111 Park Avenue
Baltimore, Maryland 21201
Contact: Lisa Stambolis, MSN
Phone: (410) 837-5533
Key words: specialized addictions team;
multi-disciplinary team; high-risk behaviors;
IV drug use
The
purpose of this study was to determine
the level of effectiveness for a specialized
addictions team in the treatment of patients
manifesting high-risk sexual behavior.
METHODS
The study population included an inception
cohort of 313 clients first encountered
for services at Health Care for the
Homeless (HCH) in September and October
1996. The study cohort at baseline,
comprised of all clients who presented
for service for the first time in September
or October 1996 (n=313), was 75 percent
male and 25 percent female. The age
range was 21 to 73 years, with a mean
age of 39 years. The racial composition
was 84 percent Black, 13 percent White,
and 3 percent other, a demographic profile
consistent with that for all HCH clientele.
This cohort was followed longitudinally
for 1 year. The study measures consisted
of an assessment of four areas of behavior,
as reported by the client and as recorded
by the provider at he time of each encounter.
These measures were indicators of high-risk
sexual behaviors; intravenous drug use;
nonintravenous drug use; and alcohol
use. Data collected at baseline, 6 months,
and 1 year were used in this study.
In April 1997, the group was sorted
into an intervention group (i.e., a
specialized addictions team group) and
a comparison group (i.e., a multi-disciplinary
team group). Clients "selfselected"
to be in the intervention or control
group, on the basis of how many times,
if any, they presented for addictions
services. There was no attempt to match
the intervention and control groups
for demographics, diagnoses, or other
factors that may have confounded the
results. An attempt was made to control
for readiness-to-change, as defined
by the Transtheoretical Model; however,
this was abandoned due to an insufficient
sample size. The model of service delivery
for the addictions team was based upon
the Transtheoretical Model of Change
(Prochaska, Velicer,
Rossi et al., 1994). In this model,
clients are assessed for their readiness
to change high-risk behaviors, and appropriate
interventions are chosen. The intervention
group for the study consisted of clients
who had three or more documented encounters
by the HCH Addictions Team over a period
of 6 months. The comparison group was
comprised of clients who had three or
more encounters by any of the other
components of the HCH Multi-disciplinary
Team (Medical, Social Service, and/or
Mental Health), but not with the Addictions
Team, during the same time period. The
risk assessment was based upon client
self-report regarding IV and non-IV
drug use, sex behaviors, and alcohol
use. The provider conducting the risk
assessment assigned a nominal indicator
(YES, NO, or NOT ASSESSED) denoting
the results of the risk assessment (i.e.,
did the patient report performing the
particular behavior within the last
6 months). The results of the assessment
were recorded on the HCH Encounter Form
by the provider, who could be a member
of any component of the multi-disciplinary
team. Each risk factor was assessed
independently of the others; therefore,
a client may have been assessed in only
one realm or up to all four realms.
All clients were assessed at baseline
and on every successive encounter for
the four risk factors. In addition,
the assessment determined if the client
had an HIV Test within the past 6 months
and the result of the test; however,
this area of assessment was eliminated
from the data analysis due to inconsistencies.
The responses recorded by the provider
were entered into the HCH Data Base
along with other data for the encounter.
All HCH staff participated in an inservice
on how to record responses on the Risk
Assessment section of the Encounter
Form just prior to the start of the
study in September 1996.
The Risk Assessment conducted at baseline
demonstrated the following distributions:
Table 1: Risk Factors Assessment
at Baseline
N=
313
At risk
Not at
risk
Not
Assessed
N=222
|
High-risksex
154
110
49 |
IV
Drug use
131
159
23 |
Non-IV
Drug Use
156
134
23 |
Alcohol
use
184
115
14 |
A large number of the inception cohort,
which numbered 313 clients, were found
to have dropped out of service by March
and April 1997. Only 25 clients were
found, by query of the data base, to
have had three or more encounters with
any provider at HCH during the interval.
Of the clients who remained under care,
the first follow-up Risk Assessment
showed the following:
Table 2: Risk Factors Present
at 6 Months
N=
25
At risk
Not at
risk
Not
Assessed
|
High-risk
sex
9
9
7 |
IV
Drug
use
3
21
1 |
Non-IV
Drug
Use
6
18
1 |
Alcohol
use
7
18
0 |
Each Risk Assessment realm was examined
separately. However, because of the
small sample size, the data across all
risk factors was pooled for statistical
analysis. Fourteen clients demonstrated
a positive change, meaning that one
or more of the Risk Factors that had
been coded NO at baseline were now coded
YES. Three clients demonstrated an ostensible
negative change, defined as a change
in the indicator for one or more Risk
Factors from NO to YES. Eight clients
demonstrated no change from baseline
to the 6-month assessment. At the same
time the query was done to determine
change, the cohort was sorted into the
control and intervention groups, based
on the number or encounters the clients
had and with whom. The following table
presents the changes demonstrated by
both the control and the intervention
groups, with all risk factors pooled
together:
Table 3: Change in Risk Factors
at 6 Months
|
Control
group
N= 22
Intervention group
N=3
Totals
|
Positive
change
13
1
14 |
Negative
change
2
1
7 |
No
change
7
1
25 |
Over half (56 percent) of the cohort
demonstrated a positive change over
the course of 6 months. This represents
significant change, as determined by
McNemars test for significance of change
(Chi square =9.000, df =1, p<0.01).
