Polygenic risk scores in the clinic: a systematic review of stakeholders’ perspectives, attitudes, and experiences

Abstract

Polygenic Risk Scores (PRS) are statistical methods estimating part of an individual’s genetic susceptibility to various disease phenotypes. Their potential clinical applications to enhance the prediction, prevention, and risk management of complex conditions motivate current research efforts worldwide. While a growing body of literature has highlighted the scientific and ethical limitations of PRS, the technology’s clinical translation will present both opportunities and challenges for the stakeholders involved. Here, a mixed-method systematic review of empirical studies was performed to gather evidence on the perspectives, attitudes, and experiences of healthcare providers, patients, and the public regarding the use of PRS in healthcare settings. The PRISMA reporting protocol was followed and 24 articles were included. Three major themes were identified. First, we reported on participants’ familiarity with the test, including their knowledge, understanding, and education on PRS’ clinical use. The second theme collects stakeholders’ motivations for taking the test and their perspectives on sensitive issues related to the return of results. Participants’ normative stances regarding the appropriate use of PRS, their benefits, and harms were presented in the third theme. The findings underscore significant knowledge gaps and challenges in the clinical interpretation of PRS among healthcare providers. On the other hand, the provision of genetic counseling benefitted patients’ understanding of PRS results and in most cases, no psychosocial burden was reported. Finally, the review highlights that stakeholders’ perspectives on the clinical use of PRS are highly context-dependent, shaped by population characteristics, disease type, and social factors, emphasizing the need for tailored approaches across diverse healthcare settings.

Introduction

Polygenic Risk Scores (hereafter: PRS) are statistical protocols that allow for the quantification of part of an individual’s DNA-based susceptibility to a phenotype [1]. First hopes regarding the potential clinical applications of PRS trace back to 2009, when researchers showed the importance of polygenic variation (multiple genes with potentially small individual effects) in the genetic susceptibility to schizophrenia [2]. Genome-wide association studies (GWAS) have since been instrumental in identifying these small-effect variants, mapping the polygenic architecture of complex diseases, and providing the foundational data for PRS development [3]. Since then, enthusiasm has grown about using PRS as a clinical precision medicine tool for complex diseases (e.g., common cancers, cardiovascular diseases, diabetes).

From a clinical point of view, the risk of developing a complex condition is attributed to multiple genetic and environmental factors. Some patients might exhibit an apparent familial predisposition due to inherited monogenic variants that act as a significant, sometimes determinant, risk factor. Genetic testing is available for these phenomena, and a substantial part of modern clinical genetics is dedicated to offering genetic testing and counseling. However, the rarity of these cases in a population does not explain the heritability of complex conditions [4, 5], suggesting that numerous genetic factors, each with a small effect, collectively contribute to the risk.

Understanding and quantifying the impact of small-effect size genetic variants on disease phenotypes via PRS is, therefore, of clinical interest both at individual and population levels. Should PRS reach clinical utility, proponents underline the benefits of enhanced disease prediction and early disease detection, targeted prevention, and risk-management strategies [6]. Current efforts are directed at integrating classic PRS methods - weighted linear combinations of singular genetic effects – into more complex risk prediction models considering the diversity of factors at stake in the etiology of complex conditions [7, 8]. Notable examples come from the Breast Cancer (BC) field, whereby PRS have been incorporated into existing clinical risk prediction models (e.g.: Tyrer-Cuzick, Can Risk) to improve risk stratification [9]. Clinical trials such as the WISDOM study, for instance, compare the efficacy of a Personalized Breast Cancer Risk Assessment (PBCRA) to the standard of care in BC screening [10]. Other large-scale research initiatives exploring the clinical use of PRS are currently on the way in the USA (e.g.: eMERGE4 network) [11]. As common complex diseases continue to be major causes of morbidity and mortality worldwide and in Europe, integrating genomic medicine approaches into clinical practice has also become a key focus of national healthcare systems. In this framework, the UK research program Our Future Health plans to generate PRS of the study participants and offer them personalized feedback [12].

While PRS clinical applications have sparked considerable interest, the technology’s clinical value is still debated due to epistemic limitations. Classic PRS, for instance, do not inform about an individual’s absolute risk of developing a specific disease, and efforts to validate PRS models in non-European ancestry groups have only started. Therefore, using PRS in clinical contexts to inform disease prediction, prevention, and treatment strategies raises important ethical and social questions for stakeholders. These questions revolve around three major themes [13]. First, clinical PRS may cause harm to individuals due to uncertainties surrounding the clinical interpretation and communication of the test relative measure and the possibility of risk shifts [6, 14, 15]. It has been suggested that interpretation failures may lead to overdiagnosis, overtreatment, or false reassurance. Second, social and economic factors are key determinants of common diseases [16]. According to part of the existing literature, public health initiatives that target such determinants should not be underscored in favor of approaches that target only a part of the population at high genetic risk and that can access genomic technologies [17]. The poor transferability of PRS to non-European ancestry groups is also a major reason for health-equity concerns around their clinical applications [18, 19]. Third, it is still unclear what best practices clinicians should endorse in consultation settings, including personal and familial counseling [13, 19].

While professional societies have developed position statements [20–22], the normative gaps surrounding the clinical use of PRS are reflected in the lack of guidelines on the appropriate use of the technology. Understanding stakeholders’ perspectives is essential to health policy, as insufficient knowledge of stakeholders’ attitudes and underlying values undermines the elaboration of guidelines and policies that are sound, empirically informed, and thus widely acceptable in pluralistic societies [23]. Considering PRS technological developments and worldwide research efforts oriented to their clinical translation, stakeholder research focusing on healthcare providers, patients, and the public has recently started. Here, we consider these three major stakeholder groups and systematically review the existing empirical literature, assessing their perspectives, attitudes, and experiences toward the clinical use of PRS as susceptibility tests for common complex diseases. To our knowledge, this represents the first systematic review conducted on the topic.

