Abstract
Objectives:
The objective of this article is to explore diagnostic decision making around psychological symptoms presenting to general practitioners (GPs) and psychiatrists, identify attitudinal and personality factors of possible relevance in these decisions, and compare GPs and psychiatrists to help identify potential educational targets.
Methods:
GPs and psychiatrists attended separate peer-facilitated workshops in which two case presentations were discussed. Decision making was explored by structured questions embedded in the workshop, with responses recorded by electronic keypad technology. Participants completed demographic questionnaires and measures of personality and attitudes to depression.
Results:
GPs and psychiatrists accorded emphasis to different elements of the history, and assigned different diagnoses based on the same set of symptoms. Both groups relied on non-pharmacological management for milder psychological symptoms; GPs were less likely to make a diagnosis of bipolar disorder. Traits of Extraversion and Agreeableness were associated with greater ease in treating depression.
Conclusions:
Differences in diagnostic decision making likely reflect the different contexts of specialist and generalist practice. Educational targets may include information about key symptoms to assist in diagnostic precision, but further information is needed to determine the best match between diagnostic processes, context and outcome. An awareness of the role of personality factors may help when designing education and support programs.
General practitioners (GPs) have been criticised for under-recognition of mental disorders and deficiencies in management strategies, but much of the relevant research has been conducted with specialist care in mind. 1 It is important to establish the nature of mental health care by GPs and how it differs from psychiatrists, because a large proportion of individuals experiencing anxiety and depression seek help from their GP rather than a psychiatrist or other mental health specialist. 2 Educational programs that seek to improve the skills of GPs in the diagnosis and management of mental disorders must be developed on a platform of awareness of current practice and the nature of the general practice setting. Therefore, it is of interest to establish how GPs and psychiatrists differ in diagnostic processes and management as well as their attitudes toward mental health, especially patients with depression and anxiety since they are highly prevalent. 2
Previous research suggests that GPs and psychiatrists assign different diagnoses based on the same symptoms, 3 and their decision-making processes differ. Experienced psychiatrists use experience and intuition as a shortcut in decision making, 4 usually guided by medical and biopsychosocial models. On the other hand, patients with mental health problems usually present to GPs with physical symptoms. GPs are aware of the possibility of underlying mental disorder but prioritise the exclusion of significant physical disorder.5–13 They are also conscious of stigma associated with psychiatric diagnoses, concerned that such a diagnosis will be unacceptable to their patients, and hence use exploration of the presenting physical symptoms to enhance patient engagement. 13 Additionally, the validity and usefulness of the major psychiatric classificatory systems in primary care is seen as poor by GPs, 14 who place more emphasis on the personal context of the patient’s presentation. 15
Personality style and attitudes toward mental health help explain physicians’ approach, interpersonal style, likeability, trust and acceptance. Personality attributes are often explored in the context of the five-factor model of McCrae and Costa.16,17 Previous research has shown that medical practitioners rate significantly higher on Extraversion, Openness and Conscientiousness than a normative community sample. 18 Additionally, experienced doctors rated higher on Agreeableness and lower on Neuroticism than medical residents. 19 Agreeableness is associated with good interpersonal skills, trust, cooperation and altruism. Openness is associated with curiosity and imaginativeness. Previous research suggests that doctors with high levels of Openness and average, as opposed to very high, levels of Conscientiousness are more likely to be trusted by patients. 20
The Depression Attitude Questionnaire (DAQ) 21 was developed to measure GPs’ attitudes to depressive illness, but has also been used with other medical personnel.22,23 Previous research shows that GPs rate themselves as less comfortable in dealing with depression and find it more difficult to distinguish depression from unhappiness compared with psychiatrists, 23 although the latter may be due to patients in general practice being usually less unwell, with shorter illnesses of lower severity than patients in specialist practice,24,25 and that by the time the patient reaches a psychiatrist, a psychological focus has been identified by the referring GP.
This study aimed to identify and compare the key factors influencing diagnostic decision making in general and specialist practice. It was hypothesised that psychiatrists would be more likely to make specific diagnoses earlier in their professional contact with a patient, that they would be influenced by different aspects of the patient history than GPs, and that they would be highly likely to use the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) or the International Classification of Diseases, Tenth Revision (ICD-10) criteria as the basis of their diagnoses.
Method
GPs and psychiatrists in the Sydney Metropolitan area were recruited to participate in workshops in which a peer facilitated two case discussions. Ethics committee approval was obtained for assistance from the field force of a pharmaceutical company when other recruitment efforts failed. Additionally, psychiatrists attending professional development activities at a public hospital were recruited; they participated in the workshop discussions and answered the electronic ‘KEEPAD’ questions, but most did not complete questionnaires.
