Effectiveness of antipsychotic treatments in a nationwide cohort of patients in community care after first hospitalisation due to schizophrenia and schizoaffective disorder: observational follow-up study
The conclusion of Tiihonen et al.: “Patients treated with
perphenazine depot, clozapine, or olanzapine have a lower risk of
rehospitalisation or all cause discontinuation of their initial treatment
than patients treated with haloperidol” is blatantly wrong, because
incomplete. The wording suggests that the differences observed are related
to the properties of the drugs, but there is no way to discern differences
between drugs from differences between patients in this study. The
conclusion would be justified if inextricably bound up with “Whether this
is the result of differences between the particular drugs or between the
particular patients receiving these drugs, can not be discerned”. It
should be added that no relevant data are available about these patient
characteristics.
This is not just a theoretical consideration. I could refer to
studies showing antipsychotics not being given at random in relation to
patients characteristics. But I will refrain from this exercise, because
it would detract from the main point: the authors should prove that there
is no difference between the groups of patients receiving different drugs.
I can not understand the BMJ publishing this study without these data.
Of course differences between groups of patients exist in the study
by Tiihonen et al., as shown by one example. In the discussion on
limitations of their study they do refer to the common problem of
selection bias in studies such as theirs and go on to mention: “Since […]
clozapine is used only in treatment resistant patients (the most severely
ill), it is unlikely that selection bias could explain the better outcome
associated with these drugs.” Of course, the treatment resistant patient
reacting to clozapine fares well. But what are the authors comparing here?
A better outcome with clozapine in patients resistant to a first choice
antipsychotic in comparison to the outcome on this first choice
antipsychotic in patients not resistant to it?
Cause for more concern than the publication of irresponsible
conclusions is the fact that this leads to generalisations like in the
editorial of the same issue of the BMJ: “In this week's BMJ Tiihonen and
colleagues show that, in practice, some older drugs such as perphenazine
are as efficacious as the newer ones. (Editorial: Psychological and social
interventions for schizophrenia)”. The clinician who considers to use
these results for treatment decisions in his own patients is ill advised.
Rapid Response:
Stop irresponsible conclusions
Dear Sirs,
The conclusion of Tiihonen et al.: “Patients treated with
perphenazine depot, clozapine, or olanzapine have a lower risk of
rehospitalisation or all cause discontinuation of their initial treatment
than patients treated with haloperidol” is blatantly wrong, because
incomplete. The wording suggests that the differences observed are related
to the properties of the drugs, but there is no way to discern differences
between drugs from differences between patients in this study. The
conclusion would be justified if inextricably bound up with “Whether this
is the result of differences between the particular drugs or between the
particular patients receiving these drugs, can not be discerned”. It
should be added that no relevant data are available about these patient
characteristics.
This is not just a theoretical consideration. I could refer to
studies showing antipsychotics not being given at random in relation to
patients characteristics. But I will refrain from this exercise, because
it would detract from the main point: the authors should prove that there
is no difference between the groups of patients receiving different drugs.
I can not understand the BMJ publishing this study without these data.
Of course differences between groups of patients exist in the study
by Tiihonen et al., as shown by one example. In the discussion on
limitations of their study they do refer to the common problem of
selection bias in studies such as theirs and go on to mention: “Since […]
clozapine is used only in treatment resistant patients (the most severely
ill), it is unlikely that selection bias could explain the better outcome
associated with these drugs.” Of course, the treatment resistant patient
reacting to clozapine fares well. But what are the authors comparing here?
A better outcome with clozapine in patients resistant to a first choice
antipsychotic in comparison to the outcome on this first choice
antipsychotic in patients not resistant to it?
Cause for more concern than the publication of irresponsible
conclusions is the fact that this leads to generalisations like in the
editorial of the same issue of the BMJ: “In this week's BMJ Tiihonen and
colleagues show that, in practice, some older drugs such as perphenazine
are as efficacious as the newer ones. (Editorial: Psychological and social
interventions for schizophrenia)”. The clinician who considers to use
these results for treatment decisions in his own patients is ill advised.
Competing interests:
Speaker fees, consultancy fees or research funding:
Astra Zeneca, Boehringer-Ingelheim, Bristol-Myers-Squibb, Eli-Lilly, GlaxoSmithKline, ICN, Janssen, Lundbeck, Organon, Pfizer, Solvay-Duphar, Synthon, Wyeth
Competing interests: No competing interests