Referrals from primary care to community mental health teams: what’s missing?

KEYwords: Referrals; communication; clinical audit; primary care; mental health

Community mental health teams in New Zealand are funded to provide care and treatment for individuals experiencing severe symptoms of serious mental illness; ~3% of the population.^1,2 Policy and guidelines locally, nationally and internationally have long promoted the importance of better integrated health-care systems where transitions between different levels of service are smooth and cause minimal disruption to individual patients’ care.^2–6

The literature is consistent in its acknowledgment of the role played by primary care practitioners in early detection of mental disorders.^7–9 With the exception of emergency care, referrals are required to facilitate transfer of care between primary care and specialist health-care providers. Referrals may be made as phone calls, hardcopy letters or electronic messages. The aim of a referral is to outline the symptoms or condition the patient is experiencing and to request specialist consultation for diagnosis, advice about acute management or longer-term management. Information contained in referrals allows recipient clinicians to triage the level of urgency or need against service entry criteria. The quality and content of a referral plays a major role in determining the referral’s outcome.

An integrative review of available literature was undertaken and subjected to thematic analysis. For the last 30 years, the literature has been consistent in acknowledging the variability in content and quality of referrals from primary care to specialist services, for both physical and mental health concerns.^10–17

Information contained in referrals falls into two categories: administrative and clinical. Previous studies note that most referrals contain adequate administrative details, allowing recipient services to process them. When details are missing, delays may occur. Omissions in clinical data are potentially of greater concern to service providers and to clinicians making and receiving referrals. Key clinical data may include the following: evidence of client consent to have the referral made; statement of the purpose of the referral; and desired outcome. Risk information and rating of urgency (despite there being clinical disagreement between referrers and specialists) have been identified as important in assisting triage and allocation processes.^14,17 The referred person’s physical health history is more likely to be missed in psychiatric than in other referrals. Finally, a case history is important regardless of whether referral is for physical or mental health, but is likely to incorporate slightly different indicators such as allergies, family history of disease, comorbid health conditions for physical health and family history of suicide or mental illness, alcohol and substance use for mental health services.

Where referral guidelines exist, they are shown to be helpful, but not consistently adhered to, although the reasons for this are unclear.^12,14 The relatively new HealthPathways system (https://www.cdhb.health.nz/about-us/key-projects-and-initiatives) has the potential to improve the quality of referrals, but there is currently no guideline within Healthpathways for mental health services.^18,19 There are also no New Zealand-based studies relating to the content and quality of referrals from primary care to specialist mental health services. The objective of this study was to evaluate content of referral letters, using an audit tool.

This was a retrospective file audit of referrals (n = 92) received by the Southern District Health Board (SDHB) Dunedin urban adult community mental health teams. Rural services were excluded because they include crisis referrals. Ethics approval was gained through the University of Auckland Human Ethics Committee (ref 019427) and locality approval (project ID 01360) through Health Research South. Maori Consultation was an integral part of this process.

Referral letters were selected according to identified inclusion and exclusion criteria. Included referrals were initial referrals to secondary mental health services for adults aged 18–65 years from primary care providers. Referrals were excluded for anyone who had previously been under the care of secondary mental health services, or if the referral came from a non-mental health secondary care service such as neurology or the pain clinic or where written advice only was requested. The SDHB’s electronic patient management system identified 469 referrals made during the study time period; 1 January – 30 April 2017. A total of 93 referrals met the inclusion criteria. One referral was unavailable for viewing, leaving a sample of 92 referrals for audit.

As there was no pre-existing audit tool, an instrument was created (Table 1) based on a review of literature and consultation with key informants (triage clinician, Māori mental health manager and a psychiatrist), who suggested adaptations such as inclusion of ethnicity recording, exclusion of mental health act status and other minor adaptations to the terminology used. The audit tool aimed to measure the presence and quality of information contained within referrals. The tool comprised six sections: (1) demographics; (2) risk information; (3) referral purpose; (4) current presentation and case history; (5) social situation; and (6) treatment to date. Data were collected on referral outcome, time between referral and first appointment, whether the referral was sent on to a primary care provider and the method of referral (hardcopy letter, electronic referral management system (ERMS) or other electronic system). The tool was piloted on a sample of 10 files, and minor changes were made to improve clarity.

