Alcohol Use Disorder and Associated Factors Among Medicaland Surgical Outpatients of University of Gondar Specialized Hospital: A Cross-sectional Study

Background: Alcohol use disorders increase the risk for physical harm, mental or social consequences for patients and others in the communities. Studies on alcohol use disorder and associated factors among medical and surgical outpatients in Ethiopia are limited. Therefore, this study is meant to provide essential data alcohol use disorder and associated factors among medical and surgical outpatients for future interventions. Methods:An institution-based cross-sectional study was conducted using the systematic random sampling technique. Alcohol use disorder was assessed using the World Health Organization’s 10-item Alcohol Use Disorder Identication Test (AUDIT) questionnaire. Data were analyzed using the Statistical Package for Social Sciences (SPSS) version 20. Bivariate and multivariate logistic regression analyses were performed, P-values less than 0.05 were considered statistically signicant in the multivariate analysisand the strength of association was measured using AOR at a 95% condence interval. Results: The prevalence of alcohol use and use disorder were 322(68.50%) and 111(34.5%), respectively. In the multivariate analysis, male sex (AOR=3.33, 95%CI: 1.40, 7.93), history of mental illness (AOR=2.68, 95%CI: 1.12, 6.38), drinking for relaxation (AOR=1.88, 95%CI: 1.02, 3.48) and history of lifetime tobacco use (AOR=5.64, 95%CI: 1.95, 16.29) were signicantly associated with the disorders. Conclusions: The prevalence of alcohol use disorders among medical and surgical outpatients was found to be high. Male sex, history of mental illness, use of alcohol for relaxation and tobacco smoking need further attention in the assessment of the disorder.

accounted only for many of the deaths although it posed complications in various medical and surgical conditions [6,7].
Alcohol use disorders were reported to be 20-40% in general medical hospital settings [8,9]. The prevalence of patients with AUDs identi ed in general hospitals was higher than that in community surveys. The reason is that people with AUDs often seek help only when they become medically ill.
However, in overcrowded clinical settings, medical staff often fail to recognize AUDs unless there are obvious physical or psychosocial effects relating to alcohol abuse [4]. Studies involving hospitalized patients showed that up to one-third of patients admitted to medical and surgical wards had alcohol related conditions. In some studies, nearly 40% of patients hospitalized for medical and surgical illnesses had AUDs [2,10]. Despite this high prevalence of AUDs in clinical settings, many studies revealed that less than a third of such individuals were identi ed and that only 5%-10% of the patients were referred to psychiatric services for further evaluation and management [11,12]. The undetected alcohol problems led to unpleasant consequences, and it is for this reason that AUDs are the major causes of morbidity and mortality among medical and surgical patients.
Previous studies have indicated that younger groups, the less-educated, cases with history of smoking, drinking fathers, male sex and patients getting treatment in internal medicine wards were factors signi cantly affecting AUDs [4,[13][14][15].
Evidences from previous studies showed that alcohol related problems among patient with general medical conditions posed greater consequences including the worsening of the prognosis of the illness.
The investigators, through their liaison services, observed a high number of patients with medical and surgical illnesses were consulted for psychiatry intervention due to alcohol withdrawal related problems.
But, the magnitudes of alcohol use and use disorders among these groups of people have not been well explored in Ethiopia, particularly in the study area. These conditions led the investigators to assess alcohol use disorders and associated factors among patients treated for medical and surgical health problems. Therefore, the results of this study could be vital input for health care providers to be vigilant of alcohol related problem and institute early interventions.

Methods And Materials
Study design, period and setting An institution based cross-sectional study was carried out from May 01 to 30, 2016. The study was conducted at the University of Gondar specialized hospital northwest Ethiopia. It is found in Gondar town, 727 Km from Addis Ababa, the capital of Ethiopia. The hospital had 550 beds with around six medical and two surgical outpatient departments (OPDs). In the last twelve months, the average adult outpatient ow to the medical and surgical outpatients was 2034 and 718 respectively.

Population
All medical and surgical outpatients aged 18 years and above and visited the hospitals during data collection were the source population, whereas patients with medical and surgical problems included into the sample and from whom information was collected were the study population.

