| |
Problem
12.1
Can the length of follow
up significantly impact interpretation of a TMD RCT? Two case
histories -- real life examples from two of our trials -- show
how longer RCTs can completely change the apparent outcome of
a study compared to results analyzed only at the end of the intervention
or treatment interval.
Case 1
An RCT comparing a group receiving education, instruction, and
practice in self-care management of TMD signs and symptoms, directed
by a dental hygienist, with no treatment provided by oral medicine
clinician-dentists, was compared to the control group of usual
treatment by oral medicine clinicians. The results -- see Figure
12.1 -- illustrate change over time:
Figure
12.1 RCT Results: Self-Care (SC) vs. Usual Treatment
(UT) for TMD: Characteristic Pain (CPI)
ANCOVA (adjusted for baseline levels & education)
|
|
|
- At the post-treatment
time point, about 3 months after baseline, the characteristic
pain intensity (the average of the highest, mean and current
pain level in the last 6 months) (CPI) of the self-care group
was almost identical to the usual care group, although both
groups showed lower pain intensities.
- However, at 6 months,
and then at 12 months, the CPI of the self-care group decreased
further to create a significant difference between the groups
at 12 months (Dworkin
et al., 2002e).
Often there can be
an induction period or lag time before the change in the outcome
measure is apparent after the intervention (Wittes,
2002).
Case 2
On the other hand, in an RCT to evaluate the possible benefit
for TMD patients of a cognitive-behavioral therapy intervention
combined with usual care compared to usual care alone, experimental
TMD patients receiving the comprehensive treatment package were
significantly improved at the end of the treatment intervention,
with regard to pain severity and pain-related psychosocial interference,
compared to the usual care group.
However, after the
intervention ended, the subjects in the experimental group did
not continue to show the same marked rate of improvement. So,
at the end of one year, both treatment and control groups had
improved significantly, but there was no longer a significant
difference between the experimental and usual care groups with
regard to the primary outcome measures.
In this case, the comprehensive
care experimental group attained their improvement much earlier
than the controls, and, on average had lower pain severity, but
this may indicate that the initial intervention was too brief,
or a short refresher intervention was needed after 6 months (Dworkin
et al., 2002j). See Figure 12.2:
| Fgiure
12.2 RCT Results: Comprehensive Care (CC) vs. Usual
Treatment (UT) for TMD: Characteristic Pain Intensity
(CPI) |
|
|
Question
12.1.1
The interpretations
of the results from both of these studies are dramatically different;
certainly for the cognitive-behavioral therapy intervention, had
the RCT ended with post-treatment data collection and no further
follow up, the conclusion would be that such interventions would
be highly efficacious.
A lesson to be drawn
from these case histories is:
|