bs_bs_banner
R E V I E W
Drug and Alcohol Review (September 2015), 34, 540–548 DOI: 10.1111/dar.12272
Screening for hazardous alcohol use among university students using individual questions from the Alcohol Use Disorders Identification Test-Consumption MEI-LING BLANK1, JENNIE CONNOR1, ANDREW GRAY1 & KAREN TUSTIN2,3 1
Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand, 2Department of Psychology, University of Otago, Dunedin, New Zealand, and 3National Centre for Lifecourse Research, University of Otago, Dunedin, New Zealand
Abstract Introduction and Aims. Abbreviated versions of the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) instrument have not been investigated among students.We compared a modified second item (AUDIT-2*) (typical quantity per occasion as the number of drinks, rather than categorical responses) and the third item (AUDIT-3) (heavy episodic drinking frequency) with AUDIT-C scores and described their associations with sociodemographic variables. Design and Methods. We analysed cross-sectional data from the 2011 baseline of the Graduate Longitudinal Study New Zealand, including respondents aged up to 25 years (n = 5082, response rate 65%). Hazardous drinking was defined as an AUDIT-C score of seven or greater for men and five or greater for women.We calculated the area under receiver operating characteristic curves, sensitivities, specificities, and positive and negative predictive values for the AUDIT-2* and AUDIT-3. Odds ratios and 95% confidence intervals were calculated to describe the associations between drinking patterns and sociodemographic factors. Results. Based on the sex-specific AUDIT-C cut-points, 36% of participants reported drinking at hazardous levels. For the AUDIT-2*, the best combination of sensitivity and specificity was obtained for a cut-point of five standard drinks. The best cut-point for the AUDIT-3 was for monthly heavy episodic drinking. Positive and negative predictive values were above 0.80 for both cut-points. Consumption was associated with age, degree level, domestic/international status, accommodation type, relationship status and employment. Discussion and Conclusions. The AUDIT-2* and the AUDIT-3 appear to be promising standalone screening items for detecting hazardous drinking in this population of heavy-drinking students. [Blank M-L, Connor J, Gray A, Tustin K. Screening for hazardous alcohol use among university students using individual questions from the Alcohol Use Disorders Identification Test-Consumption. Drug Alcohol Rev 2015;34:540–8] Key words: alcohol drinking, mass screening, students.
Introduction The 10-item Alcohol Use Disorders Identification Test (AUDIT) [1] and its shortened derivatives, including the AUDIT-Consumption (AUDIT-C), comprising the first three items (typical frequency, typical quantity, heavy episodic drinking frequency), the AUDIT-QF, consisting of the first two items, and the AUDIT-3, consisting of the third item only, have been examined in a number of populations in order to screen for hazardous drinking behaviours, alcohol abuse or dependence disorders [2,3]. Compared with the full AUDIT, the AUDIT-C, for some populations [4–6], including university students [7], appears to be a better measure for
identifying individuals who report drinking at hazardous levels. However, cut-points indicating a positive screening result need to be tailored for the population under investigation [3]. For example, AUDIT-C cutpoints of four or five for men, and three or four for women, appear to yield the best balance between sensitivity and specificity in the detection of hazardous drinking or abuse/dependence among adults [3]. Among a tertiary student population in which heavy alcohol use is normative behaviour however, AUDIT-C cut-points of seven for men and five for women have been recommended for the detection of hazardous drinking [7]. Despite being considerably shorter in length and scoring complexity than the full AUDIT, the AUDIT-C
Mei-Ling Blank MPH, Assistant Research Fellow, Jennie Connor MB, ChB, PhD, Professor, Andrew Gray BA, BCom (Hons), Senior Research Fellow, Karen Tustin PhD, Research Fellow. Correspondence to Ms Mei-Ling Blank, Department of Preventive and Social Medicine, University of Otago, PO Box 56, Dunedin 9054, New Zealand. Tel: +64 3 479 7207; Fax: +64 3 479 7298; E-mail: [email protected] Received 6 September 2014; accepted for publication 17 February 2015. © 2015 Australasian Professional Society on Alcohol and other Drugs
Brief alcohol screening in students
