- Research article
- Open Access
Identification of health-related behavioural clusters and their association with demographic characteristics in Irish university students
BMC Public Health volume 19, Article number: 121 (2019)
Students engage in risky health-related behaviours that influence their current and future health status. Health-related behaviours cluster among adults and differently based on sub-populations characteristics but research is lacking for university populations. Examining the clustering of health- related behaviours can inform our initiatives and strategies, while examining cluster members’ characteristics can help target those who can prosper most from health promotion efforts. This study examines the clustering of health-related behaviours in Irish university students, and investigates the relationship with students’ sex, age, field of study and accommodation type.
An online survey was completed by 5672 Irish university students (51.3% male; 21.60 ± 5.65 years) during 2014. Two-step cluster analysis was used to understand how health-related behaviours (physical activity, smoking, alcohol intake, drug use and dietary habits) cluster among male and female students. Binary logistic regressions were conducted to examine the likelihood of students falling into certain clusters based on their characteristics.
Five cluster groups were identified in males and four in females. A quarter of males were categorised as ideal healthy with older students and those from certain fields of study having a higher likelihood of being classified in a low physical activity and poor diet (OR = 1.06–2.89), alcohol consumption (OR = 1.03–3.04), or smoking and drug use (OR = 1.06–2.73) cluster. Forty-five percent of females were categorised as ideal healthy with older females more likely to be in a low active and smoking cluster (OR = 1.03), and less likely to be in a convenience food cluster (OR = 0.96). Females from certain fields of study were also more likely to be classified in these clusters (OR = 1.59–1.76). Students living away from their family home had in increased likelihood of being in a cluster related to a higher frequency of alcohol consumption (OR = 1.72–3.05).
Health-related behaviours cluster among this population and need to be taken into account when designing multi-health interventions and policies. These findings can be used to target student groups at risk, leading to more efficient and successful health promotion efforts. The addition of modules providing information regarding health-related behaviours are advised in all fields of study.
Unhealthy or risky health-related behaviours (HRB) are primary causes of premature morbidity and mortality [1,2,3]. Physical inactivity, alcohol abuse, smoking tobacco, and poor dietary behaviours are suggested as the four main contributors to diseases such as hypertension, diabetes and certain cancer . There is a risk that individuals will engage in risky HRB such as above, including illicit drug use, as they gain increased independence when they transition from high school to university [4, 5]. A recent study found that 36% of Irish university students (n = 8122; 49.1% male; 23.17 ± 6.75 years) reported being insufficiently active, while 22% drank alcohol at least twice a week, 21% smoke tobacco (occasionally or frequently), and 20% had previous or current illicit drug use . Earlier studies have also shown the high proportion of university populations engaging in risky HRB, which is worrying. Two thirds of Irish undergraduate students (n = 2250; 36.9% male) indicated hazardous alcohol consumption , with another study finding that a quarter of Irish students were smokers and 37% had used illicit drugs in the past year . Much like with physical activity (PA), some students also fail to maintain healthy dietary behaviours, which are associated with reduced risk of developing chronic diseases [4, 9, 10].
Studies often examine HRB in isolation of one another [5, 7, 11], but emerging research suggests that these behaviours co-exist or cluster in most populations . While associations between risky HRB have been noted , little is known about the clustering of these behaviours nor their relationship with demographic (i.e. sex, age) and environmental (i.e. living accommodation) factors in this population [2, 4, 12, 13]. Noble and colleagues (2015) conducted a review of behavioural cluster research, identifying 56 relevant articles in adult populations. The most popular cluster groups reported were a healthy cluster containing no risk factors (81%), smoking and alcohol consumption (56%), all risky health behaviours (i.e. physical inactivity, alcohol consumption, smoking, and poor dietary behaviours; 50%), and poor diet with physical inactivity (44%). Four of the articles identified in this review included university populations from the UK and the USA [1, 14,15,16], with one paper examining the general Irish population . In total, the studies examining university populations looked at 5585 students, age ranges from 18 to 24 years, with three studies looking at mixed sex populations and one looking at females only. The clusters that emerged in each study fall somewhere on a scale from unhealthy/ high-risk clusters to moderate-risk and low-risk for health. Effective interventions that lead to the adoption of healthy behaviours and to the reduction of risky behaviours are needed . There is currently a gap in our understanding of if and how these behaviours cluster and co-exist in this population, which can help target and tailor future health promotion efforts and increase their success.
