Abstract
The Alcohol Use Disorders Identification Test (AUDIT) was developed by the World Health Organization (WHO) and is able to screen separately for hazardous, harmful and dependence-related symptoms associated with alcohol [1]. The AUDIT has been used in a number of general hospital settings including patients attending Accident and Emergency [2–4] and with inpatients and outpatients attending general hospital Medical and Surgical Departments [5].
With the exception of one study undertaken by the author and colleagues which reported on the frequency of hazardous alcohol use and dependence [6], the use of the AUDIT in the general hospital psychiatric setting has not been reported. This is despite considerable evidence that hazardous alcohol use is common practice among persons with psychiatric diagnoses [6–12], and may be positively related to the severity of psychiatric sequelae [13–17].
This paper is a substudy of this larger study which reported on alcohol consumption in patients admitted to the psychiatric units of general hospitals [6] and discusses major impediments to AUDIT screening in these settings. The paper also makes some recommendations for those wishing to undertake alcohol screening among this population.
Method
Subjects
The sample was drawn from persons aged 18–64 who resided in the Perth metropolitan area, and who were admitted to the psychiatric wards of two general public hospitals in Perth (Sir Charles Gairdner Hospital and Royal Perth Hospital) in 1994 and 1995. Admissions originated from the Accident and Emergency Department, general hospital wards, or from general practitioners, psychiatrists and other acute mental health services. Patients typically were diagnosed with acute psychiatric disorders requiring short-term hospitalization. Patient diagnoses were recorded into hospital medical records according to ICD-9 codes. The project's research officer retrieved the primary discharge diagnoses from the medical records. These were later organized into diagnostic categories as outlined in the DSM-IV [18] for the purpose of data entry and analysis.
All patients aged 18–64 and meeting the criteria for one of the following diagnoses of of mood, adjustment, anxiety or psychotic disorders were targeted for AUDIT screening as part of a larger study aimed at assessing the effects of a brief alcohol intervention. The census of the psychiatric wards at each hospital was checked two or three times a week to identify new admissions. Patients were approached by the projects' research officer to complete the AUDIT following advice from the primary nurse that they were stabilized and not suffering any cognitive impairment.
Patients were excluded for the following reasons; too old or young (over 65 or under 18 years of age), lived outside of metropolitan Perth, intellectually handicapped, interpreter required or the patient was considered too aggressive.
All patients admitted were targeted for AUDIT screening. Data on reasons for non-screening of each patient were recorded under two major headings of Cognitive Dysfunction and Left Ward before Screened. Cognitive Dysfunction was divided into three subcategories of Low IQ, Language Difficulty and Diagnosis Resulting in Dementia. Left Ward before Screened was divided into four subcategories of Discharge before Screening, Transferred to a Psychiatric Hospital, Discharge against Medical Advice and Short Admission. A Short Admission was defined as a hospital stay of 3 days or less.
Data on length of hospital stay for all patients, both screened and non-screened, were also recorded. All data were entered onto an SPSS database. This data set was analysed to identify mean length of stay for screened and non-screened patients by major diagnoses of mood, adjustment, anxiety and schizophrenia/psychosis. The SPSS data set was further analysed to identify the major reasons for non-screening in the four major diagnostic groups.
Statistical analyses
To compare if there were differences in the proportion of subjects who were not screened between diagnostic groups the ‘difference’ between two populations was used. Significance levels were adjusted using Bonferroni's comparison-wise error rate for multiple comparisons. Using a t-test on log-transformed data mean (standard error) hospital days for screened and non-screened within each of the four major diagnostic groups were compared.
Results
There were 990 psychiatric admissions during the 12 month study period, 541 (54.8%) admitted to Sir Charles Gairdner Hospitaland 449 (45.2%) to Royal Perth Hospital. Women comprised 598 (60.3%) and men 394 (39.7%) of the patients. Ages ranged from 18 yearsto 64 years, with a mean of 35.4 years (SD = 11.6 years). No significant difference in age of patients was detected between the two hospitals. Mean length of stay was 15.8 ± 16.2 days, with a range from 0to 127 days.
