Identification of alcohol abuse in adults and elderly in a Dutch hospital

Introduction and background
Currently excessive alcohol use is becoming a big problem for the healthcare.
For greater part this is attributed to the drinking behaviour of adults and
elderly. Not only the number of users is increasing, but also the amounts per
person are getting bigger. Particularly last given is a problem in the cure-
and care-sector in the Netherlands[1,2]. Previous studies show that alcohol
abuse can contribute to more admissions and (re)hospitalization[3]. Especially
patients with unexplained somatic symptoms, preoperative patients or elderly,
whether or not using multiple medication, belong to this category[4]. Most
known morbidities of alcohol consumption, whether or not postoperative, are:
infections, bleeding disorders, need for ventilator support, cognitive
dysfunction [5,6,7 ]. Therefore, it is very troublesome to learn from prior
studies that alcohol abuse commonly remain undetected thereby also untreated
[8-11]. The degree of under diagnosis is attributed to diverse factors, such as
depression, dementia, physical changes associated with age, life events, late
onset of alcoholism and lack of screening. In addition, elderly patients may
find it difficult to disclose their alcohol intake and relatives may deny the
existence of the problem[4,12]. Given above information, it seems reasonable to
investigate if extensive screening on internal medicine admissions and of
preoperative patients can lead to improvement of detection rates and treatment
of excessive alcohol use.
It has been shown that the Alcohol Use Disorders Identification Test (AUDIT) is
a valid and simple way for detection of high risk patients. The AUDIT-C is a
shortened version of the AUDIT. It consists of the first 3 questions of the
AUDIT questionnaire and is equal in accuracy to the full AUDIT. This
questionnaire is applicable to: men, women, elderly, drug users, internal
medicine inpatients, other cultures and the unemployed[13-16]. A more objective
test is laboratory screening on carbohydrate-deficient transferrin (CDT), gGT
and MCV. This test is being used for detection of excessive alcohol use in de
past week or as follow up during treatment and is mostly applied in the
secondary care[16]. Clinical laboratory procedures provide objective evidence
of problem drinking.

Research question: Is routine laboratory screening on alcohol abuse (MCV, g-GT
and CD-tect) indicated to improve detection rates of alcohol abuse?

Methods
Study design
The study will be conducted at the internal medicine ward and the preoperative
assessment clinic of a general dutch hospital, performed and supervised by the
compartment of psychiatry. For this diagnostic and cross-sectional study data
will be collected during 9 weeks. We want to investigate if the combination of
a lifestyle questionnaire, including the AUDIT-C, and laboratory screening is
indicated to increase detection rates of alcohol abuse in the secondary care.
The intention is to cover up the alcohol related questionnaire with general
lifestyle questions, to not lay focus on their alcohol drinking behaviour. To
accomplish this adult patients, above the age of 18 years, admitted to internal
medicine or undergoing preoperative assessment screening will be consecutively
included. Patients, admitted to internal medicine, who didn*t stay at least 24
hours and patients with poor understanding of the dutch language will be
excluded.

Measurements
After acquiring oral consent, patients will be interviewed with the aid of a
lifestyle questionnaire.
This questionnaire contains the following 10 questions: How much social
contacts do you have? Are you environmentally responsible? How often do you
sport? Do you watch tv? If yes, how often and what kind of programs? Do you
have a cultural activity? If yes, what kind of? Do you read the paper or books?
If yes, witch one? Do you go out? If yes, what kind of activities? Do you
smoke?if yes, what and how much? Do you drink alcohol? If yes, AUDIT-C
screening follows. Do you use any kind of drugs?. Each question of the AUDIT-C
contains 5 possible answers, scoring 0-4 points per question. The maximum score
is 12. A score of >= 4 or more identifies 86% of men who report drinking above
recommended levels. A score of >= 4 identifies 84% of women who report hazardous
drinking[14] .
Official informed consent will be provided before questioning. Afterwards a
blood sample will be taken from the internal medicine patients, independent of
the results of the questionnaire, for laboratory screening on levels of gGT,
CDT and MCV. The accuracy of these markers is affected by various factors such
as non-alcoholic liver damage, use of medications or drugs, and by metabolic
disorders. gGT is the most commonly used biochemical measure of drinking.
Chronic alcohol use of 4 or more beverages a day for 4 to 8 weeks significantly
raises levels of this blood protein. CDT values become elevated substantially
earlier (1-2 weeks) in response to prolonged excessive drinking. MCV increases
with excessive alcohol intake after 4 to 8 weeks. It can detect evidence of
earlier drinking after a long period of abstinence. The sensitivity of this
marker is too low to justify its use as a single indicator. However, it has
higher specificity compared with other tests[17,18].
Evaluation of clinical records
Furthermore, the medical records referring to current admission and medical
history of alcohol abuse will be reviewed. Collected data must include: gender,
age, reason for admission or kind of surgery, earlier admissions for the same
reason, earlier pre- or perioperative complications, presence of
alcohol-related somatic disease, if and how alcohol drinking behaviour is
mentioned (units per week or otherwise graded) and presence of psychiatric
disorder(s).

Analysis
Analysis

Data of the 2 wards, internal medicine and preoperative assessment clinic, will
be analysed separately. If patients are not reported as high risk drinkers, in
their medical record, or are not given advise to withhold or moderate their
drinking, I will report these findings as acceptable drinking behaviour for the
physician. All variables and collected data will be processed using SPSS
software. A quantitative analysis will be done using descriptive statistics,
with the intention to present frequencies of alcohol use, abuse and *undetected
cases* and to describe the characteristics of the population.
A correlation analysis will be used to describe the relationship between the
outcome values of the AUDIT-C questionnaire and the anamnesis of the physician
concerning alcohol abuse. Because the data we use has an ordinal level, the
*Spearman rank order correlation* is best applicable here. For further
diagnostic analysis the sensitivity and specificity of the laboratory markers
gGT, CDT and MCV will be calculated for detecting high-risk drinking. Receiver
operating characteristic (ROC) analysis will be used to assess the performance
of the markers across the full range of potential cut-off values. ROC curves
plot a scale*s ability to detect true positives against the rate of false
positives. The area under the ROC curve is used as a measure of overall test
accuracy. For each marker (CDT, GGT, and AST), separate ROC analyses will be
conducted to compare those drinking at high-risk levels against those at
low-risk levels, measured by AUDIT-C.

nvt