Family Practice Advance Access originally published online on May 30, 2008
Family Practice 2008 25(3):181-190; doi:10.1093/fampra/cmn025
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Failure to improve appropriateness of referrals to adult community mental health services—lessons from a multi-site cluster randomized controlled trial
a Health Service and Population Research Department (HSPRD), Institute of Psychiatry, King's College London, Denmark Hill, London SE5 8AF, UK
b School of Community-Based Medicine, University of Manchester, Manchester, UK
c Park Royal Centre for Mental Health, London, UK
Correspondence to Mike Slade, Health Service and Population Research Department (Box 029), Institute of Psychiatry, King's College London, Denmark Hill, London SE5 8AF, UK; Email: mike.slade{at}lop.kcl.ac.uk
Received 9 October 2007; Revised 11 April 2008; Accepted 21 April 2008.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
Background. Non-clinical factors impact on decisions about whether to refer a patient from primary care to specialist mental health services. The aim of this study was to investigate whether introducing a standardized assessment of severity improves agreement on referrals.
Methods. Multi-site mixed-method cluster randomized controlled trial, investigating GP referrals from 73 practices (408 839 patients) to 11 community mental health teams (CMHTs). Intervention group GPs were asked to complete a Threshold Assessment Grid (TAG) rating of mental health problem severity. CMHTs rated referral appropriateness (ISRCTN86197914 [controlled-trials.com] ).
Results. Two hundred and eighty-one GPs made 1061 mental health referrals. The intervention was only partly implemented with 25% of intervention group GPs completing TAGs. No difference was found in appropriateness (OR 1.18, 95% CI 0.91–1.53) or secondary outcomes. Post-referral primary care contact rates were higher for the intervention group (IRR 1.36, 95% CI 1.07–1.73). Qualitative data identified professional and organizational barriers to implementation.
Conclusions. Asking GPs to complete a TAG when referring to CMHTs did not improve primary–secondary care agreement on referrals. Low implementation means that uncertainty remains about whether introducing a severity-focussed measure into the referral process is beneficial. Introducing local protocols to manage demand at this interface may not be successful and more attention needs to be paid to human and organizational factors in managing interfaces between services.
Keywords. Community mental health services, primary care, primary–secondary interface, referral.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
Patients with mental health problems are mostly managed in primary care with patients with more severe mental health problems receiving additional care from specialist secondary mental health services. The four-fold increase between 1971 and 1997 in levels of referral1 indicates that effectively managing the primary–specialist care interface is a policy priority. An emergent demand management strategy is to triage by severity. Patients with mild to moderate presentations (especially of common mental disorders) are managed in primary care. Patients with more severe mental health problems are managed by specialist mental health services, with input from primary care services varying between countries, but generally including physical health care and some involvement in coordinating care.
Ensuring that the right patients are seen by the right part of the mental health system remains problematic.2 Fewer than 50% of people with common mental disorders will be identified by GPs,3 and once identified, a decision has to be made whether to treat solely at primary care level or refer to specialist mental health services, and there is evidence that severity of illness is not the only factor considered. Thus, patient's gender influences the decision to refer,4 as does the GP relationship with the patient5 and the personal threshold for the individual GP.6 In general, community-based services struggle to retain a focus on the severe mental illness (SMI)7 and the imposition of referral guidelines on primary care does little to help.8
When people are referred due to factors other than clinical need, efficiency and consequently access are reduced because some people are assessed by the CMHT who do not need to be assessed and other people who would benefit from specialist mental health care may not be referred. There is disagreement about appropriateness for 20% of primary care referrals to community mental health teams (CMHTs),9 indicating that this is a priority area for innovations to improve access.