may still be too burdensome in some situations. Shorter and less complex instruments may result in greater use in circumstances where opportunistic screening and intervention may be warranted, such as hospital emergency departments or time-pressured primary care clinics. Brief validated instruments may also help reduce respondent burden and improve response rates when included in research questionnaires. The AUDIT-QF has been proposed as an alternative [2,8,9] to the full AUDIT or AUDIT-C; however, as a two-item instrument, a degree of scoring complexity remains, which may limit its utility in certain opportunistic screening situations. Like most other quantity– frequency instruments, the AUDIT-QF measures average consumption over a period of time and does not adequately capture irregular consumption that is atypically low or high compared with the individual’s normal drinking pattern [10]. This misclassification is problematic when considering individuals who typically drink at moderate levels, but who also sporadically engage in hazardous heavy episodic drinking. The AUDIT-3, assessing heavy episodic drinking (six or more drinks per occasion) frequency, is a potentially useful brief screening tool that has the advantage of a simple scoring system. The full ten-item AUDIT has previously been compared with established and novel abbreviated versions in a community-based survey of young people [11]. The AUDIT-C and AUDIT-3, which both measure consumption only, unsurprisingly performed worse than three- and four-item variations that assessed the constructs of dependence and harm in addition to consumption. However, the single-item AUDIT-3 performed almost as well as the three-item AUDIT-C [11], raising the possibility that among young people, the consumption construct of hazardous drinking may potentially be addressed with a single question. Previous research among adults suggests that using a positive screening threshold of ‘ever’ drinking six or more drinks per occasion may result in acceptable sensitivities and specificities for the identification of hazardous drinking [3]. However, there is uncertainty about what frequency of heavy episodic drinking should be considered a positive screening result among a naturally heavy-drinking population, with several researchers recommending higher thresholds for cultures where heavy drinking is common [8,9,12]. Additionally, the quantity threshold of six or more drinks per occasion may not be appropriate for women [12,13]. The AUDIT-3 has not yet been evaluated in a student population. There have also been no investigations assessing the utility of the second AUDIT item (typical quantity consumed per occasion) for detecting hazardous drinking among students. The objectives of this study were: (i) to describe the baseline distribution of alcohol consumption in a
541
nationally representative sample of final-year university students in New Zealand aged up to 25 years; (ii) to compare a modified second AUDIT item (AUDIT-2*), and the AUDIT-3, with total AUDIT-C scores; and (iii) to describe the associations of hazardous drinking patterns, as identified by these individual items, with sociodemographic variables. Methods Data were collected during the first wave of the Graduate Longitudinal Study New Zealand (GLSNZ), a new cohort study investigating the employment, health and social outcomes of university graduates. The GLSNZ baseline survey included over 400 items, including sociodemographic factors; university expectations, experiences and satisfaction; employment plans and career aspirations; academic beliefs and attitudes; current financial circumstances; physical health, disability and functional impairment; health risk behaviours; psychological well-being; personality type; social support and social integration; and local and international community participation. All of the survey items are described in the GLSNZ Extended Baseline Report [14]. Participants and procedures Participants were final-year students from each of New Zealand’s eight universities who were in a program of study that would have allowed them to graduate with a bachelor’s degree or higher after successful completion of their studies during 2011. The sampling procedure has been described in detail elsewhere [14]. However, in brief, final-year students from each university were stratified by subject of study. Within subject areas, universities provided a specified number of randomly selected students according to age groups (5-year age bands, starting from age 15), sex, self-reported ethnicity [New Zealand European, Ma¯ori, Samoan, Cook Islands Ma¯ori, Tongan, Niuean, Chinese, Indian, other (e.g. Dutch, Japanese, Tokelauan)], enrolment status (full-time, part-time), study mode (intramural, extramural), degree level (undergraduate, postgraduate) and fee status (domestic, international). All international PhD students were included, as were all students from the smallest university. Sampling weights were constructed based on the actual number of potentially graduating students in each category (international PhD or other) for each campus in 2011. These data were provided by each university in 2012 after the finalised student numbers were available. The survey was conducted between July and December 2011. Eligible students were contacted by letter and email, and given a unique study code and password to log on to the © 2015 Australasian Professional Society on Alcohol and other Drugs
542
M.-L. Blank et al.
