Cardiovascular risk communication strategies in primary prevention. A systematic review with narrative synthesis

Abstract Aim To evaluate the effectiveness of cardiovascular risk communication strategies to improve understanding and promote risk factor modification. Design Systematic review with narrative synthesis. Data sources A comprehensive database search for quantitative and qualitative studies was conducted in five databases, Cumulative Index to Nursing and Allied health Literature (CINAHL), Medical Literature Analysis and Retrieval System Online (MEDLINE), EMBASE, Applied Social Sciences Index and Abstracts (ASSIA) and Web of Science. The searches were conducted between 1980 and July 2019. Review methods The systematic review was conducted in accordance with Cochrane review methods. Data were extracted and a narrative synthesis of quantitative and qualitative results was undertaken. Results The abstracts of 16,613 articles were assessed and 210 underwent in‐depth review, with 31 fulfilling the inclusion criteria. We observed significant heterogeneity across study designs and outcomes. Nine communication strategies were identified including numerical formats, graphical formats, qualitative information, infographics, avatars, game interactions, timeframes, genetic risk scores and cardiovascular imaging. Strategies that used cardiovascular imaging had the biggest impact on health behaviour change and risk factor modification. Improvements were seen in diet, exercise, smoking, risk scores, cholesterol and intentions to take preventive medication. Conclusion A wide range of cardiovascular risk communication strategies has been evaluated, with those that employ personalized and visual evidence of current cardiovascular health status more likely to promote action to reduce risk. Impact Future risk communication strategies should incorporate methods to provide individuals with evidence of their current cardiovascular health status.


| INTRODUC TI ON
Cardiovascular disease is the leading cause of death and disability globally (World Health Organisation, 2017). An ageing population and increases in cardiovascular risk factors, such as obesity, are exacerbating the problem (Timmis et al., 2018). It is estimated that as many as 80% of these deaths are preventable (World Health Organisation, 2017), highlighting the importance of prevention in reducing the number of unnecessary deaths and the burden of cardiovascular disease. Nurses form the largest health professional group managing cardiovascular risk factors (Hayman et al., 2015) and thus have a key role to play in the prevention of cardiovascular disease. Identifying individuals who would benefit from risk factor management is challenging due to the insidious nature of atherosclerosis, which is often advanced before symptoms develop (Cooney et al., 2009). Cardiovascular risk prediction assessments aid health professionals in clinical decision-making about preventative treatment and are also used to inform individuals about their risks (Rossello et al., 2019). An abundance of research exists on how to accurately predict cardiovascular risk, but attention is needed on how best to inform individuals of that risk (Ahmed et al., 2012). Individuals can only make informed decisions around risk reduction if they fully comprehend their risk.

| Background
Risk communication is the open, real-time exchange of information, advice and opinion between experts and those at risk to improve understanding and facilitate informed decisions about clinical management (Thomson et al., 2015). It is a cornerstone of cardiovascular screening and should enhance a person's knowledge and perception of risk, allowing them to make informed decisions (Ahmed et al., 2012).
Individuals who are better informed about their cardiovascular health are more likely to adhere to preventative measures and may have better outcomes (Thomson et al., 2015). Information must be credible, clear and easy to understand (Ahmed et al., 2012) to avoid potential misinterpretation of risk and suboptimal choices about treatment.
Furthermore, poor communication can also reduce confidence in health professionals and lead to anxiety and other adverse outcomes.
Many different strategies exist to communicate cardiovascular risk to individuals including numerical formats, qualitative information, visual representation or a combination. Recently, health professionals have been providing feedback from medical imaging to communicate risk information to individuals (Hollands et al., 2010).
With such a pivotal role in managing risk factors, nurses involved need to be confident and skilled communicators. Understanding the effectiveness of available strategies will enable nurses to select the best approach for cardiovascular risk discussions. Quantitative studies Adarkwah et al. (2019), Germany