We hypothesized that this disagreement is partly due to a differing emphasis on severity—primary care clinicians focussing on supporting their patient in accessing specialist input and de-emphasising comparative severity (because every patient is important) and secondary care clinicians using severity in decision making both in order to implement policy imperatives and as a non-transparent demand management strategy (i.e. raising the severity threshold when capacity is over-stretched). The Threshold Assessment Grid (TAG)10 is a brief referrer-rated assessment of severity of mental health problems, which has been developed and evaluated for use by referrers to mental health teams.
The aim of this study was to investigate whether introducing a standardized measure of mental health problem severity (TAG) into the referral process improved agreement between primary care referrer and referred-to CMHTs on the suitability of the referral for specialist mental health services. A GP practice-level cluster design was used to address within-practice contamination.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
Design
The study was a mixed-method multi-site cluster randomized controlled trial (RCT). GP practices were randomized, and the unit of analysis was an individual referral of a patient to adult mental health services. The intervention was intended to change GP behaviour, so the design was clustered to reduce the likelihood of contamination.11 The cluster was chosen as the practice rather than the GP because of the logistical problems which would have arisen if a practice contained both intervention and control group GPs.
Four hypotheses were tested, all at the individual referral level. The primary hypothesis was that the intervention will improve the agreement between the GP and the CMHT on the ‘appropriateness’ of the referral. The secondary hypotheses were that the intervention would make it easier for the mental health team to identify (i) the urgency of the referral and (ii) the most appropriate professional to make the initial assessment and (iii) be associated with reduced referral discussion time in the CMHT referral meetings. Additionally, the interface and referral process were explored using qualitative methods, and the resource consequences of using TAG were investigated.
Preliminary evidence about effectiveness was available from previous work,9 which in turn was based on previous theory12 and modelling.10,13 The study was intended to be a definitive RCT within the Medical Research Council complex intervention evaluation framework.11 Appropriate ethical and research governance approvals were obtained for each site.
Sample and setting
The sample size for the primary hypothesis was based on 10 mental health teams each receiving 24 referrals per month for 6 months (1440 in total), of which 65% (936) come from primary care. With a participation rate of 90%, this would give 842 referrals from participating GPs, i.e. 421 referrals per arm. Assuming an average of five referrals per GP, this sample size is equivalent to an approximate effective sample size of 234 per group, assuming an intraclass correlation (ICC) for referrals from individual GPs of up to 0.2 leading to a design effect of 1.8. (The a priori power calculation was done at the design stage when the randomization was envisaged as being by a GP). When it was decided that the randomization was to be by practice, it was decided to retain a sample large enough to allow for GP clustering also, as this was likely to have a larger effect than practice clustering. A deliberately conservative estimate for the GP intraclass cluster correlation was chosen. In fact, the GP cluster design effect was overestimated but the design effect assumed (1.8) turned out to be approximately correct for the practice intraclass cluster correlation actually obtained (1.84 with an ICC of 0.06 and 15 referrals per practice). Previous work found that 20% of referrals were rated as inappropriate,9 so this sample size would allow a difference to be detected in the proportion of inappropriate referrals between the groups of 10% (i.e. a 50% drop from 20% to 10% in the intervention group) with a power of 83%, using a significance level of P = 0.05. In the event, the trial was extended to 9 months and 11 teams to obtain the needed sample size.
The sites were chosen to meet three goals: (i) to have a nationally representative population; (ii) to include high deprivation inner-city areas in which primary care services are less comprehensive, leading to increased pressure on the primary–secondary interface (since these are the areas where the TAG will have the highest potential impact); and (iii) to allow health economic analysis of the system-wide impact of the intervention on a sizeable area. To meet these goals, the sites chosen were one complete London Borough (Croydon) comprising eight adult CMHTs and three CMHTs in Manchester.
The MINI deprivation score for Croydon varies widely (range 81.7–111.1) within what is overall an area with average levels (mean MINI: 100.1) of deprivation for England, making Croydon a nationally representative location. Specialist mental health services are organized into three localities containing eight CMHTs with each CMHT having a multidisciplinary mix of professionals and access to beds. Croydon Borough is not inner city and not predominantly multi-ethnic. The three teams from Manchester were chosen to add these characteristics to the sample frame. The participating teams had MINI scores of 115 (two teams) and 125, the highest indicators for mental health needs outside London, with multi-ethnic populations and limited primary care mental health services.
