The sample

All doctors who contacted the service between February 2002 and February 2004 were eligible for inclusion (n = 123). Of these, 98.4% (n = 121) consented to take part.

Measures

Data were gathered from three standardised self-report questionnaires: the Clinical Outcomes in Routine Evaluation - Outcome Measure (CORE-OM); ^{Reference Evans, Mellor-Clark and Margison13} the BSI; ^{Reference Derogatis9} the Maslach Burnout Inventory (MBI); ^{Reference Maslach and Jackson14} and a 20-item bespoke questionnaire designed by the service to assess users’ views and experiences of having attended. At intake a CORE therapist assessment questionnaire was also completed, but this was not repeated at follow-up. The standardised self-report questionnaires were completed at two points in time (before being seen at the service and at the end of the consultation process) with the service user questionnaire administered only at follow-up. The CORE-OM is a 34-item, 5-point Likert scale questionnaire that provides scores for the subscales of subjective well-being, commonly experienced problems or symptoms, life/social functioning and risk to self and others. Standardised scores to determine clinical cut-off points on this measure were used as provided by the distributor. ^{Reference Evans, Mellor-Clark and Margison13} The BSI is a 53-item, 5-point Likert scale questionnaire that targets symptoms in 9 areas. Caseness implies the criterion for requiring a formal mental health assessment and was computed using adult gender-matched non-patient norms. Adult psychiatric out-patient norms for the BSI and the Global Severity Index (GSI) as a component of the BSI were used to measure clients’ degree of symptomatic distress. ^{Reference Derogatis9} The MBI is a 22-item, 7-point Likert scale questionnaire that evaluates work-related distress through three dimensions: emotional exhaustion, depersonalisation and personal accomplishment, each of which was scored independently. ^{Reference Maslach and Jackson14} The user questionnaire is a service-specific, 20-item, 4-point Likert scale questionnaire combined with a section of open boxes to reflect service users’ comments and suggestions. Follow-up questionnaires were administered by post.

Data regarding engagement with and referral on from the service were gathered by two of the authors through examining the notes. This included data regarding frequency of being seen at the service, whether the doctors were on medication prior to presentation to the service, the number who completed a treatment episode (in the service) and did not need further treatment, how many were referred on for additional treatment and for which type of further treatment. The treatment that doctors engaged with and were advised on were compared in terms of the subsample of suicidal and non-suicidal doctors based on CORE therapist assessment at intake as described in the original paper. ^{Reference Garelick, Gross, Richardson, von der Tann, Bland and Hale8}

Data collection and analysis at baseline are described in detail in the original paper The scores of those who completed the questionnaires at follow-up were analysed using paired sample t-tests. Quantitative data were analysed using SPSS v. 15 for Windows. Data regarding engagement, onward referral and user satisfaction were analysed thematically by two of the authors.

CORE-OM

In total 121 consented to take part in the study. Of these, 95% (n = 115) completed the CORE-OM at intake, 74% (n = 89) completed it post-treatment and 73% (n = 86) fully completed both pre- and post-treatment questionnaires. There were no inter-group differences in terms of the doctors who did and did not complete the follow-up questionnaire.

Significant reductions for all domains on the CORE-OM were identified between pre- and post-treatment, apart from the domain of risk (Table 1). There was a reduction in the proportion of doctors who scored above clinical cut-off on the CORE-OM at follow-up (>1.19 male, >1.29 female), from 63% (n = 72/115) of respondents to 34% (n = 30/89). Of the 86 individuals who completed the questionnaire both pre- and post-intervention, 52 had scored above clinical cut-off at intake, and more than half (n = 28/52) no longer had a high score at follow-up. However, six individuals with a normal score at intake were above clinical cut-off at follow-up.

Table 1

Clinical Outcomes in Routine Evaluation - Outcome Measures

			Paired samples (n = 86)
Subscale	Pre-intervention, mean (s.d.) (n = 115)	Post-intervention, mean (s.d.) (n = 89)	t-test	P
Functioning	1.41 (0.76)	1.05 (0.85)	3.98	<0.001
Problems	1.84 (0.86)	1.22 (0.93)	6.36	<0.001
Well-being	2.02 (0.97)	1.28 (1.09)	6.44	<0.001
Risk	0.32 (0.48)	0.27 (0.50)	0.418	0.677
Total score	1.44 (0.69)	1.00 (0.80)	5.58	<0.001

For the subsample of doctors with suicidal risk prior to treatment (n = 51, n = 35 completed at both time points), there remained a significant difference in scores on all subscales and the CORE-OM overall score for these domains, apart from risk. Whereas 34 of the subsample were above cut-off pre-intervention, only 16 were above cut-off post-intervention (3 new clients, 14 were no longer above cut-off).

