Research A rticle
Predicting Inpatient Detoxification Outcome of
Alcohol and Drug Dependent Patients: The Influence of
Sociodemographic Environment, Motivation, Impulsivity,
and Medical Comorbidities
Yvonne Sofin,
1
Heidi Danker-Hopfe,
2
Tina Gooren,
3
and Peter Neu
1,3
1
Jewish Hospital Berlin, Clinic for Psychiatry and Psychotherapy, Heinz-Galinski-Str. 1, 13347 Berlin, Germany
2
Competence Center for Sleep Medicine, Charit
´
e-School of Medicine, Campus Benjamin Franklin, Hindenburgdamm 30,
12203 Berlin, Germany
3
Charit
´
e-School of Medicine, Campus Benjamin Franklin, Hindenburgdamm 3 0, 12203 Berlin, Germany
Correspondence should be addressed to Peter Neu; peter.neu@charite.de
Received 22 December 2016; Accepted 13 February 2017; Published 6 March 2017
A
cademic Editor: Markus Backmund
Copyright © Yvonne Son et al. is is an open access article distributed under the Creativ e Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Aims. is prospective study aims to identify patient characteristics as predictors for treatment outcome during inpatient
detoxication treatment for drug and alcohol dependent patients. Methods. A mixed gender sample of consecutively admitted
drug and alcohol dependent patients were interviewed by an experienced physician. e impact of a variety of factors concerning
social environment, therapy motivation, impulsivity related variables, medical history, and addiction severity on treatment
outcome was examined. Results. (.%) of the patients completed detoxication treatment whereas (.%) dropped out
prematurely. Being female, living in a p ar tnership, having children, being employed, and having good education were predictive for
a positive outcome. Family, health, the fear of losing the job, prosecution, and emergency admission were signicant motivational
predictors for treatment outcome. Being younger, history of imprisonment, and the number of previous drop-outs were predictive
for a negative outcome. Conclusions. Variables concerning social environment and the number of previous drop-outs have been
identied as best predictors for treatment outcome. Socially stable patients benet from the current treatment setting and treatment
shall be adapted for patients with negative predictors. Treatment may consequently be tailored with respect to intervention type,
duration, and intensity to improve the outcome for those patients that full criteria with negative impact on treatment retention.
1. Introduction
Addictionisachronicdiseasethataectsmillionsofindi-
viduals worldwide. In Germany, alcohol dependence is the
most serious, expensive, and socially disruptive health issue
[]. In industrialized countries, alcoholism is among the
leading causes of death []. Qualied detoxication treatment
(QDT) is the rst step in inpatient substance abuse treatment.
Premature discontinuation of QDT is a serious and common
complication in the detoxication treatment []. e risk
of relapse substantially determines prognosis and mortality
of the disease []. Treatment noncompletion is generally
associated with poor success and an unfavourable long-term
outcome []. Regular completion of treatment is therefore
a key success criterion of inpatient detoxication treatment
as the discontinuation of therapy is usually accompanied by a
relapse []. For both, alcohol dependent and drug dependent
patients, high drop-out rates were reported in literature.
us,adrop-outrateof%hasbeenreportedforalcohol
dependents [], while the observed drop-out rate for illicit
drug users was even higher at % []. erefore it is of
high importance to identify determinants and risk factors of
unplanned premature discharge and to adapt the treatment
for the individual patient accordingly. Some predictors were
repeatedly identied in previous studies: for drug addiction,
levelofeducation,unemployment,anddelinquencywere
predictive for relapse [–]. For alcohol addiction, signicant
predictors were the number of prior hospitalizations for
Hindawi
Journal of Addiction
Volume 2017, Article ID 6415831, 11 pages
https://doi.org/10.1155/2017/6415831
Journal of Addiction
detoxication, dependence severity, and psychopathologic
rating [, , ]. e link between illicit drug use and crime
is well documented []. Backmund et al. [] found history
of imprisonment and currently being on probation to be
signicant predictors of completing detoxication treatment.