Participants
The inclusion criteria for general practices were that they provided care for residents in the catchment area of the participating teams.
Measures
The TAG is a one-page referrer-rated assessment of mental health problem severity.10 It is completed by making one tick to indicate level of severity in each of seven domains: (i) intentional self-harm; (ii) unintentional self-harm; (iii) risk from others; (iv) risk to others; (v) survival needs/disabilities; (vi) psychological needs/disabilities; and (vii) social needs/disabilities. The scale is "None", "Mild", "Moderate" and "Severe" (four-point scale, ranging from 0 to 3) for domains (ii), (iii), (vi) and (vii), with an extra "Very Severe" domain (score 4) possible for the remaining three domains (which may require immediate action). The TAG total score is the sum of the seven items and ranges from 0 (least severe mental health problems) to 24 (most severe). TAG was developed using innovative consensus techniques: six search workshops (n = 57) followed by a Delphi consultation (n = 58).10 The psychometric properties were then investigated in referral cohorts to 10 adult and older adult mental health teams (n = 605), showing good construct and concurrent validity, internal consistency and test–retest reliability and adequate inter-rater reliability.13 It was also shown to be feasible for routine clinical use by primary care referrers to adult mental health services, with a cut-off TAG total score of 5 maximizing sensitivity (76%) and specificity (50%) in matching mental health team view of suitability.9 Further information is available from www.iop.kcl.ac.uk/prism/tag.
A customized version was developed for each study site, since tailored interventions are more likely to be implemented.14 The TAG score sheet comprised just the one-page rating form, and together with a checklist of aspects to consider in rating the TAG10 was used in Manchester. A TAG pack was used in Croydon which included the TAG score sheet, information about local mental health services, guidance on the referral letter contents,15 a checklist of aspects to consider in rating the TAG,10 and a request to explicitly state for patients with a TAG total score of 4 or less for justification why they require referral to specialist mental health services.
A referral rating form for use by the CMHT was adapted from previous work9 and recorded sociodemographic and clinical information, TAG scores (where applicable), the primary outcome of appropriateness of referral (appropriate versus not appropriate) and the secondary outcomes of ease of identifying urgency and ease of identifying the right professional group to undertake the initial assessment (both on a five-point Likert scale, 1 = Very Difficult to 5 = Very Easy) and discussion time. All outcomes are at the individual referral level.
Randomization of practices
Allocation on the basis of GP practice clusters was performed by an independent statistician (who was blind to the practice identity) using computer-generated random numbers on permuted blocks, stratified by site (Croydon versus Manchester) and practice list size tertiles (small: <3500, medium: 3501–6000, large: >6001).
Intervention
The intervention was at the cluster level. All GPs working in ‘control group’ practices were asked to continue with their usual referral practice: either a referral letter (Croydon) or use of an existing referral form (Manchester). All GPs working in ‘intervention group’ practices were asked to continue with their existing method and in addition to complete and attach a TAG, using the TAG pack in Croydon or the TAG score sheet in Manchester.
Recruitment of practices
A local implementation group (comprising key stakeholders from local services) provided advice on running the study. The Primary Care Trust, Local Medical Committee and local development meetings were involved/informed of the study as appropriate.
In 6 months before the intervention began (in January, 2005), letters were sent to all practice managers, with telephone follow-up by the researchers. It took, on average, six telephone contacts per practice to either arrange a visit to the practice to discuss the study or to establish the practice was opting out. The study team provided lunch during the GP practice visits to encourage attendance from as many GPs and other staff as possible. During the visit, the purpose of the study was explained, and the TAG pack (in Croydon) or TAG score sheet (in Manchester) was distributed. Questions were constructively replied to but no training was provided for completing the TAG, since it is intended to be used without formal training. Study awareness and retention were maximized by letters from the Mental Health Trust Borough Director before the study began (Croydon only) and distribution of a newsletter during the data collection phase.
