Alcohol and drug use among Danish physicians. A nationwide cross-sectional study in 2014

Studies have shown that 8-15% of physicians experience a substance use disorder (SUD) [1] at some point in their careers, and alcohol use disorders are as common among physicians as they are in the general population [2, 3]. Substances are defined as all kinds of intoxicants, such as alcohol and drugs (licit and illicit). A review showed that SUD is seen among all specialties and that SUD are largely equally prevalent in different specialties [2].

A study comparing the alcohol use of general practitioners and hospital physicians in Denmark and Germany showed that significantly more Danish physicians (12.7%) than German physicians (2.5%) reported hazardous alcohol use [4].

Physicians tend to postpone treatment until their SUD reaches a critical stage [3], even though long-term disorders that go untreated may have fatal consequences [5]. A study of 108 Danish physicians registered with a drug use disorder by the Danish Health Authority in 1949-1957 revealed that after 20 years, only 40% of the registered physicians were alive as opposed to 80% among other groups of Danes with drug use disorder; and at least 25% of the dead physicians had committed suicide [6]. This dismal prognosis was confirmed in two later Danish studies [5, 7].

Danish physicians’ patterns of licit and illicit drug use are unknown. Furthermore, little is known about the management of colleagues with SUD in physician workplaces. If we are to improve prevention and treatment of SUD, we need to know more about the frequency of SUD and whether specific groups are more prone to SUD than others.

The aims of this study were to describe alcohol and drug use among Danish physicians (including illicit as well as licit drugs), the physicians’ self-reported assessment of their own alcohol and drug use, and their management of colleagues with SUD in physician workplaces.

METHODS

Study population

Data were collected using a cross-sectional survey among 4,000 physicians randomly selected among all 26,669 active members of the Danish Medical Association (DMA). We sampled 1,333 respondents from each of the following three DMA subgroups: The Danish Association of Junior Doctors (DAJD), The Danish Association of Medical Specialists (DAMS) and The Danish Organisation of General Practitioners (DOGP). Retired physicians were excluded.

Data collection

The 4,000 physicians received an e-mail with a hyperlink to an electronic questionnaire (Survey Xact) distributed by the DMA (April-June 2014). To encourage participation, a personalised hard-copy letter signed by the Chairman of the DMA was distributed to all participants before the email was sent. Additionally, the study was mentioned in the DMA journal before the emailing of the survey and between the second and third reminders. Reminders were sent out three times within eight weeks. Participation was not remunerated. The questionnaire took approx. 30 minutes to complete. Twenty-two physicians asked to be excluded from the survey because of recent retirement, work abroad and maternity leaves.

The questionnaire

Based on a pre-study consisting of qualitative interviews study with physicians who had previously been addicted/had a SUD, we identified themes for the questionnaire. Thereafter the questionnaire was developed and validated, and standardised self-rating scales were selected. All ad-hoc items were developed in the research group and pilot-tested before use. The questionnaire was pilot-tested by 30 randomly selected physicians representing the three DMA groups. Items were tested cognitively and for floor/ceiling effects and missing items. The ad-hoc items were inspired by a Danish national survey of the population’s use of intoxicants [8] and Danish and Norwegian surveys of physicians’ health and work conditions [9].

We used standardised measures of problematic alcohol use based on the internationally used Alcohol Use Disorders Identification Test (AUDIT). A modified version of the Drug Use Disorders Identification Test (DUDIT) [10] was used for the investigation of drug use. The DUDIT questions were slightly modified as our focus was on drug (licit and illicit) use for the purpose of intoxication. Both scales are based on the International Classification of Diseases, 10th version (ICD-10) definitions and developed for the World Health Organisation
[10, 11].

The ten questions in AUDIT and the 11 questions in DUDIT are scored on a five-point Likert scale from 0 (“never”) to 4 (“daily or almost daily”). The maximum score on AUDIT is 40 points. Following recommendations [12], the AUDIT score was divided into four groups defined as follows: no hazardous use (< 8), hazardous alcohol use (risky use potentially harmful and causing dependence) (8-15), harmful alcohol use (causing physical or mental harm) (16-19) and alcohol dependence (≥ 20). In the present article, we use the term “risky use” to comprise, hazardous, harmful and dependent substance use.