Methodology

Search string and outcomes

We conducted a mixed methods systematic review following the Joanna Briggs Institute (JBI) approach to mixed methods systematic reviews [24, 25] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [26]. PubMed, Scopus and Web of Science electronic databases were systematically searched on 20 November 2023. The search string comprised three categories of terms relating to ‘polygenic risk scores’, ‘attitudes’, and ‘stakeholders’. The concept-associated terms reflected the databases’ specific controlled vocabulary and relevant keywords, as found in the literature. The string syntax was adapted to each database (available as Supplement 1).

In total, we retrieved 1439 records. Results were merged, and duplicates (n  =  623) were removed via the software Rayyan [27]. A PRISMA flow diagram for the publication retrieval and screening process is provided in Supplement 2.

Eligibility criteria

To be eligible for inclusion, studies had to meet the criteria reported in Table 1.

Table 1.

Inclusion and exclusion criteria.

Included	Excluded
Qualitative and quantitative original empirical research studies	R1: Non-empirical research studies (reviews, position papers, normative guidelines) R2: Empirical observational studies
Studies that inquire about stakeholders’ (HPs, patients, the public) attitudes, views, perspectives, and decision-making factors	R3: Studies that do not mention stakeholders’ attitudes, views, perspectives, and decision-making factors in their research questions
Studies that focus on clinical Polygenic Risk Scores	R4: Studies that focus only or primarily on other types of predictive tests (carrier and presymptomatic in monogenic contexts), or whole genome sequencing
Studies that focus on the use of PRS for the prediction and treatment of common complex diseases in healthcare systems	R5: Studies that focus on • prenatal contexts or • Direct To Consumer context or • the use of PRS for the prediction of socio-behavioral traits
Peer-reviewed articles	R6: Non-academic literature
English publications	R7: Non-English publications

Quality appraisal

The first author critically evaluated the included studies using the Mixed Methods Appraisal Tool (MMAT) [28] (table available in Supplement 3). Since the quality appraisal is intended to be informative rather than evaluative, we did not exclude any study based on their methodological quality.

Data analysis and synthesis

First, the included studies were read multiple times. After having familiarized with the data, the research team concluded that the results of both qualitative and quantitative studies were similar enough to allow for a convergent integrated approach to data analysis and synthesis, as outlined in Stern et al. (2020) and reported in the JBI Manual for Evidence Synthesis [24, 25]. Therefore, data were first extracted from the results sections of each study. Afterward, quantitative data were ‘qualitized’, namely described in a narrative form. The process of ‘qualitization’ refers to the conversion of quantitative data ‘into themes, categories, typologies, or narratives’ [24, 29]. We chose the data qualitization format proposed in Heyvaert et al. (2016) [29].

Inspired by Braun and Clarke (2021), an inductive thematic data analysis of both qualitative and ‘qualitized’ data was conducted. First, extracted data were assigned a code describing the content of the data. Second, similar codes were categorized into broader categories, based on similar meanings. Finally, we developed three themes, here conceived as ‘patterns of shared meaning, united by a central concept or idea’ [30]. Each theme was further characterized into sub-themes. Conflicts about the categorization of summary topics under a specific theme or subtheme were verbally discussed by all the team members until resolution. Finally, our integrated findings were reported in a narrative form.

Results

Studies characteristics

The 24 included studies were published between 2017 and 2023. Qualitative research methods were employed in 12 studies, while 10 studies used quantitative methods, and 2 studies used a mixed-methods design (Fig. 1). Patients’ and public perspectives were investigated in 14 studies, and those of Healthcare Providers (hereafter: HPs) in 11 studies (Fig. 1). Two studies focused on a mix of stakeholders. Most studies were conducted in Australia (n = 10) and the USA (n = 9) (Fig. 2). The majority of the included studies explored stakeholders’ perspectives regarding the clinical use of PRS in the cancer clinic, with 10 studies focusing on breast cancer. In 5 publications, the use of clinical PRS was considered generally, without a focus on a particular condition, while three studies explored stakeholders’ attitudes toward psychiatric PRS (Fig. 3). Further study characteristics were collected and reported in detail in Table 2.

Fig. 1.

Study population and research methodology.

Fig. 2.

Country of research.

Fig. 3.

Conditions examined.

Table 2.

Studies’ characteristics (ordered per research method and year).

