The Great White Hope, spinning occupational outcomes re CBT-GET?
White did a presentation titled "
What helps occupational rehabilitation when the doctor cannot explain the symptoms?" Dolphin posted the URL a few weeks ago (
http://www.sou.gov.se/socialaradet/pdf/Peter Whites presentation.pdf). This reminded me of my recent post about claims by Cella et al (Sharpe & Chalder) 2011 on occupational outcomes for CBT and GET which upon investigation seem to be smoke and mirrors (
http://forums.phoenixrising.me/show...ychosocialists&p=235535&viewfull=1#post235535) so I had a look into it.
From slide 17 of 24 of White's presentation (after discussing CBT/GET for CFS) ...
But do these treatments help patients return to work?
Only cognitive behavior therapy, rehabilitation, and exercise therapy interventions were associated with restoring the ability to work.
- Even without occupation as the aim.
Systematic review: SD Ross et al, Arch Intern Med 2004
The key word here is "associated", the methodological quality of these outcomes was poor so Ross et al probably weren't able to use stronger wording (
http://archinte.ama-assn.org/cgi/reprint/164/10/1098.pdf). White conveniently failed to mention the authors also stated that "
No specific interventions have been proved to be effective in restoring the ability to work." The authors state, "
Only 4 longitudinal studies [26-29] reported employment at baseline and follow-up after intervention." The relevant information is in Table 6 (
http://archinte.ama-assn.org/content/vol164/issue10/images/medium/ioi30120t6.gif) and on p1103, below I'll briefly describe each study in a quotation box:
* Akagi et al 2001 [CBT] (
http://www.ncbi.nlm.nih.gov/pubmed/11600166 or
http://www.cfids-cab.org/cfs-inform/Cbt/agaki.etal01.pdf) : A non-RCT retrospective followup of 94 patients with a questionnaire response rate of only 61% and no control group.
* Dyck et al 1996 [rehabilitation] (
http://www.ncbi.nlm.nih.gov/pubmed/8694980) : I cannot tell from the abstract if there was a control group, however in Table 6 of Ross et el 2004 the employment outcome was based on only 2 patients where 1 of those patients became employed at 3 month followup.
* Fulcher & White 1997 [exercise therapy] (
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2126868) : The comparison of improved occupational status was uncontrolled at 12 month followup because it was a crossover study (and did not account for dropouts or compare with the controls that only did the flexibility program instead of the exercise program), the authors acknowledge this weakness but then try to dismiss it by claiming that spontaneous improvement was an unlikely explanation because it didn't occur in a "similar sample" in another study. The sparse details are on p1651. Table 6 of Ross et el points out that followup figures are "based on the number of patients enrolled" and at 15 month followup from baseline the rate of employment went from 39% to 47% (no control group).
* Marlin et al 1998 [individualized programs] (
http://www.ncbi.nlm.nih.gov/pubmed/9790492) : It appears from the abstract that the intervention group received a range of treatment in addition to CBT/GET ("optimal medical management", pharmacological treatment for psychiatric comorbidity, sleep management, participation of patients' family, etc) while the control group received absolutely nothing at all, so we don't know what effect CBT/GET had by itself. Also, in a systematic review of interventions which included CBT and GET (Whiting et al 2001 -
http://jama.ama-assn.org/content/286/11/1360.full), the methodological quality of Marlin et al 1998 was described as "very poor".
Two other studies are included in Table 6 of Ross et al 2004 as a quasi control group for natural course (Tiersky et al 2001, Vercoulen et al 1994) with the allusion that the poor occupational outcomes therein suggests that the above interventions could be effective, but as stated elsewhere have not proved to be effective in restoring the ability to work.
The Whiting et al 2001 paper I mentioned earlier points out that the reviewed studies reported occupational outcomes at baseline but not post-treatment, and argues for the importance of such outcomes eg employment hours. An updated version of that systematic review (Chambers et al 2006 -
http://jrsm.rsmjournals.com/cgi/content/full/99/10/506) merely repeats the findings of Ross et al 2004 rather than review the data themselves: "
Although the authors found some small studies of interventions (including rehabilitation, CBT and graded exercise therapy [GET]) that reported improved employment outcomes, they concluded that no intervention has been proved to be effective in restoring the ability to work."
On to the updated Cochrane 2004 systematic review for GET (
http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD003200.pub2/pdf), there is no occupation related outcomes (?), "functional work capacity" was mentioned under quality of life but is exercise related (?) and in the one included study for this outcome there was no statistically significant improvement anyway, although in the conclusions it said the improvement was "close to significance".
On to the updated Cochrane 2008 systematic review for CBT (
http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD001027.pub2/pdf), occupational outcomes was reviewed, one study showed no significant improvement in absenteeism from work, while another ("Sharpe 1993" published as Sharpe et al 1996 without the employment data? -
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2349693/pdf/bmj00523-0026.pdf) did show a significant improvement in work status at 12 months between the two groups of 30 participants each (risk ratio = 3.17 [95%CI = 1.47, 6.81]).
The CBT meta-analysis of Malouff et al 2007 (
http://www.cfids-cab.org/cfs-inform/Cbt/malouff.etal07.pdf) briefly mentions but does not clearly give occupation related outcomes. They state that "
Effects of treatment did not vary significantly between objective and subjective measures. That finding may suggest that treatment benefits extend about equally to subjective reports and to observable behavior, such as cognitive test performance and work and school attendance." Note that Table 4 where this data is presented is based on only 62 participants and it is unclear which study the data for "objective functioning" was drawn from?
Have not seen the full text of the recent CBT/GET meta-analysis of Castell et al 2011 (
http://onlinelibrary.wiley.com/doi/10.1111/j.1468-2850.2011.01262.x/full) but judging from PR threads on the paper (
http://forums.phoenixrising.me/show...ed-Exercise-for-CFS-A-Meta-Analysis-(Castell)) and accompanying editorial by Knoop (
http://forums.phoenixrising.me/show...-Fatigue-Syndrome-Where-to-Go-From-Here-Knoop) I think it is a relatively safe bet that occupational outcomes weren't presented.
Not to mention, as noted in Twisk & Maes 2009, the evaluation of the (failed) Belgium CFS clinic application of CBT/GET which showed that employment hours actually decreased after CBT/GET (
http://niceguidelines.files.wordpress.com/2009/10/twisk-maes-cbt1.pdf). The PACE Trial group have not yet published the data it collected on occupational outcomes (the results given for the "Work and Social Adjustment Scale" is not the same). Another safe bet would be, this outcome would have been proudly presented in the 2011 Lancet paper if it was clearly successful.
So apparently after 20 years of research and sweeping claims and false hope, the "best" evidence for improved occupational outcomes boils down to a single CBT study on 60 patients meeting Oxford criteria (not included in the paper that White is using to give the impression of improved occupational outcomes), with contradictory evidence or poor evidence on top of that.