Article Text

Download PDFPDF

Review: psychotherapy reduces a variety of costs when used to treat patients with the most severe psychiatric disorders
  1. Bernie J O'Brien, PhD
  1. McMaster University Hamilton, Ontario, Canada

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

OpenUrlCrossRefPubMedWeb of Science


To determine the economic effects (including the costs of psychotherapy, admission to a psychiatric hospital, mortality, disability and impaired work performance, and other medical and laboratory costs) of providing psychotherapy for patients with psychiatric disorders.

Data sources

English language studies were identified by searching Medline (limited to peer reviewed papers, 1984–94) using the terms psychotherapy; cost effectiveness; cost offset; utilisation of medical care; inpatient admissions; efficacy of individual, group and family therapy; prevention of relapse; and psychology. The newer psychotherapy journals not included in Index Medicus were also searched.

Study selection

Studies were selected if there was an evaluation of a psychotherapeutic intervention (individual, group, or family therapy) in psychiatric illness, and if there were measures of outcomes that had some implications for cost (such as use of services, relapse into illness with a likely increase in treatment costs, number and duration of admissions to hospital, number of medical visits, job performance, employment status, general healthcare costs, laboratory tests and radiographs, and suicide) with a report of actual cost accounting or data on medical care utilisation or work functioning. Review papers and meta-analyses and studies that calculated overall costs by using imputed price indexes multiplied by utilisation costs were excluded.

Data extraction

Data were independently extracted by 2 reviewers on number of patients included; inclusion and exclusion criteria; type of psychiatric disorder; type of intervention and control treatment; and cost and clinical outcomes.

Main results

Of 686 studies identified, 35 were reviewed and 18 met the selection criteria of which 10 were randomised controlled trials (RCTs). Statistical pooling methods of meta-analysis were not used because the studies included in the review varied widely in the type and number of cost categories. Of the 10 RCTs, 8 showed that psychotherapy reduces total costs. 5 studies evaluating the use of psychotherapy for patients with schizophrenia showed that family therapy led to reductions in a variety of costs. Psychotherapy also led to reductions in costs when used to treat patients with bipolar affective disorders and borderline personality disorders. Cost savings were generally realised because of fewer days in hospital for patients receiving psychotherapy. In 1 study (n=36) evaluating the effectiveness of behavioural family management consultation in patients with schizophrenia, the total cost of hospital care per year was $US 45 280 in patients in the control group (individual supportive therapy) compared with $US 4245 for those receiving psychotherapy. The only diagnostic category for which psychotherapy had little effect on cost was that of affective disorders (1 RCT involving 121 outpatients). All 8 non-RCTs showed that psychotherapy reduces costs.


Psychotherapy reduces a variety of costs when used to treat patients with the most severe psychiatric disorders.


The study by Gabbard et al is an interesting and laudable attempt to summarise a diverse literature on the economics of psychotherapy. This review is systematic because the search is comprehensive and the criteria for selection of studies are transparent and reproducible. The major analytical problem encountered by the authors is that the statistical summarising techniques of meta-analysis—much in vogue for combining clinical trials to derive a “meta answer” to a clinical question—cannot be applied to the economics of psychotherapy because studies differ widely in terms of how and why economic end points were measured. Such variation reflects 3 points. Firstly, the definition of psychotherapy used in this review is quite broad and hence the range of clinical evaluative questions addressed has been wide ranging. Consequently, large numbers of replicate economic studies on specific psychotherapy questions simply do not exist.

Secondly, for any given evaluative question there is diversity in how multiple cost and clinical outcomes from psychotherapy have been defined and attributed to therapies. Thirdly, there have been limited advances in the development of methodological and reporting standards for socioeconomic evaluations in mental health care. This latter problem is improving as more payers require economic evidence before granting reimbursement for the latest, expensive antidepressant agent.1 More careful economic scrutiny of new drug treatments is needed for non-pharmacological mental healthcare services.



  • Source of funding: no external funding.

  • For article reprint: Dr G O Gabbard, The Menninger Clinic, 5800 Southwest 6th Avenue, Topeka, KS 66606, USA. Fax +1 913 272 9577.