AUTHOR	YEAR	TITLE	COUNTRY	CONDITION	RESEARCH METHOD	DATA COLLECTION	POPULATION	SAMPLE CHARACTERISTICS	RECEIVED THEIR PRS	RECEIVED COUNSELING
Smit et al. [31]	2021	Knowledge, views and expectations for cancer polygenic risk testing in clinical practice: A cross-sectional survey of health professionals.	Australia, Canada, USA	Cancer, General	Mixed method	Cross sectional survey	HPs (N = 105)	1) Genetic counselors (84%); 2) oncologists (6%); 3) clinical geneticists (4%); 4) other (7%)	N/A	N/A
Saya et al. [41]	2022	Informed choice and attitudes regarding a genomic test to predict risk of colorectal cancer in general practice	Australia	Cancer, Colorectal	Mixed method	Questionnaire + semi-structured interviews	Patients (N1 = 150; N2 = 16)	General practice patients	YES	YES
Young et al. [36]	2017	Making Sense of SNPs: Women’s Understanding and Experiences of Receiving a Personalized Profile of Their Breast Cancer Risks	Australia	Cancer, Breast	Qualitative	Semi-structured interviews	Patients (N = 39)	Diagnosed patients who received a negative BRCA1/2 result	YES	YES
Forrest et al. [39]	2019	High-risk women’s risk perception after receiving personalized polygenic breast cancer risk information	Australia	Cancer, Breast	Qualitative	Semi-structured interviews	Patients (N = 39)	Patients with personal and family history of breast and ovarian cancer with no pathogenic variant in high BC risk genes	YES	YES
Yanes et al. [35]	2020	Women’s responses and understanding of polygenic breast cancer risk information	Australia	Cancer, Breast	Qualitative	Semi-structured interviews	Patients (N = 21)	Diagnosed and unaffected patients with family history of BC and no pathogenic variant in high-moderate BC risk genes	YES	YES
Putt et al. [40]	2020	Exploration of experiences with and understanding of polygenic risk scores for bipolar disorder	Australia	Psychiatric Disorders	Qualitative	Semi-structured interviews	Patients (N = 18)	Diagnosed patients of white European ancestry (based on genetic principal components analysis), fluent English speakers	YES	YES
Sierra et al. [49]	2021	Exploring Implementation of Personal Breast Cancer Risk Assessments	Australia	Cancer, Breast	Qualitative	Focus group	Public, patients (N = 31)	1) Unaffected women; 2) unaffected women with a strong family history of BC but no familial BRCA1 or BRCA2 pathogenic variant; 3) unaffected women with a pathogenic variant in either the BRCA1 or BRCA2 genes	NO	NO
Woof et al. [55]	2021	Introducing a low-risk breast screening pathway into the NHS Breast Screening Program: Views from healthcare professionals who are delivering risk-stratified screening.	United Kingdom	Cancer, Breast	Qualitative	Focus group; Semi-structured interviews	HPs (N = 28)	HPs working in breast screening and primary care	N/A	N/A
Willis et al. [38]	2021	Influence of lived experience on risk perception among women who received a breast cancer polygenic risk score: ‘Another piece of the pie’.	Australia	Cancer, Breast	Qualitative	Semi-structured interviews	Patients (N = 40)	Unaffected women with family history of BC	YES	YES
Lewis et al. [33]	2022	Patient and provider perspectives on polygenic risk scores: implications for clinical reporting and utilization	United States	General	Qualitative	Semi-structured interviews	1) Patients (n = 25); 2) HPs (n = 21)	1) Patients enrolled in biobank. Self-reported ancestry: Asian, Black, Hispanic/Latinx, White; 2) Primary Healthcare Providers	NO: received a mock report; N/A	NO; N/A
Terek et al. [43]	2022	Attitudes among Parents towards Return of Disease-Related Polygenic Risk Scores (PRS) for Their Children.	United States	T2D; Asthma; Obesity	Qualitative	Semi-structured interviews	Parents (N = 40)	Parents/legal guardians of children enrolled in research institution/biobank. Self-reported ancestry: Black/African American or Hispanic	NO: received a mock report	NO
Suckiel et al. [44]	2022	Perspectives of diverse Spanish- and English-speaking patients on the clinical use of polygenic risk scores.	United States	General	Qualitative	Semi-structured interviews	Patients (N = 30)	Biobank patients. Self-reported ancestry: African, African American, or Black (AA) or as Hispanic/Latinx (H/L), and at least half reporting Spanish as their preferred language.	NO	NO: were presented with a digital educational module before interview
Riddle et al. [34]	2023	The role of polygenic risk scores in breast cancer risk perception and decision-making	United States	Cancer, Breast	Qualitative	Semi-structured interviews	Patients (N = 24)	Unaffected patients with family history of BC and no pathogenic variant in high-moderate BC risk genes	YES	YES
Pacyna et al. [50]	2023	Examining the Impact of Polygenic Risk Information in Primary Care	United States	General	Qualitative	Semi-structured interviews	Patients (N = 19)	Primary care patients	NO	NO
Laza-Vasquez et al. [37]	2022	Feasibility and Acceptability of Personalized Breast Cancer Screening (DECIDO Study): A Single-Arm Proof-of-Concept Trial	Spain	Cancer, Breast	Quantitative	Questionnaire	Patients (N = 387)	Women who did not receive a previous diagnosis of breast cancer, or were undergoing a breast study, or fulfilled clinical criteria for cancer-related genetic counseling	YES	YES
Pereira et al. [32]	2022	Psychiatric polygenic risk scores: Child and adolescent psychiatrists’ knowledge, attitudes, and experiences.	United States	Psychiatric Disorders	Quantitative	Survey	HPs (N = 960)	Child and adolescent psychiatrists	N/A	N/A
Hollit et al. [48]	2022	Attitudes Toward Glaucoma Genetic Risk Assessment in Unaffected Individuals	Australia	Glaucoma	Quantitative	Survey	Patients and Public (N = 193)	1) Unaffected first- degree relatives of individuals with a known glaucoma diagnosis; 2) optometry patients; 3) members of the general community without an ocular health history	NO	NO
Lapointe et al. [47]	2022	Polygenic risk scores and risk-stratified breast cancer screening: Familiarity and perspectives of health care professionals	Canada	Cancer, Breast	Quantitative	Survey	HPs (N = 453)	Physicians (22.3%) and nurses (69.75) working in family medicine or oncology or other fields of medicine	N/A	N/A
Kamp et al. [45]	2022	Clinicians’ Perceptions towards Precision Medicine Tools for Cardiovascular Disease Risk Stratification in South Africa.	South Africa	Cardiovascular Diseases	Quantitative	Survey	HPs (N = 109)	Consultant clinicians with specialization in internal medicine, e.g.: cardiology, endocrinology (19.,3%) or non-internal medicine (52.3%) e.g.: pediatrics, clinical genetics; Trainees (24.8%)	N/A	N/A
Venning et al. [54]	2022	Preferences for a polygenic test to estimate cancer risk in a general Australian population	Australia	Cancer, General	Quantitative	Survey	Public (N = 1002)	General adult population	NO	NO
Ayoub et al. [46]	2023	Risk-Stratified Breast Cancer Screening Incorporating a Polygenic Risk Score: A Survey of UK General Practitioners’ Knowledge and Attitudes.	United Kingdom	Cancer, Breast	Quantitative	Survey	HPs (N = 109)	General practitioners	N/A	N/A
Moorthy et al. [52]	2023	How do experts in psychiatric genetics view the clinical utility of polygenic risk scores for schizophrenia?	United States	Psychiatric Disorders	Quantitative	Survey	HPs (N = 276)	1) Researcher in the field of mental and behavioral health (84.8%); 2) Psychiatrist providing direct patient care (12%); 3) Genetic counselor providing direct patient care (1%); 4) Other mental health professional (2.2%)	N/A	N/A
Vassy et al. [53]	2023	Perceived benefits and barriers to implementing precision preventive care: Results of a national physician survey.	United States	General	Quantitative	Survey	HPs (N = 367)	Primary Care Providers practicing family medicine, general practice, or internal medicine	N/A	N/A
Casauria et al. [42]	2023	Australian parental perceptions of genomic new-born screening for non-communicable diseases.	Australia	General	Quantitative	Survey	Parents (N = 40)	Parents with at least one-born child under 18 years	NO	NO

