Skip to main content
×
×
Home

Assessing the ‘true’ effect of active antidepressant therapy v. placebo in major depressive disorder: use of a mixture model

  • Michael E. Thase (a1), Klaus G. Larsen (a2) and Sidney H. Kennedy (a3)
Abstract
Background

There is controversy about the implications of relatively small average drug–placebo differences observed in randomised controlled trials of antidepressant medications.

Aims

To investigate whether efficacy is better understood as a large effect in a subgroup of patients.

Method

The mixture model was used to identify patient subgroups (patients benefiting or not benefiting from treatment) to directly model the skewness of Montgomery–åsberg Depression Rating Scale (MADRS) scores at week 8.

Results

The MADRS scores improved by 15.9 points (95% CI 15.2–16.6) among patients who benefited from treatment. The proportion of patients who benefited from escitalopram and not from placebo treatment was 19.5%, corresponding to a number needed to treat of 5.

Conclusions

This model gave a considerably better fit to the data than the analysis of covariance model in which all patients were assumed to benefit from treatment. The small average antidepressant–placebo difference obscures a much larger effect in a clinically meaningful subgroup of patients.

  • View HTML
    • Send article to Kindle

      To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

      Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

      Find out more about the Kindle Personal Document Service.

      Assessing the ‘true’ effect of active antidepressant therapy v. placebo in major depressive disorder: use of a mixture model
      Available formats
      ×
      Send article to Dropbox

      To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

      Assessing the ‘true’ effect of active antidepressant therapy v. placebo in major depressive disorder: use of a mixture model
      Available formats
      ×
      Send article to Google Drive

      To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

      Assessing the ‘true’ effect of active antidepressant therapy v. placebo in major depressive disorder: use of a mixture model
      Available formats
      ×
Copyright
Royal College of Psychiatrists, This paper accords with the NIH Public Access policy and is governed by the licence available athttp://www.rcpsych.ac.uk/pdf/NIH%20licence%20agreement.pdf
Corresponding author
Dr Michael E. Thase, University of Pennsylvania School of Medicine, Suite 689, 3535 Market Street, Philadelphia, PA 19104, USA. Email: thase@mail.med.upenn.edu
Footnotes
Hide All

The original studies were sponsored by H. Lundbeck A/S or Forest Pharmaceuticals, Inc.

Declaration of interest

M.E.T. is an advisor/consultant for H. Lundbeck A/S. During the past 5 years has been advisor/consultant for, and/or received research funding and/or honoraria for talks from: the Agency for Healthcare Research and Quality, Aldolor, Alkermes, AstraZeneca, Bristol-Myers Squibb, Cephalon, Cyberonics, Dey Pharmaceuticals, Eli Lilly, Forest Laboratories (including PGx), GlaxoSmithKline, Janssen Pharmaceutica, MedAvante, Merck (including Organon and Schering-Plough), National Institute of Mental Health, Neuronetics, Novartis, Otsuka, PamLab, Pfizer (including Wyeth), Rexahn, Sanofi Aventis, Sepracor, Shire US, Takeda and Transcept. He has equity holdings in MedAvante and has received income from royalties from American Psychiatric Publishing, Guilford Publications and Herald House. S.H.K has received grant funding and consulting honoraria from H. Lundbeck A/S. In the past 5 years he has also received grant funding or consulting honoraria from AstraZeneca, Biovail, Boehringer-Ingelheim, Eli Lilly, GlaxoSmithKline, Janssen-Ortho, Merck-Frosst, Organon, Pfizer, Servier and St Jude Medical. K.G.L. is an employee of H. Lundbeck A/S.