Procedure
The referral rating form was completed for all referrals received by the CMHTs between January 1 and September 30, 2005. It was completed when the referral was discussed, generally in the team meeting, i.e. solely on the basis of referral information, before the CMHT contacts the patient. Local arrangements were made to ensure that it was also completed for emergency referrals processed by the CMHT outside of the team meeting. CMHTs were blind to allocation status, but blinding could not be fully maintained since referrals including a TAG were known to have come from an intervention group GP.
Costs data were collected to explore the cost resources of the intervention: (i) the number of GP contacts for referred patients was recorded for 6 months prior to referral and for the period since referral and (ii) anti-depressant and atypical antipsychotic prescribing costs incurred at primary and secondary care were collected for 6 months before and after the study started. Data from the Primary Care Trust could be disaggregated by allocation status, whereas data from the secondary Mental Health Trust could not.
GPs from both intervention and control groups, mental health team leaders and consultant psychiatrists were invited to participate in semi-structured interviews to explore issues around the primary–secondary care interface, referral and use of TAG. Purposive sampling of GPs was used to ensure variation in practice size and GP gender, ethnicity and experience. All interviews were audiotaped with consent and transcribed verbatim.
Analysis
Stata version 816 was used. Outcomes were compared at follow-up using chi-squared tests and independent sample t-tests. Secondary outcomes measured on five-point scales were converted to binary variables for ease of interpretation and consistency with the primary outcome. An intention-to-treat (ITT) analysis was performed, followed by analysis of the appropriateness and ease of rating urgency and profession using a ‘complier average causal effect’ (CACE) analysis. This analysis is based on the assumption that the proportions of potential compliers are the same in both arms and that the mean for non-compliers in the control arm is the same as that for the non-compliers in the intervention arm.17 The effect being estimated is that of actually using the TAG rather than being given the opportunity to use it (as in an ITT analysis). ITT and CACE analyses were performed as follows: for binary variables, a generalized linear model that models probabilities directly; for continuous outcomes, linear regression; for rates, Poisson regression and for the CACE analysis, instrumental variable analysis (two-stage least squares). Stata commands binreg, regress, poisson and ivreg were used for these analyses, respectively. In all cases, site was included as a fixed effect and clustering by practice was included by using the cluster option, or in the case of ivreg, bootstrapping of clusters, a procedure which also corrects standard errors for misspecification of the distribution.
The cost of using the TAG was estimated using information on the material costs of producing and distributing the instrument and staff time spent completing and reading it. Primary care contact rates were compared using a Poisson regression model with the time during which contacts could be made used as the exposure variable and the group variable as the independent variable. For follow-up GP contacts, the baseline rate was entered as an additional independent variable.
In the nested qualitative study, interviews with the primary care (GP) and secondary care (community mental health team leaders, consultant psychiatrists) participants were analysed thematically by constant comparison.18 Themes emerging from one set of professionals informed the interview schedule used with the other. Analysis was completed independently by researchers with differing professional backgrounds (GP, psychiatry, nursing and psychology), with themes agreed through discussion.19
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
Practices
The trial flow diagram for allocation and involvement of GP practices is shown in Figure 1.
|
Participating GP practices provided care for 408 839 patients (297 756 in Croydon, 111 083 in Manchester).