The most common diagnoses among the 990 admissions were mood disorder (43.3%), adjustment disorder (14.3%), anxiety disorder (15.5%) and schizophrenia/psychosis (11.1%), accounting for 84.2% (n = 834) of total admissions. Primary substance-related disorders and personality disorders represented only 4.4% and 5.5% of admissions, respectively, with other DSM-IV or ICD-9 categories (e.g. delirium, dementia, amnestic or other cognitive disorders; mental disorders due to a general medical condition; somatoform disorders; factitious disorders; dissociative disorders; eating disorders; and sleep disorders) representing 5.9% collectively.
Of these 834 patients with a diagnosis of mood, adjustment, anxiety or schizophrenia/psychosis, 571 were screened and 263 were not. Nonscreening was for a number of reasons. Of the four major diagnostic groups screening was not carried out on betweenapproximately 27–42% of each diagnostic group. There was no significant differences in the proportion of patients screened versus not screened for mood, adjustment or schizophrenia/psychosis. There was, however, significantly less patients with anxiety disorder screened compared with mood disorder (p = 0.0024).
Reasons for non-screening
The majority of patients (n = 209) left the ward before screening could be undertaken. This included discharged before screened (86), transfer to a major psychiatric hospital (31), short admission defined as a hospital stay of 3 days or less (79), and discharged against medical advice (13). Twenty-seven patients were not screened due to cognitive dysfunction. This included low IQ (13), language difficulties (8) and dementia-type diagnosis (6). Twenty-seven patients refused to undertake screening, 14 because they had previously been screened at another hospital (Table 1).
Reasons for not being screened among major diagnostic categories
Leaving the ward before AUDIT screening could be undertak en was the major reason for non-screening among major diagnostic groups (mood disorder, 76%; anxiety disorder, 87.5%; adjustment disorder, 77.5%; schizophrenia/psychosis, 79.2%). Within this category, the subcategories of discharged before screened and short admission accounted for approximately 74–93% of non-screening in all four major psychiatric diagnoses.
Transfer to another facility was only noted as a major reason for non-screening in the mood (16%) and schizophrenia/psychosis (24%) groups, while discharge against medical advice only featured as a reason for non-screening for the diagnosis of adjustment disorder (11.5%).
Cognitive dysfunction accounted for only a small amount of nonscreening among major diagnostic groups (mood disorder, 9.5%; anxiety disorder, 7.8%; adjustment disorder, 12%; schizophrenia/ psychosis, 15.6%. Table 1).
Length of hospital stay
Length of stay was also recorded for 981 of 990 patients. Number of days in hospital ranges from 0 to 127 days, with a mean length ofstay 15.8 days (SD = 16.2 days). No significant difference was detected between length of hospitalization between either hospital (SCGH, 16.4 days vs RPH, 15.1 days).
Those non-screened patients in all major psychiatric groups had significantly shorter hospital stays than their diagnostic counterparts who were screened (p < 0.001) [Table 2.
Mean hospital stay (days) by reasons for not being screened among major diagnostic categories compared with those screened
Discussion
For those providing treatment services to persons with psychiatric disorders, screening for problem alcohol use and dependence is especially important. Not only does a substantial body of evidence exist to indicate that hazardous and problem alcohol use is common practice among persons with psychiatric diagnoses [6–12], but that this consumption is likely positively related to the severity of psychiatric symptoms observed [11],[13–17]. In addition, the recent use of alcohol or the presence of withdrawal symptoms such as anxiety, disorientation, hallucinations or confusion may mimic primary psychiatric symptoms, making accurate initial diagnosis more difficult, with compliance with treatment regimens likely to be compromised in those with problem alcoholuse [19–22].
