All things considered, my risk for diabetes is medium: A risk personalization process of familial risk for type 2 diabetes

Abstract Background A positive family history of type 2 diabetes (T2D) has been associated with risk awareness and risk‐reducing behaviours among the unaffected relatives. Yet, little is known about how people with a positive family history for diabetes develop and manage their personal sense of risk. Objective To characterize two key concepts, salience and vulnerability, within the familial risk perception (FRP) model among unaffected individuals, at increased familial risk for T2D. Design We conducted a mixed method study. Descriptions of salience and vulnerability were collected through semi‐structured interviews. Participant's perception of self‐reported risk factors (family history, age, race/ethnicity, medical history, weight and exercise) was measured using the Perceived Risk Factors for T2D Tool and was compared to a clinical evaluation of the same risk factors. Results We identified two components of salience: (a) concern for developing T2D and (b) risk awareness triggers, and two features of vulnerability: (a) statement of risk and (b) risk assessment devices. Although few participants (26%) were concordant between their perceived and clinical overall T2D risk, concordance for individual risk factors was higher, ranging from 42% (medical history) to 90% (family history). Discussion and conclusion Both familial and non‐familial events lead people to contemplate their T2D risk, even among people who have a positive family history. Participants often downplayed their overall risk and underestimated their overall risk compared to a clinical risk assessment of the same self‐reported risk factors. Clinicians could leverage key components of the FRP process as way to engage patients in risk reduction strategies earlier.


| INTRODUC TI ON
The systematic collection of a family medical history captures information about shared inherited, environmental and behavioural risk factors for genomically complex diseases such as cancer, cardiovascular disease and diabetes. Goals for eliciting a family history include risk classification for the purposes of early detection and counselling to support behaviour changes to prevent the disease and/or minimize health complications related to these diseases. 1 Additionally, a positive family history of T2D has been associated with developing risk awareness and engaging in risk-reducing behaviours among the unaffected relatives. [2][3][4][5][6] Yet, little is known about how people with a positive family history for complex diseases such as diabetes develop and manage their personal sense of risk. 7 Understanding this process could facilitate better collaboration between healthcare providers and patients aimed at prevention and risk reduction interventions. 7 To that end, Walter et al 8 developed the familial risk perception (FRP) personalization model. The model is comprised of four major constructs: salience, mental models, vulnerability and coping/control. Walter et al 8,9 posit that the FRP personalization process is initiated among unaffected family members when a family member is diagnosed. In FRP, the risk personalization process involves a coalescence of salience (sense of awareness of family history, experiences and disease severity), personal mental models of health (explanations of disease causation and inheritance) and notions of how alike one is to their affected family members. In turn, these factors influence a person's sense of risk and vulnerability for developing the disease.
The level of perceived risk influences coping and risk control strategies, which may or may not include behaviour changes ( Figure 1). 9 The original model was largely drawn from cases of familial cancer and coronary artery disease with very few examples of diabetes. 8,9 Thus, we planned a mixed methods study to further develop the FRP model for people with a positive family history of T2D and are, themselves, unaffected. Previously, we described beliefs about cause, genetics and inheritance for T2D among participants in this study. 10 In this article, we characterize salience and vulnerability. We had two research questions: (a) how do people at increased familial risk for T2D describe salience and vulnerability (qualitative arm) and (b) how does perceived diabetes risk compare to a clinical assessment of diabetes risk among individuals at increased familial risk for T2D (quantitative arm).

| Overview of Core Mixed Method Study
To further develop the FRP model for T2D, we conducted a mixed method study using a concurrent design. 11 In this type of mixed method approach, data are collected simultaneously. We selected this design because it allows investigators to elucidate complementary aspects of the same phenomenon and can facilitate a deeper understanding of participants' responses. The qualitative arm was the primary focus of our core project. We used a single semi-structured interview to collect data on the FRP model domains. In the quantitative arm, we collected supplemental information about each domain through a survey. Figure 2 provides an overview of participant enrolment and study flow.