Four of the 27 non-participating practices gave no reason, and some gave more than one. Reasons for opting out were ‘Too busy’ (n = 12), ‘Already have too much paperwork’ (n = 9), ‘Not interested’ (n = 4), ‘Don't see the benefit for the practice of the research’ (n = 4), ‘TAG is too complicated’ (n = 1) and ‘Not interested unless paid’ (n = 1). Opt-out GP practices provided care for 119 961 patients (56 078 in Croydon and 68 883 in Manchester), comprising 6 large practices (four Croydon and two Manchester), 6 medium practices (one Croydon and five Manchester) and 15 small practices (5 Croydon, 10 Manchester). Hence, small practices (<3500 patients) comprised 54% of opt-out practices and 29% of participating practices. The opt-out rate was higher in Manchester, for reasons explored later in the results.
Patients
In total, 1061 referrals were made by 281 participating GPs to 11 CMHTs. The characteristics of the referred patients are shown in Table 1.
|
Researchers attended 451 (89%) of 508 referral meetings held during the study period. Of the 57 unattended meetings, data were provided by a team member for 52, indicating missing data from five (1%) referral meetings about six (1%) of patients. Outcome data were therefore available for 1055 patients.
Outcomes of the intervention
Implementation was low—the TAG was used with 25% of intervention group referrals (14% Manchester, 28% Croydon, P = 0.07 for difference between sites). There was no evidence for difference in the gender or diagnosis of patients with and without TAGs attached.
Table 2 shows the primary and three secondary outcomes by trial arm on an ITT basis for the 1055 (99%) rated referrals. There were no significant differences at P = 0.05 between the two trial arms in any outcome. Logistic analyses controlling for site and practice or GP (included as random effects) showed no significant differences at P = 0.05 for any of the comparisons. The ICCs for appropriateness among referrals from the same practice and also among referrals from the same GP were 0.06 in both cases.
|
CACE analysis indicated that the estimated mean primary outcome (appropriateness of referral) for compliers before using the TAG was 49%, for rating urgency was 60% and identifying the professional was 87%. Table 2 includes the CACE analyses, showing greater (but still non-significant) treatment effects than the ITT analyses (for example, complier mean for the primary outcome increased by 16%).
Resource consequences
Resource consequences were investigated in Croydon only. Thirty of the 55 Croydon GP practices supply contact rates. Patients referred by intervention group GPs had higher post-referral primary care contact rates than control group patients (Table 3). The difference after adjusting for differences in baseline contacts and time between referral and audit was statistically significant. These rates extrapolated over a 1-year period result in an extra cost of £42 pa for the intervention group.20
|
Implementation issues
Reasons for the inadequate implementation were explored qualitatively with 35 GP referrers and 17 CMHT leaders and consultant psychiatrists. Two types of implementation block were identified: professional (for both referrer and referred-to team) and organizational. Illustrative data are shown in Table 4.
|
For GPs, forgetting to use the TAG when making a referral (as so few referrals were made that TAG use did not become routine) was not the only reason that TAG was not completed. GPs suggested that the TAG was simplistic and so did not reflect the complexity of dealing with patients with mental health problems. Some GPs expressed concern that the TAG score could be manipulated by other GPs to coerce the CMHT to accept referrals, and other GPs feared that TAG would be used by CMHTs to further restrict referrals.
For CMHT respondents, the view was expressed that GPs were neither willing to complete schedules nor reliable in their completion of TAGs. However, they also reported that TAGs accompanying referrals had not been considered in their referral meetings, so TAG scores had not, in fact, affected their decision making.
At the organizational level, the two sites used differing approaches to implementation. In Croydon, the evaluation was called a service development and directly supported by the mental health trust. In Manchester, the evaluation was not part of a service change, so was perceived as research, and practices were more able to initially refuse to participate in the study and to later opt out of using TAG. This may account for a lower GP practice participation rate and lower use of TAG in Manchester.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
This multi-site multi-method cluster RCT investigated the introduction of a standardized assessment of mental health problem severity into the referral process from primary to secondary care. The use of TAG did not impact on CMHT views about the ‘appropriateness’ of the referral, and so the intervention, while low cost, was not shown to be effective. The intervention was only implemented by 25% of intervention group GPs. The nested qualitative study identified two barriers to implementation: professional (e.g. degree of trust and interpersonal relationships) and organizational (e.g. perception that standardized referral approaches are a camouflaged approach to rationing and differing perceptions about the importance of severity).