The clustering of HRB allows us to study how groups of students engage in a range of behaviours, but few studies have explored whether specific sub-groups of populations are more or less likely to be classified in certain clusters [2, 18]. Research has shown that males and younger age groups are associated with more risky health cluster combinations . Although this is useful, we have yet to investigate how a students’ accommodation type or field of study associates with possible clusters of HRB. Living environment is a factor worth considering when examining HRB, whereby some students live in their family home while others live in student or rented accommodation [19, 20]. Students living in university accommodation report higher PA levels [20, 21] but have risky dietary habits  and an increased prevalence of binge drinking [23, 24] when compared to those living off university campuses. Risky dietary habits were seen as the increased consumption of convenience foods such as snacks, sweets, cakes and fast foods . Behavioural differences have also been observed between students studying in different fields, with biological students showing increased alcohol consumption, arts and social science students more likely to smoke and use illicit drugs , and students studying sport and health-related courses less likely to display poor dietary behaviours . These young adults are in a learning environment and are still at an age where HRB that influence future health status can be influenced and directed , but research has shown the inefficiency of health behaviour strategies with a one -for-all approach . Examining the characteristics of cluster members will help to identify students who express similar HRB, therefore, identifying potential target groups for health promotion efforts . Thus, the purpose of this study is to investigate the clustering of HRB in Irish university students and identify student types classified within these clusters.
Data for this study was collected during October and November 2014 as part of the Student Activity and Sport Study Ireland (SASSI) . Due to the all-island approach, the permission to conduct the study was granted from relevant ethical committees in the Republic of Ireland (Waterford Institute of Technology School of Health Science Research Ethics Committee; Dublin City University Research Ethics Committee), Northern Ireland (Ulster University Research Governance), and extended through recognition by all institutes involved. Participants (N = 9197) from 31 institutes of higher education around Ireland using quota based sampling considering institution size and field of study were administered the survey, while 8122 (50.9% male; 21.51 ± 5.55 years) of those completed the minimum amount required for analysis (i.e. demographic data). Relevant personnel (e.g. health promotion officer, health lectures, and health researchers) volunteered in each of the universities to recruit students and administer the survey via a Survey Monkey link during class time. Recruitment was done through emails and direct contact with lecturers and heads of departments to allow access to the class groups required based on the sampling provided. Weekly updates were provided to the relevant personnel in each university to ensure that the correct students were being administered the survey. The survey could be accessed through any smartphone, laptop or computer. The use of an online survey administered in a class, instead of through emails, was based on previous research protocols where participation rates in excess of 90% were achieved . Participation in the online survey was voluntary and anonymous. Students were informed that by completing the questionnaire they were providing their informed consent to participate. They were also informed that they could exit the survey at any time by closing the survey window on their smartphone, laptop or computer. The survey included study information, demographic information (age, sex, field of study and accommodation type), PA levels, risky HRB, and dietary habit questions. Field of study was grouped as relating to i) social, business and law; ii) health, welfare and exercise; iii) humanities and arts; iv) education; v) science, maths and computing; vi) engineering and manufacturing; vii) other. Accommodation was either recognised as living in a family home or living outside of the family home (e.g. student accommodation, renting privately etc.).
Participants PA levels over the last 7 days was measured using the International Physical Activity Questionnaire – Short Form (IPAQ – SF) . Participants were classified into ‘low’, ‘moderate’ and ‘high’ categories depending on the level of PA reported from the nine items. Various interpretations of IPAQ can be used, but for the purpose of this study it was considered that only those categorised as ‘high’ were meeting the physical activity guidelines (PAGL) . Participants were then dichotomised into meeting or not meeting the PAGL. This was based on the current PAGL, which state that adults should engage in at least 150 min of moderate-intensity aerobic PA, or 75 min of vigorous-intensity aerobic PA in bouts of at least 10-min each week . The IPAQ – SF has previously been found to have acceptable validity and reliability in university students [32, 33]. Items assessing smoking, drug use, alcohol consumption and dietary habits were taken from the Survey of Lifestyle and Attitudes to Nutrition (SLÁN) study . Smoking levels were assessed using a single item: ‘Do you now smoke every day, some days or not at all?’ Answers were dichotomised into yes (yes and sometimes) and no. Drug use was assessed using a single item: ‘Have you ever taken non-prescribed/recreational drugs?’ Answers were dichotomised into any drug use (yes and yes but have stopped) and no previous or current drug use. Alcohol consumption was assessed using a single item: ‘How often do you have a drink containing alcohol?’ Answers were dichotomised into % ≤2 times/week (never, monthly or less and 2–4 times a month) and % ≥2 times/week (2–3 times a week and 4+ times a week). Dietary habits were assessed by asking first ‘How often do you eat convenience food (i.e. fast food, takeaways, Chinese, Indian, burgers, chips etc.)’ where the answers were dichotomised into % ≤ once/week (never or less than once a week) and % ≥ once/week (1–3 times a week, 4–6 times a week or daily). A second question asked ‘How often do you prepare food from fresh ingredients rather than pre-prepared food?’ where the answers were dichotomised into % ≤4 times/week (never, less than once a week or 1–3 times a week) and % ≥4 times/week (4–6 times a week or daily). Although the measures in this study were different, dichotomisation of responses was based on a similar approach used in the research . Consequently, this also aided with the interpretation of cluster outputs.