| Participant Recruitment
The study took place in the United States, Midwest in two locations with populations of 450 000 (comprised of urban and rural communities) and 27 000 (single urban community), respectively. All participants were recruited into the core study as follows. Study recruitment posters and brochures were placed in neighbourhood restaurants, hair salons, grocery stores and a community-based agency that serves lower-income individuals. We also recruited through a rural, primary care clinic in a largely Hispanic neighbourhood and by mass emails sent to a college campus community. Interested participants directly contacted the research team. Interested individuals from the community-based organization provided contact information to their case worker, who passed the information to the research team. A research team member screened, obtained consent and enrolled the participants. Inclusion and exclusion criteria are listed in Figure 2. The study was approved by the first author's Institutional Review Board. Specific enrolment issues and data collection, analysis and results with respect to the salience and vulnerability domains F I G U R E 1 Familial risk perception (FRP) personalization model

Coping/Control=
Level of concern and the actions taken

| Participant enrolment goals
Because one of the major goals of the core study was to identify subtypes of the FRP personalization process that combined qualitative and quantitative data using qualitative cluster analysis, we enrolled a larger number of participants than typical for a qualitative study.
Based on previous qualitative studies using cluster analysis, we determined at least 100 interviews would be needed. 12,13 To achieve diversity and ability to conduct the cluster analysis, our goal was to enrol 30 participants in each non-Hispanic White, non-Hispanic Black and Hispanic groups. Although individuals from other ethnic groups were not excluded, very few people from other ethnicities lived in the study's catchment area. 10

| Data collection
We conducted a semi-structured interview based on Walter and Emery's 9 original study (refer to Table 1 for the interview guide) and collected a three-generation family health history focusing on any type of diabetes and metabolic syndrome. The interviews took place in person or over the phone and lasted between 30 and 90 minutes.
Each interview was audio-recorded and transcribed verbatim, and then uploaded into NVivo 10. 14 The family histories were recorded and stored in Progeny. 15 After completing the interview, participants received a $25 gift card. 10

| Data analysis
All transcripts were analysed using direct content analysis, a deductive process most appropriate for validating or extending an existing conceptual framework. 16 Q ualitative coding was conducted in three stages. 10

| RE SULTS
We summarized participant characteristics in Table 2. Just over half of the participants identified as female (n = 61, 55%), and over half reported their race as something other than non-Hispanic White F I G U R E 2 Core study participant enrolment and study flow (n = 64, 58%). Generally, the study participants were young and well-educated. Three people did not have a first-degree relative (FDR) with diabetes. However, their family history was consistent with metabolic syndrome and therefore high risk for T2D, or in one case a maternal grandmother had T2D and had raised the participant.

| Salience
Only five participants reported they really did not think about their risk. Twenty-one participants reported that although not a burning issue, T2D risk is 'always in the back of their mind'. The majority (n = 85) felt that T2D risk was a major concern in their life. We identified two over-arching components of salience: (1) concern for developing T2D and (2) risk awareness triggers.

| Developing T2D is concerning
Part of salience is developing a heightened sense of concern-that T2D is a serious disease.  Diabetes was not necessarily participants' primary concern. For some, overall health and disease prevention in general were key: However, some participants' concern about developing T2D was moderated by the idea that T2D is a manageable disease and not as life-threatening as cardiac diseases and cancer.
People get cancer, like all the time. And it kills them.
And people get diabetes all the time, but it doesn't seem to kill them. Diabetes it's more like, they manage it. I don't see diabetes as being fatal.

| Risk awareness triggers
Family characteristics such as a positive family history, a personal experience with a family member who has T2D and the severity of a Having a family of one's own and being exposed to content about diabetes in formal and informal educational settings also helped to create awareness of T2D risk.

| Vulnerability
We identified two features of vulnerability: (a) statement of risk perception and (b) risk assessment devices. Risk appraisal was collected by directly asking participants how they would rate their personal risk for diabetes (prompted as needed by asking participants if their risk was high, medium, low or something else). Responses were fairly equally divided into high (n = 31; 30%), medium (n = 41; 37%) and low (n = 39; 35%). If participants had not spontaneously described how they came to their stated risk, we asked participants to elaborate.
In doing so, participants named personal risk factors that either increase or decrease their risk and verbalized a rationale for the stated risk, which we categorized into different assessment devices.