Footnotes
References
Hide All
1 Moncrieff, J, Kirsch, I. Efficacy of antidepressants in adults. BMJ 2005; 331: 155–7.
2 Kirsch, I, Deacon, BJ, Huedo-Medina, TB, Scoboria, A, Moore, TJ, Johnson, BT. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med 2008; 2: 45.
3 Turner, EH, Matthews, AM, Linardatos, E, Tell, RA, Rosenthal, R. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med 2008; 358: 252–60.
4 Fournier, JC, DeRubeis, RJ, Hollon, SD, Dimidjian, S, Amsterdam, JD, Shelton, RC, et al. Antidepressant drug effects and depression severity: a patient-level meta-analysis. JAMA 2010; 303: 4753.
5 Lepola, UM, Loft, H, Reines, EH. Escitalopram (10–20 mg/day) is effective and well tolerated in a placebo-controlled study in depression in primary care. Int Clin Psychopharmacol 2003; 18: 211–7.
6 Burke, WJ, Gergel, I, Bose, A. Fixed-dose trial of the single isomer SSRI escitalopram in depressed outpatients. J Clin Psychiatry 2002; 63: 331–6.
7 Rapaport, MH, Bose, A, Zheng, H. Escitalopram continuation treatment prevents relapse of depressive episodes. J Clin Psychiatry 2004; 65: 44–9.
8 Ninan, PT, Ventura, D, Wang, J. Escitalopram is effective and well tolerated in the treatment of severe depression. Poster presented at the Congress of the American Psychiatric Association, 17–22 May 2003, San Francisco, California. (http://www.forestclinicaltrials.com/CTR/CTRController/CTRViewPdf?_file_id=scsr/SCSR_SCT-MD-26_final.pdf).
9 Alexopoulos, GS, Gordon, J, Zhang, D. A placebo-controlled trial of escitalopram and sertraline in the treatment of major depressive disorder. Neuropsychopharmacology 2004; 29 (suppl): S87.
10 Khan, A, Leventhal, RM, Khan, SR, Brown, WA. Severity of depression and response to antidepressants and placebo: an analysis of the Food and Drug Administration database. J Clin Psychopharmacol 2002; 22: 40–5.
11 Khan, A, Brodhead, AE, Kolts, RL, Brown, WA. Severity of depressive symptoms and response to antidepressants and placebo in antidepressant trials. J Psychiatr Res 2005; 39: 145–50.
12 Bech, P, Andersen, H, Wade, A. Effective dose of escitalopram in moderate versus severe DSM-IV major depression. Pharmacopsychiatry 2006; 39: 128–34.
13 Montgomery, SA, Åsberg, M. A new depression scale designed to be sensitive to change. Br J Psychiatry 1979; 134: 382–9.
14 Lavori, PW. Clinical trials in psychiatry: should protocol deviation censor patient data? Neuropsychopharmacology 1992; 6: 3948; discussion 49–63.
15 Mallinckrodt, CH, Clark, WS, David, SR. Accounting for dropout bias using mixed-effects models. J Biopharm Stat 2001; 11: 921.
16 McLachlan, GJ, Peel, D. Finite Mixture Models. Wiley, 2000.
17 Zhang, B, Mitchell, SL, Bambauer, KZ, Jones, R, Prigerson, HG. Depressive symptom trajectories and associated risks among bereaved Alzheimer disease caregivers. Am J Geriatr Psychiatry 2008; 16: 145–5.
18 Larsen, K. Joint analysis of time-to-event and multiple binary indicators of latent classes. Biometrics 2004; 60: 8592.
19 Akaike, H. Information theory as an extension of the maximum likelihood principle. In Second International Symposium on Information Theory (eds Petrov, BN, Csaki, F): 267–81. Akademiai Kiado, 1973.
20 Kasper, S, de Swart, H, Andersen, HF. Escitalopram in the treatment of depressed elderly patients. Am J Geriatr Psychiatry 2005; 13: 884–91.
21 Thase, ME. Methodology to measure onset of action. J Clin Psychiatry 2001; 62 (suppl 15): 1821.
22 Mayer, D. Essential Evidence-based Medicine: 117. Cambridge University Press, 2004.
23 Kirsch, I, Moore, TJ, Scoboria, A, Nicholls, SS. The emperor's new drugs: an analysis of antidepressant medication data submitted to the US Food and Drug Administration. Prevent Treat 2002; 5: 23.
24 Uher, R, Muthén, B, Souery, D, Mors, O, Jaracz, J, Placentino, A, et al. Trajectories of change in depression severity during treatment with antidepressants. Psychol Med 2010; 40: 1367–77.
25 National Institute for Health and Clinical Excellence. Depression: The Treatment and Management of Depression in Adults. National Clinical Practice Guideline CG90. NICE, 2009 (http://www.nice.org.uk/CG90fullguideline.pdf).
26 Cohen, J. Statistical Power Analysis for the Behavioural Sciences: 532. Erlbaum, 1988.
27 Kennedy, SH, Andersen, HF, Thase, ME. Escitalopram in the treatment of major depressive disorder: a meta-analysis. Curr Med Res Opin 2009; 25: 161–75.
28 Cipriani, A, Furukawa, TA, Salanti, G, Geddes, JR, Higgins, JP, Churchill, R, et al. Comparative efficacy and acceptability of 12 new-generation antidepressants: a multiple-treatments meta-analysis. Lancet 2009; 373: 746–58.
29 Box, GEP, Draper, NR. Empirical Model-Building and Response Surfaces: 424. Wiley, 1987.
Recommend this journal