Previous studies indicated that lower injection f requency
before admission was associated with twofold increases in the
likelihood of having favourable follow-up outcomes on illicit
drug use, alcohol use, and criminal involvement [].
Additionally, impulsivity plays a major role in substance
use disorders []. Imp ulsiveness involves behaviour char-
acterized by little or no forethought or consideration of
theconsequences[].Impulsiveactionsarethereforeoen
poorly conceived or inappropriate to the situation and result
in undesirable consequences, for example, choosing short-
term gains over long-term gains []. Suicide attempts are
oen regarded as impulsive acts []. According to W ines
Jr . et al. [], previous suicide attempts are common in
substance-dependent individuals. Nearly half of the drug
dependent patients (%) reported having attempted suicide
at some point of their life []. Pretreatment suicide attempts
are associated with a higher likelihood of relapse [].
e purpose of the present study was to identify fur-
ther predictors of premature discharge during inpatient
QDT for drug and alcohol dependent patients to thereby
achieve better outcomes in terms of treatment completion
for drug or alcohol dependent patients. With regard to
the high costs in public healthcare systems, prediction of
treatment outcome provides the opportunity to identify client
groups that achieve poorer outcomes and identify targets
in treatment to improve inpatient detoxication treatment.
Clinicians should be enabled to set realistic treatment goals
and adapt intervention duration and intensity. We therefore
conductedaprospectiveanalysistoinvestigatetheinuence
of sociodemographic as well as medical variables on QDT
outcome on patients suering from addiction.
2. Methods
2.1. Participants. During the year , consecutively
drug and alcohol dependent patients admitted to the hospital
were screened and asked for participation. patients were
included in the study. All patients fullled the DSM-IV cri-
teria for substance addiction and gave their written informed
consent to participate in this study. Exclusion criterion was
noncapacity of giving informed consent (severe organic or
psychiatric disorders like Korsakow syndrome, etc.).
2.2. Setting and Treatment Procedure. e study was con-
ducted on two specialized inpatient units for qualied detox-
ication treatment of addiction diseases in a psychiatric hos-
pital in B erlin, Germany. e treating team comprised med-
ical doctors, psychologists, specialized nurses, occupational
therapists,physiotherapists,andsocialworkers.equalied
detoxication treatment enriches detoxication treatment
with psychoeducation and relapse prevention. It consists of
three steps. While detoxication t he patients were withdrawn
from the dr ug and, w here needed, withdrawal symptoms
were treated. In the second step, the patients had to attend at
least ten group-therapy sessions and ve psychoeducational
group-sessions. In the third step, the preparation of transition
to a long-term follow-up treatment aer hospital discharge
including the attendance of ve self-help groups outside the
clinic was conducted.
e average treatment took between and days but
could last longer in case of persisting withdrawal symptoms
or particularly severe general condition.
Clomethiazole at tapered doses was used for alcohol
detoxication. Methadone at tapered doses was used for opi-
oid detoxication. For cannabis, amphetamines and cocaine
detoxication abrupt cessation without medical support was
chosen. e severity of alcohol withdrawal symptoms was
captured according to the CIWA Withdrawal Score [].
2.3. Diagnostic Criteria. For diagnosis of addiction and con-
comitant diseases Diagnostic and Statistical Manual (DSM)
edition IV was applied.
2.4. Denition of Outcome Criteria. e treatment was
considered successfully completed if t he patient remained
abstinent while hospital stay and participated in the treat-
ment program as described above until regular discharge.
e attendance to at least ten group-therapy sessions, ve
psychoeducational group-sessions, and ve self-help groups
outside the clinic was mandatory.
e treatment was considered aborted if the patient le
against medical advice or due to disciplinary early discharge.
Substance use or refusal to participate in the treatment
program led to disciplinary discharge.