Table 1. Audit tool

Data were entered into a Microsoft Excel spreadsheet (Microsoft Corporation, Redmond, WA, USA) and analysed using descriptive statistics. Overall, referral quality was rated by two experienced triage clinicians using pre-agreed criteria (poor = minimal information in referral making it unable to be processed; adequate = some information missing in referral but enough included to start triage process; good = no further information required to process referral). A test of interrater reliability was conducted.

Assessment of referral quality (Figure 1a) discovered that one-third of the 92 referrals audited (37%, n = 34) were of poor quality; that is, the triage clinician would need additional information before the referral could be processed. Each section of the instrument allowed for quality to be assessed. An overall rating of quality was also included. Analysis showed that the number of items present in referrals was positively associated with the overall quality of the referral. The same tests applied to data comparing number of items present and referral outcome was not statistically significant; neither was the relationship between referral quality and referral outcome. A total of 57.6% (n = 53) of referrals were accepted into secondary mental health services. Of the accepted referrals (n = 53), a small proportion (17.0%, n = 9) were not seen. Reasons for these people not being seen may vary, and this information was not available to the study, but based on clinical experience, the most common reason for this occurrence is the person not attending the offered appointment. Of the 92 referrals audited, 43.5% (n = 40) were not seen at all by the SDHB’s community mental health team.

Figure 1.  (a) Overall quality of information contained in referral. (b) Presence of risk information in referrals (*Ministry of health initiative addressing family violence; and **Ministry of Health initiative supporting children of parents with a mental illness). (c) Presence of information regarding current presentation. (d) Presence of information relating to treatment to date.

The data show that 13% (n = 12) of referred people were referred on to another service by the triage clinician. Only one of these cases was also accepted by the community mental health team. Of the 12 patients referred on to an alternative primary care provider, 50% (n = 6) had a face-to-face assessment by the community mental health team before the onwards referral was made.

Of the referrals audited, 56.5% (n = 52) referrals were sent via the ERMS system, 38% (n = 35) were handwritten or typed letters, and the remaining 6.5% (n = 6) were made in an alternative electronic format such as email direct to the community mental health team. ERMS referrals contained 12–25 items out of a possible 26 total, with 42 referrals having ≥16 items present (mean = 17.9 items). Referrals by hardcopy letter contained 7–20 items in total, with a one-third of these (n = 13) having ≥16 items present (mean = 14.4 items).

Some key items of clinical information were regularly not included in referral letters. Risk information was completely absent in 42% (n = 38) of cases (Figure 1b). Information pertaining to the referred person’s current presentation often failed to mention physical co-morbidities (missing in 48%; n = 45), co-existing substance or alcohol use disorders (missing in 54.4%; n = 50) or any scores from standardised rating scales recommended for use in primary care (missing in 89.4%; n = 82; Figure 1c). Evidence of an attempt to treat in primary care was absent in 21% (n = 19) of referrals. The information most frequently included was medications trialled, but this was present in only 60% (n = 54) of referrals audited (Figure 1d).

Information regarding the reason for referral (81.5%; n = 75), the case history (76%; n = 70) and social circumstances (78.2%; n = 72) was reasonably acceptable to the specialist services both in terms of frequency of inclusion and quality of the information provided. The most notable exception to this was information about client consent to the referral, with <45% (n = 42) not noting this.

Administrative data were the most frequently included, with the exception of information on ethnicity, which was recorded in <60% of referrals. However, administrative information provided was not always correct (partly due to the pre-population of fields such as address and contact number in the ERMS, which might not have been current).