Sample size determination and technique
The sample size was determined using the single population proportion formula. So far there has been no study on AUDs among surgical and medical outpatients in Ethiopia. Therefore, the initially the sample size was calculated by taking a 50% proportion at 95% level of con dence and a 5% margin of error.
Assuming a 10% nonresponse rate, a total sample of 423 was obtained. To determine the maximum sample size, alternative sample size determination was done using the risk factors of alcohol use disorder. Previous studies indicate that male sex, use of cigarettes and khat are signi cantly associated with alcohol use disorder [16]. Based on these factors, sample size was determined with the following assumptions (Table 1). Thus, using a proportional allocation, 351 and 124 patients from the medical and surgical OPDs, respectively, were included in the total sample of 475. The systematic random sampling technique was used to recruit participants. Sampling interval was determined by dividing the number of the average monthly outpatients by the proportionally allocated outpatients to each department (k = 2034/351 = 6 to medical and k = 718/124 = 6 to surgical patients). Therefore, the participants were interviewed at every sixth regular interval, and the rst participant was selected by the lottery method.

Measurements
Data were collected by face to face interviews and chart reviews conducted to know the types of diagnoses. The socio-demographic and related parts of the questionnaire were developed by reviews of literature. Alcohol use disorder was assessed by the WHO Alcohol Use Disorders Identi cation Test (AUDIT) screening tool [1]. It is a primary and most effective tool to identify problematic alcohol use at an early stage with sensitivity 94.1% and speci city 91.7%. The AUDIT rst three questions (1-3) were concerned with the quantity and frequency of alcohol consumption (Hazardous Alcohol Use); the second three questions (4-6) assessed signs of alcohol dependency while the last four (7-10) investigated alcohol-related problems (harmful alcohol use). A total AUDIT score of eight or more was used to de ne AUDs. This alcohol screening tool has been validated across different African countries [1,6,17]. The local ethanol beverages were converted into standard drinking units based on their local measurement with an equivalent standard drinking measurement [16]. The data were collected by four trained BSc degree nurses graduate supervised by one MSC in psychiatry health professional.

Data quality control
The English version questionnaire was translated to Amharic and back to English by two different native language experts. A pre-test was conducted on 5% of the sample one week before the data collection at Felegehiwot referral hospital, Bahidar, Ethiopia, to check understandability and reliability of the questionnaire. The supervisor and data collectors were trained for one day before the data collection on the objective of the study, and on how to handle ethical issues and the con dentiality of information.

Statistical analysis
Data were checked for completeness, cleaned manually, pre-coded, entered into EPI info version 7 and exported to SPSS version 20 for further analysis. For descriptive variables, frequencies, percentages, graphs, and tables were used. In the binary logistic regression analysis, variables found to have a p-value of less than 0.2 were candidates for the multivariable logistic regression analysis. In the multivariable logistic regression analysis, variables with less than 0.05 p-values were considered as signi cantly associated with the outcome variable. The strength of associations was explained with OR at 95% con dence interval (CI).    (Fig. 1).

