Fisher and Mr H Alderwick have no competing interests to declare.