BSI

Of the initial 121 doctors, 94.2% (n = 14) completed the BSI questionnaire prior to treatment, 75.2% (n = 91) completed the questionnaire at follow-up and 72.7% (n = 88) completed both pre- and post-treatment questionnaires. There were no inter-group differences in terms of the doctors who did and did not complete the follow-up questionnaire. There was a significant difference between all BSI subscales taken pre-and post-treatment, excluding phobic anxiety (Table 2).

Table 2

Brief Symptom Inventory outcomes

			Paired samples (n = 88)
Subscale	Pre-intervention, mean (s.d.) (n = 114)	Post-intervention, mean (s.d.) (n = 91)	t-test	P
Somatisation	0.44 (0.62)	0.30 (0.54)	2.50	0.014
Obsessive-compulsive disorder	1.42 (0.95)	0.94 (0.95)	4.59	<0.001
Interpersonal sensitivity	1.36 (1.01)	1.09 (1.10)	3.00	0.004
Depression	1.48 (0.98)	0.94 (0.99)	5.29	<0.001
Anxiety	1.13 (0.94)	0.76 (0.87)	3.61	0.001
Hostility	0.74 (0.75)	0.53 (0.67)	2.64	0.010
Phobic anxiety	0.45 (0.74)	0.32 (0.59)	1.85	0.068 (n.s.)
Paranoid ideation	0.92 (0.86)	0.70 (0.77)	2.57	0.012
Psychoticism	0.81 (0.67)	0.54 (0.65)	4.19	<0.001
Global Severity Index	0.98 (0.65)	0.67 (0.66)	4.97	<0.001
Positive Symptom Distress Index	26.33 (10.53)	1.53 (0.53)	4.86	<0.001
Positive Symptom, total	1.83 (0.57)	19.60 (14.20)	5.55	<0.001

Whereas 85% (n = 97/114) of responding doctors scored above clinical cut-off for non-patient norms on the BSI at intake (indicating the individual may require a formal mental health assessment), this number reduced to 51.6% (n = 48/91) at follow-up. Of those who completed the questionnaire both pre- and post-treatment, 32 (of 42) individuals who had scored above cut-off, no longer fell within this group. Only eight doctors scored above clinical cut-off for psychiatric out-patients at follow-up, a reduction of three individuals; all had scored above cut-off at intake.

MBI

Of the initial 121 doctors, 94.2% (n = 114) completed the MBI questionnaire prior to treatment, 67% (n = 81) doctors completed the questionnaire at follow-up and 65.3% (n = 79) completed both pre- and post-treatment questionnaires. There were no inter-group differences in terms of the doctors who did and did not complete the follow-up questionnaire. There was a significant change in scores for the emotional exhaustion subscale of the MBI at follow-up (t = 5.55, P<0.001), but not for the remaining subscales of depersonalisation and personal achievement (Table 3).

Table 3

Maslach Burnout Inventory outcomes

			Paired samples (n = 79)
Subscale	Pre-intervention, mean (s.d.) (n = 114)	Post-intervention, mean (s.d.) (n = 81)	t-test	P
Emotional exhaustion	31.97 (11.87)	26.14 (12.26)	5.55	<0.001
Depersonalisation	9.08 (6.73)	8.55 (6.64)	0.92	0.358
Personal achievement	34.59 (7.59)	35.07 (8.58)	– 0.69	0.493

Engagement with and referral on from the service

Of the 121 doctors included in the study, 33.9% (n = 41) were on medication, the majority on antidepressants (85%, n = 35). Most doctors were seen for up to six sessions (78.5%, n = 95); however, 16.5% (n = 20) were seen for 6-20 sessions and 5% (n = 6) for more than 20 sessions.

Doctors who presented with suicidal risk according to the CORE-OM at intake were compared with those who presented as non-suicidal with regard to their engagement with the service and referral onward. Within this analysis we examined how many sessions an individual attended and to what type of treatment they were referred on to, should they have received an onward referral (Fig. 1).

Fig 1

The cohort's treatment and referral pathways. CBT, cognitive–behavioural therapy; NHS, National Health Service.

Of the 70 (58%) doctors who were not deemed suicidal at intake, 58.6% (n = 41) were sufficiently helped by the process of being seen for consultation at MedNet alone, 1.4% (n = 1) was seen for a prolonged episode of treatment at MedNet and 34.3% (n = 24) were referred on for further treatment.

Of the 51 (42%) doctors who were deemed at some risk of suicide at the initial consultation, 18% (n = 9) were sufficiently helped by being seen at MedNet, 10% (n =5) were seen for a prolonged episode of treatment at MedNet and 66.7% (n = 34) were referred on for further treatment.

