Family-focused CBT versus psycho-education for adolescents with CFS: long-term follow-up of an RCT

[Dolphin]

Family-focused cognitive behaviour therapy versus psycho-education for adolescents with chronic fatigue syndrome: long-term follow-up of an RCT.

Behav Res Ther. 2012 Nov;50(11):719-25. doi: 10.1016/j.brat.2012.08.005. Epub 2012 Aug 31.

Lloyd S, Chalder T, Rimes KA.

Source
Department of Psychological Medicine, Institute of Psychiatry, King's College London, Weston Education Centre, Cutcombe Road, London SE5 9RJ, UK. Samantha.lloyd@kcl.ac.uk

Abstract


The aim of this study was to investigate the long term efficacy of family-focused cognitive behaviour therapy (CBT) compared with psycho-education in improving school attendance and other secondary outcomes in adolescents with chronic fatigue syndrome (CFS).

A 24 month follow-up of a randomised controlled trial was carried out.

Participants received either 13 one-hour sessions of family-focused CBT or four one-hour sessions of psycho-education.

Forty-four participants took part in the follow-up study.

The proportion of participants reporting at least 70% school attendance (the primary outcome) at 24 months was 90% in CBT group and 84% in psycho-education group; the difference between the groups was not statistically significant (OR = 1.29, p = 0.80).

The proportion of adolescents who had recovered in the family-focused CBT group was 79% compared with 64% in the psycho-education, according to a definition including fatigue and school attendance.

This difference was not statistically significant (Fisher's exact test, p = 0.34).

Family-focused CBT was associated with significantly better emotional and behavioural adjustment at 24 month follow-up compared to psycho-education, as reported by both adolescents (F = 6.49, p = 0.02) and parents (F = 4.52, P = 0.04).

Impairment significantly decreased in both groups between six and 24 month follow-ups, with no significant group difference in improvement over this period.

Gains previously observed for other secondary outcomes at six month follow-up were maintained at 24 month follow-up with no further significant improvement or group differences in improvement.

In conclusion, gains achieved by adolescents with CFS who had undertaken family-focused CBT and psycho-education generally continued or were maintained at two-year follow-up.

The exception was that family-focused CBT was associated with maintained improvements in emotional and behavioural difficulties whereas psycho-education was associated with deterioration in these outcomes between six and 24-month follow-up.

 

[Dolphin]

The cut-off of 70% or more attendance had been used in the previous trial, based on the fact that at this age many healthy adolescents are not attending school or college full time. The proportion of participants defined as recovered at 24 month follow-up was compared between groups using Fisher’s exact test.

I can't say I find this very convincing.
Also they don't give any other data, apart from whether somebody satisfied a 70% attendance and:

School attendance at 24 months was clearly bimodal and was therefore dichotomised according to whether adolescents had a good outcome for attendance. Good outcome was defined as in the previous trial as attendance of at least 70% of that expected.

70% attendance is not a good attendance rate for people who are supposedly recovered.
If there were some people who were studying part-time plus working, that could have been reported separately.

 

[Dolphin]

They measured global functioning:

Global functioning At 24 month follow-up, no significant group difference was found in the proportion of either adolescents’ or mothers’ reports that the adolescent was much better or very much better overall (Fisher’s exact test ¼ 0.62 and 1.00 respectively). McNemar tests indicated that therewas no significant change in either group in the proportion of adolescents or mothers reporting this outcome between six month and 24-month follow-up. In the adolescents in the psycho-education group, the proportion was 89% at 24 month follow-up, compared with 70% at 6-month follow-up (p = 1.00). The corresponding ratings for mothers in the psycho-education group were 90% and 75% (p = 0.50). In the CBT group 77% of adolescents reported being much or very much better at 24 month follow-up, compared to 83% at six-month follow-up (p = 0.63). For mothers in the CBT group the corresponding figures were 94% and 72% (p = 0.50).

This is presumably a seven point CGI scale.