Themes overview

The results have been conceptualized in 3 broad themes: (1) Preparedness of Stakeholders, (2) PRS in the Consultation Room and (3) Evaluating PRS. Each of the three themes is further characterized into sub-themes, as shown in Fig. 4. The first theme considers stakeholders’ familiarity with PRS from a theoretical and practical point of view and their educational needs. The second theme collects stakeholders’ attitudes (or experiences) toward concrete issues emerging in a consultation setting, such as their motivations to take the test, their preferences regarding reporting results, and their perspectives on the impact of PRS results on relatives. Finally, the third theme reports on stakeholders’ normative orientations toward the test use, its benefits, and harms at individual and population levels. A detailed description of the sub-themes is given in Table 3, while in Supplementary 4 some direct and qualitized quotes are reported to illustrate the themes/subthemes.

Fig. 4.

Themes and sub-themes overview.

Table 3.

Description of themes and subthemes.

Preparedness of stakeholders	Participants’ familiarity with PRS, their preparedness to the test clinical use
Knowledge and Interpretation	Participants’ understanding of the concept of PRS, its clinical meaning and capability to interpret a PRS graph
Confidence and Experience	HPs attitudes towards using the test in clinical practice and previous experiences
Education	HPs educational needs and educational preferences
PRS in the Consultation Room	Participants’ attitudes and/or experiences toward concrete issues emerging in consultation settings before and after taking the test
Motivations for and against testing	Motivations that would prompt a patient to accept or decline a PRS test
Emotional reaction	Patients’ emotional responses to receiving a PRS report (real or mock)
Reporting and Disclosure	Participants’ attitudes toward the reporting of risk information (absolute vs relative; binary vs continuous) and the report design. Patients and public preferences toward the modality of disclosure of results
Risk Management	Participants’ perspectives on risk management interventions following the delivery/reception of PRS results
Implications for Family Members	Participants’ understanding of and attitudes toward the relevance of PRS results to family members
Expert intermediation	Patients and public preferences regarding the intermediation of experts in the delivery of the test
Evaluating PRS	Participants’ normative orientations (value judgments and desirability statements) toward the test use at individual and population level
Appropriate use	Participants’ perspectives toward the value of the test in clinical practice
Perceived benefits, potential harms, and disparities	Participants’ opinions regarding clinical and personal benefits of the test, potential harms, and contribution to health disparities
Desirability in Population Risk Stratification	Participants’ attitudes toward the use of PRS as population risk stratification tools to inform public health screening programs

Preparedness of stakeholders

Knowledge and interpretation

Healthcare providers

The knowledge and interpretation skills of HPs were assessed in 7 of the included records. Most HPs in 4 studies had low knowledge and familiarity with the concept of PRS or risk-stratified screening (1–4). In the study of Kamp et al., however, the knowledge score of almost all participants improved after watching an educational lecture, and previous genetic training was found to be an important influencing factor in HPs’ knowledge (4). Better levels of understanding of the concept of PRS were found in 3 studies [31–33]. In one study, almost all HPs could correctly define a PRS. Nonetheless, none of the participants was familiar with the technical aspects of a PRS, such as calculation and assessment. Poor knowledge of polygenic inheritance was also reported [31]. Interpretation challenges emerged in Pereira et al., whereby only a minority of HPs could interpret a PRS graph (1). Where HPs understood and correctly interpreted the measure of relative risk used in the risk report, however, participants used different heuristics to interpret the score holistically, resulting in different conclusions and recommendations [33].

Patients and public

The theme recurs in 12 records. When patients received counseling, a good understanding of polygenic risk and its implications was reported [34–41]. This finding was consistent among patients who received their PRS as part of a Personalized Breast Cancer Risk Assessment (PBCRA) [34–39]. In 2 studies, women understood the mechanisms of polygenic inheritance and were able to recall the transmission of SNPs as both maternally and paternally inherited [35, 36]. Being aware of the multifactorial nature of their disease helped participants contextualize PRS information within their family history [35, 38].

Low levels of familiarity with the concept of PRS and interpretation failures were reported in other studies [33, 42–44]. Parents or legal tutors presented with a hypothetical PRS report of their children understood that a high PRS corresponded to increased genetic susceptibility and that the conditions tested were multifactorial [43]. However, absolute risk percentages were not universally understood, with some interpreting high-risk results as low-average [43]. Similarly, not-high-risk results were commonly misinterpreted as low-risk in another study [33]. Furthermore, the percentile risk information was mistaken for percent chance also by high numeracy patients [33].