F I G U
Compare the effects of presenting a cardiovascular risk to patients and their subsequent adherence to intervention using 10-year risk illustration in the decision aide software Arriba (emoticons) and newly developed time to event illustration Prospective randomized trial n = 294 patients who GPs wanted to discuss behaviour change with Bonner et al. (2015), Australia Test the effect of heart age on psychological and behavioural outcomes compared with 5-year absolute risk is low-(i.e., 5year absolute risk of a CVD event <10%) to moderate-risk (10%-15% absolute risk) patients Randomized 2 × 3 factorial design n = 469, non-diabetic, not known to be high risk of CVD, no anti-hypertensives or lipid lowering medication Damman et al. (2018), Netherlands (1) Evaluate the effects of infographics about qualitative risk dimensions either with or without risk numbers on risk comprehension (2) Investigate what type of qualitative risk dimension (causes, timeline or consequences) can be best emphasized in infographics.
(3) Test effects of heart age compared with traditional risk number on risk comprehension Controlled experimental 2 × 2 n = 727, target population of cardiovascular risk calculators Domenech et al. (2016), Spain Test the hypothesis that knowledge of the genetic risk score (GRS) in uncontrolled hypertensive patients would improve BP control A randomized, singleblind cohort study in two parallel groups n = 67 patients with uncontrolled ambulatory BP (24 h-ABPM >130/80) Fair et al. (2008), UK Test the hypothesis that responses to coronary heart disease (CHD) risk estimates are heightened by the use of ratio formats, peer group risk information and long time frames.
Cross-sectional, between factors design n = 1480, general population French et al. (2004), UK Examine the emotional and cognitive impact of personal and social comparison information about health risk Observational factorial design n = 970, 40-60 years with no history of heart disease Frileux et al. (2004), France Explore the impact of the preventive medical message on the intention to change behaviour.
Observational factorial design n = 150 unpaid volunteers with no history of heart disease Johnson et al. (2015), USA

| Aims
The aim of this systematic review was to identify existing cardiovascular risk communication strategies and to evaluate their acceptability and effectiveness to improve understanding and promote risk factor modification in asymptomatic individuals without known cardiovascular disease.

| Design
A systematic review of quantitative and qualitative studies was chosen to permit a more complete analysis and maximize findings. The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) (Moher et al., 2009) and the Enhancing Transparency in Reporting the Synthesis of Qualitative Research (ENTEREQ) (Tong et al., 2012) guidelines (Appendix S2). The review protocol was registered with an international register of systematic reviews (PROSPERO ID: CRD42020204797).

| Search methods
The systematic review search was guided by the population, intervention, comparison and study design (PICOS) criteria (Appendix S3).

| Search outcome
After duplicates were removed, a total of 16,613 titles and abstracts were identified and screened. The full-text manuscripts for 210 studies were reviewed and a total of 31 (20 quantitative and 11 qualitative) met the inclusion criteria ( Figure 1).

| Quality appraisal
A quality assessment was undertaken for included studies by one reviewer (SS) and a second reviewer (AF) appraised 30% of the studies to ensure consistency (supplementary material online). The quality appraisal was based on criteria from the Critical Appraisal Skills Programme (CASP) tools for qualitative, randomized controlled trials, case-controlled studies and cohort studies.

| Data abstraction
Separate quantitative and qualitative data extraction forms were developed to collect data from eligible studies and included: location, study design, participant characteristics, data collection methods, risk communication strategy and outcome data. The data were then visually presented in a table.

| Data synthesis
Due to the heterogeneity in study designs and outcomes, a metaanalysis was not feasible. A narrative synthesis, which adopts a textual approach to summarize the findings of systematic reviews, was performed for the quantitative studies in accordance with the Economic and Social Research Council (ESRC) methods programme guidance on narrative synthesis in systematic reviews (Popay et al., 2006). The data extraction forms were used to produce a descriptive summary, organizing data by communication strategy. The study outcomes and results were then tabulated (  Figure 2 and Table S4).