secure survey website. Non-responders and noncompleters were sent multiple reminder emails and were contacted up to four times by trained call centre staff. Overall, 65% (n = 8719/13 343) of the students who were invited agreed to participate and completed the survey.The present analysis is restricted to all of the respondents who were aged up to 25 years (n = 5082). We restricted the age range to reduce the heterogeneity in the sample as ‘mature students’ may have widely different life experiences before and during university study compared with students who matriculated soon after completing high school.These different life experiences may subsequently impact on an individual’s drinking pattern. The study was approved by the New Zealand Multi-region Ethics Committee (MEC/11/ EXP/049). Measures Alcohol consumption. We assessed alcohol consumption using the AUDIT-C [4], which consists of the first three items from the full AUDIT [1] measuring typical frequency, typical quantity per occasion and heavy episodic drinking (six or more drinks per occasion) frequency. Unlike the original AUDIT-C, respondents were not asked to restrict their responses to the previous 12 months (Supporting Information Table S1). Scores for the AUDIT-C range from 0 to 12, with higher scores indicative of a more hazardous drinking pattern [4]. Participants who reported never drinking alcohol were classified as abstainers.They were not asked the quantity and heavy episodic frequency questions, and were assigned an AUDIT-C score of zero. Although the original AUDIT-C does not specify a standardised measure of alcohol, respondents were shown illustrations of New Zealand standard drinks (10 g of pure ethanol) for a range of alcoholic beverages and provided with examples (e.g. ‘a jug of beer equals three drinks’). We included a greater number of response options for each item than the original instrument to allow for more fine-grained analyses of respondents’ drinking patterns. The modified response options were able to be collapsed into the original categories (Supporting Information Table S1). Importantly, for the second item about typical quantity, respondents were able to select a whole number of standard drinks ranging from 1 to 25+, rather than the original ordered classification (e.g. 1–2 drinks, 3–4, etc.). This modified second item is called AUDIT-2*. The AUDIT-3 consists of the third item about heavy episodic drinking frequency. Hazardous drinking was defined as an AUDIT-C score of seven or greater for men and five or greater for women. These are validated AUDIT-C cut-points for identifying university students aged 18 to 25 years who report drinking at hazardous levels [7]. © 2015 Australasian Professional Society on Alcohol and other Drugs
Social and demographic variables. Self-reported information was collected on a range of factors that have been identified as being relevant to an individual’s alcohol consumption during early adulthood, including age, sex, accommodation type, relationship status, parenthood and paid employment [15–22]. Information was also collected on whether the participant was the first person in their immediate family to attend university and the education level of the participant’s most highly educated parent/caregiver. Analysis Descriptive statistics were calculated for the demographic characteristics and the alcohol use patterns of the participants. For participants who answered all three items, we created boxplots of AUDIT-C scores for each possible response to the individual AUDIT-C items.We also constructed receiver operating characteristic (ROC) curves and calculated the area under the curves (AUCs) for the AUDIT-2* and AUDIT-3 against the sex-specific AUDIT-C cut-points used to define hazardous drinking. We calculated sensitivities, specificities, positive predictive values (PPVs) and negative predictive values (NPVs) for a range of cut-points for the AUDIT-2* and AUDIT-3 against the sex-specific AUDIT-C scores. In the absence of any justification to treat misclassifications differently, Youden’s index (J), which places equal weighting on the sensitivity and specificity of a screening test (J = Sensitivity + Specificity − 1), was used to guide the initial decisions regarding appropriate cut-points [23]. The associations between