341 East 12th Avenue
Eugene, Oregon 97401
Contact: Bob Dritz
Phone: (541) 342-8255
Key words: loss of SSI/SSDI; substance
abuse; housing status;
social contacts; mental and physical
health;
contacts with the criminal justice system
The purpose of this study was to determine
the effect of losing Supplemental Security
Income (SSI) or Social Security Disability
Insurance (SSDI) on homeless patients
who lost these benefits as a result of
their drug or alcohol addiction and changes
in eligibility criteria for SSI and SSDI.
METHODS
Due to Federal legislation enacted in
1996, beginning January 1, 1997, recipients
of SSI and SSDI were designated as no
longer being eligible to receive these
benefits if the basis of their eligibility
derived from either alcohol or another
type of drug addiction. That is, unless
an individual had another qualifying disability
other than a substance abuse problem and
could be reclassified as such, he or she
was no longer eligible to receive SSI
or SSDI.
Recognizing the significance of such a
change in public policy, White Bird Clinic
(WBC) hose to follow a sample of its patients
to determine the effects this new policy
on its clients. Through its case managers,
WBC recruited 60 clients who had been
receiving SSI/SSDI benefits prior to January
1, 1997, 33 of whom were to be declared
ineligible for SSI/SSDI and would, therefore,
lose their benefits. Baseline information
was obtained in January 1997; the two
groups were then seen on a monthly basis
for the next 6 months.
In addition to demographic information
collected the first month, all visits
included data collection (primarily through
clients_ completion of a questionnaire
as they were aided by a case manager or
interviewer) on the following eight subject
areas: 1) physical health; 2) mental health;
3) general outlook on life; 4) time with
family/friends; 5) housing status; 6)
alcohol and drug use; 7) social services
applied for and the result; and 8) contact
with the criminal justice system. Most
responses required patients to complete
a short fill-in-theblank or choose an
option from a modified Likert scale. Of
those who lost their benefits, 6 (18.2
percent) were female; 27 (81.8 percent)
were male; 29 (87.9 percent) were White;
and 5 (15.2 percent) were Native American
or Alaskan Native. Ten (37 percent) of
those who retained their benefits were
female; 17 (63 percent) were male; and
24 (92.3 percent) were White. The median
age for both groups was 42 years.
Multiple approaches to statistically testing
group differences were used: statistical
tests included Chi square tests of differences
between group percentages; t-tests for
differences between group means; two-way
repeated measures of analysis of variance
for self-reported general health measures;
and repeated measures anova to test for
a time effect, group effect, and a time-by-group
interaction.
Group differences at baseline were practically
nonexistent. The only difference at baseline
was the SSI/SSDI qualifying disability.
Twenty-four (75.0 percent) of those who
lost their benefits were alcohol dependent
and 3 (9.4 percent) were drug dependent;
no alcohol or drug dependency was reported
in the group that retained their benefits.
These data were primary or secondary disabilities
reported in the application for benefits.
This difference was consistent with the
grouping variable that was defined indirectly
as a function of whether participants
reported alcohol and/or drug dependency
as a disability.
The two-way anova indicated that all three
health measures were lower for those who
lost benefits (significant group effect).
Regarding physical and mental health,
both groups declined over time (significant
time effect); this was not the case regarding
general outlook on life. Both groups changed
at the same rate for all three measures
(insignificant interaction effects). Although
the groups changed over time, the group
that retained its benefits was higher
on all three measures during all 6 months
of follow-up. Plotting the group means
confirmed that the group that lost its
benefits reported significantly lower
measures and that both groups failed to
improve over the 6- month follow-up period.
An analysis of individual change over
time was performed using measurements
of slope and contingency table analyses
(odds ratios associated with positive
change were estimated). The results of
these estimates were the following: 1)
physical health3 (9 percent) of the 33
participants who lost benefits reported
positive change, while 7 (26 percent)
of the 27 who maintained their benefits
reported positive change; 2) mental health_4
(12 percent) of the 33 participants who
lost benefits reported positive change,
while 8 (30 percent) of the 27 who maintained
their benefits reported positive change;
and 3) general outlook on life 5 (15 percent)
of the 33 participants who lost their
benefits reported positive change, while
9 (33 percent) of the 27 who maintained
their benefits reported positive change.