The key findings of the audit were that referrals were of inconsistent quality; important information was often absent and the quality of the referral was not an indicator of acceptance of the referral, as in some cases, triage clinicians sought further information from the referrer. These findings can be considered in light of the current systems in place for referral management, triage processes and local and national policy for delivery of mental health care in both primary care and specialist services, and the integration of the two.

The results show that there are gaps in referral information across the identified categories. Although a referral containing all required information fields may not meet criteria for acceptance into a secondary service, it will help the triage clinician to guide the referrer to a more appropriate service. The inclusion of information does not necessarily make for a better-quality referral, as without a clear statement of the consumer’s background, the information may be difficult to interpret. Changes to the referral system should be part of a wider effort towards systems integration, thereby ensuring that referrers are more confident of when to refer as well as what information is required. If at the point of referral, the range of treatment options and services available across both primary and secondary sectors is understood, the number of inappropriate or incomplete referrals may decrease.

The audit found that referrers often fail to include information regarding treatment to date, including medications trialled and use of primary care or other providers, with 21% (n = 19) referrals containing no such information. The time taken to gather additional information impacts on the wait time for the person being referred. The point of transition between services is an identified time of increased risk for patients, as the responsibility of care is held between providers. The 2017 New Zealand Auditor General’s report highlighted this risk, particularly in relation to individuals being discharged from inpatient services without identified follow up.²⁰

Integration of primary and secondary care service for patients needing treatment for mental disorders has been identified as a priority in multiple government strategy documents such as The Primary Health Care Strategy,⁴ the Blueprint documents,^21,22 in addition to the World Health Organization’s recommendations.³ Despite policy commitments to improving services so that the needs of the people are met, the recent New Zealand mental health inquiry suggests that the reality is somewhat different and there is considerable room for improvement.²³

We found that specific information that should be routinely gathered in general practice (family violence, presence of children in homes where a parent is experiencing mental health difficulties and people’s drug and alcohol use) according to the Ministry of Health,^5,24,25 appears to be infrequently included in referrals. It is unknown whether this omission is due to the referrer not gathering such information or not believing it is relevant to the referral. The inclusion of such information in mental health referrals is an aid to triage and also benefits patients.^10–16

The strategic directions of national and local health-care services both identify improvements in technology systems as being key to the improved integration of services, ultimately improving patient experience of health care and leading to better outcomes.²⁶ As referrals are a central part of the health-care journey, any system aiming to improve referral quality will impact on both health outcomes and patient experience. This audit indicates that the development of systems is important and that the quality of referrals can be improved by better systems, including the use of mental health-specific referral guidelines.

A limitation of this audit is that it involved a single site, so the results may not be applicable to rural services or other district health boards across New Zealand where alternative referral systems are used.

This audit provides a New Zealand perspective on the challenges found in referral communications that is consistent with international findings. The outcome of the 2018 Mental Health Inquiry²³ and the Government’s recent response to this indicates the goal of improving service integration is now being financially backed through budget allocation. This audit does not address the triage process occurring between receipt of referrals and decisions made about acceptance into a specialist service. One way of exploring this would be to measure the time spent from receipt of a referral to initial assessment. Further study of referrals to rural community mental health teams and emergency psychiatric services would improve understanding of referral patterns and identify areas for improvement across the whole region. A comparison of the efficacy of different triage models may also assist in better understanding the role triage has to play in ensuring that all referrals are actioned in an appropriate and timely manner.

A study of referral patterns and numbers may ascertain the need for referrer education regarding resources available to people with mental health conditions, including non-government providers. Additionally, a reminder to general practices regarding acceptance criteria may be helpful in reducing the number of referrals deemed inappropriate. Future research could involve general practitioners (GPs) in developing and testing an audit instrument, and exploring GPs’ experiences of referrals to mental health services. Systems such as Healthpathways should incorporate prompts to include key referral information to guide GP practice.

The authors declare that they have no competing interests.

The study was completed with funding support from the Te Pou o te Whakaaro Nui, a national centre of evidence-based workforce development for the mental health, addiction and disability sectors.