| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Psychiatric Bulletin | Advances in Psychiatric Treatment | All RCPsych Journals |
|
|
|
||||||||||
|
||||||||||||
School of Psychiatry and Clinical Neurosciences, University of Western Australia, QEII Medical Centre, Nedlands, Australia
Correspondence: Dr Robert Tait, School of Psychiatry and Clinical Neurosciences, University of Western Australia, D Block, QEII Campus, Nedlands, WA 6009. E-mail: rjtait{at}cyllene.uwa.edu.au
Funding detailed in Acknowledgements.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
Aims To assess in adults with less severe psychiatric morbidity the relationship between alcohol consumption and subsequent 7-year hospital admissions, and the development and recurrence of alcohol use disorders.
Method Follow-up data were assembled via a population-based hospital record-linkage system.
Results Baseline alcohol use groups were: dependent (n=31), harmful (n=114), moderate (n=621) and abstinent (n=249). The moderate but not the abstinent group had fewer mental health admissions and a longer time to first admission than the harmful and dependent groups. Both the moderate and the abstinent groups had longer times to ‘all-cause’ admissions than the dependent group. Many of those with alcohol use disorders at baseline relapsed (66%) but few (14%) developed a first-time alcohol use disorder.
Conclusions Overall, moderate alcohol consumption among those with less severe psychiatric morbidity was not associated with more mental health admissions; those with alcohol dependence had poorer health outcomes than the remaining categories.
|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
|
|
METHOD |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
Design and data collection
The study used a prospective cohort design to analyse the relationship
between alcohol consumption level assessed as part of a general hospital
psychiatric admission in 1994–1996 and subsequent hospital admissions.
Admission data were assembled using the Western Australian Data Linkage
System. This can assemble administrative health data from a number of sources
and covers the entire population of Western Australia from 1980 onwards
(Holman et al, 1999).
It has been used in a wide range of studies, from identifying post-operative
complications (Valinsky et al,
1999) to assessing death rates from ischaemic heart disease among
psychiatric patients (Lawrence et
al, 2003).
In 2003 we submitted identifying details on the cohort for record linkage. The requested data covered deaths, specialist mental health unit in-patient and general hospital admissions, including psychiatric admissions. The Western Australian Linkage System contains diagnostic data coded using ICD–9–CM from 1988 to mid-1999 (World Health Organization, 1986) and ICD–10 from mid-1999 onwards (World Health Organization, 1992).
Participants
The participants were 1017 people who had agreed to take part in an earlier
study (Hulse et al,
2000; Hulse & Tait,
2002). They were aged 18–65 years (median 39.7 years,
interquartile range (IQR) 31.3–50.0 years) when they were admitted to a
psychiatric ward at one of three participating general hospitals between
September 1994 and October 1996. Those with poorer courses of disease (more
severe conditions or who failed to stabilise) were transferred to a specialist
psychiatric hospital and were ineligible for the baseline study and this
follow-up study. In Western Australia the specialist hospital is a dedicated
facility which includes locked and open wards for long-stay admissions and
contains the state forensic psychiatric unit.
In the 5 years before the index admission, 479 participants (47.2%) had had no mental health admissions; the median total days in hospital with a mental health disorder was 1 day (IQR 0–20 days). The median length of stay for the index admission was 14 days (IQR 8–26 days). Participants gave written informed consent, and the study received university and hospital ethics approval. For the current study, access to the Western Australian Linkage System required additional institutional human research and ethics approvals.
Measures
For the current study, alcohol consumption categories of harmful use or
dependence were determined by inspecting baseline hospital discharge diagnoses
(ICD–9–CM). Those with a diagnosis of dependence or harmful use
were assigned to these categories. As part of the baseline study, the level of
alcohol consumption by patients was assessed using the Alcohol Use Disorders
Identification Test (AUDIT; Saunders et al,
1993a,b)
to determine their eligibility for a brief alcohol intervention (Hulse &
Tait, 2002,
2003). Those without an
alcohol use disorder at baseline were classified as abstinent (non-users) if
they scored zero on the AUDIT. All other sub-clinical levels of alcohol use
were defined as moderate.
We defined mental health admissions as any in-patient admission where the discharge diagnosis included a code from the ICD chapter V (mental and behavioural disorders) (World Health Organization, 1986, 1992). The term ‘alcohol-related admissions’ included all conditions with an aetiological fraction of one (English et al, 1995). This includes conditions such as alcoholic gastritis and alcoholic liver cirrhosis.