| Quantitative arm: comparison of clinical risk assessment and perceived risk for T2D
To further characterize salience and vulnerability in the FRP, we also compared participants' perceptions about individual risk factors and overall risk to a clinical assessment of individual risk factors and overall risk for T2D. For this analysis, we identified a subset of participants from the core study who completed assessments of both perceived and clinical measures of risk for T2D (n = 153).

| Data collection
We collected perceived and clinical measures of risk for T2D as part of a survey (refer to Figure 3). After completing the survey, participants received a $25 gift card. Study data were collected and managed using REDCap electronic data capture tools 17 hosted at The University of Iowa.

| Perceived risk factors for type 2 diabetes (PRF-T2DM)
The PRF-T2DM is a measure of perceived personal risk for T2D. 18 Using a four-point Likert scale (0 = I do not know, 1 = there is no effect on risk, 2 = decreases the risk, 3 = increases the risk), participants rated the effects of each of 12 risk factors on their risk for developing T2D. Participants assessed the following risk factors: age, weight, race/ethnicity, personal medical history, family medical history, diet habits, exercise habits, financial resources, support resources, neighbourhood resources, community resources and work/ school conditions. The sum of the responses makes up the total score, and higher scores represent heightened perception of risk factors (salience). PRF-T2DM has high internal consistency (0.81) and reliability (0.83) based on major risk factors for T2D, demonstrating construct validity. 18 We found the overall internal reliability (α = 0.68) and validity of the PRF-T2DM to be acceptable in our study population. 19 Perception of overall risk for diabetes (vulnerability) was assessed by asking 'What is your overall risk to develop type 2 diabetes?' Participants rated their overall risk as no risk, low risk, moderate risk or high risk. 18 This question was placed immediately following the 12 risk factors on the PRF-T2DM. 19

| Exercise
We also collected data on activity levels using the International Physical Activity Questionnaire (IPAQ) 21   know effect of risk factor and were created, as described in Table 4.

| Analysis
The four perceived risk-factor groups were stratified by overall risk perception (under-, concordant-or over-estimators factor was perceived between those who underestimated their overall risk and those who concordantly estimated their overall risk.

| RE SULTS
These analyses were conducted on 153 participants who completed the PRF-T2DM, health status questionnaire and the IPAQ. Fifty-six of the 153 had also completed a qualitative interview.

| Overall risk
Most participants (n = 113; 74%) were discordant between their perceived overall risk and their clinical overall risk; most underestimated their risk, and only two perceived their overall risk for diabetes as higher than their clinical overall risk. Forty participants (26%) were concordant between their perceived overall risk and their clinical overall risk for diabetes. Of these 40 participants, 12 were at moderate risk for diabetes and 28 were at high risk. The only demographic a Numeracy was assessed with a six-item numeracy questionnaire that assesses numeracy skills. We combined two, 3-item questionnaires. 46,47 The score equals the total number of correctly answered questions.
difference among the three groups was BMI classification. There was near significance in meeting exercise requirements between the three groups. There were no differences in age, numeracy, sex, race/ethnicity, education or marital status among the three groups (Table 5).

| Individual risk factors
Although just 26% of participants were concordant between their perceived overall risk and clinical overall risk for diabetes, the proportion of participants that were concordant on individual risk factors was higher, ranging from 42% (personal medical history) to 90% (family medical history). Participants most often overes-