Therefore, those who retained their benefits
were: three times more likely to show
positive change in their physical health
(p<0.10); two and a half times more
likely to show positive
change in their mental health (p<0.10);
and more than two times more likely to
show positive change in their general
outlook on life (p<0.10). It should
be noted, however, that despite these
significant results, a relatively small
proportion of all the studys participants
showed a positive change. Data regarding
contact with friends and family were collected
along with information about living arrangements
for each of the 6 months during follow-up.
For some questions such as contacts with
family/friends, the researcher attempted
to collect monthly counts of how often
contacts were made. This data did not
provide much variation and the ability
of participants to recall information
was a concern. Therefore, instead of using
the counts, researchers collapsed the
data for each measure into an indicator
of whether or not contacts had occurred
that month. These results indicated that
those who lost benefits stayed on the
streets or with relatives/friends significantly
more, while renting significantly less.
There was very low report of alcohol and
drug use. The statistical analyses of
group differences failed to find any statistically
significant group differences for alcohol,
pharmaceutical drug use, or any other
drug use. It is important to note that
this data was self-reported by participants,
who may have been reluctant to report
alcohol and drug use.
Not surprisingly, the group that lost
benefits applied for significantly more
services (e.g., food stamps, medical services),
and occasionally received significantly
fewer services. In addition, those who
lost benefits had significantly more contact
with the criminal justice system (i.e.,
arrested or jailed). Ironically, the group
that maintained its benefits applied less
often for other benefits, yet received
services more often.
This 6-month study may be viewed as an
initial tracking of what happens when
a group of homeless people with severe
behavioral health problems lose a monthly
financial stipend. Starting from an already
low level of health, they deteriorate.
They search more for assistance from other
governmental or charitable sources, but
ironically have less success in obtaining
help as compared to their counterparts
who retain SSI/SSDI benefits.
After SSI/SSDI benefits were discontinued,
there was no indication of an increase
in the participants ability to support
themselves. If people lose their benefits,
they may turn to family and friends, often
with little success. They are likely to
show up more frequently on the streets,
in courts, and in jail. The task of dealing
with them is then shifted to the state
and local human services providers or
the criminal justice system. With the
odd exception of prison, these resources
typically do not replace the SSI/SSDI
that was lost, therefore their health
outlook and housing situation decline
further.
The long-term implications of SSI/SSDI
benefits being lost could be examined
by extended longitudinal studies. For
example, further tracking could show that
people who lose these benefits continue
to decline, possibly leading to hospitalization,
imprisonment, or death. Another possible
avenue of investigation would be to further
research whether or not two groups such
as those in this study truly differ at
baseline only with regard to their SSI/SSDI
status (i.e., loss of this status due
to their primary disability being related
to their substance abuse history).
Trends of the group that maintained benefits
are noteworthy. Even though they were
consistently better off than the group
that lost benefits, they too showed a
decline in physical health, mental health,
and general outlook on life during the
6-month study period. The implication
may be that SSI/SSDI benefits do not stabilize
recipients decline, but rather moderate
it instead.
Some comments should be made regarding
the measures used. The average correlation
of 0.80 for the general health indicators
suggests that the three measures are interrelated.
Using time with friends/family as an indicator
was somewhat ambiguous. Alcohol and drug
use as indicators were the least reliable
indicators, primarily because of participants
fear of being completely honest in their
self-reporting of this information. The
indicators on availability of health,
mental health, dental care, and substance
abuse services were probably skewed.
As clients of WBCs Health Care for the
Homeless program, participants had safety-net
access even after losing Medicaid. Even
though the food stamp indicator revealed
a connection between loss of SSI/SSDI
and loss of food stamps, more detail about
how this occurs would be useful. Trying
to collect any of this type of information
can be very difficult, especially given
the social and mental condition of any
group of patients similar to those in
this study.
The outcome of redefining disability to
exclude substance abusers was negative
for the homeless people in this study,
and shifted additional burdens and costs
on to local communities. The general implications
of this study appear to be that stopping
SSI/SSDI benefits for dually-diagnosed
homeless people results in a further decline
in their already marginal health and social
status. Furthermore, this type of policy
results in an increased demand on other
entitlement programs as well as social
service and correctional agencies.