Separate workshops were run for GPs and psychiatrists using the same cases, which reflected common presentations in general practice. In Case 1, a young mother presented with symptoms of mixed anxiety and depression that would not meet DSM-IV or ICD-10 criteria for disorder, but could have been classified under a number of International Classification in Primary Care version 2 (ICPC-2) codes. In Case 2, a young woman presented with initial symptoms of depression, later returning with symptoms of hypomania that would have met criteria for bipolar disorder. Throughout the cases, scripted questions were posed that explored decision-making processes. Answers were collected and stored using a KEEPAD system. A peer-facilitated discussion around these questions was recorded. Qualitative data from the GP workshops have been reported previously. 13
Prior to the workshop, participants completed the following questionnaires:
Demographic questionnaire
This included age, gender, full-time or part-time work, number of years in practice and details of postgraduate training.
DAQ 26
Four principal components have been identified: 1. Treatment attitudes (higher scores indicating stronger belief in a biochemical basis of severe depression); 2. Professional ease (higher scores indicating discomfort in working with depressed patients); 3. Depression malleability (higher scores indicating stronger belief that depression is not amenable to change); and 4. Depression identification (higher scores indicating stronger belief that depression is difficult to distinguish from normal unhappiness).
NEO-FFI form S 17
This provides a measure of personality traits according to the five-factor model, giving scores for Neuroticism, Extraversion, Openness, Agreeableness and Conscientiousness.
Data were analysed using SPSS Version 19. The chi square test of independence was used on all KEEPAD variables that were nominal in nature. Independent samples t test was used to compare groups on continuous variables, and Pearson correlation was used to explore the relationship between continuous variables.
Results
Forty-four GPs (M age = 53.6 years) and 37 psychiatrists (M age = 49.8 years) participated in the study. Most GPs were female (n = 29, 66%), employed full time (n = 27, 63%) and had worked as a GP for more than 15 years (n = 35, 79%). For the psychiatry sample, KEEPAD data were available for all participants, but questionnaire data were obtained from only 19; of those, the majority were male (n = 11, 58%), employed full time (n = 15, 78.9%) and about half had worked as a psychiatrist for more than 15 years (n = nine, 47%). No significant gender differences in questionnaire responses were identified between GPs and psychiatrists.
Diagnostic decision-making processes
Data are reported from all participants. Psychiatrists (n = 37) indicated that they rely mostly on clinical experience and medical notes when diagnosing a patient (Figure 1). The majority of psychiatrists (n = 27, 73%) stated that it is important for them to establish a diagnosis after a patient’s first visit. Sixty-five per cent (n = 24) of the psychiatrists indicated that they change their initial diagnosis less than 25% of the time. Reasons for changing include further observation of the patient’s mental state (n = 11), further history from the patient (n = 10) or collateral history from family/carer (n = 11), or following the patient’s response to treatment (n = 5).
Factors identified as important to psychiatrists when diagnosing a patient, reported as proportions.
Case 1
Psychiatrists specifically diagnosed depression more often (34% vs 10% of GPs), and less often diagnosed mixed anxiety and depression (29% vs 54% of GPs). However, similar proportions of psychiatrists and GPs assigned a diagnosis of anxiety, depression or the combination (77% of psychiatrists vs 82% of GPs). A greater proportion of psychiatrists (20%) compared to GPs (13%) would not have assigned any diagnosis. There is a significant relationship between diagnosis and specialty, χ2 = 15.03, df = 6, p < 0.05.
Both groups preferred non-pharmacological management strategies, with GPs relying more on advice or education (44% vs 31% of psychiatrists), and more psychiatrists electing to provide counselling or psychotherapy themselves (36% vs 18% of GPs). There was no significant association between medical management strategies and specialty, χ2 = .97, df = 5, p = .97, or non-pharmacological management and specialty, χ2 = 8.64, df = 4, p < 0.07.
Case 2
More than half of the psychiatrists initially diagnosed the patient with major depression, compared to only a quarter of GPs (Table 1). Most GPs diagnosed the patient with mixed anxiety and depression. However, after the patient returned with symptoms of hypomania, three-quarters of the psychiatrists but only a quarter of GPs diagnosed bipolar disorder. The majority of GPs gave other diagnoses, including ICPC-2 diagnoses of anxiety (13%), situational stress (18%) and anger/irritability (24%).