Confidence and experience

Healthcare providers

The confidence and experience of HPs with clinical PRS were assessed in 6 studies. In general, HPs did not feel confident using PRS in clinical consultations. HPs felt the least confident recommending the test, discussing insurance implications (4), or interpreting results (4). However, in 2 of these studies, HPs felt comfortable initiating discussions about PRS with patients or explaining polygenic inheritance (2). For instance, more than half of HPs had discussed the possibility of polygenic testing with cancer patients in one study [31]. In another context, involvement in medical research and previous experiences with risk stratification positively influenced HPs’ confidence in PRS [45]. In general, only a few HPs in two studies have requested or returned a PRS to a patient (1,5).

Education

Healthcare providers

Many HPs did not feel prepared for the integration of PRS in clinical care and felt the need for additional training to improve their interpretation skills and confidence in using PRS in their practice [31, 33, 46, 47]. Online learning formats were generally preferred [46, 47], focusing on the interpretation of PRS and the basics of risk stratification rather than the ethical, legal, and social aspects. Almost half of the participants in Smit et al. highlighted the primary role of primary providers in the implementation of clinical PRS and suggested that genetic health professionals act as trainers of non-genetic professionals [31].

PRS in the consultation room

Motivations for and against testing

Patients and public

Participants’ motivations to undertake or refuse the test were found in 5 studies [35, 36, 40, 41, 48]. Overall, participants were favorable to testing. In 2 records, patients would seek PRS information based on the assumption that knowledge is power, regardless of the content of the information [35, 40]. In 3 studies, affected patients would get tested to find an explanation for their diagnosis of Bipolar Disorder (BD) or BC [35, 36, 40], which some recalled as a ‘traumatic event’ [35]. In Putt et al., patients who received a diagnosis of BD were motivated by the possibility of discovering the truth about themselves [40]. Some patients felt that testing was the right thing to do as a matter of individual and familial responsibility [35, 41] as well as to take action to reduce risk and aid research [41]. The fear of developing a disease was also found to be a strong motivation to undergo the test in unaffected individuals [48].

Many participants in one study reported privacy concerns and insurance implications as motivations for not testing [41]. Some appealed to the right not to know negative things about the future, while others referred to the poor predictive value of the test compared to other clinical metrics [41]. Disinterest in the test was also associated with practical burdens such as follow-up appointments, potential costs, and psychological impact [48]. In one study, a few believed that a psychiatric PRS was not useful for older people or people without children [40]. Other minority motivations for not testing were the lack of enough evidence to propose a genetic model of BD and the negative impact of PRS on self-identity and recovery. For a few participants, PRS information would reinforce patients’ identification with a diagnostic label [40].

Returning results: emotional reactions

Patients and the public

In 6 studies, patients reacted to their personal PRS results [34–36, 38–40]. These studies found that receiving high PRS had no significant negative impact on patients’ emotional states. Initial negative emotional reactions, on the other hand, were generally followed by a positive experience as patients came to accept the new information [35, 40]. For some, high-risk results confirmed previous suspects or provided relief, psychological, and moral reassurance [36, 40]. In one study, for instance, some participants felt ‘emotionally satisfied’ to learn that SNPs are inherited both maternally and paternally, as ‘it relieved women of the (sole) responsibility for transmitting [BC] risk to their children’ [36]. Similar results were reported by Putt et al., whereby - after having had the time to process high-risk results - some participants reported feelings of relief, removal of blame, and guilt, as the test confirmed their diagnosis [40]. Finally, the emotional impact of receiving PRS results was found to be stronger in BC patients, as the high-risk report made them re-experience the trauma of their diagnosis and confirmed their increased risk for contralateral BC [35].

Only a minority of participants had an emotional reaction to a mock risk report [33]. Parents who received a mock PRS report of their children felt initially overwhelmed, as they did not have prior experience with the test. Seeing the words ‘high risk’ and the colors associated with the report, made participants feel a high level of concern for their children [43]. However, once they were able to read through the report, several reported feeling at ease and empowered to help their child reduce their risk.

Returning results: reporting and disclosure

Healthcare providers

The sub-theme is investigated in one study [33]. Here, HPs expressed a strong preference for risk reports displaying absolute risk estimates [33]. In the report, HPs would also want to see a distinction between the genetic attributable risk and other risk factors. Very few HPs preferred a binary risk report. Many were concerned that the cut-off typical of a binary report could cause false reassurance in patients just below the risk threshold. Furthermore, they thought it was paternalistic not to share continuous results if the information was available. Many also expressed the desire for a report showing multiple risk categories. Several HPs, finally, emphasized the importance of the report design to structure the doctor-patient interaction and help HPs “walk [patients] through it” [33].

Patients and the public

Patients expressed their preferences for an absolute risk report, as they found absolute risk information more familiar and relatable than relative risk [33, 43]. Almost all patients in one study preferred a continuous risk report over a binary one, which, for some, lacked nuance [33]. Continuous reports were perceived as empowering, with most participants stating that knowing their exact risk number would enable them to alter it [33]. Finally, patients were positive about using visual communication tools to facilitate understanding of the report [33, 39].

Most patients in 2 studies would prefer to have their PRS results disclosed during a face-to-face consultation, as it allows time for reflection, support, and counseling [40, 44]. Electronic disclosure methods were chosen for being time and cost-efficient [44, 48]. However, patients also specified that their preferred method would vary according to the level of risk reported, with electronic disclosure methods being acceptable in case of low risk, and face-to-face consultation available on request [44, 48]. In Sierra et al., patients who had previously undertaken a monogenic test reflected on their personal experiences and highlighted the cons of in-person disclosure, such as showing vulnerability or being too shocked to ask relevant questions. Therefore, some suggested delivering PBCRA first via letter or phone call, followed by an in-person consultation [49].