DUDIT includes both questions about illicit drugs and commonly abused prescription medication; and it identifies dependence, hazardous use and harmful use of drugs with a maximum score of 44 points [10]. DUDIT is a relatively new tool; and there are ongoing discussions concerning the cut-off points for scores. In line with other studies [13], the present study chose 1 as the cut-off for risky use (potentially harmful and causing dependence) for both sexes as Denmark has a zero tolerance of drug use for intoxication purposes.

Statistical analyses

The analyses were performed using STATA software (version 13.1) and the included survey package. Differences between subgroups of physicians concerning substance use were tested using Pearson’s chi-squared test when number and percentage (%) were reported. The 27 medical specialties were categorised into seven groups (see Appendix). When mean and standard deviation were reported, one-way analysis-of-variance (ANOVA) was used. All reported estimates including proportions were weighted when appropriate according to the original proportion of physicians in the three DMA groups.

Ethics

The data collection was approved by the Danish Data Protection Agency (case no. 2013-41-1996). The physicians’ identities are only known by the DMA, who has no access to the data. A funding agreement ensured the authors’ independence in designing the study, interpreting the data, writing and publishing the report.

Trial registration: not relevant.

RESULTS

A total of 1,943 (48.6%) physicians completed the questionnaire. The respondents’ socio-demographic characteristics are shown in Table 1. The overall proportions of physicians reporting risky alcohol and drug use (hazardous, harmful and dependent use) were 18.9% and 3%, respectively (Table 1).

A total of 108 (6.4%) physicians had used drugs (licit or illicit) at some point in time during their medical career. A total of 58 physicians (3%) had used drugs with the purpose of intoxication (34 (4.1%) males and 24 (2.5%) females).

During the year preceding the survey, 95 (4.7%) physicians had used sleeping medicine, and 61 (3.2%) physicians had used prescribed painkillers (other than over-the-counter medicine) without medical indication. During the past year, 184 (7.9%) physicians reported using tranquillisers less than once a week, and 71 (3.4%) had used sleeping medicine without medical
indication.

Table 2 shows physician characteristics in relation to substance use. Among male physicians, 25.1% reported risky alcohol use as opposed to 14.4% of the female physicians (< 0.001).

Within the three sub-associations, approx. 2.5% reported harmful or dependent use of alcohol. Among the emergency medicine specialty, a score indicating harmful or dependent use of alcohol was seen for 5.6% (Table 2). The specialties with the highest prevalence of risky alcohol use were internal medicine and emergency medicine both with (24%), and the lowest score was recorded for general practice (16%). In total, 8 (0.4%) respondents had been in treatment for SUD.

The main reported reasons for alcohol or drug use among physicians with a risky substance use were to
enjoy the taste (74.4%) and to relax efficiently after work (54.6%) (Table 3). Among the 383 respondents with risky substance use, 76.9% characterised their substance use as unproblematic.

In case a colleague showed signs of SUD, 57.7% of the physicians reporting risky substance use and 55.5% of those reporting unproblematic use stated that they would offer their help and encourage their colleague to seek treatment (Table 4). A total of 60 physicians (2.5%) reported that SUD was discussed openly at their workplace.

DISCUSSION

We found that nearly one-fifth of physicians had engaged in risky alcohol use, with similar results within the three sub-associations of the DMA. The highest proportions of physicians with a risky alcohol use were found in internal and emergency medicine and in surgery. Male physicians were statistically significantly more likely to have a problematic use of alcohol than female physicians. Overall, 3% reported a hazardous use of drugs. About three quarters of those reporting a risky substance use considered that their use was unproblematic. If a colleague showed signs of SUD, the preferred action was personal contact.

Strengths and limitations

This is the first national survey of its kind in Denmark. We used two standardised screening tools, AUDIT and DUDIT. For AUDIT, a cut-off point at 8 has previously been tested, and this cut-off point yields a sensitivity of 98% and a specificity of 94% for hazardous alcohol use [11]. Denmark has a zero tolerance policy for use of drugs for intoxication purposes. For DUDIT, we therefore decided to use 1 as the cut-off point for both sexes, which indicates that any use of drugs for intoxication purposes is risky; a choice, which is in line with e.g.
[13].

