Is the Risk Perception Attitude Framework Associated with the Accuracy of Self-Reported vs Actual Cardiometabolic Risk and Physical Activity in Young Adults with Overweight/Obesity?

Abstract

We examined the accuracy of perceived vs actual cardiometabolic risk and physical activity within the Risk Perception Attitude Framework (RPA). We analyzed baseline data from 343 young adults (23.3 ± 4.4 years) participating in a weight management clinical trial. Cardiometabolic risk factors were measured according to standard clinical procedures. A cardiometabolic risk score was created from five biomarkers according to whether or not a standard clinical risk cut point was exceeded. Physical activity was determined by ActiGraph and self-report. Perceived risk and physical activity self-efficacy were assessed by validated measures. The Proactive cluster (low perceived risk/high self-efficacy) was most accurate regarding actual vs perceived risk awareness (54%), while the Responsive cluster (high perceived risk/high self-efficacy) was the least accurate (16%). All RPA clusters underestimated their actual physical activity, self-reporting less than half the moderate-to-vigorous physical activity that was captured via accelerometry. The RPA Framework can identify young adults unlikely to be aware of their cardiometabolic risk. Given the growing prevalence of metabolic syndrome, efforts early in adulthood to increase knowledge and awareness of cardiometabolic risk, and behaviors necessary to reduce that risk, can have substantial impact on future health.

Keywords

physical activity cardiometabolic risk risk perception attitude framework young adults

“Participants in all four clusters recorded a high level of weekly MVPA with the ActiGraph.”

Introduction

Between 1988 and 2012 the overall prevalence of the metabolic syndrome increased from 25.3% to 34.2% in the United States (U.S.) and this increasing prevalence has become especially evident in young adults (18-34 years).¹ Previous studies have indicated that between 20 and 89% of young adults have one or more components of the metabolic syndrome, with greater prevalence of components seen in samples with overweight/obesity.^2-5 Despite these increasing trends, knowledge and awareness of cardiometabolic risk is low.^6-8 Indeed, data from a large population of young adults living in the United States indicate that 30%, 37%, and 43% of those with diabetes, hypertension, and hypercholesterolemia, respectively, were unaware of their risk.⁷

Physical activity is an important component of cardiometabolic risk prevention in all age groups, including young adults. Current U.S.⁹ and global public health recommendations¹⁰ call for adults (including the young adult age range) to achieve at least 150 minutes of moderate-intensity physical activity per week; yet, young adults’ knowledge and awareness of these recommendations are low. In fact, less than 40% of young adults in the United Kingdom¹¹ and the United States^12,13 are able to correctly define current physical activity guidelines. Moreover, agreement between self-reported and device-measured physical activity is modest at best (r = .16 to r = .49),^14,15 suggesting that young adults may lack awareness about their own risk reduction behaviors.

Awareness of risk is a necessary factor for making health behavior changes. The Risk Perception Attitude (RPA) framework¹⁶ categorizes individuals by their risk perceptions, along with their level of self-efficacy in practicing the behaviors necessary to mitigate their risk. Within this framework risk perceptions are considered “motivators” while efficacy beliefs are “facilitators” of behavior change. People who believe they are at risk for disease and also that they can take preventive actions to mitigate the risk are classified as Responsive, since they may be most likely to take action on their own behalf (Figure 1). In contrast, those having both low risk perception and low efficacy beliefs are classified as Indifferent because they may have neither the motivation nor the ability to act. People with high risk perceptions and low efficacy beliefs are classified as Avoidant, while those with low risk perceptions and high efficacy beliefs are classified as Proactive. Importantly, these four clusters have been shown to distinguish differences in attitudes and behavioral intentions for several disease preventive strategies.^17-19 To our knowledge, however, this framework has not been linked to accuracy of perceived vs actual risk status in young adults. To the extent that disease prevention behaviors are driven by people’s perceptions about risk²⁰ as is also posited by the Health Belief Model²¹ and the Protection Motivation Theory,²² it is important to know how well people are able to gauge their own risk to various diseases.