Returning results: risk management

Healthcare providers

Participants in 2 studies were asked to consider possible interventions in front of a high PRS [32, 33]. Some HPs stated that they would act on evidence-based guidelines, which are currently missing [33]. Most psychiatrists would increase monitoring or would evaluate for symptoms [32]. In both studies, only a few HPs would consider referring patients to a clinical geneticist, and the majority were skeptical about prescribing medications [32, 33]. In the study of Pereira et al., furthermore, it was found that HPs who had good objective knowledge of PRS—assessed via the interpretation of a PRS graph—were less likely to prescribe medications [32].

Patients and the public

Most participants in 3 studies stated they were willing to act based on PRS information [43, 50] to reduce their disease risk or follow the recommendations following a PBCRA [37]. Similarly, for other patients receiving PRS results increased their awareness of the importance of ‘keep doing the good job’, even for low-risk patients [38, 40]. This attitude was often justified by the multifactorial nature of the disease in question (BC) and the inherent uncertainties regarding its development.

In the included studies, 3 risk-reducing interventions were considered. Behavioral change was regarded as a desirable outcome of PRS testing [50] and the most preferred intervention by parents to prevent the onset of Non-Communicable Diseases at an early age [42]. The impact of PRS on lifestyle behaviors elicited reflections on the role of autonomy in risk mitigation [49]. In this case, patients concluded that should PRS information be available to the population at large, individuals would still preserve their decisional power regarding their lifestyle choices, similarly to other risk information. Finally, personal, financial, and access barriers to behavioral change were highlighted in two studies [43, 44]. Taking medications to reduce risk was regarded as undesirable in the studies examining the topic [34, 50]. Patients expressed a general skepticism towards the sufficiency of PRS to justify the intervention and concerns about potential side effects. Some considered taking preventive medications as ‘too extreme’ and argued that the risks might outweigh the potential benefits of risk reduction via other means. The possibility of prophylactic surgery was examined in 3 studies [39, 49, 50]. In 2 of them, the genetic information provided by a PRS is not considered a sufficient metric to decide on prophylactic surgery. On the contrary, BRCA1 or BRCA2 carriers stated that the information provided by the PBCRA may influence their decisions to undergo a prophylactic bilateral mastectomy and oophorectomy [49].

Returning results: implications for family members

Healthcare providers

Views and perceptions regarding the implications of PRS findings to family members were not deeply investigated in HPs’ studies. In one study, one participant showed concerns about the unclear implications of PRS for disease inheritance; similarly, uncertainties regarding the communication of this type of risk information in a family context were reported [31]. At the time of decision making, almost no HPs would advise testing the relatives of an asymptomatic patient with a high PRS, while half of them would consider the option for family members of diagnosed patients [32].

Patients and the public

In 3 of the included studies, patients had a general understanding of SNPs inheritance as ‘randomly inherited by each parent’ [35, 36, 40] and were aware that other family members’ risks could not be directly inferred from their test results. In one study, many participants were comforted by this knowledge and hoped that relatives would have a lower BC risk [51]. The personalized nature of the risk assessment was not as evident to participants in other studies [34, 43]. In Riddle et al., the implications of PRS results to family members were often unclear to participants, who explored the possibility that ‘family members could have the same thing’ [34]. Similarly, parents considered PRS’ results of one child beneficial to understanding the potential risk for siblings [43].

Returning results: expert intermediation

Patients and the public

Most participants in 4 studies expressed preferences for testing with or receiving their PRS results from their Primary Healthcare Providers (PHPs) [42–44, 50]. Participants stated their trust in their PHPs and their skills [43]. The importance of receiving appropriate counseling via expert intermediation was also underlined [37, 43]. Some reservations regarding the preparedness of PHPs were shown in one study [50]. Finally, some patients reported a mismatch between the recommendation of taking preventive medications received in the context of a clinical trial, and the unwillingness of their PHPs to prescribe them. For this reason, patients considered consulting a specialist [34].

Evaluating PRS

Appropriate use

Healthcare providers

According to many HPs in 2 studies, PRS have a complementary value [33, 52]. In Lewis et al., HPs considered PRS ‘a natural extension of their practice’ and thus would use them similarly to other risk information. In the same study, some HPs mentioned that PRS could have an instrumental value and be used to nudge patients into virtuous health behaviors. Others, however, questioned the acceptability of such a practice. In 2 of the included records, HPs thought that PRS could support decision-making in its ‘gray areas’ [33, 53], such as adjusting screening frequency in high-risk patients. In psychiatry, many experts thought that PRS could improve other clinical metrics in their field [52] and believe that PRS will be clinically useful in the future [32, 52]. Despite the optimism, family history was generally considered a better risk predictor [52].

Patients and public

In one study, most participants believed that PRS are valuable disease prevention tools under certain conditions: first, PRS should provide information that cannot be ascertained otherwise; second, their results should be actionable and possibly corroborated by other (non-genetic) tests [50]. Some participants tied the value of PRS to the availability of risk-reducing interventions and disease type, arguing that some PRS may be more personally useful than others [44]. The type of disease predicted was also found to be a valuable dimension of the test [54].

Perceived benefits, potential harms and disparities

Healthcare providers

In 2 studies, most HPs believed that the benefits of PRS would outweigh their harms by enhancing early disease detection, more personalized monitoring, and treatment [31, 33]. A consideration of the harms of PRS appears in 7 records [31–33, 45, 46, 52, 53]. In general, HPs in these studies were concerned about potential harms deriving from the unclear implications of PRS for individual risk, the lack of practice guidelines to inform decision-making, the possible negative psycho-social consequences on patients, insurance discrimination, and the lack of transferability of PRS across populations. Some psychiatrists mentioned the possibility that PRS elicit negative societal consequences such as stigma and discrimination [32]. While psychiatric experts in another study were less concerned about stigmatization, many highlighted the uncertainties regarding the impact of PRS on family relationships [52]. Finally, in two studies, the transferability of PRS across different ancestry groups was found to be a top concern of clinicians [45, 52], with most HPs stating that PRS should not be implemented until they are validated in different populations [52].