Despite the high level of suicidal ideation in the cohort, only one doctor was admitted to hospital. This was an emergency informal admission to a National Health Service (NHS) hospital because of the degree of suicidal risk at presentation. There were eight (6.6%) further referrals to NHS services: psychiatric out-patient clinics (six), an eating disorder service (one), and a therapeutic community (one).

Service user satisfaction

Of the initial 121 doctors, 83.5% (n = 101) returned the questionnaire, 78.5% (n = 95) completed the full questionnaire and 5% (n = 6) completed only part of the questionnaire. In terms of the clinical service provided, doctors were asked to rate their satisfaction-dissatisfaction with the service on a four-point Likert scale on a number of service-specific questions as shown inTable 4.

Table 4

Service user questionnaire outcomes

How satisfied/dissatisfied were you that…	Very satisfied n (%)	Satisfied n (%)	Dissatisfied n (%)	Very dissatisfied n (%)
The sessions were allowing you to talk about your problems	70 (70)	27 (27)	2 (2)	1 (1)
The sessions were helping you understand your problems	51 (50.5)	41 (41.4)	6 (6.1)	1 (1.1)
The sessions were enabling you to cope with your problems	38 (38.8)	49 (50)	9 (9.2)	2 (2)
The sessions were addressing important issues to you	48 (48)	44 (44)	6 (6)	2 (2)
The clinician treated you with respect	83 (82.2)	17 (16.8)	-	1 (1)
Information about you and your problems would be confidential	58 (61)	34 (37.8)	3 (3.1)	-

Furthermore, they were asked for their views in terms of the practical side of the service provision such as transport links and accessibility of the service, comfort of the waiting room, helpfulness of reception staff, etc., and were asked to comment on these matters in open boxes and offer suggestions in terms of ways in which the service could be improved. The main themes that emerged from the open boxes are described inBox 1.

MedNet outcomes

In this naturalistic study, the instruments used demonstrated a significant difference between before and after treatment scores on all questionnaires.

When we examined the results of the suicidal sample, on the measures used there was no significant improvement in terms of risk. However, clinically, in this group, during the time of the study, no suicide attempts or overdoses occurred. The most common clinical dynamic in this group is of doctors who were perfectionists and had set high standards for themselves, with a self-punitive stance if these were not achieved. In psychodynamic terms a possible understanding of the high rate of suicidal ideation is that these doctors have a rejecting or attacking response to their own vulnerabilities, which they see as indicative of failure and weakness. Frequently an associated belief is that evidence of vulnerability or of a need for help is an impediment on their ability to maintain high standards, in particular in their professional lives. In general, these are high-functioning individuals who tend to keep working very well, at the expense of their personal relationships and despite significant mental distress. Often these doctors would not present as globally depressed and have waited a long time to seek help. Our cohort consisted, on the whole, of doctors who had had either none or minimal contact with services prior to being seen at Mednet. We would suggest that the type of broad assessment provided at MedNet, within an analytic framework, helped these doctors to be open about difficulties they had previously not declared or discussed. Being able to express, discuss and understand thoughts felt to be forbidden, including suicidal thoughts, contributed to a reduction of mental distress. In addition, becoming able to have a conceptual understanding of their difficulties gave them a view that they could be helped. It is important to recognise that symptoms - which showed marked change in terms of formal outcomes over the period of the study - may need longer-term psychotherapeutic work if improvement is to be maintained and resilience to further difficulty promoted (as discussed under care pathways).

With regard to the MBI, there was a significant change only for the emotional exhaustion subscale. This is the domain that we would expect to show change first, both in the development of burnout and its improvement, rather than the more entrenched domains of personal achievement and depersonalisation. A paper by Rø et al ^{Reference R⊘, Gude and Aasland11} reported similar findings when researching a group of doctors participating in the Norwegian Villa Sana counselling programme. In this paper, emotional exhaustion was seen as the primary dimension in the subsequent development of burnout. The apparent discrepancy in the Sana doctors between reporting high levels of emotional exhaustion, and at the same time not demonstrating reduced empathy or work capacity was thought to suggest that either they take action in time to prevent burnout or stay at work as ‘hazardous heroes’, even with considerable distress, in order to fulfil their obligations and to maintain their self-esteem. ^{Reference R⊘, Gude and Aasland11}

Engagement with and referral on from the service

In terms of the pathways of treatment and referral, about three-quarters of doctors were seen for up to 6 sessions, a further 15% were seen for up to 20 sessions and 5% were seen for more than 20 sessions.

Two-thirds of doctors who were not suicidal were helped by the treatment provided at MedNet and found it sufficient, whereas the remainder were referred on for additional treatment. For the suicidal subsample, the service was sufficient for about a quarter. For two-thirds the service was helpful in terms of managing the presenting crisis, feeling listened to, and as a stepping-stone in terms of longer treatment that they were advised about and in many cases facilitated in accessing. These results show that the doctors who, according to the CORE-OM and BSI were more depressed and had more significant problems in terms of suicidal ideation, did require more long-term interventions and were therefore more likely to receive longer treatment at Mednet and/or be referred on for further treatment.