Patients on general hospital psychiatric wards typically have a shorter hospital stay and a less extensive history of mental health disorders than those treated at specialist psychiatric hospitals where long-term in-patient treatment and extensive history of disease are common. For heuristic purposes, these groups are respectively described as having ‘less severe’ or ‘severe’ mental health disorders. The term ‘less severe’ also serves to differentiate this cohort from the cohort described by Drake and co-workers (Drake et al, 1989) as having ‘severe’ disorders.
Analysis
At all levels of alcohol consumption a proportion of people are likely to
have had previous alcohol use disorders. This may be of particular importance
for those in the abstinent group, as treatment for alcohol dependence
typically recommends zero alcohol use. To reduce the potential for bias
associated with this group, we used the Western Australian Linkage System to
identify previous diagnoses of alcohol psychosis, dependence or withdrawal
(e.g. ICD–9–CM codes 291.0–291.9, 303.00–303.99). A
sensitivity analysis was conducted with and without participants classified as
abstinent at baseline.
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
|
As shown in Table 1, there was a significant association between alcohol use group and gender. There was also a significant association between alcohol use group and age. An inspection of the means suggested that there was not a group by gender interaction. However, violation of the assumptions for analysis of variance (ANOVA) prevented statistical assessment.
Baseline primary diagnoses were grouped via discharge diagnostic codes and tabulated according to alcohol use groups (Table 2). The ‘other mental health’ disorders included 41 poisonings, of which 40 were classified as self-harm or suicide attempts. Notably the dependent group had 12 ‘other mental health’ diagnoses, of which 5 were self-harm poisonings.
|
Mental health admissions
By 12 months there was a significant difference between groups in the
proportion that had had at least 1 mental health admission
(Table 1), with the moderate
group having the lowest proportion of admissions (
2=10.2 (3),
P<0.025). Over 7 years there was a significant difference in total
mental health admissions (abstinent group, median 3, IQR 1–9; dependent
group, median 8, IQR 4–19; Kruskal–Wallis 15.5 (3),
P<0.001). The abstinent and moderate groups had a similar level of
admissions (Mann–Whitney U=75 948, P=0.68), but the
moderate group had fewer admissions than the harmful group (U=30 880,
P<0.05), whereas the abstinent group did not differ significantly
from the harmful group (U=12 660, P=0.097).
The survival time to first mental health admission was assessed with Kaplan–Meier survival curves (Fig. 1) and log-rank tests. There was a significant difference in survival times across the alcohol use categories (Table 1). By 7 years, only 6% of the dependent group had not had a mental health admission compared with 24% of the moderate group (mean survival times 361 and 899 days respectively). Between groups, the dependent group had significantly shorter survival times than all the other categories. Notably, the only other significant difference between groups in survival times was between the moderate and harmful groups (log-rank test 3.9 (1), P<0.05). There was no reliable difference between the abstinent and harmful groups (log-rank test 1.6 (1) P=0.21). Six cases were censored owing to death of the participant. After adjusting for gender, the same pattern of survival results was obtained.
|
Survival times to first alcohol-related admission decreased for each increasing level of consumption; the respective mean survival times for the abstinent and dependent categories were 2411 and 511 days (Table 1). A between-groups assessment of survival times found significant differences between all the groups, including the abstinent and moderate categories (log-rank test 18.2 (1), P<0.001). There were 59 cases censored owing to death of the participant, of which 22 were in the abstinent and 34 in the moderate group. After adjusting for gender, the same pattern of survival results was obtained.
All-cause hospital admissions
At 12 months there was no reliable difference between the proportion of
people in each group who had had a hospital admission
(Table 1), with an
unstandardised rate of 62.1 admissions per 100 person-years. At 7 years there
was a significant difference between groups in the total number of hospital
admissions (Kruskal–Wallis 10.9 (3), P<0.025), with the
median number of admissions ranging from 6 (IQR 2–14) for the moderate
users to 10 (IQR 6–22) for the dependent group.
There was a significant effect of alcohol use group on survival times to the first hospital admission of any type, with survival time decreasing from moderate through to dependent use (overall log-rank test 12.0 (3), P<0.01) (Table 1). The moderate group had a significantly longer survival time than both the dependent (log-rank test 7.5 (1), P<0.01) and harmful groups (log-rank test 5.1 (1), P<0.05). The abstinent group also had a longer survival time than the dependent group (log-rank test 6.3 (1), P<0.05). By 7 years most people had had at least one admission, the proportion ranging from 97% in the dependent to 81% in the moderate group. Three cases were censored owing to death of the participants. After adjusting for gender, this pattern of survival times was replicated except in that the abstinent group also had a longer survival time than the harmful group (log-rank test 4.4 (1), P<0.05).