Patients and public

In a few studies, patients were asked to reflect on the prospective benefits of PRS [40, 44]. Their answers referred to both clinical and personal benefits. Clinical benefits such as the possibility for early and accurate diagnosis, appropriate treatment, and personalized disease management were cited [40, 44]. In this context, some patients reflected on their negative experiences of misdiagnosis and mismanagement [40]. Among the personal benefits, patients reported reducing self-blame and stigma within psychiatric disorders [40] as well as potential family and community benefits [44]. For instance, some participants thought that PRS could shed light on the risk associated with diseases prevalent among certain racial and ethnic groups [44].

Insurance discrimination, superannuation, and consequences on employment opportunities were identified as potential harms of PRS by patients [40, 49]. Similarly, the Australian public expressed clear preferences for a test that did not impact life insurance eligibility and premiums [54]. In Suckiel et al., on the other hand, most health insurance concerns regarded coverage and possible access disparities [44]. Out-of-pocket costs were a top concern for other participants, together with psychological distress [48]. The potential for PRS and PBCRA to cause anxiety and fear was a topic of consensus in two studies [44, 49]. Some participants believed they could experience fear in front of a high PRS [44]; others considered that it would be ’better not to know’ their PBCRA, depending on an individual’s personality [49]. Self-fulfilling prophecies were also mentioned as an unintended consequence of PRS, although by a minority of participants, in two studies [40, 44]. Structural barriers such as socio-economic factors affecting equal access to clinical PRS, the availability of post-test medical care, and the ability to follow PRS-related medical recommendations were considered important barriers in one study [44]. Concerns regarding the predictive accuracy of PRS in non-European ancestries were mentioned by two participants [44].

Desirability in population risk stratification

Healthcare providers

In 3 studies, HPs showed positive attitudes towards the use of PRS as a risk-stratification tool in population screening [45, 46, 55]. HPs in these studies agreed that high-risk patients should have earlier and more frequent screening. However, none of them were comfortable with the possibility of reduced screening pathways for low-risk patients. HPs in Woof et al. were concerned about missing interval cancers and appealed to the responsibility of physicians to prevent such harm [55]. Despite positive attitudes towards the theoretical possibility of risk-stratified screening, none of the participants in one study felt that their National Healthcare System (NHS) was ready to implement risk-stratified screening [46]. HPs have reported a series of implementation concerns, such as the need to enhance the NHS by increasing the number of GPs and genetic counselors and investing in the professional training of HPs [45, 46, 55]. Other structural barriers such as implementation costs [45] and the need to address socioeconomic disparities in access to screening were also mentioned [46]. Finally, in the study of Smit et al., HPs’ opinions on the acceptability of incorporating PRS into prevention and screening population programs were divided. The country of HPs’ practice was found an influencing factor in their attitudes, with Australian HPs being the most supportive, compared to their peers in the USA and Canada [31].

Patients and public

The role of PRS in risk stratification to inform public health screening initiatives is examined in 5 of the included records. In these studies, participants were generally favorable to the idea of extending PRS testing to the public to personalize population screening programs [34, 35, 37, 49, 50] in the context of a PBCRA. However, participants unanimously disagreed with the possibility of reducing screening frequency for individuals at low risk and argued that baseline screening should be offered to everyone [35, 37, 49]. When compared to existing screening programs, women with a family history of BC were open to having a PBCRA [49]. However, the same women also felt satisfied with the current screening protocols. On the contrary, most women in one study expressed their preference for a personalized screening over the current standard of care [37]. In Riddle et al., some participants thought that additional screening (MRI) to the annual mammography for individuals at high risk was not always necessary, underlying potential risks and possible barriers such as insurance coverage [34].

Discussion

We conducted a mixed-method systematic review of empirical studies to gather the best evidence regarding the attitudes, perspectives, and experiences of HPs, patients, and the public, all relevant stakeholders involved in the prospective clinical use of PRS.

First, our results show that many HPs do not feel prepared to integrate PRS in clinical practice. Poor preparedness of HPs was reflected in low knowledge, interpretation, and confidence with the scores. These findings are consistent with analogous research on physicians’ genomic literacy and familiarity with new genomic tests [56, 57]. Our results also highlight a difference between (theoretical) knowledge and interpretation capabilities of HPs. Knowledge of the definition of PRS did not automatically correspond to the capability of interpreting a PRS graph and inferring appropriate clinical conclusions. The included studies of Lewis et al. and Pereira et al. have shown the practical implications of the interpretation challenge, as HPs’ clinical recommendations and risk management choices varied depending on their level of preparedness or their preferred interpretation strategy [32, 33]. At present, evidence to use PRS to guide medical management is limited. In the absence of clinical guidelines, the American College of Medical Genetics and Genomics states that ‘clinicians using PRS should make efforts to standardize their management approach and, when possible, contribute to efforts that analyze clinical outcomes of prospective PRS-based management’ [20].

Although the preparedness of health personnel was sub-optimal in the included studies, patients and public preferences for testing lie principally with non-genetics specialists such as general practitioners and primary healthcare providers. Such findings highlight the critical importance of training a variety of professional figures involved in the clinical management of common complex diseases while genomics becomes increasingly more mainstream. Siermann et al. have recently emphasized the need to educate non-specialists as a matter of shared collective responsibility across all the phases of a test clinical use, from the decision to order a test to the disclosing of the test results [58]. As pointed out by the authors, evidence indicates that ‘physicians with a higher understanding of genetics seem to be more hesitant to order a genetic test’ [58] This hesitancy may stem from heightened scrutiny of the specific circumstances under which genetic testing is considered appropriate and useful, indicating that deeper knowledge may lead to a more cautious approach in clinical decision-making [59].