Strengths and limitations
Strengths include sample size and design. The participating sample was large: 73 GP practices providing care for 408 839 patients, i.e. 0.8% of the population of England. The cluster design is an appropriate approach to minimize within-practice contamination. The use of a multi-method approach gives an understanding of why implementation was limited.
The study had several weaknesses. The impossibility of fully blinding participating CMHTs to allocation status points to the need for objective and unbiased measures of outcome, and the use of a subjective outcome of appropriateness as the primary outcome increases increase the possibility of bias. This outcome was chosen to have maximum validity, since the goal of the intervention was to improve agreement. The general issues of the low level of implementation and the difficulty in using trial methodology to evaluate complex interventions are discussed below.
Managing the primary–secondary care interface
Several barriers to accessing both primary and secondary mental health care have been identified, including no primary care contact, poor recognition and in-system access barriers.21 There is evidence of inequitable primary to secondary mental health care access, with non-clinical factors impacting on decisions about whether to refer, e.g. gender4 and ethnicity.22 When people are referred due to factors other than clinical need, an inefficient system results: unnecessary referrals are made, leading to reduced access for necessary referrals. The situation is further complicated by emerging evidence of the lack of benefit from early referral of common mental disorders over watchful waiting in primary care.23 Active management of the primary–secondary care interface is needed to ensure equity of access.
Our study can inform efforts to improve primary–secondary care communication in two ways. First, caution should be exercised over the introduction of a new process such as a referral form. Prior to our study, the TAG had been carefully developed over a 10-year period within an externally funded research programme to develop a standardized mental health referral form. Four previous research grants had funded a systematic review, Delphi consultations, expert consensus workshops and a 10-site prospective cohort study evaluating the TAG. The rationale for its use was explained in our study through visits by researchers to 60 of the 72 participating practices. Since most new processes will be less tested before introduction and less explained when implemented, the likelihood of benefits arising may be even lower.
Second, the narratives of both GP referrers and referred-to team leaders and psychiatrists concentrated on the relationships between the health professionals and how this influenced the referral process and outcome for both patient and professional. This indicates that formal referral processes (i.e. the paperwork) are embedded in a rich interpersonal context.6,8 More attention to these processes (such as identifying and minimizing implementation barriers) in the TAG development phase (rather than our focus on psychometric properties and feasibility of the measure itself) might have led to a higher rate of implementation. Future research will need to use methods which investigate formal process changes as only one part of a multi-level intervention to improve communication and mutual understanding across the interface.
Treatment fidelity in complex interventions
Our study raises a general methodological issue. The uptake of the intervention was low, so uncertainty remains about its effectiveness.24 In other words, even a rigorous trial based on current best practice in complex intervention evaluation25 may not yield clear-cut results. Inclusion of a ‘process evaluation’ (collection of information to understand how the intervention is implemented and received26) indicated the relevance of multiple contextual factors.
The importance of context for complex interventions is becoming apparent.27 Interventions which are tailored to the setting are more effective,14 but varying the intervention conflicts with the goal of minimizing variation in intervention implementation (i.e. treatment fidelity). This tension can be addressed using two approaches.
The first approach to considering context involves amending the intervention to the minimal degree necessary to allow implementation in each site. Our study used this approach by having TAG either as a stand-alone single-page adjunct to an existing referral form or as a multi-page elaborated pack given to referrers. However, we showed that not just the setting but also multiple contextual aspects from interpersonal relationships between individual participants to organizational beliefs were relevant to implementation. Our study design took no account of the impact of these moderators28 for example using different approaches with GPs who were more or less favourable towards the importance of severity, who had or did not have an existing positive relationship with their CMHT, etc.