Email your librarian or administrator to recommend adding this journal to your organisation's collection.

The British Journal of Psychiatry
  • ISSN: 0007-1250
  • EISSN: 1472-1465
  • URL: /core/journals/the-british-journal-of-psychiatry
Please enter your name
Please enter a valid email address
Who would you like to send this to? *
×

Metrics

Altmetric attention score

Assessing the ‘true’ effect of active antidepressant therapy v. placebo in major depressive disorder: use of a mixture model

  • Michael E. Thase (a1), Klaus G. Larsen (a2) and Sidney H. Kennedy (a3)
Submit a response

eLetters

The use of metaphor and the validity of rating scales

David H Marjot, Consultant Psychiatrist
08 February 2012

The use of metaphor dominates psychiatric thinking (Lakoff and Johnson 1980).One such is happiness in up, sad is down. When sad we are down or we say we are metaphorically depressed. Such metaphorical depression is a metaphorical declivity in a metaphorical surface rather like a puddle or pond. Sadness has intensity so a metaphorical depression has metaphorical depth. ( The only literal dimension in depression is TIME) As we have no scale for metaphorical depth we try to create a metaphorical scale whereby symbols such as 012345.... are used as if they were real numbers. But they are adjectival symbols such as 0 = absent, 1 =slight 2 = moderate etc. to signify this metaphorical depth. These symbols012345... are ideograms standing for adjectival statement and are not numbers that can be added subtracted etc. They are indissolubly part of a code and as such are not available for manipulation as a literal numbers. It follows that rating scales as opposed to symptom check lists such as the Montgomery-Asberg depression scale are logically invalid - the result of confusing a metaphor with literalness. (N.B. Literal means without metaphor, metonomy etc but recently by a change of popular useage a passive metaphor such as 'he was filled with anger' can become active by such saying such as 'he was so angry he literally (i.e. metaphorically) exploded with rage'). It is no wonder that the understanding of clinical responses is so confused if we cannnot distinguish metaphor from literalness. Lakoff G, Johnson M. Metaphors We Live By. 1980. University of Chicago Press. Chicago and London. Montgomery S A, Asberg M. A new depression scale to be sensitive to change.1979 British Journal of Psychiatry. 134;382-389

... More

Conflict of interest: None declared

Write a reply

Understanding responses to antidepressants

Daniel McQueen, Consultant Psychiatrist
13 January 2012

We read with interest the paper by Thase et al., which attempts to delineate two groups of patients with depression, treatment responders andtreatment non-responders.

We suggest that are two main concepts that can describe the superiority of antidepressant over placebo: One is that the degree of improvement with treatment is stochastic, but that antidepressant treatment causes the mean depression score on a given rating scale to reduce to a greater degrees than placebo: The other is that there is a subgroup of patients who are insensitive to the effects of antidepressantsleaving another group in whom the antidepressant-associated benefits are to be found. To our reading, Thase et al have not specified, or tested, a hypothesis about which of these concepts is operating.

The authors have demonstrated that a two-group model fits the data better than a single group model. However, when generating mathematical models it is generally true that the more groups in the model the better the fit to the data. In the limiting condition of the number of groups being equal to the number of subjects the fit will be perfect. If they hadrun competing models with different numbers of groups and demonstrated that there was no gain in accuracy above two groups that would provide support for their 'two group' thesis. Not to have done so sets a substantial limitation on the interpretation of the study. Although Thase et al have demonstrated that the data can be modelled in a certain way only a weak inference can be based upon this observation. We are unclear whether something has been said about the inherent nature of depression under treatment or only that that something has been said about how to be analyse clinical trial data. We would welcome disambiguation on this point.

