Chapter 5    Pg 1 | Pg 2 | Pg 3 | Pg 4 | Pg 5 | Pg 6 | Pg 7

Evaluation and Research Findings

Resource Allocation
An important question for psychiatrists and family physicians establishing a new collaborative service is, “How much psychiatrist time is enough?” A service that offers consultation once a month (Subotsky and Brown 1990) is likely to have considerably less impact on patient outcomes and family physician behaviours than an on-site program that is available 5 days a week (Jackson and others 1993). Similarly, a service staffed by residents (Carr and others 1997c) may produce qualitatively different collaborative relationships than a service provided by a senior consultant who remains committed to the service over a long period. In the studies reviewed, the amount of psychiatrist time spent in the primary care setting was usually reported as number of sessions weekly or monthly. In most cases, a session was considered to be 3 to 4 hours. The range in these studies was 10 sessions yearly to 2 sessions weekly. The populations served ranged from 12 000 to 170 000. There are no scientific justifications cited for methods of allocating psychiatric resources, but one program in Hamilton, Ontario (Kates 1997c), has attempted to standardize psychiatrist allocation based on the size of the primary care practice.

Volume of Service
Five studies reported data on referrals to psychiatrists in sufficient detail to permit calculation of the number of referrals per family physician per year. Despite large differences in the size of these shared care programs and the size of the populations served, the range of referrals was surprisingly narrow. Meadows (1998) reported a frequency of 2.8 referrals per family physician per year, Jackson and others (1993) reported 4.2 referrals per family physician per year, Kates (1997b) and Biderman and others (1999) each reported 5 referrals per family physician per year, and Carr (1997) reported 8.1 referrals per physician per year.

Access to Care
Many programs excluded patients with alcohol and substance abuse. Several had age exclusions and provided service only to patients aged 16 to 65 years. In some services that did accept patients over the age of 65, referrals from this age group were low (Biderman 1999; Kates 1988) This raises issues about equity of access for children, adolescents, and the elderly. For adults, however, the existence of a shared care program appears to increase the likelihood of referral to a psychiatrist or another mental health professional. Kates and others (1997b) report that family physicians referrals rose from 8 yearly prior to the beginning of their service to 73 in the first year following startup. Jackson and others (1993) compared 2 groups of matched general practices in Manchester: one with access to a primary care psychiatry team and one with access only to the traditional outpatient service. In the intervention group of practices, the number of patients who began to receive psychiatric care during the study period (the inception rate) was twice the number in the control group. (P < 0.01). The inception rate for depressive illness was more than 4 times greater in the intervention group (P < 0.01), and the inception rates for mania and schizophrenia were 1.7 times the rate in the control group (just below significance). The effect was greatest, however, for anxiety disorders (5.7 times that of control subjects) and adjustment disorders (5.6 times that of control subjects). Although the data are limited, they suggest that collaborative programs may offer a significant advantage in terms of the accessibility of psychiatric care to primary care patients.

Patient Typologies
Most studies report a significantly higher proportion of female patients referred for psychiatric assessment in the primary care setting. Mood disorders and anxiety disorders were the 2 most commonly diagnosed conditions in patients referred to psychiatrists. The frequency of mood disorders ranged from 7.6% to 59%, and the frequency of anxiety disorders ranged from 11% to 36%. The frequency of schizophrenia and other psychotic disorders in these studies ranged from 1% to 21% of patients, substance abuse ranged from 1% to 18%, organic conditions ranged from < 1% to 3%, and personality disorders ranged from 5% to 26%.

One of the criticisms of shared care programs has been that they tend to deal with the less severe end of the diagnostic spectrum, compared with traditional outpatient clinics (Jackson and others 1993; Corney 1996). Jackson and others describe a “conflict for the staff between the demands for anxious and unhappy people with overwhelming social and domestic difficulties on the one hand, and the more formally ill but often less demanding patients on the other” (p 377). Low and Pullen (1988) reviewed the diagnoses of all patients seen in primary care psychiatry clinics, general hospital psychiatry clinics and psychiatric hospital clinics in the city of Edinburgh for a 5-year period during the 1980s. They found that primary care clinics dealt with a larger proportion (25%) of patients receiving the labels “adjustment reaction” and “no psychiatric diagnosis,” when compared with either the general hospital clinics (15%, P < 0.01) or psychiatric hospital clinics (10%, P < 0.01). In contrast, Ferguson and others (1992) found that there were no significant differences in diagnoses between the patients referred to a primary care clinic in Nottingham and those seen in the area’s traditional outpatient service. Paulsen (1996), on the other hand, describes the patients seen in a well-established program in Massachusetts: “Diagnostically, we see the entire spectrum, but anxiety, somatization, and depressive disorders predominate. Psychopathology such as bipolar disorder or schizophrenia, tends to find its way to psychiatry outpatient treatment” (p 120). Kates and others (1997b) report that 25% of their referrals were given diagnoses of adjustment disorder, V codes, or no diagnosis. Another study by Kates (1988) reports that 30% of patients referred fell into these categories. Carr and others (1997c) report that 34.4% of patients were diagnosed with adjustment disorders, V codes, or no diagnosis. Turner and de Sorkin (1997) and Nickels and McIntyre (1996) each report only 12% of patients falling into the category of adjustment disorder or no diagnosis. Strathdee (1990) reports 13% of patients in these categories.

These results suggest wide variations in the policies that determine the types of patients served by collaborative programs. Ideally, service policies will be determined by local need. In a community with few resources to deal with SMI patients, it may be difficult to justify allocating major psychiatric resources to the milder end of the illness spectrum. At the same time, the significant morbidity created by high-prevalence disorders that are considered less severe cannot be ignored. Shared mental health care initiatives that begin with adequate local-needs assessment and open consultation are likely to be most successful in negotiating an appropriate balance.

Pg 1 | Pg 2 | Pg 3 | Pg 4 | Pg 5 | Pg 6 | Pg 7