Table 2 Recommendations in designing PRS reports. Integrating perspectives from both patients and PCPs, we offer the following points to consider for PRS clinical report design. Many of these recapitulate best practices from the risk communication literature; some are PRS-specific. See Discussion for the highlighting of certain points, in particular for a continuous (versus binary) sharing of risk information, the clear advantage of sharing absolute risk information, and the handling of differential performance of PRS by population group

 Both patients and PCPs were frustrated in the absence of granular risk information.

Give either a continuous estimate or category of risk, e.g., amongst a five-category framework.

 Both patients and PCPs expressed a desire for absolute risk information. Absent absolute risk, PCPs will “do the calculation,” multiplying odds ratio with disease prevalence to come up with an approximate absolute risk.

Give absolute risk if possible. While an absolute risk incorporating all known risk factors is preferable, just incorporating age and sex and appropriate base rates is an improvement over giving relative risk.

 Many patients and some PCPs interpreted the percentile as an absolute risk.

Be wary of potential misinterpretations of percentile.

 Both patients and PCPs expressed the desire to know the amount of overall risk captured by all genetic factors, and the amount of risk captured by the PRS.

Give the attributable risk for the PRS in question, and all genetic factors.

 Both patients and PCPs wanted to know how PRS information related to family history information, e.g. is it independent, or additive? The desire for this information holds even if an integrated score is used, to aid in overall understanding of the significance of PRS information.

Explain the relationship to other risk factors, in particular family history.

 Population categories described in terms of continental ancestry categories are likely to be interpreted as racial groups by both patients and PCPs; Patients and PCPs were both uncertain about how the results would apply to someone who didn’t fit neatly into one of the groups; PCPs wanted to understand why the numbers were different for the different population groups; PCPs had confusions about how these differences related to different disease prevalence rates in different racial groups.

Don’t use population categories if the clinical implications of the information are the same for each category. If using population categories, (a) make a statement that not everyone falls into one of these categories, and what should be done in this case, (b) clarify why these numbers are different, and c) clarify how this information relates to different disease prevalences.

 Both patients and PCPs had questions including whether they should have less confidence outside of European ancestry individuals, and how likely the PRS is to be underestimating risk in certain populations

Any statement about limitations of the information in different ancestry groups should make clear what the implications are for patients.

 Both PCPs and patients wanted to know what the implications of PRS information was for family members, including under what conditions, if any, family members should be tested.

Explain implications for family members.

 Many patients responded well to a graphical representation of the data, and PCPs found a graph to be a useful talking point.

Use graphical representations.

 Patients anticipated feelings of anxiety and concern for receiving high risk information, and also of feelings of being “off the hook” if they were not identified as at high risk. PCPs were concerned that their patients would have these same reactions.

Talk to the patient about potential emotional responses to a report during pre-test counseling and informed consent.

 Many patients mentioned the emotional valence associated with the color red, it can create unnecessary patient anxiety by placing a judgment on a piece of information. It can also prompt action. Some PCPs mentioned that red can also prompt possible PCP overreaction.

Red should be used carefully or avoided entirely.

 Some patients did not understand the meaning of any of the numbers shared. Others understood all the quantitative information and wanted more. Many just engaged with one component of the report.

Plan for diversity in patient ability to understand, and desire for, quantitative information. In particular, plan for selective engagement with the report; it should not be possible to come away with the “wrong” view if they just looked at one bit.

 Some PCPs utilize rules of thumb; others want to know pre-test probabilities; others will only act on professional guidelines.

Plan for diversity in how PCPs think about individual risk factors.

 PCPs view the report as more than a way to convey risk information. The report can act as a tool to structure the flow of the conversation.

Design report with the way it will structure the conversation between the patient and PCP in mind.

 Many PCPs wondered why monogenic information was not included, and wouldn’t have been able to answer patient questions on this point.

Ideally, monogenic risk should be incorporated into an aggregate genetic risk score. If this is not possible, the report should explain why monogenic risk information is not incorporated into the PRS.

 Some PCPs stressed that after the conversation with the PCP, the patient should have materials that they can understand.

Ensure availability of accessible, easy-to-understand patient materials, in multiple languages.

 Some PCPs expressed a preference for materials to help them understand what a PRS is and provide succinct explanations suitable for the short time period allotted for clinical visits.

Provide materials to help physicians get comfortable with their “spiel”; these points should help guide the flow of conversation.

 PCPs wanted to know answers to such questions as: Do professional societies approve of the use of this type of information? Do they actively disapprove? Are listed recommended actions those that guidelines already recommend for someone of average risk?

Given that there are not currently published guidelines for healthcare based on PRS, clarify relationship of recommended actions to existing guidelines (e.g., if recommendations given follow standard guidelines for those of average risk, state this)