the participants’ sociodemographic characteristics and the hazardous drinking patterns (typical quantity and heavy episodic frequency) determined from the preceding analyses were modelled using multinomial logistic regression following evidence of non-proportionality from ordinal logistic regression models. Pairwise comparisons between levels of categorical independent variables were only performed where the Wald test was statistically significant for that variable. All analyses incorporated sampling weights and used Huber–White robust standard errors to accommodate the clustering effects of campuses.Two-sided P < 0.05 was considered statistically significant in all cases. stata version 13.1 was used for all analyses (StataCorp, College Station,Texas, USA). Results Table 1 shows the characteristics of the 5082 students who were aged up to 25 years and completed the survey (unadjusted 63% women). Although the mean ages for men and women were very similar, there were meaningful differences for several of the other characteristics
Brief alcohol screening in students
543
Table 1. Characteristics of the participants
Demographic factors Age, mean (SD) Undergraduate, n (%) Domestic student, n (%) Accommodation type, n (%) Parents/guardian Shared house/apartment Residence hall Other Relationship status, n (%) Single Relationship not living together De facto/married/civil union Parent, n (%) First in family to attend university, n (%) Employed, n (%) Alcohol use Drinking pattern, n (%) Abstainer Moderate drinkera Hazardous drinkerb
AUDIT-C, median (IQR) Drinking measure, n (%) Typical frequency Occasionallyd Monthlye Weeklyf Typical quantity (standard drinksg) ≤5 6–10 ≥11 Heavy episodic drinking (≥6 standard drinksg) frequency Never Occasionallyd Monthlye Weeklyf
Men (n = 1869)
Women (n = 3213)
Total (n = 5082)
22.4 (1.4) 1292 (68.4) 1660 (89.1)
22.1 (1.4) 2455 (77.2) 2985 (93.1)
22.2 (1.4) 3747 (73.9) 4645 (91.6)
653 (37.8) 955 (49.0) 42 (2.1) 215 (11.1)
1093 (36.5) 1584 (47.1) 75 (2.3) 459 (14.1)
1746 (37.0) 2539 (47.8) 117 (2.2) 674 (13.0)
1138 (62.1) 500 (26.5) 223 (11.5) 21 (1.0) 607 (31.5) 921 (48.1)
1587 (50.3) 991 (30.8) 625 (18.9) 52 (1.7) 1091 (33.6) 1937 (60.7)
2725 (54.7) 1491 (29.2) 848 (16.1) 73 (1.4) 1698 (32.8) 2858 (56.0)
169 (9.5) 1073 (58.9) 619 (31.6)
292 (9.6) 1631 (51.7) 1281 (38.7)
461 (9.6) 2704 (54.4) 1900 (36.1)
n = 1692c
n = 2912c
n = 4604c
5 (3–8)
4 (2–6)
4 (3–7)
334 (20.9) 485 (29.0) 873 (50.1)
707 (25.1) 1036 (35.5) 1169 (39.5)
1041 (23.5) 1521 (33.1) 2042 (43.5)
986 (60.2) 411 (23.8) 295 (16.1)
2080 (72.8) 736 (24.0) 96 (3.2)
3066 (68.0) 1147 (23.9) 391 (8.0)
253 (15.9) 631 (38.4) 377 (21.4) 431 (24.3)
565 (20.4) 1313 (44.9) 638 (21.6) 396 (13.2)
818 (18.7) 1944 (42.5) 1015 (21.5) 827 (17.3)
Numbers are raw frequencies; percentages, means and medians are weighted for the sampling design. aMen: AUDIT-C > 0 and 0 and
R E V I E W
Drug and Alcohol Review (September 2015), 34, 540–548 DOI: 10.1111/dar.12272
Screening for hazardous alcohol use among university students using individual questions from the Alcohol Use Disorders Identification Test-Consumption MEI-LING BLANK1, JENNIE CONNOR1, ANDREW GRAY1 & KAREN TUSTIN2,3 1
Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand, 2Department of Psychology, University of Otago, Dunedin, New Zealand, and 3National Centre for Lifecourse Research, University of Otago, Dunedin, New Zealand
Abstract Introduction and Aims. Abbreviated versions of the Alcohol Use Disorders Identification Test-Consumption (AUDIT-C) instrument have not been investigated among students.We compared a modified second item (AUDIT-2*) (typical quantity per occasion as the number of drinks, rather than categorical responses) and the third item (AUDIT-3) (heavy episodic drinking frequency) with AUDIT-C scores and described their associations with sociodemographic variables. Design and Methods. We analysed cross-sectional data from the 2011 baseline of the Graduate Longitudinal Study New Zealand, including respondents aged up to 25 years (n = 5082, response rate 65%). Hazardous drinking was defined as an AUDIT-C score of seven or greater for men and five or greater for women.We calculated the area under receiver operating characteristic curves, sensitivities, specificities, and positive and negative predictive