There is no analysis to identify characteristics of the hypothetical groups (except outcome), and the model has not been tested prospectively. Without having defined these subgroups prospectively they have 'picked thewinners' post hoc, and then claimed that this is evidence of superior effectiveness of antidepressants in an undefined subgroup.

We question the production of NNTs for "responders" and "non-responders", because membership of the groups has only been defined retrospectively based on outcome. This is tautologous; subjects who respond to antidepressants respond better than subjects who do not.

The authors conclude: "These analyses indicate that small mean differences obscure large and clinically meaningful responses for a subgroup of people with depression." An alternative and parsimonious explanation would be that there is a significant amount of random variation in the data.

As it stands the paper shows that some people respond to antidepressants more than others, and whether or not an individual will respond to an antidepressant has not been shown to be identifiable other than retrospectively by observing their response to antidepressants. Giventhe circular nature of this conclusion we feel our understanding of the report may be at fault and would invite clarification.

Thase ME, Larsen KG, Kennedy SH. Assessing the 'true' effect of active antidepressant therapy v. placebo in major depressive disorder: useof a mixture model. Br J Psychiatry 2011;199:501-7.

... More

Conflict of interest: None declared

Write a reply

Can a "true" effect be built on a "wrong" model?

Florian Naudet, PhD Student
21 December 2011

Thase et al. use a sophisticated model to assess the "true" effect ofactive antidepressant therapy versus placebo (1).

Health authorities generally evaluate the efficacy of new medicationsfrom Randomised Controlled Trials (RCTs) versus placebo which are well documented and rely on such a simple statistical paradigm that they can resist the major financial conflicts of interest inherent in the evaluation of pharmaceuticals. Concerning antidepressants, these studies generally identify small average drug-placebo differences (2).

Using statistical modeling, authors addressed the question of outcomemeasurement (3) and found that the efficacy is better understood as a large effect in a subgroup of patients. This is consistent with a common clinical impression.

However their model leads to a curious phenomenon: everything happensas if some patients were considered as non benefiters whereas their final score is less severe than some patients considered as benefiters. As they state, "essentially, all models are wrong but some are useful". Can a "true" effect of active antidepressant versus placebo be built on such a "wrong" model?

Surely not for an health authority. Nevertheless it could be "useful"for searchers and clinicians as it generates hypotheses on the manner in which antidepressants are different from placebo. In this optic, it is necessary to go further and to compare the characteristics of benefiters to non benefiters with the add of two perspectives:

- To perform RCTs in benefiters populations in order to maximize the signal and to minimize the noise. It could help to limit the number of "negative studies";

- To use antidepressants only in this subpopulation of treatment benefiters and to propose alternatives to other patients (psychotherapy, rTMS, electroconvulsive therapy, etc...).

Finally their model is based on RCTs which application in Major Depressive Disorder raises any fundamental questions regarding internal (4) and external validity (5). Even if a "true" effect of active antidepressant exists, I'm not sure that it could be derived from RCTs.

References :

1) Thase ME, Larsen KG, Kennedy SH. Assessing the 'true' effect of active antidepressant therapy v. placebo in major depressive disorder: useof a mixture model. Br J Psychiatry 2011;199:501-7.

2) Kirsch I, Deacon BJ, Huedo-Medina TB, Scoboria A, Moore TJ, Johnson BT. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. PLoS Med 2008;5(2):e45.

3) Moncrieff J, Kirsch I. Efficacy of antidepressants in adults. Bmj 2005;331(7509):155-7.

4) Ioannidis JP. Effectiveness of antidepressants: an evidence myth constructed from a thousand randomized trials? Philos Ethics Humanit Med 2008;3:14.

5) Naudet F, Maria AS, Falissard B. Antidepressant Response in Major Depressive Disorder: A Meta-Regression Comparison of Randomized ControlledTrials and Observational Studies. PLoS One 2011;6(6):e20811.

... More

Conflict of interest: I was peer reviewer for the first draft of the manuscript. I have made the following comments which were not added in the published paper.

Write a reply

×

Reply to: Submit a response


Your details


Conflicting interests

Do you have any conflicting interests? *