The second approach to context involves treating the intervention as a collection of options all based on a single coherent theoretical base. This means that the content of one implementation of the intervention may overlap totally, partially or minimally with another instance. This approach has been used with patient-level interventions, such as the development of manuals for psychological therapies. It has not been used with service-level interventions, perhaps because of the difficulty in describing their theoretical basis. In our study, this might involve the use of TAG as one of several elements of an overarching package of interventions to improve primary–secondary care communication. The higher proportion of small list size practices who opted out may point to the need for a different type of intervention, depending on practice size.
The complexity of complex interventions lies on a continuum. Those at the more complicated end are concerned with services or systems rather than patients, address problems characterized by polarized disagreement (e.g. how important is severity in deciding to refer?) and require attitudinal change for implementation. There is a need for methodological development to combine the strengths of clinical trials with a recognition of this contextual complexity. One approach is to modify standard trial designs to account for patient preference, although systematic differences may exist between patients expressing or not expressing a preference or between those willing or unwilling to be randomized.29 Where an adequate theoretical basis can be established, further modifications to trial design may be needed to investigate systematically varied interventions. For complex interventions in complicated contexts, it may however become necessary to employ evaluative methodologies which treat context as an opportunity rather than a threat, such as the realistic evaluation approach to investigating how mechanisms acting in contexts produce outcomes.30
|
Declaration of interest |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
Funding: NHS Service Delivery and Organisation (SDO) Research and Development Programme (SDO/71/2003). The views expressed in this paper do not represent those of the funders.
Ethical approval: This study was approved by Metropolitan Multi-Centre Ethics Committee (04/MRE11/8) with Local REC approval in London and Manchester, and research governance support from the Manchester and Croydon PCTs and South London and Croydon Maudsley NHS Trust and Manchester Mental Health and Social Care Trust.
Conflicts of interest: None.
|
Acknowledgments |
|---|
The study was funded by the NHS Service Development and Organisation R&D Programme. We are grateful to the participating primary care and mental health service staff in Croydon and Manchester.
|
Notes |
|---|
Slade M, Gask L, Leese M, McCrone P, Montana C, Powell R, Stewart M, Chew-Graham C. Failure to improve appropriateness of referrals to adult community mental health services—lessons from a multi-site cluster randomized controlled trial. Family Practice 2008; 25: 181–190.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclaration of interestReferences |
|---|
1 Verhaak PF, Van De Lisdonk EH, Bor JH, Hutschemaekers GJM. GPs' referral to mental health care during the past 25 years. Br J Gen Pract (2000) 50:307–308.[Web of Science][Medline]
2 King C. Severe mental illness: managing the boundary of a CMHT. J Ment Health (2001) 10:75–86.[CrossRef]
3 Vázquez-Barquero JL, Herran A, Simon JA. Epidemiology of mental disorders in the community and primary care. In: Common Mental Disorders in Primary Care—Tansella M, Thornicroft G, eds. (1999) London: Routledge. 3–16.
4 Ross S, Moffat K, Mcconnachie A, Gordon J, Wilson P. Sex and attitude: a randomized vignette study of the management of depression by general practitioners. Br J Gen Pract (1999) 49:17–21.[Web of Science][Medline]
5 Evans A. A study of the referral decision in general practice. Family Pract (1993) 10:104–110.
6 Chew-Graham CA, Slade M, Montana C, Stewart M, Gask L. A qualitative study of referral to community mental health teams in the UK: exploring the rhetoric and the reality. BMC Health Serv Res (2007) 7:117.[CrossRef][Medline]
7 Harrison J, Kisely SR, Jones JA, Blake I, Creed FH. Access to psychiatric care. J Public Health Med (1997) 19:69–75.
8 Chew-Graham CA, Slade M, Montana C, Stewart M, Gask L. The loss of doctor-to-doctor communication: lessons from the reconfiguration of mental health services. J Health Serv Res Policy (2008) 13:6–12.