values for the AUDIT-2* and AUDIT-3. Odds ratios and 95% confidence intervals were calculated to describe the associations between drinking patterns and sociodemographic factors. Results. Based on the sex-specific AUDIT-C cut-points, 36% of participants reported drinking at hazardous levels. For the AUDIT-2*, the best combination of sensitivity and specificity was obtained for a cut-point of five standard drinks. The best cut-point for the AUDIT-3 was for monthly heavy episodic drinking. Positive and negative predictive values were above 0.80 for both cut-points. Consumption was associated with age, degree level, domestic/international status, accommodation type, relationship status and employment. Discussion and Conclusions. The AUDIT-2* and the AUDIT-3 appear to be promising standalone screening items for detecting hazardous drinking in this population of heavy-drinking students. [Blank M-L, Connor J, Gray A, Tustin K. Screening for hazardous alcohol use among university students using individual questions from the Alcohol Use Disorders Identification Test-Consumption. Drug Alcohol Rev 2015;34:540–8] Key words: alcohol drinking, mass screening, students.
Introduction The 10-item Alcohol Use Disorders Identification Test (AUDIT) [1] and its shortened derivatives, including the AUDIT-Consumption (AUDIT-C), comprising the first three items (typical frequency, typical quantity, heavy episodic drinking frequency), the AUDIT-QF, consisting of the first two items, and the AUDIT-3, consisting of the third item only, have been examined in a number of populations in order to screen for hazardous drinking behaviours, alcohol abuse or dependence disorders [2,3]. Compared with the full AUDIT, the AUDIT-C, for some populations [4–6], including university students [7], appears to be a better measure for
identifying individuals who report drinking at hazardous levels. However, cut-points indicating a positive screening result need to be tailored for the population under investigation [3]. For example, AUDIT-C cutpoints of four or five for men, and three or four for women, appear to yield the best balance between sensitivity and specificity in the detection of hazardous drinking or abuse/dependence among adults [3]. Among a tertiary student population in which heavy alcohol use is normative behaviour however, AUDIT-C cut-points of seven for men and five for women have been recommended for the detection of hazardous drinking [7]. Despite being considerably shorter in length and scoring complexity than the full AUDIT, the AUDIT-C
Mei-Ling Blank MPH, Assistant Research Fellow, Jennie Connor MB, ChB, PhD, Professor, Andrew Gray BA, BCom (Hons), Senior Research Fellow, Karen Tustin PhD, Research Fellow. Correspondence to Ms Mei-Ling Blank, Department of Preventive and Social Medicine, University of Otago, PO Box 56, Dunedin 9054, New Zealand. Tel: +64 3 479 7207; Fax: +64 3 479 7298; E-mail: [email protected] Received 6 September 2014; accepted for publication 17 February 2015. © 2015 Australasian Professional Society on Alcohol and other Drugs
Brief alcohol screening in students
may still be too burdensome in some situations. Shorter and less complex instruments may result in greater use in circumstances where opportunistic screening and intervention may be warranted, such as hospital emergency departments or time-pressured primary care clinics. Brief validated instruments may also help reduce respondent burden and improve response rates when included in research questionnaires. The AUDIT-QF has been proposed as an alternative [2,8,9] to the full AUDIT or AUDIT-C; however, as a two-item instrument, a degree of scoring complexity remains, which may limit its utility in certain opportunistic screening situations. Like most other quantity– frequency instruments, the AUDIT-QF measures average consumption over a period of time and does not adequately capture irregular consumption that is atypically low or high compared with the individual’s normal drinking pattern [10]. This misclassification is problematic when considering individuals who typically drink at moderate levels, but who also sporadically engage in hazardous heavy episodic drinking. The AUDIT-3, assessing heavy episodic drinking (six or more drinks per occasion) frequency, is a potentially useful