Article Text


Review: family intervention added to standard treatment for schizophrenia reduces relapse
  1. Stephen M Lawrie, MRCPsych, MPhil
  1. Royal Edinburgh Hospital Edinburgh, UK

    Statistics from

 QUESTIONS: In patients with schizophrenia, does family intervention increase the efficacy of standard treatment, and what are the characteristics of successful intervention?

    Data sources

    Studies were identified by searching Medline (1976–98), scanning reviews, and handsearching journal titles (1976–97) listed in Evidence-Based Mental Health.

    Study selection

    English language studies were selected if they were clinical trials of psychosocial treatment aimed at families of patients with schizophrenia and related disorders, used a standardised diagnostic system, were randomised controlled trials (RCTs) that compared family intervention with a control group or alternative family interventions, and systematically described and measured outcomes.

    Data extraction

    Data were extracted on number of patients in treatment groups, type of family intervention, patient characteristics, setting, outcome measures, and results.

    Main results

    25 studies (1744 patients) met the selection criteria. 15 studies compared relapse rates, which were defined in different ways across studies. Statistically significant lower relapse rates were reported for the intervention group than for the control group in 7 of 10 studies at 6–12 months. The median relapse rates across studies was 18% and 44% at 1 year and 33% and 64% at 2 years for the intervention and control groups, respectively. 1 study published in 1997 reported higher relapse rates for the intervention group than for the control group at 6–12 months (42% v 25%). 6 studies compared changes in level of pathology in intervention and control groups. 1 study showed higher remission rates for the intervention group than for the control group (56% v 22% at 9 mo and 67% v 17% at 2 y); 2 studies showed moderate treatment effect sizes of 0.34 and 0.38 on global psychopathology. Data from the other 3 studies were not adequate for drawing conclusions. 2 of 9 studies assessing social functioning outcome showed a moderate to substantial treatment effect size for family intervention (0.54 [estimated in non-relapsed patients only] and 1.43 [in 1 of 2 measures used]); 2 studies reported a small treatment effect size, 2 studies reported no statistically significant differences, and 3 studies had unclear or incomplete data. 14 studies examined the effect of family intervention on family-related variables; various measures were used, which limits the strength and comparability of findings. Most studies assessing psychoeducational and/or behavioural approaches had positive effects, whereas studies assessing other models showed negative results or had substantial methodological weaknesses. Studies in which the intervention was aimed at relatives without the patients present consistently showed negative results. Similarly, studies in which family intervention was provided in the short term (≤10 sessions over <6 mo) did not show long lasting results.


    In patients with schizophrenia, the addition of family intervention to standard treatment reduces the risk of clinical relapse. Insufficient evidence exists on the effect of family intervention on mental state and social functioning or on family related variables. Successful intervention characteristics are the inclusion of the patient in treatment, long duration, and the use of information and education about the illness.


    Evidence is increasing for the efficacy of psychosocial interventions in schizophrenia, and some patients (and families) prefer them over medications.1 There can, however, be difficulties with implementation in everyday clinical practice where resources are scarce, trained staff may not be available or may not have the time to see many patients, and providing such interventions over many years is different from being part of a team doing a randomised controlled trial. Indeed, variations in precisely what is given, by whom, and where and when may account for some of the inconsistencies in these studies. Many of the approaches combine overlapping elements, and if we knew what worked for which patients, it might be easier to provide a service.

    Barbato and D'Avanzo in their review of family interventions replicate and extend an earlier version of an existing Cochrane review.2 They show that family interventions reduce relapse rates but also highlight several recent studies that have failed to find this. Many reasons are possible for this finding, but the most likely is that more recent studies are better controlled for non-specific therapeutic elements. The other possibilities worth mentioning are that earlier studies may have included more families with “high expressed emotion” or that patients just spend less time with their families than they used to. They also show that at least some of the effect of family intervention is to increase compliance with medication.

    Barbato and D'Avanzo suggest that family interventions should provide information about the illness, include the patient, and be long term to have long term benefits. The latter result could reflect an important dose response effect. Alternatively, it may mean that non-specific therapeutic elements such as long term support are as important as anything else. The latter interpretation is compatible with what families actually say is beneficial about these approaches.3

    The review by Merinder is arguably a more important contribution because it focuses on one element of many psychosocial interventions—psychoeducation, although I don't see why it can't be called “education.” Information certainly appears to increase knowledge, but it is less clear that it changes attitudes to the illness or its treatment, improves compliance, or reduces relapse. The Cochrane review and meta-analysis of 10 of the RCTs included here, however, finds an overall effect on relapse.4 There is also one rather curious finding in the Merinder review, which may have substantial clinical relevance.

    It appears that short (<10 sessions) and long interventions have no differential effect on improving knowledge or compliance. This could mean that only a certain amount of information can be given usefully to patients. It is also possible that family intervention effects result from more than just education.

    Where does all this leave the practising clinician? The best available evidence shows that various psychosocial interventions can be beneficial in schizophrenia, but that the common elements of information about the illness and long term support may be as good as anything else. Patients clearly need to be involved, and it may only be necessary to involve their family if they are particularly critical or hostile (provided, of course, the patients have contact with them and everyone is amenable to this approach). A rolling group, with intermittent educational sessions, may be all that is required. It should be borne in mind that any effective intervention can have adverse effects and information could conceivably reduce morale and increase hopelessness. Finally, perhaps the greatest service one could give to one's patients and the global community of patients and clinicians would be to routinely audit the implementation of an approach, ask the patients and relatives to indicate what they find useful, and publish the results.


    View Abstract


    • Source of funding: no external funding.

    • For correspondence: Dr A Barbato, Department of Mental Health, Azienda Ospedaliera Salvini, Via Monte Grappa 40, 20051 Limbiate Milano, Italy. Fax +39 02 995 13838.

    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.