| Study quality
Two of the qualitative studies scored low overall using the CASP risk appraisal tool and the remaining studies were of medium quality. The components with the poorest scores were rigour of data analysis, clarity of statement of findings and design. Two of the studies (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014;Hill et al., 2010) failed to describe the qualitative approach taken. Overall, most of the quantitative studies addressed a clearly focussed issue, considered all important outcomes and had results in keeping with existing evidence. Studies scored lower in recruitment strategies and generalizability of the results. This was attributable to several biases including recruitment of a majority or all-male sample (Powers et al., 2011;Ruiz et al., 2013Ruiz et al., , 2016) and participants at low cardiovascular risk (Bonner et al., 2015;Lopez-Gonzalez et al., 2015). A small proportion of studies scored low in minimizing bias while measuring outcomes (French et al., 2004) and reasons include using research-developed tools as opposed to validated ones (Adarkwah et al., 2019).

| Numerical formats
This describes quantitative risk information provided as percentages, risk ratios or heart age scores and was investigated in eight
• No differences in health behaviours, blood pressure, medication adherence or smoking.
• Participants with a younger heart age are more likely to recall risk (80% heart age vs. 63% percentage p = 0.009*) than those with an older heart age (both 61%; p > 0.999*). Percentages and control (no risk score) • Reduction in smoking (1.8% heart age vs. 0.4% percentage) and weight (−0.8 kg heart age vs. −0.2 kg percentages) at 12 weeks.
• At 12 months Framingham risk scores increased in the control group (+0.24%) and decreased in the risk percentage group (−0.2%) and the heart age group (−0.4%). • Risk perception highest in bar graph group at 3 months (p = 0.032)*.
• Between baseline and 3 months, risk perceptions decreased in the bar graph group (p = 0.02)*. There was no change in the icon array group. *Student's t test Navar et al. (2018) N/A each other • 22% of participants shown icon array reported a 10-year risk of 15% to be high compared with 36% shown no icon and 35% shown a bar graph (p < 0.001)*.
• 5%-6% more participants were willing to take preventive treatment when shown bar graph compared with icon array *Two-tailed test    • The genetic risk score group reported moderate weight loss in high-risk participants (−2.3 kg ± 3 vs. 0.0 kg ± 3, p = 0.002*).

Cardiovascular imaging
Coronary artery calcium scores Johnson et al. (2015) No comparison • 68% of participants could accurately identify their risk score based on their coronary artery calcium score.
• 24% of high-risk participants identified that they were in the high-risk group.
*ANOVA Orakzai et al. (2008) No comparison • Initiating aspirin therapy, increasing exercise and modifying diet increased with increasing coronary artery calcium scores (all p < 0.001* for trends).
• 56% high-risk participants modified their diet and 67% increased exercise. *t test and Mann-Whitney rank-sum test Kalia et al. (2006) No comparison • Statin compliance at 3 (±2) years was highest in the group with the highest coronary artery calcium scores (91%) and lowest in the low-risk group (44%).
• 65% increased exercise. Korcarz et al. (2008) No comparison • Higher levels of plaque led to increased intentions to take cholesterol-lowering medication (p = 0.02*) and an increased likelihood of having heart disease (p = 0.004*) and developing heart disease (p < 0.001*).
• Normal scans also lead to increased motivation to exercise (p = 0.003*).
*multiple linear regression model (Continues) studies (five quantitative and three qualitative). In a randomized trial, participants were more likely to agree that risk information was presented clearly and more helpful and reported less decisional conflict in choosing their preferred risk reduction method when presented with their Framingham risk score as opposed to risk factor education only (Powers et al., 2011). There were no differences in health behaviours, blood pressure or medication adherence and perceived risk declined in both groups at 3 months. Damman et al. (2016) identified that providing estimates of the percentage of similar individuals who will have a cardiovascular event at a given time period failed to heighten risk perception as some participants believed a risk score below 50% implied low risk. Participants also had problems recalling percentages, especially when provided with multiple numbers from their assessment.
This was reiterated in a second qualitative study (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014), where participants had difficulties remembering and understanding their risk percentage.
Only one quantitative study (Fair et al., 2008) investigated risk ratios by comparing them to percentages. Risk ratios increased risk perceptions and intentions to make lifestyle changes, however, also increased levels of worry.
The final numerical strategy evaluated was the use of an estimated heart age, which calculates an individual's heart age based on their risk profile, and was investigated in three quantitative studies. Bonner et al. (2015) compared the effect of providing individuals with their heart age against a percentage event rate at 5 years on behavioural and psychological outcomes. There was no significant difference in intention to change lifestyle or in risk perceptions. At 2 weeks, recall was highest in the heart age group but had significantly decreased since the intervention. Those with a younger heart age were more likely to recall their risk than those in the percentage group, however, there was no difference in recall between groups in participants with an older heart age. Participants in the heart age group, however, perceived the results to be less credible and had less of a positive emotional response.
Another study (Lopez-Gonzalez et al., 2015) compared the effect of the heart age on modifiable cardiovascular risk factors against a percentage event rate and a control group, who received conventional medical advice only. At 12 weeks, there was a significant decrease in weight and smoking in both experimental groups compared with the control group but was accentuated in the heart age group. At 12 months, Framingham risk scores had increased in the control group but decreased in the heart age and percentage