We hypothesized that those in the Responsive cluster will have the highest concordance between their risk awareness and actual risk, as well as between their self-reported and accelerometry-measured physical activity. Our underlying rationale for this was that the Responsive cluster’s engagement in preventive behaviors (which tend to be higher than that of the other clusters)^17-19 is likely the result of high risk perceptions (which have motivated them to act) and strong efficacy beliefs (which have facilitated their behavior change). Similarly, we also hypothesized that the Indifferent cluster would also be accurate for the opposite reason: this group’s lack of action likely reflected low motivation (because of low risk perceptions) and low ability (because of low efficacy beliefs). We estimated that the other two clusters (Proactive and Avoidance) would display lower levels of accuracy because of the conflicting information from low efficacy and high risk or vice versa.

Methods

Study Population

Study participants (N = 459) were enrolled in the Healthy Body Healthy U (HBHU) randomized controlled trial that tested digital intervention strategies to promote weight loss and management among young adults between 2015 and 2018.^23,24 Main eligibility criteria were: (1) 18–35 years of age and enrolled in a college or university in the District of Columbia or Boston area; (2) a body mass index (BMI) of 25 to 45 kg/m²; (3) an active Facebook user with regular text message access; and (4) generally healthy enough to participate in physical activity. Among the N = 459 randomized participants, N = 456 were included in the identification of RPA clusters as previously described.²⁵ In the current analyses, only participants with baseline self-reported physical activity and accelerometry data (N = 405), as well as complete baseline cardiometabolic risk factor data (abdominal circumference, blood pressure, fasting glucose, HDL-C, and triglycerides) were included, resulting in a final sample size of 343 participants. Study participants provided written informed consent and all study procedures were approved by the Institutional Review Boards of the George Washington University and the University of Massachusetts-Boston.

Cardiometabolic Risk

Detailed information on the measurement of the cardiometabolic risk variables is contained elsewhere.²³ Briefly, height, body weight, abdominal circumference, and blood pressure were measured according to standard clinical procedures. Following an overnight fast, a venous blood sample was collected for the measurement of total- and HDL-cholesterol and triglycerides concentrations. Samples were processed and analyzed according to standard procedures in the Medical Laboratory at the University of Virginia. A capillary blood sample was also obtained to assess fasting glucose (OneTouch® Ultra® 2) and hemoglobin A1c (HbA1c; A1cNow+, PTS Diagnostics). We then calculated a cardiometabolic risk score (CRS)² based on whether a standard risk cutpoint was exceeded (0 = no; 1 = yes) for abdominal circumference [>102 cm (men) or >88 cm (women)], HbA1c (≥5.7%), HDL-C [<40 mg/dL (men) or <50 mg/dL (women)], SBP (≥130 mmHg), and DBP (≥85 mmHg).^26,27 Scores for each individual risk factor were summed to create the CRS, with scores ranging from 0 to 5.

Physical Activity

Physical activity was measured by both accelerometry and by self-report. Participants were instructed to wear an ActiGraph accelerometer (wGT3X-BT) for 7 days, with valid wear time counted as 4 days of 10 hours per day.²⁸ Established thresholds of accelerometer counts were used to define moderate-to-vigorous physical activity (MVPA), with MVPA defined as ≥1952 counts per minute.²⁹ To adjust for variability in number of days the ActiGraph was worn, the average daily total was multiplied by seven to obtain an average weekly total of MVPA. Participants also self-reported weekly physical activity using the seven-item International Physical Activity Questionnaire (IPAQ).³⁰ The IPAQ queries the frequency (days/week) and duration (mins/day) of vigorous- and moderate-intensity physical activity, as well as walking and sitting performed over the previous 7 days. Responses were processed and summed using the IPAQ guidelines to derive a measure of MVPA (min/week).³¹

Cardiometabolic Risk Awareness Index

Four questions were used to assess metabolic risk awareness:^8,32 (1) Have you ever heard of the term metabolic risk; (2) Do you know what your blood pressure is; (3) Have you ever had blood work to check for cholesterol; (4) Have you ever had blood work to check for high glucose? The RAI was calculated by summing responses (0 = no; 1 = yes) across the four questions. Scores ranged from 0 to 4, with higher scores indicating higher awareness of risk.

Physical Activity Self-Efficacy

Self-efficacy for physical activity was assessed by having participants rank on a 5-point Likert scale how confident they felt exercising under five distinct conditions (bad weather, poor mood, fatigue, on vacation, too busy).³³ Scores range from 5 to 25, with higher scores indicating greater self-efficacy.