| D ISCUSS I ON
Consistent with Walter et al 8,9 and others 3,24 we found family characteristics such as a positive family history, a personal experience with a family member who has T2D and the severity of a family member's disease stimulated people to think about their own risk. However, given the major inclusion criteria was a positive family history, these results are not surprising. We found that in addition to family, non-familial events and personal milestones that are encountered in and over the course of peoples' everyday lives pique risk awareness and can lead to concern about developing T2D.
Walter et al 8,9 defined vulnerability as the outcome of processing the salient features of one's family history and experiences into a sense of personal individual risk. Further, these authors posit the FRP process is an intermittent dynamic process based on on-going family events. However, our data show that risk perception was based on an interpretation and balancing of individual risk factors that included family history rather than only on-going family events.
Participants' expression of vulnerability included identifying and taking into consideration multiple risk factors and was consistent with the public's understanding that T2D is a complex disease. 10,25 However, their conclusions about the impact of an individual risk factor and/or their overall risk assessment were not always consistent with how these same self-reported risk factors were assessed using a clinical algorithm (MyFamily Optimistic bias could be contributing to the overall underestimation of diabetes risk. This is a cognitive process that leads people to believe they are less likely to suffer from a negative event (ie develop a disease) and more likely to experience a positive outcome than the data suggest. [26][27][28][29] We believe that participants expressed optimistic bias during qualitative interviews. For example, they seemed to downplay their overall risk, determining their overall risk to be medium or low and justifying their conclusion by comparing themselves to others and past versions of themselves and counterbalancing risk factors (eg although family history increased risk, exercise may be viewed as reducing risk; therefore, overall risk is estimated to be medium). The possibility of optimistic bias is also supported by the quantitative data when participants correctly perceived relevant risk factors as significant to their overall risk but judged their overall risk to be lower compared to the clinical risk algorithm (MyFamily).
Misperceptions about the impact of individual risk factors on their overall risk also support the possibility of optimistic bias (eg, perceiving exercise to lower risk and self-reporting less than 150 minutes of moderate to vigorous exercise per week). Dickerson et al 28 found that college students, similar in age to this study's participants, minimized their perceived overall risk for diabetes by downplaying the effect of lifestyle factors and basing their risk more heavily on noncontrollable factors.

| Applicability of the findings
Both provider and patient explanations of health and risk for disease provide a clinical reality, and divergence between clinician and patient explanations of these processes can impede patients' uptake of healthful behaviours. 30,31 Health-care professionals understand risk from a technical and statistical perspective, while patients may have a more experiential, personal and affective risk perspective. [32][33][34][35] However, clinicians can influence patients' risk perception adjustments. 4,36,37 As such, this study involved a relatively young group of participants, and key events that led to salience about T2D were reported prior to being diagnosed as pre-diabetic or with T2D.
Given our findings, we assert conversations about T2D risk do not happen soon enough or often enough. If misperceptions about risk persist, patients may naturally bias their risk assessment towards explanations of risk that reinforce their perspective and then delay engaging in risk reduction behaviours. [38][39][40] The act of taking a family history can be a catalyst for the FRP personalization process and if done earlier it could be leveraged on more proximal salient events before risk perception become engrained. By purposely guiding patients through the FRP personalization process, clinicians and patients could collaboratively identify patients' T2D risk and develop tailored risk reduction strategies earlier. In this way, the family history can be a tremendously advantageous tool for risk stratification as well as an intervention tool. 2,41-44

| Limitations
The MyFamily T2D risk assessment algorithm, like many clinical risk evaluation tools, relies on self-report, so risk estimations are only as good as the information supplied. Risk perception is also dynamic; as such, participants' perceptions of risk are subject to change.
Regardless, our findings showed high discordance of T2D risk estimates between patients' perception and a clinical tool. Even when participants' perceived risk is concordant with a clinical risk assessment, the process through which this conclusion is reached may vary. Family history was rarely underestimated (or overestimated) as a risk factor. However, the study was not inclusive of people with low familial risk.

| CON CLUS ION
Findings from this study improve our understanding of how people personalize and process their risk for T2D and provide important insight as to when and how ideas of risk are forming and when clinicians could collaborate with patients in this process. More research is needed to understand the relationship between how people process risk and their engagement in actions to mitigate risk.

ACK N OWLED G EM ENT
We would like to thank Amy Schumacher PhD for statistical consultation.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.