brief screening tool that has the advantage of a simple scoring system. The full ten-item AUDIT has previously been compared with established and novel abbreviated versions in a community-based survey of young people [11]. The AUDIT-C and AUDIT-3, which both measure consumption only, unsurprisingly performed worse than three- and four-item variations that assessed the constructs of dependence and harm in addition to consumption. However, the single-item AUDIT-3 performed almost as well as the three-item AUDIT-C [11], raising the possibility that among young people, the consumption construct of hazardous drinking may potentially be addressed with a single question. Previous research among adults suggests that using a positive screening threshold of ‘ever’ drinking six or more drinks per occasion may result in acceptable sensitivities and specificities for the identification of hazardous drinking [3]. However, there is uncertainty about what frequency of heavy episodic drinking should be considered a positive screening result among a naturally heavy-drinking population, with several researchers recommending higher thresholds for cultures where heavy drinking is common [8,9,12]. Additionally, the quantity threshold of six or more drinks per occasion may not be appropriate for women [12,13]. The AUDIT-3 has not yet been evaluated in a student population. There have also been no investigations assessing the utility of the second AUDIT item (typical quantity consumed per occasion) for detecting hazardous drinking among students. The objectives of this study were: (i) to describe the baseline distribution of alcohol consumption in a
541
nationally representative sample of final-year university students in New Zealand aged up to 25 years; (ii) to compare a modified second AUDIT item (AUDIT-2*), and the AUDIT-3, with total AUDIT-C scores; and (iii) to describe the associations of hazardous drinking patterns, as identified by these individual items, with sociodemographic variables. Methods Data were collected during the first wave of the Graduate Longitudinal Study New Zealand (GLSNZ), a new cohort study investigating the employment, health and social outcomes of university graduates. The GLSNZ baseline survey included over 400 items, including sociodemographic factors; university expectations, experiences and satisfaction; employment plans and career aspirations; academic beliefs and attitudes; current financial circumstances; physical health, disability and functional impairment; health risk behaviours; psychological well-being; personality type; social support and social integration; and local and international community participation. All of the survey items are described in the GLSNZ Extended Baseline Report [14]. Participants and procedures Participants were final-year students from each of New Zealand’s eight universities who were in a program of study that would have allowed them to graduate with a bachelor’s degree or higher after successful completion of their studies during 2011. The sampling procedure has been described in detail elsewhere [14]. However, in brief, final-year students from each university were stratified by subject of study. Within subject areas, universities provided a specified number of randomly selected students according to age groups (5-year age bands, starting from age 15), sex, self-reported ethnicity [New Zealand European, Ma¯ori, Samoan, Cook Islands Ma¯ori, Tongan, Niuean, Chinese, Indian, other (e.g. Dutch, Japanese, Tokelauan)], enrolment status (full-time, part-time), study mode (intramural, extramural), degree level (undergraduate, postgraduate) and fee status (domestic, international). All international PhD students were included, as were all students from the smallest university. Sampling weights were constructed based on the actual number of potentially graduating students in each category (international PhD or other) for each campus in 2011. These data were provided by each university in 2012 after the finalised student numbers were available. The survey was conducted between July and December 2011. Eligible students were contacted by letter and email, and given a unique study code and password to log on to the © 2015 Australasian Professional Society on Alcohol and other Drugs
542
M.-L. Blank et al.