Sensitivity analysis: pre-index alcohol dependence
We identified 54 participants (5.3%) who had a previous diagnosis of
alcohol dependence at the time of the baseline screening. The number in each
alcohol group is shown in Table
1 (penultimate row). After excluding the 12 people in the
abstinent group, the above three survival analyses were rerun. In each case
the pattern of findings was replicated except in relation to total hospital
admissions, where the abstinent group had a significantly longer survival time
than the harmful group before adjusting for gender.
Development and recurrence of alcohol use disorders
Of the 145 persons with an ‘active’ alcohol use disorder at
index, 96 (66%) had a subsequent admission with an alcohol use disorder. Among
the 64 people with pre-index alcohol use disorders, 32 (50%) were in remission
at the index admission, of whom 20 (62.5%) were in the moderate group.
Subsequently, 12 (37.5%) redeveloped an alcohol use disorder, of whom 10
(83.3%) were from the moderate group. Of the 838 without a disorder before or
at index, 117 (14%) were later diagnosed with an alcohol use disorder, 102
(87%) of whom were in the moderate group. Overall, there was a significant
difference in the proportion of people with ‘active’,
‘non-active’ and ‘no previous alcohol use disorders’
who developed an alcohol use disorder by 7 years (2=200.1 (2),
P<0.01).
The final row of Table 1 shows that there was a significant difference in the proportion of each alcohol consumption group that had at least one admission with an alcohol use disorder (dependent or harmful use). Surprisingly, although most of those with dependence had a readmission (28, 90.3%), few had readmission with an alcohol use disorder in the first 12 months (5, 16.1%).
Mortality
There were 93 deaths (48 women (52%)), and those who died were
significantly older than those who did not die (median 43 (IQR 35–57)
v. 39 (IQR 31–50) years, Mann–Whitney U=36 139,
P=0.012). The dependent group had the greatest proportion of deaths
(6, 19.4%), but there was no significant association between alcohol use group
and death (2=6.7 (3), P=0.08). Suicide, as assessed
from ICD codes, was recorded for 31 deaths (33%). There was no significant
association between suicide and gender (2=0.19 (1),
P<0.66) and the association between alcohol use group and suicide
could not be reliably assessed because of the number of cells with few
cases.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
Development of alcohol use disorders and relapse
The ability to sustain moderate drinking is of importance in this cohort,
as it has previously been found that those with severe mental health disorders
are unlikely to sustain moderate levels of consumption, with about a quarter
developing an alcohol use disorder (Drake
& Wallach, 1993). In the current study, less than 2% of
abstinent and about 12% of moderate users at baseline developed a first-time
alcohol use disorder in the subsequent 7 years. By way of comparison, the
estimated lifetime prevalence of alcohol use disorders is 13.5% in the US
population and 34.1% in US mental hospitals
(Regier et al, 1990).
Therefore, those in the cohort without a previous alcohol use disorder have a
prevalence of these disorders over 7 years similar to that of the general US
population lifetime estimate.
For those with an alcohol use disorder in remission at baseline, 37.5% had a recurrence within the follow-up period, with over 80% of these being from the moderate use group. In those categorised with alcohol dependence, only 16% had a recurrent alcohol use disorder within 12 months, but by 7 years this rose to 90.3%. In the harmful use group the respective figure was about 60%. In contrast, previous research has shown that about 50% of those with a severe mental health disorder and a past substance use disorder are likely to have a recurrence of their substance use disorder within 12 months (Dixon et al, 1998). Accordingly, moderate alcohol use should be discouraged in those with but not those without a history of alcohol use disorders.
Moderate use v. abstinence
A special focus for this study was the possible difference in morbidity
between those with moderate levels of alcohol use and non-users. We found no
evidence to suggest that those using moderate levels of alcohol were more
likely to have mental health or general health admissions or increased
mortality than were those abstaining from alcohol. Furthermore, survival to
either first mental health or first all-cause hospital admission was similar
for the two groups. It was only with respect to alcohol-related morbidity that
those in the abstinent group had better outcomes (e.g. longer survival times)
than the moderate use group. This result needs to be tempered with the
knowledge that only 10% of moderate consumers had an alcohol-related admission
in the first year.