This was not used as part of the recovery definition i.e. one could say one was the same or only a little better but be counted as recovered.

 

[Dolphin]

Recovery
There is no universal definition of recovery from CFS, with the definition used varying across studies in adults and adolescents with CFS (e.g. Knoop, Bleijenberg, Gielissen, van der Meer, & White, 2007; White et al., 2011; Nijhof et al., 2012; Deale, Husain, Chalder, & Wessely, 2001). In this study, recovery was defined by a score of 18 or below on the Chalder Fatigue Questionnaire (according to the Likert scoring system) and school attendance of 70% or above. The definition was based on fatigue being the principle complaint of adolescents with CFS. The criterion score of 18 or above on the Chalder Fatigue Scale was chosen in line with the cut-off for fatigue recovery defined by White et al. (2011) in their large-scale trial. School attendance was included in the definition of recovery because this was our primary outcome measure and was considered an important target for treatment given the association between CFS and school absenteeism in children and adolescents and the potentially negative impact which this can have on educational and social development.

So a Chalder Fatigue Questionnaire score of 18 or less and 70%+ attendance rate and one was counted as recovered. Not a very rigorous definition of recovery.

 

[Dolphin]

Here are some overall figures:

Table 3
Secondary outcomes over time.
Measure and time point Family-focused CBT mean (SD) Psycho-education mean (SD) Effect F-statistic (df ¼ 1) p Value

Fatigue (Chalder Fatigue Scale)
Baseline 22.26 (5.71) 29.43 (4.66) Group 0.43 0.51
6 Month follow-up 13.31 (5.90) 14.16 (8.42) Time 0.03 0.86
12 Month follow-up 10.40 (5.70) 12.15 (4.79) Group  time 0.03 0.87

Physical functioning (SF-36)
Baseline 51.25 (26.34) 41.67 (24.34) Group 1.05 0.32
6 Month follow-up 80.36 (20.19) 64.00 (36.38) Time 0.10 0.75
24 Month follow-up 76.79 (29.81) 71.20 (27.99) Group  time 2.71 0.12

So we have an average Chalder Fatigue Scale that is less than 11. I maintain these scores show the problems with this scale particularly with CBT studies. These people are saying they have less fatigue than before they were ill effectively. Healthy people score 11 out of 33.

 

[Dolphin]

Small error, I think:

Impairment was assessed using the Social Adjustment Scale (Mundt, Marks, Shear, & Greist, 2002), which measures the extent to which fatigue impairs school, home, social, leisure activities and relationships. Impairment in each domain is measured on a Likert scale from zero ‘not at all impaired’ to eight ‘very severely impaired,’ with a total score out of 40 calculated. The scale was adapted for use with younger people and has been shown to be reliable and valid (Cella, Sharpe, & Chalder, 2011; Mundt et al., 2002). Reliability in this sample was excellent, Cronbach’s a ¼ 0.86.

I think they gave scores out of 8, not 40

as:
Measure and time point Family-focused CBT mean (SD) Psycho-education mean (SD) Effect F-statistic (df ¼ 1) p Value
Impairment (SAS)a Median (IQR) Median (IQR) Z score
Baseline 4.90 (3.45e5.60) 5.00 (4.40e6.50) Group 

 

[Valentijn]

70% attendance would result in students typically being flunked where I went to school, and that was true for high school, university, and law school.

 

[Dolphin]

Probably last observation.
30% didn't respond, although no particular evidence this would bias the results

Participants
63 11 to 18-year olds referred to King’s College Hospital London and fulfilling either the Oxford or CDC criteria for CFS (Fukuda et al., 1994; Sharpe et al., 1991) took part in the original trial and were randomly assigned to family-focused CBT (n = 32) or psychoeducation (n = 31). 44 of the participants from the original study (24 in the CBT group and 20 in the psycho-education group) agreed to take part in the two year follow-up. Of the remaining participants who took part in the original trial, 14 declined to take part, whilst five were not contactable. The follow-up rate at 24 month follow-up was 70% of those participating in the original trial. Table 1 shows baseline demographics for those participating in the follow-up study.