SPSS Inc., Chicago IL, version 23 was used for all analyses. Participants who did not complete all of the items needed for the cluster analysis were removed from the study. Descriptive statistics were calculated for demographic data and each of the behaviours. Pearson’s Chi-square test for independence was performed to note any significant differences for each behaviour between sexes. Engagement in HRB was different for males and females [35,36,37,38] and previous studies have investigated the clustering of behaviours separately based on sex [14, 15]. Thus, a two-step cluster analysis was used as an explanatory tool to identify specific behavioural clusters in male and female students separately. This method was designed to handle large data sets and enables the input of categorical variables . The number of clusters was based on the log-likelihood distance and Schwarz Bayesian criterion . The cluster analysis procedures were repeated in an internal random sample of 50% of the total study sample for each sex and a kappa statistic was used to assess reliability of the cluster solutions . The cluster outputs were given descriptive names based on the behaviours evident in each one. An ANOVA with Bonferroni (or Games-Howell when lack of homogeneity) post hoc was used to test the difference between clusters for mean age. Pearson’s Chi square was assessed to test for differences in student characteristics (field of study and accommodation type) between the clusters. In order to assess the number of students from certain fields of study and accommodation types falling into each cluster, the adjusted residual (AR) was observed . When the AR rises above 2.0 it is presumed that a significantly higher proportion of students are in a certain cluster than what is expected. When the adjusted residual falls below − 2.0 it is presumed that a significantly lower proportion of students are in a certain cluster than what is expected . Further analysis was completed using binary logistic regressions to identify those students who had a higher likelihood of being categorised in a cluster containing a risky HRB. The health, welfare and exercise students and the students living in their family home were used as the reference categories since they are seen to engage in less risky HRB [22,23,24, 26]. Results are presented as Odds Ratios (OR) and 95% Confidence Intervals (CI).
After data cleaning, the analytical sample comprised of 5672 participants (51.3% male; 21.60 ± 5.65 years). The final sample were older (t (7611) = 2.30, p < 0.05) than those excluded, with no difference for sex (X2 = 0.57 (1), p = 0.33). Baseline characteristics for the study sample are shown below (Table 1). Compared to females, a higher proportion of male students met the PAGL, had a higher frequency of alcohol consumption, smoked, reported current or past drug use, and consumed convenience foods more/ than once a week. A higher proportion of females reported consuming fresh food at least four times a week.
Cluster outputs and characteristics
Cluster analysis revealed five distinct clusters for males and four for females. There was a very good agreement between the cluster solution derived from the full sample and the random subsample (50%; males: kappa = 0.82, p < 0.01; females: kappa = 1.00, p < 0.01). The distribution of behaviours (i.e. characteristics) within each cluster is shown for both male and female students (Table 2). For example, Cluster 1 in males, labelled ‘Ideal Healthy’, was characterised by meeting the PAGL, low risk relating to smoking, drug use, frequency of alcohol and convenience food consumption, and the highest proportion of students that prepared food using fresh ingredients. For females, the ‘Ideal Healthy’ cluster had similar behaviours identified, except the proportion meeting the PAGL was lower (63.4% vs 100.0%) and the proportion consuming fresh foods at least four times a week was higher (69.6% vs 65.1%). Other clusters saw engagement in a range of behaviours but they were given a descriptive name based on any predominant risky HRB.
Proportion of students categorised in each cluster
The differences between clusters based on age, field of study and accommodation type can be seen for male and female students (Table 3). Compared to the ‘Ideal Healthy’ clusters, males were significantly older in the ‘Low PA & Poor Diet’ and ‘Smoking & Drug Use’ clusters, while females were older in the ‘Low PA & Smoking’ cluster and younger in the ‘Convenience Food’ cluster. A Pearson Chi-square showed a significant difference between the students’ field of study and cluster placement for males and females. A Pearson Chi-square also identified differences between the students’ accommodation type and their cluster placement in males and females.