Numerical formats
Numbers 'Going to make me go online or make an appointment with a doctor who can make it clearer' (Ancker et al., 2009) Percentages 'oh that's only half of the risk! Let's take a look…your risk is 42%. Then it could have been worse' (Damman et al., 2016) 'I have 2%...what does that mean…does that mean 2 days out of 100 I'm at risk?'  Heart age 'I hate this 74 and 72, that's not real… The only one who can say what my heart age would be is the cardiologist when he goes in and has a look at my heart' (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014) 'I mean, I already feel that I am healthy-ish for my age.. to me that says yeah you're ok' (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014) 'Wow this is very good…It's an eye-opener… oh yeah I'm overweight and this and that but never thinking that it (would) have such an impact on my heart' (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014) 'I'm thinking that it's kind of overwhelming. It's intimidating for a man to come in who is 52 and find out he's got a heart age of 79. I think it's going to be very upsetting. He's gonna be really shaken' (Goldman et al., 2006) 'I think the idea of [cardiovascular risk-adjusted age] made it personal. Because this is your age. It brought you into it' (Goldman et al., 2006) Graphical formats Bar graphs 'well I'm not above the 50%, I'm in the red zone but the lower part of it' (Damman et al., 2016) Icon arrays 'It's a lot to look at' (Ancker et al., 2009) 'clearer that you're talking about human beings and not statistics' (Ancker et al., 2009) 'It can give a false reading' (participant talking about random sequencing) (Ancker et al., 2009) Game interaction 'It's like a game because you're playing around with it. That's what I like about it because you learn too' (Ancker et al., 2009) Genetic Risk Score 'if you have a high genetic risk it's in your genes… deprived yourself of all your nice treats but you've had the same end result, you might as well have enjoyed it and gone!' (Shefer et al., 2016) 'If it's going to run in the family you've got to accept it haven''t you? If it's your turn to, if your number comes up you can't do nothing about it' (Middlemass et al., 2014) 'I was sure there was something in the family make up but it's nice to know that's not the case' (Middlemass et al., 2014) 'The lifestyle I have led puts me at a greater risk than the person who didn't live my lifestyle' (Middlemass et al., 2014) (Korcarz et al., 2008) "Cardiovascular risk just isn't on their agenda, they're more worried about mental health issues." (Damman et al., 2017) "I already knew that I'm not as risk an eh there's nothing wrong with me, since you'd have to have complaints so really, yes for me this has no relevance" (Korcarz et al., 2008) "I feel like doctors are intimidating… they kind of rush you." (Goldman et al., 2006) "I think I'd be lower than that in reality" participant who clicked default (Damman et al., 2017) "You don't want to seem stupid, so you don't ask." (Damman et al., 2017) "Gives them a sense of empowerment, a bit of control" in relation to positive language "I like to put the fear into them… if they don't pull up their socks bad things can happen to them" (Damman et al., 2017) "You have to judge the people, at the time you have to make an informed decision as to how much information is going to sink in." (Damman et al., 2017) "this is quite good because it actually gives me targets for my BMI and what sort of weight I should be" reference to heart age (Damman et al., 2016) "How can you come up with a credible risk profile if factors like family history, exercise, stress not part of the calculation?" (Korcarz et al., 2008) "Well I did the test and it turns out because of my family" (Korcarz et al., 2008) "because of stress" (Shefer et al 2016) Acceptance of communication strategy groups. When the heart age was compared with either a percentage or natural frequency, the heart age improved intentions to become more physically active and to visit a general practitioner (Damman et al., 2018).
Two qualitative studies addressed the acceptability of the heart age. High-risk participants were less accepting of their results and questioned their credibility (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014). Those with a heart age that closely reflected their own age were also more reassured. In the study by Goldman et al. (2006), participants felt that the heart age score would be more memorable, but warned that receiving an older heart age may increase anxiety in individuals. A further qualitative study (Ancker et al., 2009) revealed that although some individuals were accepting numerical risk information, others found it too impersonal.