Attrition

Appropriate analyses were carried out in order to examine potential differences between those taking part in the 24 month follow-up and those who did not, according to baseline characteristics. Chi-squared tests found no difference between those completing measures at 24 month follow-up and those lost to follow-up in terms of dichotomous school attendance at baseline (Fisher’s exact test = 1.00). Using independent t-tests, no significant differences were found between the two groups on age, fatigue, physical functioning, child SDQ or maternal SDQ (all t = <1.82). A ManneWhitney U test showed no difference between the two groups on the social adjustment scale at baseline (z =0.58).

Further analyses were also carried out in order to investigate potential differences in outcome at six month follow-up between those participating in the 24 month follow-up and those lost to follow-up, in order to investigate attrition bias. No significant difference was found at six month follow-up between those who completed measures at 24 month follow-up and those lost to follow-up on attendance (Fisher’s exact test ¼ 0.80), fatigue, physical functioning, child SDQ, maternal SDQ (all t = <1.38), or social adjustment (z = 0.62).

 

[Simon]

Some thoughts:

• School attendance was self-reported, as opposed to monitored and reported by the schools, as it has been in other studies. That leaves scope for self-report bias: ask a teenager how much they've been attending school as a result of the programme they are on - and you might not get entirely accurate responses, you might even get overstated attendance.
• The results don't really seem to add up. There was no control group so there can be no robust conclusion on whether or not CBT or Psycho-education actually works. That said, some of the gains were awesome, with teenagers on CBT reporting that on average they had even less fatigue than before they were ill (as Dolphin says above). Psycho-education too got people back to normal in 6 months. The fact that 4 hours of anything can get those with severe fatigue back to normal in 6 months makes me very sceptical. It could mean that natural recovery rates are very high (and treatments have little effect), or they have found the answer to CFS. Either way, such strange results make me doubt the findings.

Less important

• Hard to tell if the high attrition rate biased results. The study was very small with a 30% loss to follow-up at 24 months. That limits the power of the study to detect baseline differences between those who were followed up (44) and those who were not (19) - and of course they wouldn't know if people had lost contact because they had become severely-affected.
• The small study size also makes it hard to detect any real differences between CBT & Psycho-education.
• Note that Psycho-education were much more fatigued at baseline than CBT (29.4 vs 22.3), which was significant, and more impaired (SF36 41.6 v 51.2), which was not significant but with such small sample differences have to be huge to be significant. So the most-impaired group got the least treatment (4 hours) while the least impaired got 13 hours, making for a lopsided experiment.

 

[alex3619]

70% attendance would result in students typically being flunked where I went to school, and that was true for high school, university, and law school.

To get through a science degree at my university, there was a mandatory near 100% attendance for labwork. You could miss a couple with a doctors certificate, but not many. This was why one subject I really wanted to do but had a very very heavy lab component was put in the too hard basket.

 

[Dolphin]

The fact that 4 hours of anything can get those with severe fatigue back to normal in 6 months makes me very sceptical. It could mean that natural recovery rates are very high (and treatments have little effect), or they have found the answer to CFS. Either way, such strange results make me doubt the findings.

The good prognosis/recovery rates in children make me consider the results of research of interventions for them less reliable. Get a reasonable percentage of adults who has been diagnosed* an average of five years, say, (and none less than say three years) back to good functioning and I can find that impressive enough - there seems a good chance it might not be due to chance. Do this for the same percentage of children (or even a much higher percentage) and I still wonder was it natural recovery as well as advice in the early stages of the illness.

*I make a distinction with length ill because if you have not known what was wrong with you, there's quite a good chance you weren't managing the illness well for one reason or another (including pressure from outside sources e.g. expected to work or study full-time); after getting diagnosed, many people, particularly those able to and inclined to read up on it, will improve a bit.