Sixteen evaluative studies were included in the review. There was one controlled trial (n=107), three studies with comparison groups (n=463), one quasi-experimental study (n not reported), two pre- post-intervention studies (n=125), four surveys (n=369), one survey and focus group (n=153), one time series (n=99), one correlation study (n=508) and one economic analysis (n not reported).
The sample size ranged from 3 to 508. All of the studies used convenience samples of participants. None of the studies used a power calculation to estimate the sample size. Few of the 10 studies using questionnaires reported on their reliability and validity. The response rates ranged from 23 to 92% (median 39%). Studies with low response rates did not compare responders with non-responders.
Less than 50% of the studies sought the client's opinion of the service.
Overall, the studies found respondents were satisfied with LMHS. Clinicians were satisfied with the speed of response, quality of assessments, documentation and outcome (1 survey of 29 clinicians).
General nurses were satisfied with liaison nurses' work with families and the ease of referral to LMHS (1 survey of 75 nurses). General nurses in this same study were least satisfied with the documentation and outcome recommendations of LMHS. Ward nurses valued liaison nurses for their availability, objectivity, and good counselling and supportive skills (1 focus group study with 3 nurses).
Physicians most valued a multidisciplinary team for advice on patient management and follow-up visits, but least valued their teaching and conflict resolving (1 survey of 250 physicians). Senior staff valued LMHS more than junior staff.
Clients were satisfied with care provided by liaison nurses (1 pre- post-intervention study with 95 patients), information provided about their treatment (1 survey of 15 clients), and reported high overall satisfaction with the service (1 survey of 57 users and 96 clinicians). The introduction of a liaison nursing service reduced the length of stay at a psychiatric facility (1 comparison of two liaison services with 148 users and 195 clinicians).
A multidisciplinary LMHS doubled the number of psychosocial referrals and reduced the number of people with mental health problems attending A and E departments (1 study with 30 patients comparing a period of liaison with no liaison). Attaching a community psychiatric nurse to A and E departments increased the clients' use of mental health services and reduced the use of other services (1 time series analysis of 99 users).
LMHS made little change to the number of sitters required for mental health clients in general wards (1 controlled trial with 107 users) or to interdisciplinary cooperation (1 comparative study with 148 users and 195 clinicians).
LMHS in A and E departments reduced attendances in A and E by patients with mental health problems, but did not improve attendance at psychiatric appointments (1 quasi-experimental study; number of participants was not reported).