| Graphical formats
The risk was visually represented in graphical formats in eight studies (six quantitative and two qualitative) including bar graphs and icon arrays. Risk recall was lower when presented as an icon array compared with a numerical format immediately post-intervention, with no differences between the groups at 2-3 weeks on recall or understanding of risk (Ruiz et al., 2013). There were also no significant differences in perceptions of seriousness, intentions to change lifestyle, follow medical treatment or overall satisfaction.
Moreover, there were no differences in clarity or helpfulness of the information between the groups. Additionally, participants who received their risk in a graphical format (bar graph or icon array) compared with a numerical format (percentage and risk ratio) reported lower levels of worry but were not more reassured (French et al., 2004).
In a randomized trial, where icon arrays were compared directly with bar graphs (Adarkwah et al., 2019), no differences were observed in the recall of interventions agreed on with a general practitioner at 3 months. Risk perceptions were highest in the bar graph group at 3 months but decreased in comparison with baseline, whereas it remained consistent in the icon array group. Additionally, in a second study, participants who were shown their risk as a bar graph had a higher perceived risk and were more likely to take preventative treatment than those who received an icon array (Navar et al., 2018). Conversely, a qualitative study (Damman et al., 2016) highlighted individuals may misinterpret the severity of their results when using bar graphs because high scores such as 20% appear to be in the lower portion of the graph.
The type of icon used also influenced perceptions and risk recall. The icons which performed better in risk recall were restroom icons and photographs with blocks and faces performing the worst . Mean perceived cardiovascular risk perceptions did not significantly differ and were moderately correlated with the actual risk information that was presented.
Random rather than sequential positioning of negative icons to portray the chance of suffering a cardiovascular event was associated with better alignment between risk estimates and perceptions but reduced lifestyle intention scores .
Random dispersal was also reported as more realistic in a qualitative study. (Ancker et al., 2009). Participants found icon arrays with stick icons more personal and relatable than those which use shapes.

| Avatars
Avatars are digital representations of people used to promote social interaction (Ruiz et al., 2016) and were addressed in two quantitative studies. Avatars improved overall risk perceptions among participants and alignments between risk estimates and intentions to see a doctor . Conversely, no differences in risk recall or understanding were found when avatars were compared against voice and text alone (Ruiz et al., 2016). There were no significant differences in worry, disturbance or confidence to follow medical treatments, however, the avatar was favoured in intentions to make lifestyle changes.

| Qualitative information
Qualitative information can be provided to individuals to help structure how a lay person thinks about risk. Qualitative information was investigated in one quantitative study. Participants were more likely to provide correct answers for risk recall and subjective risk comprehension questions when qualitative information was used to communicate risk compared with infographics.

| Infographics
Infographics are sophisticated visualizations compromised of imagery, charts and text to provide an overview of a topic. They also provide additional narratives such as information about risk factors. One quantitative study investigated the use of infographics and found that infographics negatively influenced the recall of risk causes and subjective risk comprehension and more correct answers were given when qualitative information was used (Damman et al., 2018). Health literacy also influenced results as participants with adequate health literacy were more likely to consider infographic information useful.

| Game interactions
Game-like interactions are those which permit individuals to interact with the information provided to them (Ancker et al., 2009). Only one qualitative study (Ancker et al., 2009) investigated the use of an interactive game format to portray cardiovascular risk and involved clicking different icons to reveal which individuals would be affected by cardiovascular disease.
Some participants enjoyed the interactive component which made it more like a game, whereas others found the process time-consuming.

