For mental health patients, quality of life (QoL) can include many things like self-esteem, autonomy, satisfaction with the care they’re receiving and the staff who are caring from them, reduced symptoms, minimal side-effects from treatment and good overall functioning.

Measuring QoL can pick up on difficulties a patient is having that wouldn’t have otherwise come to light, and reflects a holism of care that we all aspire to in mental health. Despite that, we don’t measure it enough in daily practice, probably because we’re too busy, not trained in it, or aren’t aware of how useful it might be.

Also, there isn’t all that much evidence: before the study that I’m blogging about today, only one previous RCT (Slade et al, 2006) had compared measuring QoL to treatment as usual. It was small, short and sadly didn’t find QoL assessment to be useful. So it was helpful that in June of this year the British Journal of Psychiatry published a randomised controlled trial performed by the French team of Boyer et al.

The team investigated whether assessing and feeding back QoL scores to the care teams of patients with schizophrenia could improve patient satisfaction.

Methods

This prospective, open-label, randomised controlled trial was conducted at a single site, a psychiatric day hospital in Marseille.

The inclusion criteria were:

• Aged over 18
• Diagnosis of schizophrenia as per DSM-IV-TR criteria
• “Stable disease status” (no major change for 2 months prior to recruitment)
• French speaking

And the exclusion criteria were:

• Reduced capacity to consent
• Another DSM-1V-TR Axis 1 diagnosis
• “Acute decompensation of organic disease” (this phrase was not defined by the authors)
• Mental retardation

Participants were seen by their care team at 0 months (T₀), 3 months (T₁) and 6 months (T₂). They were randomised to be seen consistently in 1 of 3 ways:

• Standard psychiatric assessment involving multidisciplinary interview, clinical examination and completion of some standardised tools:
  • Positive and Negative Syndrome Scale (PANSS)
  • Calgary Depression Scale for Schizophrenia (CDSS)
  • Extrapyramidal Symptoms Rating Scale (ESRS)
  • Global Assessment of Functioning (GAF)
• QoL feedback was identical to the standard group except that participants also completed an S-QoL (Schizophrenia Quality of Life) questionnaire before their assessment, which was given to their care team
• QoL assessment, in which the S-QoL questionnaire was completed by the participants but the care team weren’t told of the result

The primary evaluation criteria was:

• Global satisfaction at T₂

The secondary evaluation criteria were:

• Global satisfaction at T₁
• Satisfaction with care structure at T₁ and T₂
• Satisfaction with care staff at T₁ and T₂
• Psychotic symptoms (PANSS)
• Depressive symptoms (CDSS)
• Side effects (ESRS)
• Global functioning (GAF)
• Overall disease severity (CGI)

Results

Of the participants:

• 124 participants were enrolled from the 142 who were eligible
• They were evenly distributed between the groups (42/40/42)
• Only 2 (from the standard assessment group) dropped out during the trial
• The mean age was 41.1 years
• 67.7% were male
• 21.8% had had at least 12 years education
• With a mean PANSS score of 63.0, they were mildly ill
• Patient characteristics did not differ significantly between groups at baseline

Of the outcomes:

The authors state that:

[At T2] a significantly larger percentage of patients reported high levels of satisfaction in the QoL feedback group compared with the standard psychiatric assessment and QoL assessment groups… In particular, global satisfaction was significantly higher in the QoL feedback group (72.5% patients had high levels of satisfaction) compared with the standard psychiatric assessment (67.5%) and QoL assessment groups (45.2%; P = 0.025).

They add:

This trend towards higher satisfaction in the QoL feedback group was also found at the 3-month follow-up visit with regard to global satisfaction and satisfaction/trust with the staff/care. A total of 75%, 68.3% and 50.0% of patients in the QoL feedback, standard psychiatric assessment and QoL assessment groups, respectively, reported high levels of global satisfaction (P = 0.049).

The authors also mention non-significant trends in PANSS, DCSS, ESRS, GAF and CGA scores in favour of the QoL feedback group.

Discussion

The statement “global satisfaction was significantly higher in the QoL feedback group compared with the standard psychiatric assessment and QoL assessment groups” is not correct.

The difference in global satisfaction between the QoL feedback group and QoL assessment group (29 vs 19 out of 40 participants were “very satisfied”) was certainly significant, but with a sample of this size, it would be surprising if the difference between standard psychiatric assessment and QoL feedback groups (29 vs. 27 out of 40, or 72.5% vs. 67.5% rated as “very satisfied”) was statistically significant – and indeed based on our own analysis, it was not.

Evidence table produced using the CATmaker software

Outcome	Time to Outcome	CER	EER	RRR	ARR	NNT
Very satisfied	6 months	0.675	0.725	-7%	-0.050	-20
	95% Confidence Intervals:			-37% to 22%	-0.251 to 0.151	NNT = 7 to INF; NNH = 4 to INF

 

In short, there was no difference between assessing and feeding back QoL, and standard psychiatric assessment.

In light of this, the conclusions the authors draw seem entirely misplaced:

Our findings provide strong support for integrating QoL assessment and feedback with standard psychiatric assessments.

Priority should be given to strategies to implement QoL measurements in routine practice, including providing systematic feedback for clinicians.

In addition to the statistical misrepresentation, the trial was relatively small, was of only 6 months duration, was open-label and was undertaken at just one centre (predisposing heavily to experimenter bias). The patients would have known what was going on and may have given satisfaction ratings accordingly (Hawthorne bias).

Also, they studied only mildly ill patients suffering from one form of illness and employed an unvalidated QoL measure analysed in a very dichotomous way (“very satisfied” or “unsatisfied to mildly satisfied”).

The trends noted in the secondary measures all have p values of over 0.22, so in my book that’s pushing it a bit too.

The one useful lesson this study can teach us can be found in the difference between the standard assessment and QoL assessment groups: assessing a patient’s QoL and doing nothing leads to far less satisfaction than if you never assessed it in the first place. That makes sense.

Summary

The conclusions drawn in this study are at stark odds with the raw results given. Though utilising QoL measures may hold potential as a therapeutic endeavour, we will need better quality research before we can recommend them for people with schizophrenia.

Links

Boyer L, Lançon C, Baumstarck K, Parola N, Berbis J, Auquier P. Evaluating the impact of a quality of life assessment with feedback to clinicians in patients with schizophrenia: randomised controlled trial. Br J Psychiatry. 2013 Jun;202:447-53. doi: 10.1192/bjp.bp.112.123463. Epub 2013 May 9. [PubMed abstract]

Slade M, McCrone P, Kuipers E, Leese M, Cahill S, Parabiaghi A, et al. Use of standardised outcome measures in adult mental health services. Randomised controlled trial (PDF). Br J Psychiatry 2006; 189: 330–6.