Overall, about half of the doctors were helped sufficiently by the treatment that the service provided and of the other half who did need further treatment only one was admitted to hospital. This suggests that the therapeutic relationship had a holding effect and that the service and the approach used managed this special client population effectively.

The MedNet consultation model

The specific patient population of doctors and the type of problems they present calls for a specialised approach when treating their psychological difficulties. In a similar spirit, Harrison argues in a recent paper that the changing role of doctors and new concepts of medical professionalism call for a bespoke model of assessment of doctors’ health and fitness to practise that can take all the relevant aspects into account. ^{Reference Harrison15}

In terms of understanding the complex interplay between work, stress and mental health for doctors, there are three main components to be considered. First, there are the specific components of clinical anxiety and the impact of constantly managing ill patients, as well as coping with the complex organisational dynamics of the working environment. These can generate stress and anxiety that can escalate significantly and may also lead to depression. Second, this brings up questions of the work-life balance that impinge on the personal relationships of doctors, with partners, children and wider family. Third, there is the component of underlying personality, particularly perfectionist traits. A degree of obsessiveness is common among doctors and can help to make them both conscientious and committed but if excessive can become problematic. ^{Reference Wallace, Lemaire and Ghali4} All three components need to be considered when dealing with doctors’ difficulties and problems with their health.

For the large majority of doctors, working as a doctor provides satisfaction and is an integral part of their personal identity and self-esteem. These aspects of work protect and enhance mental health and the capacity to bear the painful realities that practitioners encounter in their work. The opportunity to seek understanding of one's role as a doctor and factors that lead to professionals losing touch with this aspect of their work may also be part of the consultation process. ^{Reference Hochwalder and Begsten Brucefors16}

The MedNet service therefore promotes a narrative-based, holistic approach to doctors’ problems, embracing the different issues they may face in terms of clinical practice, organisational stressors, work-life balance, personal characteristics, intrapersonal conflicts and interpersonal difficulties. In this ethos, the service provides consultations in a broad-ranging way in contrast to a more focused diagnostic-based medical model approach. The process of a series of consultations enables engagement and the gradual evolution of an in-depth understanding of a doctor's presenting difficulties.

In order to provide this type of consultation, we believe that expertise in medical and psychiatric fields as well as psychodynamic knowledge of intrapsychic processes and organisational processes is crucial. Doctors often appreciate seeing someone whom they can expect to understand their working life and pressures from ‘the inside’.

Service user satisfaction

The self-report questionnaires show that service users were happy overall with the approach used, and highlighted the importance of the existence of such a service. The questionnaires confirm that the doctors were hoping to have a greater understanding of what generated their difficulties so that they could feel more in control, and more confident that they could manage difficulties when confronted with them again. Wanting to feel in charge of one's own destiny is a particularly strong characteristic in doctors, which is also reflected in the way they are trained. The lengthy pursuit of medical knowledge reinforces the idea that knowledge confers power, prestige, self-worth and a means to autonomy. ^{Reference Halpert17} Traditionally, doctors have avoided mental health services because of attitudes formed in medical school, a perceived lesser importance of mental health as well as a tendency to treat oneself. ^{Reference Hassan, Ahmed, White and Galbraith18} Accordingly, about a third of the doctors attending MedNet had made their own diagnosis, initiated medical treatment in the form of antidepressants and only presented at the service when this approach was ineffective.

The questionnaires also show that information about the confidentiality of the service is an important area for doctors. This is further supported by a recent survey published by the Royal College of Physicians that shows that only one in five doctors would seek professional advice for mental health issues they experience and that the apparent lack of confidence in the current system protecting doctors’ confidentiality may act as a barrier to seeking help. ^{Reference Hassan, Ahmed, White and Galbraith18}

Limitations

Limitations to our study are that there is no control group in terms of the approach used and service provided at MedNet; response bias; and that data were based on self-report measures only. However, we hope that despite this we have described usefully the experience of ‘doctors in difficulty’ through a specialist service and outlined the underlying principles informing the model provided.

Implications

The observational data show that the mental health of this cohort of doctors has improved. In terms of treatment and referral pathways we found that those who were deemed at some risk of suicide were more likely to have been referred on for further treatment than those who were not at risk, and that only one doctor required admission to hospital. For the more severely unwell doctors, the service was helpful in terms of managing the presenting crisis and feeling heard, and the process was a stepping-stone in terms of longer treatment. Furthermore, we found that the bespoke consultation model provided at MedNet is valued highly by the doctors as service users. This paper highlights the efficacy, need and importance of specialist services for doctors in difficulty.