The response rate of 48.6% could lead to potential selection bias. A recent Austrian e-mail-based survey had a response rate of 18% [14], and a Danish/German postal survey had a response rate of 74%. Considering this and the delicate nature of the topic, we consider the response rate to be satisfactory. The selection bias due to non-response may imply that respondents had fewer substance use problems than non-responders because physicians with substance use problems would be more reluctant to reveal these to themselves and to their colleagues. An international review has suggested underreporting to vary between 40% and 60% in studies of alcohol use [15]. This indicates that our study may underestimate the prevalence of risky substance use among physicians.

We adjusted our analyses for the weighted sample to ensure that the prevalence rates reported within each group and overall were comparable and reported the actual figures. We used disproportional sampling to ensure a high statistical precision for every stratum, which was recorded.

Comparison with other studies

We found that around 19% of the physicians reported risky alcohol use. This percentage is higher than percentages reported in international studies where rates span from 10% to 15%. However, these differences may be due to differences in measurement tools. In comparison with similar screening methods Cut Down, Annoyed, Guilty, Eye-Opener (CAGE) and Michigan Alcohol Screening Test (MAST), AUDIT appeared to be the best screening tool to identify hazardous use and/or dependence as defined by the Diagnostic and Statistical Manual of Mental Disorders (DSM) [1, 16]. A survey from 2013 showed that 20.6% and 8.5% of the Danish population exceeded the limits (recommended by the Danish Health and Medicines Authority) of 7-14 units of alcohol per week for females and 14-21 for males, respectively [17]. Risky alcohol use was significantly more prevalent in male than in female physicians, which corresponds to international findings [2].

Physicians with risky substance use had a low degree of problem recognition as nearly three quarters believed that their use was unproblematic. As international research has shown that physicians tend to treat their patients in correspondence with their own health conduct, this may be an important finding [4]. Additionally, only seven respondents had received treatment. Similarly, results from other countries show that physicians rarely seek treatment on their own initiative [18]. The direct implication for patient safety is not well-known, but a recent US survey found that a high proportion (78%) of the surgeons reporting a medical error in the previous three-month period had alcohol abuse or dependency [19].

Research focusing on the reasons for substance use among physicians is scarce, one exception being Merlo et al [20]. In line with this study, we found that the majority of physicians with risky substance use reported that they used substances to enjoy the taste and to relax efficiently after work.

To our knowledge, there are no quantitative studies of the workplace management of colleagues with SUD before they enter treatment. Our study shows that when a colleague shows signs of SUD, half of the physicians reported that they would encourage this colleague to seek treatment, and around one third of the DAMS and DAJD physicians would inform their managers. DOGP members, who own their own practices, were almost twice as likely as physicians from DAMS and DAJD to report that they would use the Network of Physician Colleagues. The conditions and attitudes towards workplace management of SUD thus seem to vary.

CONCLUSION

The prevalence of problematic alcohol and drug use was 19% and 3%, respectively. Males had SUD significantly more often than did females. Three quarters of physicians reporting risky substance use did not recognise their risky use of substances. Very few found that there was an explicit procedure and openness about SUD in workplaces, and most would have a personal talk with a colleague showing signs of SUD. Our study indicates a need for more openness about SUD among physicians. Our study also indicates a need for prevention, monitoring and explicit procedures for managing and treating SUD, which seems to be a prevalent issue among physicians. Besides the human implications for physicians, such risky alcohol and drug use is important for patient safety. More research is needed to get a thorough understanding of the associations between this issue and both psychosocial and work cultural factors affecting it to direct prevention and intervention measures expediently.

Correspondence: Johanne Korsdal Sørensen, Section of General Practice, Department of Public Health, University of Aarhus, 8000 Aarhus C, Denmark. E-mail: jks@ph.au.dk

Accepted: 18 June 2015

Conflicts of interest:Disclosure forms provided by the authors are available with the full text of this article at www.danmedj.dk

Acknowledgements: We would like to express our gratitude to the physicians who participated in the present study. The DMA has been extremely supportive in making the study possible. We thank senior researcher, Judith Rosta (The Institute for Studies of the Medical Profession, Norway), who generously commented on drafts of the paper. Additionally, we extend our gratitude to senior researcher and DUDIT developer, Anne Berman (Karolinska Institutet, Sweden) for advice on the use of DUDIT.

Download Appendix here
Dan Med J 2015;62(9):A5132