Creating the RPA Framework

The four RPA Framework clusters then were created from the RAI score (range 0 to 4) and the physical activity self-efficacy score (range 5 to 25). Because the two variables were measured in different units and may have had unequal variances, they were standardized to a mean of 0 and a standard deviation of 1. A cluster analysis then was performed on the standardized data to create four RPA groupings based on these two variables (Figure 1).

Statistical Analysis

Univariate analyses (mean ± SD, frequencies [%]) first were generated on all study variables in order to determine their distributions within the study population and within the RPA clusters. The unadjusted association between device-measured and self-reported physical activity was determined using the Pearson Correlation Coefficient.

To determine the accuracy of participant cardiometabolic risk awareness vs their actual risk status, scores on the RAI and the CRS were categorized into tertiles (low, medium, high) based on their distribution in the study population. Participants then were categorized as Overestimating (RAI > CRS); Underestimating (RAI < CRS); or On Target (RAI = CRS). The frequency of these three categories was then compared among the four RPA Framework clusters using chi-square tests.

The accuracy of self-reported vs accelerometry-measured physical was determined by creating a difference score between the two measures (ActiGraph MVPA min/week—IPAQ MVPA min/week). A negative score would indicate that participants overestimated their actual physical activity, while a positive score would indicate an underestimation of their actual activity. Physical activity accuracy scores then were compared across the RPA Framework clusters using ANOVA. All analyses were performed using SAS (9.4) at an alpha level of .05 (two-sided).

Results

Participants were 23.3 ± 4.4 years and primarily female (78%). The majority of participants identified as Non-Hispanic White (49%), while 42% identified as Non-White and 9% were of unknown race/ethnicity. On average, the study sample had obesity (BMI = 31.0 ± 4.4 kg/m²) with a CRS of 1.63 ± .92. At baseline, participants performed 309 ± 161 min/week of MVPA based on accelerometry, but reported a substantially lower amount of MVPA on the IPAQ (128 ± 112 min/week). There was a weak, yet statistically significant correlation between the two measures of physical activity (r = .25; P < .001).

Accuracy of Cardiometabolic Risk Awareness by RPA Framework Cluster

Overall, the majority of the participants overestimated (41%) their actual cardiometabolic risk, while about 38% were on target with regard to concordance between actual and perceived risk awareness. The Proactive cluster was most likely to be on target (54%); however, the Responsive cluster appeared to have the lowest accuracy (16%) (Table 1). The Indifferent cluster was most likely to underestimate their actual risk (45%), while the proportion of those in the Responsive and Avoidant clusters who underestimated their risk was negligible (2%, 0%, respectively). In contrast, the Responsive cluster was most likely to overestimate risk (82%), followed closely by the Avoidant cluster (74%). Overall, the four clusters are statistically different from one another (P < .001). Pairwise comparisons of risk awareness are statistically significant between all clusters (P ≤ .01) except for between the Avoidant and Responsive groups (P = .178).

Table 1.

Accuracy of Cardiometabolic Risk Awareness by Risk Perception Attitude Framework Cluster.

	All(N = 343)	Responsive(n = 49)	Proactive(n = 112)	Avoidant(n = 87)	Indifferent(n = 95)
On target (%)	37.6	16.3	53.6	26.4	40.0
Underestimating risk (%)	21.3	2.0	25.9	0.0	45.3
Overestimating risk (%)	41.1	81.6	20.5	73.6	14.7

Note. Overall ANOVA P < .001. Two-way comparisons: Indifferent v. Responsive, Indifferent v. Avoidant, and Proactive v. Avoidant P < .001; Indifferent v. Proactive P = .014; Responsive v. Avoidant P = .178.

Accuracy of Physical Activity Reporting by RPA Framework Cluster

All RPA clusters underestimated their actual physical activity by reporting less than half the amount of MVPA on the IPAQ than what was captured via accelerometry. Underreporting was highest in the Indifferent cluster (207.0 ± 147.2 min/week) and lowest in the Proactive cluster (162.3 ± 195.7 min/week) although these differences were not statistically significant (Table 2).

Table 2.