There was tentative evidence that moderate alcohol use compared with abstinence was associated with a reduction in mental health admissions. Specifically, moderate users had fewer mental health admissions and a longer survival time before admission than the harmful group. The abstinent group was not reliably different from the harmful group on these measures, although the difference approached significance with respect to the number of admissions. For cardiovascular disease there is now extensive evidence that moderate amounts of alcohol confer benefits, compared with either non-use or heavy use of alcohol (Doll, 1998). Given the known exacerbation of symptoms with even moderate alcohol consumption among those with severe mental health disorders, a finding of beneficial out-comes as suggested here for those with less severe mental health disorders needs to be interpreted with great care. These data are, however, consistent with other data from the general population, where moderate drinkers have a lower risk of symptoms of depression and anxiety than either non-users or heavy drinkers (Rodgers et al, 2000). They are also compatible with data on dementia in older adults (65+ years), where abstainers and heavy drinkers have a greater risk than moderate (1–6 drinks/day, defined as 12 oz of beer, 6 oz of wine or 1 shot of spirits) drinkers of developing dementia (Mukamal et al, 2003). Additional longitudinal data show that as few as 1–3 drinks/day (amounts not defined) reduce the risk of dementia in those aged 55 years and over (Ruitenberg et al, 2002).
There are a number of possible mechanisms by which moderate alcohol consumption may affect mental health: for example, via improved general health (cardiovascular health (Doll, 1998)) or general psychological well-being (Baum-Baicker, 1985). However, these authors caution that moderate alcohol consumption may simply act as a marker for an as yet unidentified causal variable such as social stability. If this were correct, then it would be in-appropriate to view moderate alcohol use as having a protective effect for psychiatric morbidity or to recommend moderate consumption by those with psychiatric morbidity. On the other hand, we suggest that moderate alcohol use should not be actively discouraged as it may assist in maintaining the social fabric. The need for a conservative approach is emphasised by the findings of a recent longitudinal study. This reported unadjusted benefits in relation to depression for moderate alcohol users compared with long-term abstainers, ex-drinkers, heavier-moderate and heavy drinkers, but found that adjustment for a range of socio-demographic and health variables left only the heavy drinking groups at a disadvantage (Paschall et al, 2005).
The inconsistency between our findings and those of a previous study, in which even moderate alcohol use exacerbated symptoms for those with severe mental health disorders (Drake et al, 1989), probably reflects differences in the study groups. In Drake et al’s study all of the participants had a primary diagnosis of schizophrenia and, although they were recruited as out-patients, they had an extended history of mental health hospital admission (lifetime mean 4.4 years, s.d.=4.9). In contrast, nearly half of the participants in our study had no mental health admission in the 5 years before baseline and about 10% did not have a primary psychiatric diagnosis, although all were admitted to psychiatric wards. We therefore contend that the current cohort depicted here as having less severe illness is markedly different from cohorts previously described as having severe disorders.
Suicides
It has previously been found that people who have been psychiatric
in-patients have a suicide rate more than 10 times that of the general Western
Australian population (Lawrence et
al, 1999). Data from the USA have also shown that suicidality
is a feature that distinguishes people with alcohol problems with and without
depression (Cornelius et al,
1995). Therefore the prevalence of suicide in the cohort, although
high, is not unexpected. Nevertheless, it does suggest that further research
is required to develop appropriate discharge and community support strategies
for this population. Surprisingly, in the current study the proportion of
suicides among men and women were similar, whereas previous research has shown
that men have an elevated risk (3.4 times that for women)
(Lawrence et al,
1999).
|
|
Clinical Implications and Limitations |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
LIMITATIONS
|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONClinical Implications and...ACKNOWLEDGMENTSREFERENCES |
|---|
Cornelius, J. R., Salloum, I. M., Mezzich, J., et al
(1995) Disproportionate suicidality in patients with comorbid
major depression and alcoholism. American Journal of
Psychiatry, 152, 358
–364.
Dixon, L., McNary, S. & Lehman, A. F.
(1998) Remission of substance use disorders among psychiatric
inpatients with mental health illness. American Journal of
Psychiatry, 155, 239
–243.
Doll, R. (1998) The benefit of alcohol in moderation. Drug and Alcohol Review, 17, 353 –363.[CrossRef][Medline]
Drake, R. E. & Wallach, M. A. (1993)
Moderate drinking among people with severe mental illness. Hospital
and Community Psychiatry, 44, 780
–782.
Drake, R. E., Osher, F. C. & Wallach, M. A. (1989) Alcohol use and abuse in schizophrenia. A prospective community study. Journal of Nervous and Mental Disease, 177, 408 –414.[CrossRef][Medline]
English, D. R., Holman, C. D. J., Milne, E., et al (1995) The Quantification of Drug Caused Morbidity and Mortality in Australia. Canberra: Commonwealth Department of Human Services and Health.