The current paper represents a substudy of a larger one reporting on the frequency of hazardous alcohol use and dependency in psychiatric patients admitted to general hospital psychiatric wards [6]. This larger study indicated that among those with mood disorders, 25.4% of men were alcohol dependent compared with 16.3% of women, while 34.5% of men with anxiety disorder were alcohol dependent compared with 25% of women. Both gender differences were statistically significant. The difference between men and women was even greater for adjustment disorder (44.4% vs 14.5%) and psychosis (29.2% vs 4.2%). More men than women with anxiety disorder were also classified as hazardous (24.1% vs 11.7%) or harmful (11.8% vs 3.3%). The study concluded that there is a high rate of excessive alcohol consumption in people with psychiatric disorders, especially males. It further concluded that the potential for decreasing both the severity of psychiatric symptoms and number of hospital admissions following cessation or reduction in alcohol consumption is considerable, with the AUDIT being a simple screening instrument for identifying problem alcohol use and dependence.
However, as highlighted in the current paper, approximately one-third of all 990 psychiatric admissions were not screened in an environment where 100% screening of all patients was the objective. The failure to screen such a large number of patients can be largely attributed to short hospital stays. Screening during hospitalization was not possible due to the specific characteristics among patients with psychiatric disorders, compared with other types of hospital admissions, and the service provided within a general hospital psychiatric setting. These issues and recommendations to improve alcohol screening in this setting are discussed below.
WHO guidelines for AUDIT questionnaire administration recommend screening when the patients' conditions have stabilized and they are accustomed to the health setting where the interview is to take place [1]. Studies reporting on the use of the AUDIT across a broad range of hospital and other primary health-care settings have not reported difficulty in meeting this advice. Even in the Accident and Emergency setting, where acute crises are commonplace, no difficulty was reported [3]. Similarly, despite the procedures and treatments associated with hospitalizations in the medical or surgical setting AUDIT screening was not reported to be onerous or difficult [5]. This scenario is of course not difficult to accept, given that for the majority of patients the initial reason for medical or surgical admission is dealt with quickly leaving the patient, in most instances, lucid and readily able to answer AUDIT questions.
The patient with a psychiatric disorder however, differs somewhat from these other general hospital populations in a number of significant ways. First, the patient with a psychiatric disorder is often admitted with psychiatric symptoms that may take several days or weeks to resolve. This situation contrasts dramatically to the medical or surgical admission where, with the exception of residual postoperative pain, major sequelae are largely resolved in 2–3 days. Second, unlike pharmaceutical treatments commonly directed at stabilizing physical sequelae in medical or surgical wards, those used in the stabilization of psychiatric sequelae often cause significant cognitive impairment and become an additional problem for AUDIT screening. For example, antidepressants administered to severely depressed patients may initially result in sedation with longer administration being associated with improved cognitive function [23]. AUDIT screening is difficult before adequate resolution of initial psychiatric symptoms, and acclimatization to pharmaceutical treatment regimens.
Once major psychiatric symptoms have abated and pharmacological effects stabilized, it might be assumed that the patient could be screened. A number of additional features of the persons with psychiatric morbidity, however, need to be taken into consideration. For example, in persons with schizophrenia, decreased cognitive functioning particularly affects workingmemory [24]. In addition there is limited normalization of cognitive functioning which may indicate a residual deficit intrinsic to schizophrenia [25]. Similarly, among those suffering depression, there are deficits in intellectual functioning including slowing of thought, poor concentration, indecision and memory problems. A core group demonstrate cognitive dysfunction which persists much longer than acute depressive symptomatology [26].
So, even in the stabilized patient, cognitive dysfunction ensures that a proportion of patients will experience difficulties completing the AUDIT without assistance, even after the initial psychiatric symptoms and pharmaceutical stabilization period. In the current study 10% of all cases were not screened due to cognitive dysfunction.
One important implication of these findings is that although the AUDIT was designed to be either self or staff administered, the latter may be more appropriate to the psychiatric inpatient population. Staff administration affords the opportunity to check that patients have a clear understanding of AUDIT questions and to correct any misunderstanding.
In long-term inpatient psychiatric facilities, such as major psychiatric hospitals, where patients frequently remain for several weeks, identification of patients suitable for AUDIT screening (following resolution of acute psychiatric sequelae and medication stabilization) may present no problem. The emphasis of most general hospital psychiatric wards is, however, on short-term stay with early return to the community. In these settings the window period between the presence of major cognitive impairment associated with acute psychiatric sequelae and/or pharmaceutical treatments and discharge is often only 1 or 2 days.