Another relevant aspect emerging from our results is the impact of genetic counseling on patients’ complex understanding of PRS. Our findings suggest, prima facie, that receiving appropriate genetic counseling before testing benefits patients’ capability to ponder the informational value of a PRS and its implications for personal and familial risk. These results, however, should be read in light of the characteristics of the study sample. In 7 out of 8 studies, participants who received genetic counseling were highly educated, with a family or personal history of disease and/or had been previously exposed to genetic testing [34, 36, 38–41, 51]. As PRS are proposed as risk stratification tools, further research is needed to explore the reception of PRS and genetic counseling services in the general population, with attention to low-literacy and inexperienced patient groups.

Another point worth consideration is the psychosocial impact of receiving PRS results. Contrary to what was hypothesized in the normative literature [13], most patients in the included studies did not experience significant psychological harm following the return of their PRS. Such findings align with the empirical knowledge on the psychosocial implications of predictive genetic and genomic testing [60, 61], whereby severe symptoms such as depression, increased anxiety, and suicidal tendencies have only been found in Huntington Disease gene carriers. As pointed out by Wade, however, ‘acknowledging that adverse psychological responses are likely to be rare and moderate should not negate the responsibility of clinicians to remain vigilant for negative effects on their patients’ [61]. Negative emotions such as confusion and anxiety, on the other hand, were often reported in studies conducted in DTC contexts, whereby PRS results were delivered without the intermediation of a HP [62, 63]. These findings are consistent with those here reported, suggesting that genetic counseling is an important mitigator of potential adverse psychosocial effects related to the reception of polygenic risk information.

Based on our results, further generalizations about the psychosocial impact of PRS are complicated by the heterogeneity of populations and conditions considered in the included studies. This is especially true considering that some disease characteristics - such as severity, actionability, onset, and the distinction between somatic and psychiatric conditions – shape the psychosocial burden of receiving risk susceptibility information [64]. Therefore, future research should clarify the moral ramifications of clinical PRS per groups of diseases that share similar clinically and morally relevant features [64]. Disease characteristics might also play an important role in the ethical evaluation of PRS under the premise of genomic contextualism [65], the idea that ‘new genomic tests are evaluated in terms of their similarities and differences from existing tests, rather than simply being considered as a genetic test’ [41]. In this framework, it has been proposed that PRS of diseases with, for instance, little implications for lifetime risk differ greatly from predictive tests for high-pathogenic variants. Instead, in specific contexts, PRS could be compared to other (phenotypic) risk information and thus elicit similar moral considerations [41, 65]. Finally, moral considerations and attitudes related to the clinical use of PRS may also vary depending on non-clinical but socio-economic, geographical, and regulatory contexts. For instance, the limited transferability of PRS and the need for risk models tailored to African populations were found to be key concerns of South African HPs [45]. Similarly, harm-related and equity concerns due to transferability issues in people of Pacific and Maori descent were reported among surveyed HPs in New Zeeland [31]. The presence or absence of a system of universal care may also function as an influencing factor of people’s views and concerns. USA patients who received a PBCRA, for instance, reflected on the costs of additional screening to conclude that they would not change their screening regimen unless covered by their insurance [34].

The last reflection prompted by our results regards the scope of PRS. In several studies, it was observed that some patients placed the personal utility of PRS in the moral and psychological reassurance provided by the test results as they confirmed their diagnosis. Similar experiences were described in Lowes et al. by consumers who sought their PRS for confirmatory or explanatory purposes, notably in the context of a history of or diagnosed psychiatric conditions. In the same setting, furthermore, many participants declared to have resorted to DTC-PRS because of ‘unmet healthcare needs’ and dissatisfaction with a healthcare system failing to address their symptoms [62].

While these results shed light on test uses that differ from the predictive and preventive goals typically ascribed to PRS, they highlight the need for greater scrutiny of the purpose of several uses within the healthcare system and across a variety of clinical scenarios. The reviewed literature exhibits a deficiency of critical reflections in this sense. Discussions tend to assume the test purposes or focus on practical applications; a notable example is the consensus that PRS should not be used in isolation but rather in conjunction with other clinical risk factors [33, 53]. However, determining the scope of a new genetic test is a normative process involving value judgments – grounded in scientific evidence—regarding its purpose and appropriate use. In the case of PRS, future research efforts should be directed to make these considerations explicit and timely to inform clinical guidelines.

Limitations

The present review includes only publications written in the English language. The included studies, furthermore, have been predominantly conducted in Western countries. The review did not discuss how the geographical area, the socio-historical context, and the participants’ background and/or demographics might have shaped their views. The included studies are those retrieved until the latest database search update, corresponding to the starting of data analysis. Therefore, further advancements in the field made after data collection might have been overlooked. Finally, database search and screening of the records were conducted only by one author (LA).

Author contributions

LA: Conceptualization; methodology; data analysis; data curation; writing—original draft; review and editing. HP: Conceptualization; data analysis; writing—review and editing. KVS: Conceptualization; data analysis; writing—review and editing. Kris Dierickx: Conceptualization; methodology; data analysis; writing—review and editing.

Funding

This work has received funding from the European Union Horizon 2020 research and innovation program: “Precision Medicine at the interface of Translational Research and Systems Medicine (TranSYS)” (H2020-MSCA-ITN-2019).

Data availability

The data generated and analyzed during the review are included in the article and its supplements.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41431-024-01747-z.