For males the ‘Ideal Healthy’ cluster contains a significantly higher proportion of students from health, welfare and exercise related courses (AR = 5.15) and living in their family home (AR = 2.29). A significantly lower proportion of social, business and law students (AR = − 2.44), and humanities and arts students (AR = − 2.01) were seen in this cluster. A ‘Low PA & Poor Diet’ cluster contains a higher proportion of students from science, maths and computing related courses (AR = 4.01) and living in their family home (AR = 3.35). This cluster contains a lower proportion of health, welfare and exercise related students (AR = − 4.03). The ‘Convenience Food’ cluster contains a higher proportion of students from education related courses (AR = 2.31) and living in their family home (AR = 3.07). An ‘Alcohol Consumption’ cluster contains a higher proportion of students from social, business and law courses (AR = 2.76) and living away from their family home (AR = 9.78). The ‘Smoking & Drug Use’ cluster contains a higher proportion of humanities and arts students (AR = 2.91) but a lower proportion of education (AR = − 2.18) and science, maths and computing related students (AR = − 2.06).
For females, the ‘Ideal Healthy’ cluster contains a significantly higher proportion of health, welfare and exercise related (AR = 2.66) and education related students (AR = 5.25). A lower proportion of students from social, business and law related courses (AR = − 3.85) and humanities and arts (AR = − 2.29) were classified in this cluster. A ‘Low PA & Smoking’ cluster contains a higher proportion of social, business and law related (AR = 2.31) and humanities and arts students (AR = 3.33). This cluster also contains a lower proportion of health, welfare and exercise related (AR = − 2.80) and education (AR = − 6.21) students. The ‘Convenience Foods’ cluster contains a higher proportion of students living in their family home (AR = 4.18) while the ‘Alcohol Consumption’ cluster contains students living away from their family home (AR = 4.89).
Likelihood of students being categorised in each cluster
Binary logistic regressions revealed the students with a higher likelihood of falling into clusters containing risky HRB, when compared to the ‘Ideal Healthy’ cluster for males and females (Table 4). Cluster placement was considered as the dependent variable while students’ age, field of study and accommodation type were seen as the independent variables. For males, the regression models were significant for the ‘Low PA & Poor Diet’ (X2(8) = 67.891, p < 0.01; R2 = 7.2%), ‘Alcohol Consumption’ (X2(8) = 107.451, p < 0.01; R2 = 12.8%) and ‘Smoking & Drug Use’ (X2(8) = 52.414, p < 0.01; R2 = 5.6%) clusters. A one year increase in age increased the likelihood of males being classified in the ‘Low Active & Poor Diet’ (OR = 1.06, CI = 1.04–1.09), ‘Alcohol Consumption’ (OR = 1.03, CI = 1.00–1.05), and ‘Smoking & Drug Use’ (OR = 1.06, CI = 1.03–1.08) clusters. Males from all fields of study, except education had a higher likelihood of falling into the ‘Low PA & Poor Diet’ cluster (OR = 2.03–2.89, CI = 1.25–5.27) when compared to the health, welfare and exercise students. Those from all fields of study, except education and engineering and manufacturing had a higher likelihood of falling into the ‘Alcohol Consumption’ cluster (OR = 1.81–3.04, CI = 1.25–5.52). Males studying in social business and law (OR = 2.10, CI = 1.43–3.07), humanities and arts (OR = 2.72, CI = 1.74–4.27), and science and computing (OR = 1.46, CI = 1.06–2.02) related courses had a higher likelihood of being categorised in the ‘Smoking & Drug Use’ cluster, when compared to the reference category. Males living away from home had an increased likelihood of being classified in the ‘Alcohol Consumption’ (OR = 3.05, CI = 2.34–3.98) cluster.
For females, the regression models were significant for all clusters when compared to the ‘Ideal Healthy’ cluster (X2(8) = 26.739–83.552, p < 0.01; R2 = 2.8–6.6%). A one year increase in age increased the likelihood of being categorised in the ‘Low PA & Smoking’ cluster (OR = 1.03, CI = 1.01–1.04), but showed a decreased likelihood of being placed in the ‘Convenience Food’ cluster (OR = 0.96, CI = 0.93–0.81). Females in social, business and law (OR = 1.76, CI = 1.29–2.39), humanities and arts (OR = 1.70, CI = 1.23–2.35), and science maths and computing (OR = 1.59, CI = 1.15–2.21) courses had a higher likelihood of being classified in the ‘Low PA & Smoking cluster’ when comparted to the health, welfare and exercise students. Contrastingly, females studying education courses had a decreased likelihood of being categorised in this cluster (OR = 0.32, CI = 0.19–0.54). Females in social, business and law courses also had a higher likelihood of falling into the ‘Convenience Food’ cluster (OR = 1.60, p < 0.01, CI = 1.16–2.20). Females living away from home had a decreased likelihood of being classified in the ‘Convenience Food’ (OR = 0.75, CI = 0.61–0.93) cluster, but an increased likelihood of being classified in the ‘Alcohol Consumption’ (OR = 1.72, CI = 1.31–2.26) cluster.
