This
page presents a summary of data from the small pilot study described in
the NIH proposal, "EEG Biofeedback Treatment of Anxiety in Methadone Patients."
Posted March 21, 2003.
Attendance
Sessions were postponed twice due to heavy snowstorms when non-essential
services of the University were closed.
Subject attendance varied somewhat. Of the seven persons who signed
informed consents, two (Patients 5 & 7) failed to attend any actual
training sessions. Patients 1, 2 & 3 began at the first scheduled
session and all completed. All three could have attended a maximum of 42
sessions over 70 calendar days. Patient 1 missed eight sessions, including
three in a row. She was ill for some of the time, experienced a number
of severely stressful events, including the return from prison of both her
husband and son, yet was able to complete the project. Patient 2 missed
only four of the 42 sessions, despite her paraplegic daughter's frequent hospitalization
for respiratory distress. Patient 3 missed eleven of the 42 sessions
for various reasons, including three in a row, but did complete. Patient
4 began several weeks after others had started, attended seventeen sessions
in a row, and dropped out, citing problems with housing and finances.
Finally, Patient 6 attended eleven sessions, missing three, before dropping
out, citing conflicts with his wife, who did not want him on methadone or
receiving the treatment. In summary, of seven patients who agreed to
participate, three completed, and two more attended at least ten sessions.
Overall, we found the attendance adequate and unsurprising given the subject
population. It is possible that patients starting late did not feel as
committed to the treatment as those who began at the first session. If
and when this research is repeated, we will try to avoid adding patients
to ongoing groups.
Preliminary Results
We quickly discovered that the subjects disliked spending
more than a few minutes per week on testing. As a result, different tests
were administered on different days, and graphs below reflect this. Because
of the inconsistency from patient to patient and test to test, the statistical
analyses mentioned below were based simply on the first and last available
scores. In future research, we will ensure that testing is done more systematically.
The results
are based mainly on only four or five subjects on whom we have multiple
test results. We are well aware that this is a weak foundation on which
to claim success. In particular, the results are entirely pre-post without
a control group and are subject to alternative interpretations, especially
that subjects were simply responding to the attention they were receiving.
Nevertheless, we find the results interesting and encouraging.
Beck
Anxiety Inventory scores showed a marked decline for all subjects over
the course of treatment. Despite the very small N, the average decline
was highly significant by t-test (t (df=4)=
8.40, p< .001). (Here and below, 2-tailed paired-samples t-tests were
used.)
Beck Depression Inventory scores
did not decline among subjects tested with the BDI (t(df=2)=
0.29, not significant). This is consistent with the hypothesis that improvement
is specific to anxiety, and not a general improvement in functioning.
State-Trait Anxiety Index: The State subscale showed
a reduction for all four subjects. This was significant (t(df=3)=
3.67, p< .05), suggesting that day-to-day anxiety levels were declining.
State-Trait Anxiety Index: Not surprisingly, the trait subscale showed
no clear pattern. Mean change was not significant (t(df=3)=
0.28).
Peak skin temperature varied greatly from session to session, probably
because of variations in ambient temperature.
Conclusions
Overall, we were pleased with the results. We have gained a great deal
of experience even from this very small pilot study, and we continue to feel
that the approach is worth further study.