Factors associated with alcohol use disorders
In the bivariate analysis, sex, educational status, diagnosis of injury, history of mental illness, occupation, lifetime tobacco use, peer pressure and drinking for relaxation were candidates for the multivariable logistic regression analysis at p-value < 0.2. In the multivariable logistic regression analysis, male sex, history of mental illness, drinking for relaxation and lifetime tobacco use were signi cantly associated with AUDs at p-value < 0.05.   [19], and 32.9% in Brazil among hospitalized patients [2]. On the other hand, the current result is higher than those of two different studies in Ethiopia 21% [20], and 3% [16], 9.7% in Nigeria among patients attending family medicine [21], 25.1% and 10.8% in Kenya [22], 18.9% and 27.6% in two different institutional studies on hospital outpatients in South Africa [23,24], 4.1% in a community-based survey and 5% in a facility study on men in Uganda [25], 9.5% in another institutional study on high risk sexual behavior outpatients in Uganda [26], 20.3% on tuberculosis patients [27], and 10.9% on primary health center male outpatients in south India [28], 15% in northern Ireland [29], 4.1% in Eastern Mediterranean region [30], and 7.3% in a study done at a primary health facility in Nepal [31]. The possible reasons for the discrepancy might be variations in study populations. For instance, a community survey was carried out in Ethiopia, while only TB patients, who had chances to abstain due to their illnesses, were sampled in India. Furthermore, general outpatients were dealt with in Nigeria and the medication respondents used were focused on in Kenya, while both in and outpatients were studied in Ireland. Besides, differences in ndings also relate to the tools used. For example, the Fast Screening Test was put to use in Ethiopia, whereas the short Alcohol Dependence Data Questionnaire was utilized in Brazil. Research results could also vary owing to study designs. For example, the Randomized Control Trial with follow ups ranging from six to twelve months was employed in South Africa. Moreover, investigation outcomes might also differ due to the socio-cultural practices of participants.
On the other hand, the prevalence of alcohol use disorder in this work is lower than that of a study conducted on medical and surgical outpatients aged 45-64 years in Nigeria (41.4%) [32], 53.5% on America Veteran Affairs outpatients [33] and 40.5% in Nepal [34]. The possible reasons for the difference might be the tools. In Nigeria, for instance, a structured clinical interview Diagnostic Statistical Manual-IV was employed, while the International Classi cation of Disease-9 code for alcohol use disorder which has a high sensitivity for assessing such problems was used in America. When it comes to study populations; Veteran Affairs outpatients in America, while men in the 45-64 age group more prone to consume alcohol which is likely to increase prevalence were interviewed in Nigeria. In Nepal, only medical outpatients were considered. As a matter of fact, socio-cultural variations are also responsible for differences in study results.
In our study, the prevalence of AUDs was 85.6% among men and 14.4% for women (male: female ratio 5.9:1). This showed that the prevalence rates were much narrower than was reported in China 66:1 [35], and slightly higher than those same epidemiological surveys in the United states 5:1 [36].
In this study, male sex was signi cantly associated with AUDs. The nding is supported by studies in Taiwan [14], the Republic of Ireland [3], Brazil [2], India [37], Tanzania [38], and Kenya. Alcohol drinking is more socially acceptable among males than females, predisposing men to AUDs.
In the current study, history of tobacco use was signi cantly associated with AUDs. The result is similar to ndings in South Africa [23], Sri Lanka [39], and India [37]. The possible explanation might be that smokers used alcohol to stop the stimulation of the nicotine after they smoked.
History of mental illness was signi cantly associated with AUDs. The nding of our study was similar to those of Ethiopia [20], Nigeria [21], and in South Africa [23]. The possible explanation could be that patients were using alcohol as a self-treatment.
Participants who were using alcohol for relaxation were statistically signi cant predictors of AUDs.
Psychological distress and stressful life events were risk factors for the use of alcohol a study reported in Sri Lanka [39]. Users of alcohol for relaxation have chances to increase the dose of alcohol to get the desired effect. Therefore, they are prone to develop alcohol use disorders.

Limitations
The study is, we hope, effective in that it has used a standard tool for assessing the disorders. Despite its capacity to provide valid evidence however it has some limitations in that it is subject to social desirability bias. Besides, since data were gathered by an interviewer administered questionnaire, respondents might have tended to reply in ways favorable to others by either under or over reporting. Moreover, the cross-sectional design we used has prevented us from reporting casual effect relationships.
The other limitation is, the association of socio-demographic factors and the types of illnesses. That is the majority of the patients sought help for physical illness relating to AUDs rather than for the latter alone. Thus, further studies that include in-depth variables and other designs are needed.

Conclusion
The prevalence of alcohol use disorders among medical and surgical outpatients was found to be high.
The habit of screening and managing the problems in medical and surgical departments is low. The comorbidity of alcohol use problems with physical illnesses may affect the prognosis by further complicating physical illnesses due to diminishing the immunity and withdrawal effects. Therefore, collaborative consultation-liaison works between psychiatrists and other health professionals need to be strengthened. Training of clinicians on how to detect, manage and refer patients with alcohol withdrawal and use disorders is required. Moreover, male sex, history of mental illness, use of alcohol for relaxation and tobacco smokers need further attention for the assessment of the disorder. Ethical Approval was obtained from the Ethical Review Board of the University of Gondar. Letter of permission was obtained from the University of Gondar specialized hospital. The purpose of the study explained and informed written consent was received from study participants. Con dentiality was maintained by omitting personal identi ers.

Consent to publish
Not applicable