To the best of the authors’ knowledge, this is the first study to examine the clustering of HRB based on sex and their associations with students’ characteristics in Irish universities. The results of this study show that HRB cluster in this university population, much like previous research in the general Irish population  and university students in other countries [1, 14,15,16, 35]. Cluster outputs for male and female students were similar with the only difference being that drug use was found to be a prominent behaviour in the male clusters only. The co-existence of behaviours in this population is complex, with engagement in HRB varying in each cluster group. As Noble and colleagues (2015) found, our study identified an ‘Ideal Healthy’ cluster and a number of clusters containing a combination of healthy and risky HRB. For example, the risky clusters included meeting the PAGL combined with higher alcohol consumption in both sexes. Various associations between the individual behaviours have been observed in the research with positive associations noted for PA and alcohol consumption , PA and fresh food consumption [42, 43], and an inverse association for PA and smoking . The combination of healthy and risky behaviours may even be explained by a ‘work hard, play hard’  or ‘sensation-seeking lifestyle’  theory. These results help to understand which HRB cluster indefinitely, somewhat or not at all, aiding the creation of interventions that tackle clustered HRB, which have been more effective and less costly in the past .
Students’ age, field of study and accommodation type were shown to influence the likelihood of being categorised in a cluster containing risky HRB, as opposed to the desired ‘Ideal Healthy’ cluster. These findings allow us to understand which sub-groups of the university population are more likely to engage in risky HRB helping direct the appropriate interventions to the populations most in need. Increases in age were shown to elevate the likelihood of students being classified in the ‘Low PA & Poor Diet’, ‘Alcohol Consumption’ and ‘Smoking & Drug Use’ clusters in males and in the ‘Low PA & Smoking’ cluster in females. Research has reported a decline in PA as individuals age , and increases in alcohol consumption for students in later years of study , but is yet to note any association between age and smoking or drug use in this population. A suggestion for this may be that older students have had more years of independent living and exposure to risky HRB, such as smoking and drug usage, which has increased their likelihood of engagement. It may be important to use interventions preventing the initiation of these behaviours during the adolescent and early adult years, with interventions designed to cease engagement in these behaviours more applicable for older students who have had increased exposure to both independent living and risky HRB. In contrast, an increase in age decreased the likelihood of females being classified in the ‘Convenience Food’ cluster, with research showing that age and female sex were positively associated with indicators for healthy dietary habits in a representative sample of Norwegian adults .
Cluster members also varied based on their field of study and accommodation type. Descriptive analysis allowed us to observe the proportion of students classified in each cluster, while regression analysis examined the influence of students’ characteristics on the likelihood of being placed in a cluster containing risky HRB. The ‘Ideal Healthy’ cluster contained a significantly higher proportion of health, welfare and exercise related students in males and females. This field of study often contains learning modules that increase the knowledge of exercise and health, which is a known determinant of PA and other HRB . In comparison to the health, welfare and exercise students, these results show that students studying certain fields of study have an increased likelihood of being classified in clusters containing risky HRB. Interfaculty differences have been noted in the past, with the prevalence of smoking and drug use increased in arts, public relations, public administration, and communications courses when compared to degrees containing health, welfare and exercise components . In addition, Valera-Mato and colleagues (2012) found no interfaculty differences for alcohol consumption in 985 Spanish university students (32.6% male), but we have identified this to be increased in males studying social, business and law related, and humanities and arts courses when compared to the health, welfare and exercise students. It is recommended, that all university courses include a module or workshop in their curricula providing information on the risks of engagement in certain behaviours. This can be used to promote healthy behaviours while preventing engagement in risky HRB among all students, and not just those studying health, welfare and exercise related courses.
For accommodation, female students living in their family home had an increased likelihood of being categorised in a cluster containing a higher frequency of convenience food consumption. This contrasts previous findings that students living in their family home display healthy dietary behaviours when compared to those living outside of the family home [22, 50]. Students living in university accommodation are more likely to eat in campus facilities , where more fresh foods and healthy options are being offered in recent times. This may have a positive influence on students’ dietary behaviours while living away from home and should continue to be encouraged in university food outlets. Living away from the family home was found to increase the likelihood of both male and female students being categorised in clusters with an increased frequency of alcohol consumption. The association between alcohol consumption and living away from home has been reported [23, 24], with White and colleagues finding (2006) that leaving home was a stronger predictor of increased drinking behaviour than was university attendance. There is a need to publicise the risks of drinking alcohol past moderation among Irish students, especially those living away from home. Unfortunately, a high frequency of alcohol consumption tends to be accepted in Irish students, where it is integrated into the social norms of university life. Successful interventions in the past have altered the beliefs or social norms of students so that high levels of alcohol consumption are not seen as normal behaviour . Norman and colleagues (2017) used theory of planned behaviour-based messages targeting key beliefs about binge drinking in students, three weeks before attending university. These focused on the beliefs that students can have fun without binge drinking, that binge drinking can have a negative impact on studies, and that being a student does not mean you have to binge drink alcohol. Students who received the messages engaged in binge drinking less frequently and had less harmful patterns of alcohol consumption during the first 6 months of university . Overall, understanding the individuals within each cluster allows us to identify students who are at risk and may be potential targets for such health promotion efforts.