Accuracy of Moderate-to-Vigorous Physical Activity Reporting by Risk Perception Attitude Cluster.

Cluster	Mean ± SD
Responsive (n = 49)	179.6 ± 173.5
Proactive (n = 112)	162.3 ± 195.7
Avoidant (n = 87)	177.9 ± 160.9
Indifferent (n = 95)	207.0 ± 147.2

Note. Means are the difference in MVPA measured by accelerometry vs self-report. A positive score indicates underestimation of actual physical activity.

Discussion

We used the RPA framework to categorize young adults according to their perceived cardiometabolic risk, along with their level of self-efficacy in practicing a specific behavior (i.e., MVPA) to mitigate their risk. Contrary to what we hypothesized, the highest concordance between risk awareness and actual risk was observed within the Proactive cluster (low perceived risk/high self-efficacy), while the Responsive cluster (high perceived risk/high self-efficacy) had the lowest concordance. The Indifferent cluster was on target 40% of the time, which was the second most accurate cluster. The two clusters that displayed most accuracy are both defined by low risk perceptions. This suggests that, among young people (whose current levels of cardiovascular risks are, indeed, lower than that of older adults), roughly half of them accurately gauge their (low) risk. What was more difficult to explain, however, is the finding that the Responsive cluster had the lowest level of accuracy. This is typically the group that engages in most preventive behaviors, which, combined with the finding in Table 1 showing a systematic overestimation of risk, suggests that the Responsive cluster’s behaviors are driven by their overestimation of risks. One possible interpretation of this finding is that those who have high efficacy for performing a behavior might feel they have more agency over their health, and are more aware of the links between their behavior and health risk. Both the Responsive and Proactive clusters have high efficacy beliefs; however, they differ in their risk perceptions, as the Responsive cluster has high risk perceptions while the Proactive cluster has low perceptions. Those in the Proactive cluster likely had high accuracy of risk assessment because they were already engaging in risk-reducing behaviors and (correctly) inferred their risk on the basis of those behaviors.

The simple correlation between self-reported and accelerometer-measured MVPA was low, but statistically significant. A correlation coefficient of r = .25 is within the range of values typically reported between self-reported and device-measured measures of physical activity in young adults.^14,15 What was striking, however, is that participants reported 181 fewer min/week of MVPA than what they actually performed, which is in stark contrast to evidence suggesting that MVPA is usually over-reported in this age group.¹⁵ Accuracy in reporting MVPA was highest in the Proactive cluster and lowest in the Indifferent cluster, but these differences were not statistically significant. Again, those in the Proactive cluster, having high efficacy beliefs (regardless of their actual risk), may be more aware of what they do to mitigate their future risk of cardiovascular disease. In contrast, those in the Indifferent cluster (low risk perception/low efficacy beliefs) have neither the motivation nor the ability to act on behalf of their own health, and thus are less aware of their risk preventive behaviors.

The strengths of this investigation include both the self-reported and device-measured measurement of cardiometabolic risk and physical activity. Nonetheless, these were cross-sectional data and therefore, the temporal sequencing between risk knowledge and awareness and actual risk status is difficult to discern. The high risk awareness observed in the Responsive and Avoidant clusters could be an artifact of the questions we used to create the RAI, which were focused on risk awareness (i.e., knowledge of terms, and awareness of their own individual values) vs one’s own personal risk perception. The HBHU participants (those with overweight/obesity) may have been more likely to be screened compared with those of normal weight due to a family history of cardiometabolic disease. Indeed, Zhang and Moran³⁴ reported that more frequent health care visits were associated with greater awareness of hypertension risk. Participants in all four clusters recorded a high level of weekly MVPA with the ActiGraph. Although accelerometry is a precise and valid objective measure of physical activity, it may not have captured the participants’ typical level of activity. Also, self-reported physical activity from the previous 7 days on the IPAQ may not have been representative of typical behavior, thereby explaining the marked difference in device-measured vs self-reported physical activity among the RPA clusters. Finally, all participants had overweight/obesity and had enrolled in a weight management trial, suggesting some level of concern regarding their weight and/or health. We may have seen different results had individuals in the normal weight BMI category been studied. For example, the Responsive cluster might have higher accuracy in a sample of young adults with healthy weight status. Future research can explore this hypothesis.