secure survey website. Non-responders and noncompleters were sent multiple reminder emails and were contacted up to four times by trained call centre staff. Overall, 65% (n = 8719/13 343) of the students who were invited agreed to participate and completed the survey.The present analysis is restricted to all of the respondents who were aged up to 25 years (n = 5082). We restricted the age range to reduce the heterogeneity in the sample as ‘mature students’ may have widely different life experiences before and during university study compared with students who matriculated soon after completing high school.These different life experiences may subsequently impact on an individual’s drinking pattern. The study was approved by the New Zealand Multi-region Ethics Committee (MEC/11/ EXP/049). Measures Alcohol consumption. We assessed alcohol consumption using the AUDIT-C [4], which consists of the first three items from the full AUDIT [1] measuring typical frequency, typical quantity per occasion and heavy episodic drinking (six or more drinks per occasion) frequency. Unlike the original AUDIT-C, respondents were not asked to restrict their responses to the previous 12 months (Supporting Information Table S1). Scores for the AUDIT-C range from 0 to 12, with higher scores indicative of a more hazardous drinking pattern [4]. Participants who reported never drinking alcohol were classified as abstainers.They were not asked the quantity and heavy episodic frequency questions, and were assigned an AUDIT-C score of zero. Although the original AUDIT-C does not specify a standardised measure of alcohol, respondents were shown illustrations of New Zealand standard drinks (10 g of pure ethanol) for a range of alcoholic beverages and provided with examples (e.g. ‘a jug of beer equals three drinks’). We included a greater number of response options for each item than the original instrument to allow for more fine-grained analyses of respondents’ drinking patterns. The modified response options were able to be collapsed into the original categories (Supporting Information Table S1). Importantly, for the second item about typical quantity, respondents were able to select a whole number of standard drinks ranging from 1 to 25+, rather than the original ordered classification (e.g. 1–2 drinks, 3–4, etc.). This modified second item is called AUDIT-2*. The AUDIT-3 consists of the third item about heavy episodic drinking frequency. Hazardous drinking was defined as an AUDIT-C score of seven or greater for men and five or greater for women. These are validated AUDIT-C cut-points for identifying university students aged 18 to 25 years who report drinking at hazardous levels [7]. © 2015 Australasian Professional Society on Alcohol and other Drugs
Social and demographic variables. Self-reported information was collected on a range of factors that have been identified as being relevant to an individual’s alcohol consumption during early adulthood, including age, sex, accommodation type, relationship status, parenthood and paid employment [15–22]. Information was also collected on whether the participant was the first person in their immediate family to attend university and the education level of the participant’s most highly educated parent/caregiver. Analysis Descriptive statistics were calculated for the demographic characteristics and the alcohol use patterns of the participants. For participants who answered all three items, we created boxplots of AUDIT-C scores for each possible response to the individual AUDIT-C items.We also constructed receiver operating