Diagnosis of patient in Case 2
GP: general practitioner.
Despite these differences, management was similar, and approximately half of each group indicated they would choose monotherapy with a mood stabilizer (50% of GPs, 42% of psychiatrists). An antipsychotic was chosen by 5% of GPs and 14% of psychiatrists. A combination of medications was chosen by 11% of GPs and 27% of psychiatrists. In each group, three persons (8% of psychiatrists, 7% of GPs) reported they would not use medication.
Personality traits
Psychiatrists had significantly higher Openness, t(56) = 2.02, p < 0.05, and lower Conscientiousness scores, t(56) = −2.40, p < 0.05, than the GPs. When compared with normative values for a community sample, 18 both groups were high on Openness, the GPs were high on Extraversion and Agreeableness, and the psychiatrists were low on Conscientiousness (Table 2).
NEO-FFI means (standard deviations) and comparison to community norms
p < 0.05.
DAQ
Psychiatrists reported greater satisfaction working with depressed patients than GPs, t(60) = −2.14, p < 0.05, were more positive toward the illness and believed depression to be more amenable to change, t(60) = 3.81, p < 0.001 (Table 3). GPs and psychiatrists did not differ significantly in the strength of their beliefs that they could identify depression, or in their beliefs in a biochemical basis for depression.
Means (standard deviations) for the four components of the Depression Attitude Questionnaire (DAQ)
GP: general practitioner; ap < 0.05.
Intercorrelations were examined using the five personality factors and the four components of the DAQ for GPs and psychiatrists separately, shown in Tables 4 and 5, respectively. GPs who scored highly on Neuroticism were more likely to believe that depression is not amenable to change. Both GPs and psychiatrists who scored highly on Extraversion reported finding treating patients with depression more satisfying and believed that depression can be treated. Additionally, psychiatrists who scored highly on Agreeableness were more likely to report finding identifying depression easier. Openness and Conscientiousness did not correlate with any of the variables on the DAQ.
General practitioner (GP) sample: relationship between personality traits and attitudes toward depression
p < 0.05; bp < 0.01.
Psychiatrist sample: relationship between personality traits and attitudes toward depression
p < 0.05.
Discussion
The findings from our novel study suggest that psychiatrists and GPs diagnose mental ill health differently, but use similar management strategies. Both appear to rely on heuristics, but GPs weigh family history and personal context most strongly; 13 psychiatrists’ weightings perhaps reflect the more cross-sectional nature of the assessment setting. The tendency of GPs toward non-specific diagnoses may reflect difficulty teasing apart depressive and anxiety symptoms, a view that non-specific diagnoses are less stigmatising, or a desire to avoid early diagnostic closure. It is more difficult to discriminate a case from a non-case in general practice when patients present with more somatic and less overt psychological symptoms. 27 However, since a high proportion of individuals experiencing anxiety and depression seek help from their GP rather than a psychiatrist or other mental health specialist, 2 it is important for GPs to be able to differentiate between anxiety and depression when necessary. GPs prefer to come to their diagnoses over time and repeated assessment, 13 and so the cross-sectional nature of this assessment may have underestimated GPs’ diagnostic skills.
The apparent greater emphasis of psychiatrists on an early and specific diagnosis may reflect situational contingencies such as referrer requests or perceived expectations. The potential stigma of a psychiatric diagnosis may carry less weight for psychiatrists in influencing diagnostic decision making, perhaps since patients have already been oriented to a psychiatric diagnosis by their GP. Training may also be relevant, for example, for the emphasis on using diagnostic classificatory systems and making specific diagnoses. GPs have compellingly described their reasons for taking time to make a specific diagnosis, including that symptoms often spontaneously resolve, a concern about stigma and whether patients will be receptive to a psychiatric diagnosis and engage with treatment, and priority given to excluding a medical illness.7,13,28
The importance placed on diagnosis and better developed illness scripts for mental illness amongst the psychiatrists may have contributed to the rapid change in preferred diagnosis in Case 2 as they recognized the pattern of a depressed patient switching to a manic or hypomanic state. 29 The lower rate of diagnosis of bipolar disorder among GPs may result from a lack of confidence by GPs in making this diagnosis, perhaps related to its low prevalence in primary care, or it may simply be a reflection of the greater experience of bipolar disorder by psychiatrists, enabling faster recognition of key features, 30 and could be interpreted as suggesting a need for educational programs addressing these issues. Alternatively, this may reflect ongoing reluctance to ‘label’ a patient, especially with an even more stigmatising diagnosis, or a view that bipolar disorder is a difficult diagnosis best made by a specialist.