Nonetheless, this study was the first of its kind to examine accuracy of perceived vs actual cardiometabolic risk status in young adults using the RPA framework. By identifying young adults who may misperceive their risk, our findings suggest that prevention programs targeted to young adults need to address not only health risk literacy, but also self-efficacy in practicing the behaviors that will mitigate their risk. The promotion of cardiometabolic screenings in the clinic, community, and university settings (advertised using social media or other digital messaging) can raise awareness of one’s individual risk profile and address the cognitive-bias of young adults believing they are not at risk.³⁵ Novel methods of risk communication using social norms-based messaging may be particularly effective in this age group.³⁶ Health education messages also are critical to raise awareness of the importance of physical activity for disease prevention, and commercially-based tracking devices can reduce the disconnect between perceived vs actual physical activity behavior. Given the growing prevalence of metabolic syndrome at younger ages, efforts early in adulthood to increase knowledge and awareness of one’s own cardiometabolic risk and the behaviors necessary to reduce that risk can have a substantial impact on future health and quality of life.

Footnotes

Acknowledgments

The authors recognize the project team members who contributed to the data collection especially Erika Blankenship, Catherine Cameron, Jamie Faro, Rachel Ingersoll, Meghan Maverdes, Benjamin Shambon, and Timothy Tsung. The authors also would like to acknowledge Jeanne Jordan, PhD, and her laboratory team.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research reported in this manuscript was supported by National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health under award number R01DK100916 to MA Napolitano.

Ethical Approval

All study procedures were approved by the Institutional Review Boards of the George Washington University and the University of Massachusetts-Boston.

Informed Consent

Participants provided informed consent before taking part in this study.

Data Availability

Deidentified data from this study will be made available (as allowable according to institutional review board standards) to researchers submitting a specific request and data sharing agreement to the corresponding author.

Disclaimer

The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

ORCID iDs

Caitlin P. Bailey

Melissa A. Napolitano

References

Moore

Chaudhary

Akinyemiju

. Metabolic syndrome prevalence by race/ethnicity and sex in the United States, National Health and Nutrition Examination Survey, 1988-2012. Prev Chronic Dis. 2017;14:24.

DiPietro

Zhang

Mavredes

, et al. Physical activity and cardiometabolic risk factor clustering in young adults with obesity. Med Sci Sports Exerc. 2020;52(5):1050-1056.

Huang

Kempf

Strother

, et al. Overweight and components of the metabolic syndrome in college students. Diabetes Care. 2004;27(12):3000-3001.

Huang

Shimel

Lee

Delancey

, et al. Metabolic risks among college students: Prevalence and gender differences. Metab Syndr Relat Disord. 2007;5(4):365-372.

Sumner

Sardi

Reed

. Components of the metabolic syndrome differ between young and old adults in the US population. J Clin Hypertens. 2012;14(8):502-506.

Antwi

Lavin

Sullivan

Bellavia

. Perception of and risk factors for type 2 diabetes among students attending an upstate New York college: A pilot study. Diabetol Metab Syndrome. 2020;12:25.

Bucholz

Gooding

de Ferranti

. Awareness of cardiovascular risk factors in U.S. young adults aged 18-39 years. Am J Prev Med. 2018;54(4):67-77.

Sarpong

Curry

Williams

. Assessment of knowledge of critical cardiovascular risk indicators among college students: Does stage of education matter? Int J Environ Res Public Health. 2017;4(3):250.

Piercy

Troiano

Ballard

, et al. The physical activity guidelines for Americans. JAMA. 2018;320:2020-2028.

10.

Bull

Al-Ansari

Biddle

, et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br J Sports Med. 2020;54(24):1451-1462.

11.

Knox

ECL

Esliger

Biddle

SJH

Sherar

. Lack of knowledge of physical activity guidelines: Can physical activity promotion campaigns do better? BMJ Open. 2013;3:003633.

12.

Hyde

Omura

Watson

, et al. Knowledge of the adult and youth 2008 physical activity guidelines for Americans. J Phys Activ Health. 2019;16(8):618-622.

13.

Piercy

Bevington

Vaux-Bjerke

, et al. Understanding contemplators’ knowledge and awareness of the physical activity guidelines. J Phys Activ Health. 2020;17(4):404-411.