While discharge before screening accounted for non-screening in 21, 32 and 33% of patients with schizophrenia/psychosis, adjustment or mood disorder, respectively, these patients recorded mean hospital stays of between 4 (adjustment and mood) to 6.5 days (schizophrenia/psychosis). In other general hospital settings such as medical or surgical, this period of hospitalization would likely allow time for patient stabilization and AUDIT screening to take place. Such was not the case with the inpatient with a psychiatric disorder. Anxiety disorder patients discharged before screening recorded the lowest mean hospital stay of 3 days. This reduced hospital stay may help explain why significantly fewer patients with this disorder were screened compared with those with schizophrenia/psychosis.
It should also be noted that 16% and 24% of nonscreened mood and schizophrenia/psychosis patients, respectively, were transferred to a major psychiatric facility before screening could take place. While schizophrenia/ psychosis patients stayed in the general hospital setting for a mean of only 2 days, mood disorder patients resided in the general hospital setting for a mean of approximately 6 days. The transfer of these patients to a major psychiatric facility denotes that their symptoms did not stabilize within the general hospital setting and a longer period for recovery was envisaged. Clearly, any attempt to screen these groups of patients while in the general hospital setting would have been difficult and of little value.
In the current study the research officer responsible for alcohol screening visited the psychiatric wards two or three times a week. New admissions were noted and the primary nurse for each patient was asked to assess the patient's status until circumstances were appropriate for AUDIT administration. A drawback of this screening procedure was that for a significant number of patients, their severity of psychiatric symptoms resolved and led to discharge in the period between research officer screening visits. This applied to 79% of psychiatric admissions not screened. Clearly a weekly screening regimen similar to those reported in other hospital environments (e.g. [3]), would be even less effective than observed in this study and fail to capture a greater proportion of patients with a psychiatric disorder.
Even if screening had been achieved with all targeted patients in the current study, the brief hospital stay associated with many patients would leave the remaining window of opportunity in which an alcohol intervention could have been implemented very small. This again contrasts to other hospital settings where an alcohol intervention may be feasible for several days after screening.
In the psychiatric ward alcohol intervention needs to be available immediately following identification of hazardous or harmful consumption. Given the complexity of issues associated with the psychiatric patient, AUDIT screening and if necessary brief intervention may not always be possible if both screening and intervention are not available ad libitum. Success in screening and intervention is therefore more likely where both are incorporated as part of general care and management procedures or as part of a predischarge plan.
A number of recommendations can be made to increase the number of psychiatric patients screened for alcohol in the general hospital psychiatric setting. First, it should be emphasized that the AUDIT provides a simple procedure for screening hazardous and harmful alcohol consumption and dependence among persons with psychiatric morbidities. In those patients who were screened there was little difficulty encountered in using the AUDIT. Second, the optimal time to administer the AUDIT is following the subsidence of major psychiatric symptoms, adjustment to medication and improvement of cognitive functioning. Third, in contrast to the current study procedure where a research officer visited the ward two or three times a week and consequently missed screening a significant number of patient due to discharge between visits, the AUDIT should be incorporated into the routine assessment and medical management of psychiatric units. The primary nurse or attending doctor, upon determining suitable stabilization of the patient, could incorporate the AUDIT questionnaire within the context of a regular ward consultation. The brevity of the AUDIT makes this possible. Fourth, although the AUDIT is designed to be both self or staff administered, our experience was that staff administration was more suited for the psychiatric inpatient setting and helped overcome problems in patient understanding of AUDIT questions. In this respect, the primary nurse who, following admission, commonly has the opportunity to build a rapport with the patient, as well as being aware of the patient's condition and the impact of medication, may be the most suitable person to administer the AUDIT. This timely administration of the AUDIT would in turn maximize the amount of time remaining to implement an alcohol intervention to those found to be consuming alcohol at problem levels. The primary nurse may also be the most suitable person to initiate this brief intervention.
Footnotes
Acknowledgements
This work was funded through a research grant from the Western Australian Health Promotion Foundation.