This study addresses a topic where limited research has focused on the Irish university population and could be considered the most important strength. The study also employed a data-driven approach to determine behavioural clusters and used empirical measures to minimise subjectivity in deciding the number of clusters. A limitation of this study was that the HRB included in the cluster analysis consisted of self-reported responses and did not assess the quantity (e.g. units of alcohol), only the frequency (excluding the IPAQ-SF). Future studies should look at both the frequency and the quantity of different HRB in this population to gain a greater understanding into the participation levels. Similarly, the dietary behaviour measurement tool presents similar problems, limiting the information gathered. It is advised that future studies look to employ more in-depth measurement tools  in order to understand the dietary behaviours of students. The behavioural clusters found in this university sample were determined using exploratory cluster analysis and therefore may not be generalizable to other populations. In addition, if other HRB (e.g. sexual practices) were included, different cluster groups may have arisen, while additional variables (e.g. household income) added to the regression analysis may have altered the results. Lastly, this study was cross-sectional, which means that the data only provides a snapshot of how HRB cluster amongst the population.
Understanding the behaviours of university students’ is important as both the increase of independent living they experience and the multiple stressors of university life create an environment that supports the engagement of risky HRB . Health professionals should take note of how HRB cluster when designing multi-health interventions. For example, the targeting of smoking and low PA levels together in females, as opposed to individually, which can have accumulative health effects and be less costly . Similarly, from investigating students’ field of study, modules promoting healthy behaviours may be beneficial in all fields of study and not just for those studying health, welfare and exercise related courses. In addition, it is recommended to target students based on certain characteristics, such as the inclusion of interventions to prevent or cease high frequencies of alcohol consumption in students living away from their family homes. However, more research is needed to investigate why certain sub-groups of students are highly represented within clusters involving risky health or poor dietary behaviours.
International physical activity questionnaire - short form
Physical activity guidelines
Student Activity and Sport Study Ireland
Dodd LJ, Al-Nakeeb Y, Nevill A, et al. Lifestyle risk factors of students: a cluster analytical approach. Prev Med (Baltim). 2010;51:73–7.
Morris LJ, D’Este C, Sargent-Cox K, et al. Concurrent lifestyle risk factors: clusters and determinants in an Australian sample. Prev Med (Baltim). 2016;84:1–5.
Noble N, Paul C, Turon H, et al. Which modifiable health risk behaviours are related? A systematic review of the clustering of smoking, nutrition, alcohol and physical activity (‘SNAP’) health risk factors. Prev Med (Baltim). 2015;81:16–41.
El Ansari W, Stock C, John J, et al. Health promoting behaviours and lifestyle characteristics of students at seven universities in the UK. Cent Eur J Public Health. 2011;19:197–204.
Dinger MK, Brittain DR, Hutchinson SR. Associations between physical activity and health-related factors in a national sample of college students. J Am Coll Heal. 2014;62:67–74.
Murphy M, Murphy N, MacDonncha C, et al. Student Activity and Sports Study Ireland (SASSI). 2015. http://www.studentsport.ie/wp-content/uploads/2016/02/SASSI-Full-Report-Without-Appendices..pdf. Accessed 31/05/18.
Davoren MP, Shiely F, Byrne M, et al. Hazardous alcohol consumption among university students in Ireland: a cross-sectional study. BMJ Open. 2015;5:e006045.
Hope A, Dring C, Dring J. College Lifestyle and Attitudinal National (CLÁN) Survey. 2005. http://www.drugsandalcohol.ie/4327/1/2670-2853.pdf. Accessed 31/05/18.
Ali HI, Jarrar AH, Abo-El-Enen M, et al. Students’ perspectives on promoting healthful food choices from campus vending machines: a qualitative interview study. BMC Public Health. 2015;15:512.
Plotnikoff RC, Costigan SA, Williams RL, et al. Effectiveness of interventions targeting physical activity , nutrition and healthy weight for university and college students : a systematic review and meta-analysis. Int J Behav Nutr Phys Act. 2015;12:1–10.