2.5. Data Analysis. Data on the patient’s social environment
consisting of information on their living and domestic situa-
tion, children, graduation, employment, and native language
were collected. Additionally, the patient’s therapy motivation
was asked upon hospital admission (Table ). Answers are
comprisedoffearoflosingthepartnerorfamily,harminghis
or her health, fear of losing the job and/or residence, making
a therapy instead of imprisonment, the aim of abstinence,
and other motivations. Some patients did not specify their
motivation.
e patient’s impulsiveness was measured by data on
experience of violence, ag gressive behaviour, suicidal ten-
dency, and information on constraints in terms of judicial
proceedings, probation, and imprisonment. Further, the
impact of intravenous drug use and the eect of genetic pre-
disposition on impulsiveness related behaviour expressed by
addiction and suicidal tendency in relatives were elaborated.
All patients were admitted electively for qualied detoxi-
cation treatment except for emergency admissions.
Data on medical history comprised the addiction diag-
noses and, if applicable, addiction associated disorders, for
example, central nervous system damage.
All data were captured by an experienced physician
during structured face to face admission interview. Statistical
analyseswerecarriedoutusingSAS(statisticalanalysis
system) soware by SAS Institute. It was separately examined
Journal of Addiction
for all variables, whether there was a dierence between the
patients with and without premature treatment completion.
For nominal and ordinal scaled variables, the examination
was carried out with log likelihood Chi square test. For inter-
val scaled data, the relationship between t he respective vari-
able and premature discharge was analyzed with a 𝑡-test for
independent samples (if normal distribution was assumed) or
with a Wilcoxon -sample test (if normal distribution could
not be assumed). Normal distribution was tested using the
Kolmogorov-Smirnov test with a two-sided signicance level
of 𝑝 < 0.01. Tests on group dierences were examined with
a two-sided signicance level of 𝑝 < 0.05. 𝑝 values of .
or less were considered statistically signicant. Furthermore
logistic regression analyses were performed separately for the
clusters of variables: () sociodemographic determinants,
() motivational and addiction associated determinants, ()
impulsiveness related characteristics of patients, and ()
determinants from the patient’s medical history to identify
signicant predictors of the treatment outcome. In the logistic
regression the probability of a premature treatment comple-
tion was modelled.
3. Results
3.1. Patient Characteristics. patientswereincludedinthe
study. (%) patients were male and (%) were
female. e mean age was (±) years. Sociodemographic
details of the sample are given in Table . e patient
sample was characterized by a high number of patients
without partnership (.%) and .% participants that were
dependent on welfare. .% were unskilled and more than
one in ten patients was homeless. Asked about their therapy
motivation, health and family were frequent answers with
.and.%,respectively(seeTable).Ofthepatients
included, the most frequent diagnosis was alcohol addiction
with .% followed by .% opioid abuse. While .%
sueredofonlyoneaddiction,.%hadmorethanone
diagnosed addiction (Table (a)).
3.2. Determinants for Premature T reatment Drop-Out. Over-
all, (.%) of the patients completed detoxication
treatment whereas (.%) dropped out of the program.
e individuals of the drop-o ut group comprised
patients (.%) that prematurely terminated the treatment
on their own initiative while (.%) were discharged due
to disciplinary reasons. Patients in the treatment drop-out
group were signicantly younger ( years) than the patients
who completed the treatment ( years) and men dropped
out more oen than women (.% versus .%).
In this study, all tested sociodemographic pretreatment
variables showed a signicant inuence on the treatment
outcome. Patients that were female, lived in a partnership,
or were at least together with other individuals, had children
and were employed, were well-educated, and spoke German
as native language were more likely to nish the treatment
successfully. Having children had a positive impact on the
treatment outcome. At least .% of the patients with
children completed the treatment regularly, whereas only
.% of the childless patients completed the treatment. In
our study, the increasing number of chil dren did not correlate
with an increasing probability of treatment completion. e
higher the p atient’s graduation and occupational training
was, the higher the probability to complete the treatment
regularly was (Table (a)). Logistic regression revealed that
being younger and being unemployed signicantly increased
the risk of a unplanned, premature discharge. But it did not
conrm the inuence of the gender on treatment outcome
(Table (b)).