| Timeframes
Time-based risk formats allow for timeframe manipulations and can portray the accrual of risk over time. Timeframes were addressed in two quantitative studies. Participants who were shown their lifetime cardiovascular risk reported higher incidences of worry and were less reassured than participants who received a 10-year cardiovascular risk (Fair et al., 2008). Similarly, in a study (Frileux et al., 2004), where five different timeframes (ranging from 5 to 20 years) were investigated, the shorter timeframes performed better. Higher intentions to adopt preventative behaviour were also reported when shorter timeframes were used.

| Genetic risk scores
Genetic risk scores use statistical measures of genome variations that increase an individual's probability of developing cardiovascular disease (Shefer et al., 2016). Four studies (two quantitative and two qualitative) investigated the use of providing individuals with feedback from a genetic risk score to communicate cardiovascular risk. Receiving a genetic risk score was associated with improved hypertension control at 16 weeks compared with participants who received no risk information in a randomized controlled trial. (Domenech et al., 2016). When comparing the Framingham risk score only to the Framingham risk score plus genetic risk score, there was no effect on low-density lipoprotein cholesterol at 3 or 6 months. High-risk participants in the genetic risk score group did, however, report a moderate loss in weight (Knowles et al., 2017). Shefer et al. (2016) identified that when participants were provided with both their Framingham risk score and genetic risk score, they often only remembered one score and were unable to recall which one it was. Furthermore, participants often misunderstood their genetic risk score, believing if their risk was 'in their genes' it could not be modified. This misconception was also highlighted in a second qualitative study (Middlemass et al., 2014). One participant, however, recognized the importance of the gene and environment interaction and that lifestyle modification could reduce an increased genetic risk. Participants with a low genetic risk score felt reassured by their results, particularly those with a family history of cardiovascular disease (Middlemass et al., 2014).

| Cardiovascular imaging
The results from cardiovascular imaging, including coronary artery calcium scoring and carotid ultrasound measurements, can be used to provide feedback to individuals about their current cardiovascular health and risk of a future event. Cardiovascular imaging was addressed in five quantitative studies. Three quantitative studies investigated the use of coronary artery calcium scores. The first (Johnson et al., 2015) provided participants with their coronary artery calcium score and verbal risk category. Overall, over two-thirds of the participants could accurately identify their risk category based on their coronary artery calcium score. A significantly lower proportion of high-risk participants identified that their score placed them in the high-risk category. All five risk groups showed improvements in health-promoting behaviour; however, there were no changes in risk perceptions over time. Worry was highest in the low-risk group at baseline but highest in the moderate-and high-risk groups at 3 months. In the other two studies, participants were provided with verbal information about their coronary artery calcium scores. The studies were similar in design and may include the same participants but reported different outcomes. Orakzai et al. (2008) determined that the number of participants who reported initiating aspirin therapy, increasing exercise and modifying their diet increased with increasing coronary artery calcium scores. Over half of the participants in the highest risk category reported modifying their diet and increasing exercise. In the second study (Kalia et al., 2006), statin compliance at 3 years was highest in participants with high coronary artery calcium scores and lowest in those with low-risk scores.
Overall, a large proportion of participants reported increasing exercise levels, stopping smoking and making dietary modifications.
In two studies, participants were provided with the results of their carotid ultrasound scan. In a randomized controlled trial, Framingham risk scores decreased at 1 year among participants who received a visualization of their scan result and increased in the group who received a risk score only (Näslund et al., 2019). Additionally, there was a greater reduction in low-density lipoprotein cholesterol concentrations and smoking rates in the group who received a visualization of their carotid ultrasound. Korcarz et al. (2008) provided participants with verbal information about their carotid ultrasound.
Participants with increased levels of plaque reported increased perceptions of having or developing heart disease and intentions to take cholesterol-lowering medication.

| Pre-assessment factors
Three pre-assessment factors were identified from the review including previous knowledge of cardiovascular disease. Some participants did not understand what cardiovascular disease encompassed and it was perceived as less frightening than cancer (Damman et al., 2017).
Some believed that they had adequate knowledge of cardiovascular disease and risk factors, deeming themselves not at risk. This links to the second factor, which is motivation to undergo cardiovascular risk assessment. Some participants who believed that they were of low risk were unlikely to undergo screening, particularly when they had no physical complaints (Damman et al., 2017). The appropriateness of a cardiovascular assessment was also highlighted by healthcare professionals. They believed that if patients had more pressing health concerns, discussing their cardiovascular risk would place an additional burden on them (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014).