The control group demonstrated more
positive change (59 percent) than the
intervention group (33 percent). Though
the sample size was small, Chi square
analysis was done to determine if there
were significant differences between
the control and the intervention groups.
The analysis failed to show significant
differences (Chi square=1.607594, df=2,
p=0.9003). The same analysis was done
in September 1997, 1 year after the
baseline assessment. Some of the inception
cohort who had dropped out of care during
the first 6 months reentered care, but
at the same time, there were additional
drop outs. The sample cohort at the
1-year interval again totaled 25 clients.
The following table presents the change
demonstrated on the Risk Assessment,
in one or more realms, for the 1-year
cohort:
Table 4: Change in Risk Factors
After One Year
| Control
group
Intervention group
Totals
|
Positive
change
11
5
16 |
Negative
change
2
0
2 |
No
change
5
2
7 |
The majority of clients (64 percent) did
demonstrate at least a minimal reduction
(a positive change in at least one area
of assessment) in risk factors. This was
significant change, as determined by the
McNemar test for significance (Chi square
=9.000, df =1, p<0.01). Positive change
was seen more often in the intervention
group (71 percent) than in the control
group (61 percent). However, no statistically
significant differences were found between
the control group and the intervention
group (Chi square =0.862, df =2, p =0.972).
The large number of clients who _dropped
out_ and were not included in the 6-month
or the 1-year follow-up assessments is
of concern in terms of sample size. One
might also assume that those who dropped
out were those at greatest risk at baseline,
and that they demonstrated no change in
risk behaviors. However, the implications
of this issue from a clinical standpoint
are even more critical. Are there certain
characteristics that are predictive of
attrition from services at HCH? Demographic
and Risk Assessment data were examined,
with Chi square analysis done to determine
significance of any factor as related
to attrition. Clients who were assessed
at baseline to be at risk in the realm
of intravenous drug use were found to
be slightly more likely to drop out of
care than those who did not inject drugs
(p=0.06). Alcohol use, high-risk sex behaviors,
and non-intravenous drug use were all
found to be unrelated to drop out status
at significant levels.
The most difficult issue encountered in
the analysis of this study was the small
sample size of clients who remained in
the cohort for the 6-month and the 1-year
time frames. The small sample size makes
it difficult to make any meaningful inferences.
Because of the small sample size, the
data was pooled across risk factor realms
for purposes of data analysis. This lowered
the sensitivity of the Risk Assessment
instrument in terms of analysis. By pooling
the data, more clients demonstrated change
than was seen when each factor was analyzed
separately. From a Harm Reduction standpoint,
any reduction in risks is a favorable
outcome and is to be celebrated; it would
be valuable, however, to demonstrate measurable
successful outcomes in terms of each discrete
risk factor. The large number of clients
who were coded in the Data Base at the
start of the study as "not assessed"
was of concern. Queries of the HCH Data
Base were conducted to determine which
clients had not been assessed, and which
provider(s) had seen them. Feedback was
given to individual providers. About one
third of the clients were found to have
only one documented encounter, which may
be inadequate for establishing the rapport
necessary for risk assessment. About one
quarter were found to have been seen only
in an outreach setting, which may be inappropriate
for confidential conversations between
client and provider. Providers were most
consistent in assessing risks related
to alcohol use, and the least consistent
on assessing risk related to sexual behaviors.
Social Service Providers had more difficulty
conducting the risk assessment than did
Medical or Addiction Team Providers. The
study failed to take into account the
critical construct of readiness to change.
The original research proposal included
the construct in matching control versus
intervention groups. This was abandoned
because of sample size. It may well be
that the intervention group was not as
ready to change as the control group,
or vice-versa. The inclusion of this construct,
however, did enhance the provision of
clinical services to the study cohort,
as well as the general population served
over the course of the study.
Although more than half of the clients
demonstrated positive statistically
significant changes in high-risk behaviors,
no differences in changes in risk factors
were found between the control and the
intervention groups. Therefore, although
the study was based on a small sample,
it appears that a model of service delivery
that includes a multi-disciplinary approach
to risk reduction interventions may
be as effective as one that concentrates
on interventions by a specialized team.
Changes in high-risk behaviors undoubtedly
occur as a result of a variety of factors,
only one of which was examined in this
study.