Family, health, the fear of losing the job, prosecution, and
emergency admission were signicant motivational predic-
tors for QDT outcome. For the patients that did not specify
a certain motivation or named abstinence as treatment moti-
vation, no signicant inuence was shown. Individuals with
no prior detoxication signicantly more oen completed
the treatment regularly. Also, we found that patients with
no previous treatment drop-outs signicantly more oen
completed QDT (𝑝 = 0.0001). e duration of the longest
period of the patients’ abstinence (𝑝 = 0.0874)wasnot
predictive for an early treatment drop-out (Table (a)).
Logistic regression with all motivation and treatment
variables revealed that treatment motivation is not signicant
fortreatmentoutcomebutthatthenumberofpreviousdrop-
outs was the best predictors for outcome. Subjects with one
or two previous early discharges had a .-fold increased risk
(%-CI: .; .) and for subjects with three or more previous
premature discharges the risk increased even to . (%-
CI: .; .). e number of previous premature drop-outs
was hence the best predictor of all clusters examined in our
study. Logistic regression further conrmed that duration of
abstinence is not predictive for treatment retention, although
the longest duration of abstinence was twice as long in
patients that completed the treatment as in patients that
dropped out (Table (b)).
Table (a) illustrates the impact of personality in terms
of impulsiveness related variables on treatment outcome.
Among the impulsiveness related variables, experiences of
violence, aggressive behaviour towards third parties, his-
tory of imprisonment, and intravenous drug use inuenced
treatment outcomes negatively. For subjects with prior sui-
cide attempts the number of drop-outs was not statistically
signicant. Similarly, genetic predisposition did not predict
treatment outcome, neither concerning relatives of rst or
second degree with addiction nor for relatives of rst or
second degree with suicidal behaviour.
ese ndings were veried by logistic regression anal-
ysis. In particular, patients without a history of imprison-
ment(OR:.;%-CI:.;.)andpatientswithout
intravenous drug use (OR: .; %-CI: .; .) have a
signicantly reduced risk of premature treatment completion
(Table (b)).
Data on the patient’s medical history were analyzed
(Table (a)). e presence of an addiction related infection
(𝑝 = 0.0145) or a central nervous system disorder (𝑝=
0.0416) was predictive for treatment outcome. On the other
hand, comorbid gastrointestinal disorders (𝑝 = 0.0554)or
peripheral central nervous system damage (𝑝 = 0.7909)
was not predictive for treatment outcome. C onsidering the
Journal of Addiction
T
(a) Results of likelihood Chi square test for sociodemographic determinants of premature treatment drop-out.
Characteristics
Total Treatment completed Dropped out of treatment
p
𝑁 = 832 𝑁 = 525 𝑁 = 307
Age (years) . . . 0.0001
Sex 0.0365
Male (.%) (.%) (.%)
Female (.%) (.%) (.%)
Partnership 0.0020
Living in a partnership (.%) (.%) (.%)
No partnership (.%) (.%) (.%)
Living situation 0.0469
Living alone (.%) (.%) (.%)
Living with other(s) (.%) (.%) (.%)
Children 0.0026
No children ( .%) (.%) (.%)
One child (.%) (.%) (.%)
Two children (.%) (.%) (.%)
ree or more children (.%) (.%) (.%)
Graduation 0.0001
High school ( years of school) (.%) ( .%) (.%)
Realschule ( years of school) (.%) (.%) (.%)
Hauptschule ( years of school) (.%) (.%) (.%)
No graduation (.%) (.%) (.%)
Occupat ional training 0.0001
Academic studies (.%) (.%) (.%)
Apprenticeship (.%) (.%) (.%)
Unskilled (.%) (.%) (.%)
Employment 0.0001
Employed (.%) (.%) (.%)
Pensioned (.%) (.%) (.%)
Welfare (.%) (.%) (.%)
Unemployed (.%) (.%) (.%)
Residence 0.0001
Living in own residence (.%) (.%) (.%)
Assisted living (.%) (.%) (.%)
Other (.%) (.%) (.%)
Homeless (.%) (.%) (.%)
Mother tongue 0.0041
German ( .%) (.%) (.%)
Foreign mother tongue (.%) (.%) (.%)
(b) Results of logistic regression analysis with sociodemographic determinants of premature treatment drop-out.