| Mode of assessment
Self-assessment and assessment by healthcare professionals were the two modes of cardiovascular assessment identified. Some participants felt that their doctor was too busy or intimidating to conduct an assessment (Ancker et al., 2009). They were also too embarrassed to request clarification if they did not comprehend their risk (Ancker et al., 2009). Online calculators allowed individuals to conduct their own assessment, but many underestimated their risk factor values and entered incorrect data (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014). Furthermore, if they did not understand their risk, they were unable to seek clarification from a healthcare professional.

| Communication of risk
As well as choosing which communication strategy to use, healthcare professionals highlighted that they also use different communication styles. Some opt for paternalistic communication styles promoting fear, whereas others prefer to empower patients (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014).
Furthermore, they identified that they chose which strategy to use based on how much information they believed patients could process. For example, they believed that colourful charts were most beneficial in patients with low health literacy (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014).

| Post-assessment
Three post-assessment factors were identified including perceived reliability of assessment, rationalizing risk and reducing risk. Participants questioned the reliability of some probabilistic risk scores because they expected to provide more information about their health and behaviour (Damman et al., 2016). Many went on to rationalize their risk, blaming factors out of their control including stress or family history (Damman et al., 2016). One participant felt that the heart age score encouraged risk reduction as it provided a target to work towards (Bonner, Jansen, McKinn, et al., 2014;Bonner, Jansen, Newell, et al., 2014). Another participant identified that they would be more willing to take risk reduction measures if they had decided on them in partnership with their doctor (Sheridan et al., 2009).  (French et al., 2017). It is also important to consider the limitations associated with using cardiovascular imaging to communicate cardiovascular risk. Imaging is more time-consuming and expensive than traditional methods and imaging technologies may not be readily available or accessible in developing countries. Moreover, it may provide false reassurance as normal scans in young individuals with unhealthy lifestyles may incorrectly suggest that they can continue with their current lifestyle.

| DISCUSS ION
Several risk communication strategies, such as percentages, bar graphs and icon arrays, which provide patients with a probability, fail to heighten risk perceptions (Damman et al., 2016). Many of the most commonly used cardiovascular risk scores, such as the Framingham score, classify a 20% risk of developing cardiovascular disease over the next 10 years as high. As 20% appears in the lower portion of the graphs, these scores can be interpreted as low risk. The same is true for icon arrays as a large number of positive icons makes it easy for participants to believe that they would be one of the individuals not affected (Ancker et al., 2009).
The genetic risk score also provides individuals with a probabilistic risk score. One study claimed that providing a genetic risk score improved blood pressure control (Domenech et al., 2016). It is difficult to establish if this was the result of receiving a genetic risk score or an improved awareness of risk because the control group received no risk information. Participants believed that their genetic risk could not be modified and consequently were less motivated to make lifestyle improvements. The genetic risk assessment and gene and lifestyle interaction require more explanation to individuals.
In Responding to emotions is also crucial and those involved in risk discussions must recognize the potential negative impact of being identified as high risk. Those designated as being at risk who previously viewed themselves as healthy are now faced with a revised health status despite lacking the associated symptoms (Gillespie, 2015). This may bring new social manifestations and health regimes similar to those who are ill. Care must be taken when broaching high-risk discussions and professionals must offer suitable support. 2. drafting the article or revising it critically for important intellectual content.

ACK N OWLED G EM ENTS
Marshall Dozier, Academic Support Librarian, for providing assistance and advice during the preparation of the search strategy.
PROSPERO Registration Number: CRD42020204797. Open access funding enabled and organized by ProjektDEAL.

CO N FLI C T O F I NTE R E S T
No conflict of interest has been declared by the author(s).

PEER R E V I E W
The peer review history for this article is available at https://publo ns.com/publo n/10.1111/jan.15327.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.