On-site mental health workers in primary care: effects on professional practice.

BACKGROUND:

Mental health problems are common in primary care and mental health workers (MHWs) are increasingly working in this setting. In addition to treating patients, the introduction of on-site MHWs may lead to changes in the clinical behaviour of primary care providers (PCPs).

OBJECTIVES:

To assess the effects of on-site MHWs in primary care on the clinical behaviour of primary care providers (PCPs).

SEARCH STRATEGY:

We searched the Cochrane Effective Practice and Organisation of Care Group specialised register (18-22 June 1998), the Cochrane Controlled Trials Register (18-22 June 1998), MEDLINE (1966 to 1998), EMBASE (1980 to 1998), PsychInfo (1984 to 1998), CounselLit (18-22 June 1998), NPCRDC skill-mix in primary care bibliography, and reference lists of articles.

SELECTION CRITERIA:

Randomised trials, controlled before and after studies and interrupted time series analyses of MHWs either replacing PCPs as providers of mental health care ('replacement' models) or providing collaborative care/support to PCPs in managing patients' mental health problems ('consultation-liaison' models). The participants were primary care providers. The outcomes included objective PCP behaviours such as diagnosis, prescribing and referral.

DATA COLLECTION AND ANALYSIS:

Two reviewers independently extracted data and assessed study quality.

MAIN RESULTS:

Thirty-eight studies were included involving more than 460 PCPs and more than 3880 patients. There was some evidence that 'replacement' model MHWs achieved significant short-term reductions in PCP psychotropic prescribing and mental health referral, but the effects were not reliable. Consultation rates were also reduced, but with even less evidence of a consistent effect. There were no indirect effects in prescribing behaviour on the wider population and no consistent pattern to the impact on referrals. 'Indirect' effects on PCP consultation rates were not assessed. There was some evidence that 'consultation-liaison' model MHWs had a direct effect on PCP prescribing behaviour when used as part of complex, multifaceted interventions. Few studies examined the 'indirect' effects of such interventions, and those that did failed to provide evidence that 'direct' effects were generalisable to the wider population or endured once the 'consultation-liaison' intervention was removed.

REVIEWER'S CONCLUSIONS:

This review does not support the hypothesis that adding MHWs to primary care provider organisations in 'replacement' models causes a significant or enduring change in PCP behaviour. 'Consultation-liaison' interventions may cause changes in psychotropic prescribing, but these seem short-term and limited to patients under the direct care of the MHW. Longer-term studies are needed to assess the degree to which demonstrated effects endure over time.