Characteristic OR % CI Wald Chi 𝑝
Age 0.96 0.95–0.98 19.32 <0.0001
Sex
Male .
Female . .–. . .
Partnership
Living in a partnership .
No partnership . .–. . .
Living situation
Living alone .
Living with other(s) . .–. . .
Journal of Addiction
(b) Continued.
Characteristic OR % CI Wald Chi 𝑝
Children
No children .
One child . .–. . .
Two children . .–. . .
ree or more children . .–. . .
Graduation
High school ( years of school) .
Realschule ( years of school) . .–. . .
Hauptschule ( years of school) . .–. . .
No graduation . .–. . .
Occupat ional training
Academic studies .
Apprenticeship . .–. . .
Unskilled . .–. . .
Employment
Employed .
Pensioned . .–. . .
Welfare . .–. . .
Unemployed 1.97 1.30–2.98 6.00 0.0143
Residence
Living in own residence .
Assisted living . .–. . .
Other . .–. . .
Homeless . .–. . .
Mother tongue
German .
Foreign mother tongue . .–. . .
diagnosed addiction, the rst as well as the second addictive
disorders were signicantly related to treatment outcome
whereas the third diagnosed addiction was not (𝑝 = 0.0865).
Of patients diagnosed with only one addiction, .%
completed the treatment successfully. In our study, patients
with rst addiction diagnosis of alcohol addiction or patho-
logical gambling completed the treatment in % and %,
respectively, of the cases. In contrast, more than half of the
individuals with cannabis, opioid, or multiple drug abuse
dropped out the treatment.
Having an additional, nonaddiction related diagnosis
had a signicant benecial eect on treatment outcome
(𝑝 = 0.0001). Logistic regression conrms that having
no disorder other than the addictive disorder doubled the
risk of premature treatment completion signicantly (OR
., % OR: .–.). Furthermore, it revealed that
having alcohol dependency as the rst diagnosed addictive
disorder increased the risk of premature treatment drop-out
signicantly (see Table (b)). Opioid dependency increased
the risk to . (% OR: .–.), while subjects dependent
from cannabis, sedatives/hypnotics, cocaine, pathological
gambling, or multiple drug use have twice the risk of alcohol
addicted people (OR ., % OR: .–.) to drop out
from treatment.
4. Discussion
e aim of the present study was to identify predictors
of premature discharge during inpatient QDT for alcohol
and drug dependent patients. is study showed that drug
dependent patients bare an elevated risk of premature treat-
ment drop-out compared to alcohol dependent patients.
ese ndings are consistent with past research described in
literature. Braune et al. described drop-out rates of .% for
alcohol dependent and .% for drug dependent patients
[]. Our results further suggest t hat, for patients with multiple
addictions, the main addiction, as well as the second, if
applicable, has an inuence on treatment outcome, but not if
they suer of more than two addictions. Further investigation
should be carried out to verify if there is indeed no distinction
in patients using more than two substances. In our study,
% of the pathologic gamblers succeeded, but furt her
investigations on the inuence of pathologic gambling on the
likelihood of relapse shall be conducted to verify this nding,
as in the present study only four patients with pathologic
gambling were included which limits the generalisability of
the nding.
In our study, being female was a predictor for a better
treatment outcome. But the result of logistic regression