Glover, J., Harris, K. M. & Tennant, S. (1999) A Social Atlas of Australia. Adelaide: Public Health Information Development Unit.
Holman, C. D. J., Bass, A. J. & Rouse, I. L. (1999) Population-based linkage of health records in Western Australia: development of a health services research linked database. Australian and New Zealand Journal of Public Health, 23, 453 –459.[Medline]
Hulse, G. K. & Tait, R. J. (2002) Six-month outcomes associated with a brief alcohol intervention for adult inpatients with psychiatric disorders. Drug and Alcohol Review, 21, 105 –112.[Medline]
Hulse, G. K. & Tait, R. J. (2003) Five-year outcomes of a brief alcohol intervention for adult inpatients with psychiatric disorders. Addiction, 98, 1061 –1068.[CrossRef][Medline]
Hulse, G. K., Saunders, J. B., Roydhouse, R. M., et al (2000) Screening for hazardous alcohol use and dependence in psychiatric in-patients using the AUDIT questionnaire. Drug and Alcohol Review, 19, 291 –298.
Lawrence, D. M., Holman, C. D. J., Jablenski, A. V., et
al (2003) Death rate from ischaemic heart disease in
Western Australian psychiatric patients 1980–1998. British
Journal of Psychiatry, 182, 31
–36.
Lawrence, D. M., Holman, C. D., Jablensky, A. V., et al (1999) Suicide rates in psychiatric in-patients: an application of record linkage to mental health research. Australian and New Zealand Journal of Public Health, 23, 468 –470.[Medline]
Mukamal, K. J., Kuller, L. H., Fitzpatrick, A. L., et
al (2003) Prospective study of alcohol consumption and
risk of dementia in older adults. JAMA,
289, 1405
–1413.
Paschall, M. J., Freisthler, B. & Lipton, R. I.
(2005) Moderate alcohol use and depression in young adults:
findings from a national longitudinal study. American Journal of
Public Health, 95, 453
–457.
Regier, D. A., Farmer, M. E., Rae, D. S., et al (1990) Comorbidity of mental disorders with alcohol and other drug abuse. Results from the Epidemiologic Catchment Area (ECA) Study. JAMA, 264, 2511 –2518.[Abstract]
Rodgers, B., Korten, A. E., Jorm, A. F., et al (2000) Non-linear relationships in associations of depression and anxiety with alcohol use. Psychological Medicine, 30, 421 –432.[CrossRef][Medline]
Ruitenberg, A., van Swieten, J. C., Witteman, J. C. M., et al (2002) Alcohol consumption and risk of dementia: the Rotterdam Study. Lancet, 359, 281 –286.[CrossRef][Medline]
Saunders, J. B., Aasland, O. G., Amundsen, A., et al (1993a) Alcohol consumption and related problems among primary health care patients: WHO collaborative project on early detection of persons with harmful alcohol consumption – I. Addiction, 88, 349 –362.[CrossRef][Medline]
Saunders, J. B., Aasland, O. G., Babor, T. F., et al (1993b) Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption – II. Addiction, 88, 791 –804.[CrossRef][Medline]
Swofford, C. D., Kasckow, J. W., Scheller-Gilkey, G., et al (1996) Substance use: a powerful predictor of relapse in schizophrenia. Schizophrenia Research, 20, 145 –151.[CrossRef][Medline]
Valinsky, L. J., Hockey, R. L., Hobbs, M. S., et al (1999) Finding bile duct injuries using record linkage: a validated study of complications following cholecystectomy. Journal of Clinical Epidemiology, 52, 893 –901.[CrossRef][Medline]
World Health Organization (1986) International Classification of Diseases and Health Related Problems: 9th Revision: Clinical Modification. Geneva: WHO.
World Health Organization (1992) International Statistical Classification of Diseases and Health Related Problems: 10th Revision. Geneva: WHO.
Received for publication November 17, 2004. Revision received October 19, 2005. Accepted for publication November 1, 2005.
Related articles in BJP:
This article has been cited by other articles:
|
|
 |
|
  H. O'Connell Moderate alcohol use and mental health The British Journal of Psychiatry, December 1, 2006; 189(6): 566 - 567. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Psychiatric Bulletin | Advances in Psychiatric Treatment | All RCPsych Journals |
, harmful-censored;
–, dependent (n=31);