14.

Napolitano

Borradaile

Lewis

, et al. Accelerometer use in a physical activity intervention trial. Contemp Clin Trials. 2010;31(6):514-523.

15.

Slootmaker

Schuit

Chinapaw

MJM

, et al. Disagreement in physical activity assessed by accelerometer and self-report in subgroups of age, gender, education and weight status. Int J Behav Nutr Phys Activ. 2009;6:17.

16.

Rimal

Real

. Perceived risk and efficacy beliefs as motivators of change: Use of the risk perception attitude (RPA) framework to understand health behaviors. Hum Commun Res. 2003;29(3):370-399.

17.

Rains

Hingle

Surdeanu

, et al. A test of the risk perception attitude framework as a message tailoring strategy to promote diabetes screening. Health Commun. 2019;34(6):672-679.

18.

Rimal

Juon

H-S

. Use of the risk perception attitude framework for promoting breast cancer prevention. J Appl Soc Psychol. 2010;40(2):287-310.

19.

Rimal

Bose

Brown

, et al. Extending the purview of the risk perception attitude framework: Findings from HIV/AIDS prevention research in Malawi. Health Commun. 2009;24(3):210-218.

20.

Weinstein

Nicolich

. Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors. Health Psychol. 1993;12(3):235-245.

21.

Rosenstock

. The health belief model and preventive health behavior. Health Educ Monogr. 1974;2(4):354-386.

22.

Rogers

. A protection motivation theory of fear appeals and attitude change. J Psychol. 1975;91(1):93-114.

23.

Napolitano

Whiteley

Mavredes

, et al. Using social media to deliver weight loss programming to young adults: Design and rationale for the Healthy Body Healthy U (HBHU) trial. Contemp Clin Trials. 2017;60:1-13.

24.

Napolitano

Whiteley

Mavredes

, et al. Effect of tailoring on weight loss among young adults receiving digital interventions: An 18 month randomized controlled trial. Transl Behav Med. 2021;11(4):970-980.

25.

Napolitano

Tjaden

Bailey

DiPietro

Rimal

. What moves young people? Applying the risk perception attitude framework to physical activity behavior and cardiometabolic risk. Transl Behav Med. 2022;12(6):742-751.

26.

Norberg

Eriksson

Lindahl

, et al. A combination of HbA1c, fasting glucose and BMI is effective in screening for individuals at risk of future type 2 diabetes: OGTT is not needed. J Intern Med. 2006;260(3):263-271.

27.

Alberti

Eckel

Grundy

, et al. Harmonizing the metabolic syndrome: A joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation. 2009;120(16):1640-1645.

28.

Troiano

Berrigan

Dodd

, et al. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008;40(1):181-188.

29.

Freedson

Melanson

Sirard

. Calibration of the computer science and applications, inc. accelerometer. Med Sci Sports Exerc. 1998;30(5):777-781.

30.

Craig

Marshall

Sjöström

, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35(8):1381-1395.

31.

IPAQ . Guidelines for the Data processing and Analysis of the International Physical Activity Questionnaire. https://sites.google.com/site/theipaq/scoring-protocol. Published 2005. Accessed February 11, 2020.

32.

Seo

Kim

Park

Ryu

. Assessment of the awareness and knowledge of cancer survivors regarding the components of metabolic syndrome. PLoS One. 2018;13(6):0199142.

33.

Marcus

Selby

Niaura

Rossi

. Self-efficacy and the stages of exercise behavior change. Res Q Exerc Sport. 1992;63(1):60-66.

34.

Zhang

Moran

. Trends in the prevalence, awareness, treatment, and control of hypertension among young adults in the United States, 1999 to 2014. Hypertension. 2017;70(4):736-742.

35.

Skøt

Nielsen

Leppin

. Who perceives a higher personal risk of developing type 2 diabetes? A cross-sectional study on associations between personality traits, health-related behaviours and perceptions of susceptibility among university students in Denmark. BMC Public Health. 2018;18:972.

36.

Tran

Zimmerman

Kupzyk

, et al. Cardiovascular risk factors among college students: Knowledge, perception, and risk assessment. J Am Coll Health. 2017;65(3):158-167.