9 Slade M, Cahill S, Kelsey W, Leese M, Powell R. Threshold 4: an evaluation of the Threshold Assessment Grid as an aid to mental health referrals. Prim Care Ment Health (2002) 1:45–54.
10 Slade M, Powell R, Rosen A, Strathdee G. Threshold Assessment Grid (TAG): the development of a valid and brief scale to assess the severity of mental illness. Soc Psychiatry Psychiatr Epidemiol (2000) 35:78–85.[CrossRef][Web of Science][Medline]
11 Medical Research Council. In: Cluster Randomised Trials (2002) London: MRC.
12 Slade M, Powell R, Strathdee G. Current approaches to identifying the severely mentally ill. Soc Psychiatry Psychiatr Epidemiol (1997) 32:177–184.[CrossRef][Web of Science][Medline]
13 Slade M, Cahill S, Kelsey W, Powell R, Strathdee G. Threshold 2: the reliability, validity and sensitivity to change of the Threshold Assessment Grid (TAG). Acta Psychiatr Scand (2002) 106:453–460.[CrossRef][Web of Science][Medline]
14 Lomas J. Diffusion, dissemination and implementation: who should do what? Ann N Y Acad Sci (1993) 703:226–257.[Web of Science][Medline]
15 Strathdee G. Delivery of psychiatric care. J R Soc Med (1990) 83:222–225.[Abstract]
16 Statacorp. In: Stata Version 8.2 (2005) College Station: TX.
17 Dunn G, Maracy M, Dowrick C, et al. Estimating psychological treatment effects from a randomised controlled trial with both non-compliance and loss to follow up. Br J Psychiatry (2003) 183:323–331.
18 Strauss A. Qualitative Analysis for Social Scientists (1987) Cambridge: Cambridge University Press.
19 Henwood KL, Pidgeon NF. Qualitative research and psychological theorizing. Br J Psychol (1992) 83:97–111.[Web of Science][Medline]
20 Curtis L, Netten A. Unit Cost of Health and Social Care (2005) Canterbury: Personal Social Services Research Unit.
21 Rosen R, Florin D, Dixon J. Access to Health Care—Taking Forward the Findings from the Scoping Exercise (2001) London: NCCSDO.
22 Shepherd M, Gunnell D, Maxwell B, Mumford D. Development and evaluation of an inner city mental health team. Soc Psychiatry Psychiatr Epidemiol (1998) 33:129–135.[CrossRef][Web of Science][Medline]
23 Kendrick T, Simons L, Mynors-Wallis L, et al. Cost-effectiveness of referral for generic care or problem-solving treatment from community mental health nurses, compared with usual general practitioner care for common mental disorders. Br J Psychiatry (2006) 189:50–59.
24 Rychetnick L, Frommer M, Hawe P, Shiell A. Criteria for evaluating evidence on public health interventions. J Epidemiol Commun Health (2002) 56:119–127.
25 Campbell J, Fitzpatrick R, Haines A, et al. Framework for design and evaluation of complex interventions to improve health. Br Med J (2000) 321:694–696.
26 Oakley A, Strange V, Bonell C, Allen E, Stephenson J. Process evaluation in randomised controlled trials of complex interventions. Br Med J (2006) 332:413–416.
27 Campbell NC, Murray E, Darbyshire J, et al. Designing and evaluating complex interventions to improve health care. Br Med J (2007) 334:455–459.
28 Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. J Pers Soc Psychol (1986) 51:1173–1182.[CrossRef][Web of Science][Medline]
29 King M, Nazareth I, Lampe F, et al. Conceptual framework and systematic review of the effects of participants' and professionals' preferences in randomised controlled trials. Health Technol Assess (2005) 9(35):1–186.[Web of Science][Medline]
30 Pawson R, Tilley N. Realistic Evaluation (1997) London: Sage.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||