Although psychiatrists and GPs differed with respect to diagnostic specificity, the management strategies of these two groups were very similar. When a patient presented with more serious symptoms, psychiatrists as well as GPs appeared to rely on pharmacotherapy, regardless of the diagnosis made. This could be problematic if the diagnosis made is inaccurate, being potentially discouraging (and unsafe) for the patient and leading to poor treatment adherence and follow-up. 31 However, the current findings suggest that a reluctance to make a specific diagnosis in general practice may not preclude identification of appropriate pharmacotherapy, as occurred with Case 2, consistent with findings elsewhere that context may have more impact on diagnosis than on therapy. 32 Findings are also consistent with the lack of diagnostic specificity of psychopharmacology in general, and could suggest that specific diagnoses may not be essential to good management.
The finding that psychiatrists were higher on Openness than both the GP sample and a normative sample is consistent with a recent study reporting that medical students who considered themselves highly likely to specialise in psychiatry were high on Openness. 33 Openness is characterised by traits of having broad interests, preferring variety, and being independent and liberal. 16 It could be argued that liberal-mindedness might be a particular asset for a psychiatrist, who must establish a therapeutic alliance with patients who may report unusual thoughts and behaviours. Being independent in outlook might be an advantage when one is drawn to a specialty that does not enjoy a high level of prestige within the medical profession.33–35 Previous research suggests that patients are more trusting of physicians who are high on Openness and lower on Conscientiousness. 20 Lower Conscientiousness scores (as found in the psychiatrists in our study) could indicate that an individual is less rule-bound or rigid, which could assist in making an individual more open to alternative views (e.g. to change diagnosis when clinically indicated), but conscientiousness is also associated with being organised and persistent, and so a lower score on this trait is potentially unhelpful, for example in complex cases or with difficult historians. The GPs in our study were in the average range on Conscientiousness but have previously been reported to have high scores. 36 It was interesting that Openness and Conscientiousness did not correlate with any of the variables on the DAQ. Rather, across both groups, it appears that high Extraversion and Agreeableness traits are most beneficial for working with depressed patients. Extraversion may be helpful in dealing with mental illness because physicians characterised by this trait are more sociable and optimistic. 17 Extraversion and Agreeableness may assist to differentiate depression from unhappiness because the qualities of sociableness, cooperation and optimism allow a detailed exploration of a person’s unhappy emotional state, without the physicians feeling emotionally overwhelmed themselves.
Personality likely makes only a small contribution to differences in diagnostic processes, with the context of practice probably most important.
The study has several limitations. Firstly, only a small number of psychiatrists completed the questionnaire measures. Secondly, all participants were volunteers and may not be representative of their professional groups more broadly. As volunteers, it is possible that the group as a whole could be higher on measures of Agreeableness. Additionally, the GPs who volunteered for this study may have a greater interest in mental health issues, which could tend to minimise differences between the groups. Finally, case studies are only a proxy for studying what actually happens in practice, leading to limitations to how the data can be applied.
Conclusions
Building on our previously reported data, the current study makes an important contribution to understanding diagnostic processes in general practice and specialist psychiatric practice by exploring clinicians’ reactions to the same clinical cases, considering also personality and attitudes. Whilst it is likely that some of the identified differences have been influenced by context (for example, the prominent presentation of physical illness in general practice), and may offer context-specific advantages, the possibility that changes in approach may result in better treatment and outcome should be explored. In particular, the role of early vs delayed diagnosis, specific vs non-specific diagnoses, and weight given to various factors in the patient history could be studied. Further data in the context of diagnostic decision making may be helpful in determining how best to support all practitioners in optimising their approach to diagnosis and management in their practice setting. An understanding of the relationship between personality variables and ease in treating patients with mood disorders may add a valuable dimension to medical education and support programs.
Footnotes
Funding
This work was supported by the McGeorge Bequest, Department of Psychiatry, University of Sydney. Following independent planning of the project and development of the workshops by the authors, ethics approval was obtained for pharmaceutical company sponsorship in implementing the workshops. Lundbeck Australia provided assistance in recruiting participants, finding venues and paying for refreshments for the participants. Lundbeck Australia had no involvement in the study design, data collection, analysis, or interpretation. Lundbeck Australia had no involvement in the preparation of this manuscript or the decision to submit it for publication.
Disclosure
The authors report no conflict of interest. The authors alone are responsible for the content and writing of the paper.