Berg CJ, Ling PM, Hayes RB, et al. Smoking frequency among current college student smokers: distinguishing characteristics and factors related to readiness to quit smoking. Health Educ Res. 2012;27:141–50.
Spring B, Moller AC, Coons MJ. Multiple health behaviours: overview and implications. J Public Health (Bangkok). 2012;34:3–10.
Watts P, Buck D, Netuveli G, et al. Clustering of lifestyle risk behaviours among residents of forty deprived neighbourhoods in London: Lessons for targeting public health interventions. J Public Heal (United Kingdom). 2016;38:308–15.
Greene GW, Schembre SM, White AA, et al. Identifying clusters of college students at elevated health risk based on eating and exercise behaviors and psychosocial determinants of body weight. J Am Diet Assoc. 2011;111:394–400.
Laska MN, Pasch KE, Lust K, et al. Latent class analysis of lifestyle characteristics and health risk behaviors among college youth. Soc Prev Res. 2009;10:376–86.
Quintiliani L, Allen J, Marino M, et al. Multiple health behavior clusters among female college students. Patient Educ Couns. 2010;79:134–7.
Conry MC, Morgan K, Curry P, et al. The clustering of health behaviours in Ireland and their relationship with mental health, self-rated health and quality of life. BMC Public Health. 2011;11:692.
Hobbs M, Duncan MJ, Collins P, et al. Clusters of health behaviours in Queensland adults are associated with different socio-demographic characteristics. J Public Health (Bangkok) 2018: (online).
Keating XD, Guan J, Piñero JC, et al. A meta-analysis of colle students’ physical activity behaviors. J Am Coll Heal. 2005;54:116–26.
Johnson LG. Physical activity behavior of university students: an ecological approach. LSU Doctorol Dissertations, 2006. https://digitalcommons.lsu.edu/gradschool_dissertations/1556/. Accessed 31/05/18.
Dinger MK. Physical activity and dietary intake among college students. Am J Health Stud. 1999;15:139.
El Ansari W, Stock C, Mikolajczyk RT. Relationships between food consumption and living arrangements among university students in four European countries - a cross-sectional study. Nutr J. 2012;11:28.
Dantzer C, Wardle J, Fuller R, et al. International study of heavy rinking: attitudes and sociodemographic factors in university students. J Am Coll Heal. 2006;55:83–90.
White HR, McMorris BJ, Catalano RF, et al. Increases in alcohol and marijuana use during the transition out of high school into emerging adulthood: the effects of leaving home, going to college, and high school protective factors. J Stud Alcohol. 2006;67:810–22.
Webb E, Ashton H, Kelly P, et al. Patterns of alcohol consumption, smoking and illicit drug use in British university students: interfaculty comparisons. Drug Alcohol Depend. 1997;47:145–53.
Ferrara C, Nobrega C, Duflan F. Obesity, diet, and physical activity behaviours of students in health-related professions. Coll Stud J. 2013;47:560–5.
Ball K, Timperio AF, Crawford DA. Understanding environmental influences on nutrition and physical activity behaviors: where should we look and what should we count? Int J Behav Nutr Phys Act. 2006;3:33–40.
Haase A, Steptoe A, Sallis JF, et al. Leisure-time physical activity in university students from 23 countries: associations with health beliefs, risk awareness, and national economic development. Prev Med (Baltim). 2004;39:182–90.
Craig CL, Marshall AL, Sjöström M, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003;35:1381–95.
Department of Health and Ispos MRBI. Healthy Ireland Survey 2015: Summary of Findings. 2015. https://health.gov.ie/wp-content/uploads/2015/10/Healthy-Ireland-Survey-2015-Summary-of-Findings.pdf. Accessed 31/05/18.
Department of Health. National guidelines on physical activity for Ireland. 2009. http://health.gov.ie/healthy-ireland/physical-activity/recommended-guidelines/.
Dinger MK, Behrens TK, Han JL. Validity and reliability of the international physical activity questionnaire in college students. Am J Heal Educ. 2006;37:337–43.
Murphy JJ, Murphy MH, MacDonncha C, et al. Validity and reliability of three self-report instruments for assessing attainment of physical activity guidelines in third level students. Meas Phys Educ Exerc Sci. 2016;21:1–8.
Morgan K, McGee H, Watson D, et al. SLAN 2007: survey of lifestyle, attitudes & nutrition in Ireland: main report. 2007.
Ye Y, Wang P, Qu G, et al. Associations between multiple health risk behaviors and mental health among Chinese college students. Psychol Health Med. 2016;21:377–85.