characteristic (ROC) curves and calculated the area under the curves (AUCs) for the AUDIT-2* and AUDIT-3 against the sex-specific AUDIT-C cut-points used to define hazardous drinking. We calculated sensitivities, specificities, positive predictive values (PPVs) and negative predictive values (NPVs) for a range of cut-points for the AUDIT-2* and AUDIT-3 against the sex-specific AUDIT-C scores. In the absence of any justification to treat misclassifications differently, Youden’s index (J), which places equal weighting on the sensitivity and specificity of a screening test (J = Sensitivity + Specificity − 1), was used to guide the initial decisions regarding appropriate cut-points [23]. The associations between the participants’ sociodemographic characteristics and the hazardous drinking patterns (typical quantity and heavy episodic frequency) determined from the preceding analyses were modelled using multinomial logistic regression following evidence of non-proportionality from ordinal logistic regression models. Pairwise comparisons between levels of categorical independent variables were only performed where the Wald test was statistically significant for that variable. All analyses incorporated sampling weights and used Huber–White robust standard errors to accommodate the clustering effects of campuses.Two-sided P < 0.05 was considered statistically significant in all cases. stata version 13.1 was used for all analyses (StataCorp, College Station,Texas, USA). Results Table 1 shows the characteristics of the 5082 students who were aged up to 25 years and completed the survey (unadjusted 63% women). Although the mean ages for men and women were very similar, there were meaningful differences for several of the other characteristics
Brief alcohol screening in students
543
Table 1. Characteristics of the participants
Demographic factors Age, mean (SD) Undergraduate, n (%) Domestic student, n (%) Accommodation type, n (%) Parents/guardian Shared house/apartment Residence hall Other Relationship status, n (%) Single Relationship not living together De facto/married/civil union Parent, n (%) First in family to attend university, n (%) Employed, n (%) Alcohol use Drinking pattern, n (%) Abstainer Moderate drinkera Hazardous drinkerb
AUDIT-C, median (IQR) Drinking measure, n (%) Typical frequency Occasionallyd Monthlye Weeklyf Typical quantity (standard drinksg) ≤5 6–10 ≥11 Heavy episodic drinking (≥6 standard drinksg) frequency Never Occasionallyd Monthlye Weeklyf
Men (n = 1869)
Women (n = 3213)
Total (n = 5082)
22.4 (1.4) 1292 (68.4) 1660 (89.1)
22.1 (1.4) 2455 (77.2) 2985 (93.1)
22.2 (1.4) 3747 (73.9) 4645 (91.6)
653 (37.8) 955 (49.0) 42 (2.1) 215 (11.1)
1093 (36.5) 1584 (47.1) 75 (2.3) 459 (14.1)
1746 (37.0) 2539 (47.8) 117 (2.2) 674 (13.0)
1138 (62.1) 500 (26.5) 223 (11.5) 21 (1.0) 607 (31.5) 921 (48.1)
1587 (50.3) 991 (30.8) 625 (18.9) 52 (1.7) 1091 (33.6) 1937 (60.7)
2725 (54.7) 1491 (29.2) 848 (16.1) 73 (1.4) 1698 (32.8) 2858 (56.0)
169 (9.5) 1073 (58.9) 619 (31.6)
292 (9.6) 1631 (51.7) 1281 (38.7)
461 (9.6) 2704 (54.4) 1900 (36.1)
n = 1692c
n = 2912c
n = 4604c
5 (3–8)
4 (2–6)
4 (3–7)
334 (20.9) 485 (29.0) 873 (50.1)
707 (25.1) 1036 (35.5) 1169 (39.5)
1041 (23.5) 1521 (33.1) 2042 (43.5)
986 (60.2) 411 (23.8) 295 (16.1)
2080 (72.8) 736 (24.0) 96 (3.2)
3066 (68.0) 1147 (23.9) 391 (8.0)
253 (15.9) 631 (38.4) 377 (21.4) 431 (24.3)
565 (20.4) 1313 (44.9) 638 (21.6) 396 (13.2)
818 (18.7) 1944 (42.5) 1015 (21.5) 827 (17.3)
Numbers are raw frequencies; percentages, means and medians are weighted for the sampling design. aMen: AUDIT-C > 0 and 0 and