Makino M, Hashizume M, Tsuboi K, et al. Comparative study of attitudes to eating between male and female students in the People’s republic of China. Eat Weight Disord - Stud Anorexia, Bulim Obes. 2006;11:111–7.
Weber EU, Blais A-R, Betz NE. A domain-specific risk-attitude scale: measuring risk perceptions and risk behaviors. J Behav Decis Mak. 2002;15:263–90.
von Bothmer MIK, Fridlund B. Gender differences in health habits and in motivation for a healthy lifestyle among Swedish university students. Nurs Health Sci. 2005;7:107–18.
Everitt B. Cluster analysis. Qual Quant. 1980;14:75–100.
Landsberg B, Plachta-Danielzik S, Lange D, et al. Clustering of lifestyle factors and association with overweight in adolescents of the Kiel obesity prevention study. Public Health Nutr. 2010;13:1708–15.
Agresti A. Categorical data analysis. Berlin Heidelberg: Springer; 2003.
Gillman MW, Pinto BM, Tennstedt S, et al. Relationships of physical activity with dietary behaviors among adults. Prev Med (Baltim). 2001;32:295–301.
Jago R, Nicklas T, Yang SJ, et al. Physical activity and health enhancing dietary behaviors in young adults: Bogalusa heart study. Prev Med (Baltim). 2005;41:194–202.
Musselman JRB, Rutledge PC. The incongruous alcohol-activity association: physical activity and alcohol consumption in college students. Psychol Sport Exerc. 2010;11:609–18.
French MT, Popovici I, Maclean JC. Do alcohol consumers exercise more? Findings from a national survey. Am J Health Promot. 2009;24:2–10.
Busch V, Van Stel HF, Schrijvers AJ, et al. Clustering of health-related behaviors, health outcomes and demographics in Dutch adolescents: a cross-sectional study. BMC Public Health. 2013;13:1118.
Hallal PC, Andersen LB, Bull FC, et al. Physical activity levels of the world ’ s population Surveillance progress , gaps and prospects. Lancet. 2012;380:247–57.
Johansson L, Thelle DS, Solvoll K, et al. Healthy dietary habits in relation to social determinants and lifestyle factors. Br J Nutr. 1999;81:211–20.
Bauman AE, Reis RS, Sallis JF, et al. Correlates of physical activity: why are some people physically active and others not? Lancet. 2012;380:258–71.
Varela-Mato V, Cancela JM, Ayan C, Martín V, Molina A. Lifestyle and health among Spanish university students: differences by gender and academic discipline. Int J Environ Res Public Health. 2012;9:2728–41.
Devine P, Lloyd K, Gray AM. University Student Food Attitudes and Behaviour Survey. 2006. http://www.ark.ac.uk/services/nistudentsurvey2005.pdf. Accessed 31/05/18.
Norman P, Cameron D, Epton T, Webb TL, Harris PR, Millings A, Sheeran P. A randomized controlled trial of a brief online intervention to reduce alcohol consumption in new university students: combining self-affirmation, theory of planned behaviour messages, and implementation intentions. Br J Health Psychol. 2018;23:108–27.
Rehm CD, Peñalvo JL, Afshin A, Mozaffarian D. Dietary intake among US adults, 1999-2012. Jama. 2016;315:2542–53.
The authors would like to thank all members of the SASSI research team and the team at IPAN, Deakin University.
This study was funded by Student Sport Ireland. The funding body played a role in the development of SASSI but had no input into the writing of this study.
Availability of data and materials
The datasets generated and/or analysed during the current study are available in the ISSDA repository, http://www.ucd.ie/issda/data/sassi/.
JM: Postgraduate Researcher. CMD: Lecturer and Researcher. MM: Dean of Postgraduate Research. NM: Lecturer and Researcher. AT: Alfred Deakin Professor. RL: Alfred Deakin Postdoctoral Research Fellow. CW: Chair in Physical Activity and Health.
Ethics approval and consent to participate
Due to the all-island approach, ethical clearance from relevant ethical committees in the Republic of Ireland (Waterford Institute of Technology School of Health Science Research Ethics Committee; Dublin City University Research Ethics Committee) and Northern Ireland (Ulster University Research Governance) were obtained and extended through recognition by all institutions involved. All participants were advised that they were providing informed consent by completing the online survey and that participation could be terminated at any time. This protocol was used due to the number of participants surveyed and was approved by the relevant ethics committees.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
About this article
Cite this article
Murphy, J.J., MacDonncha, C., Murphy, M.H. et al. Identification of health-related behavioural clusters and their association with demographic characteristics in Irish university students. BMC Public Health 19, 121 (2019). https://doi.org/10.1186/s12889-019-6453-6
- University students
- Health-related behaviours