Diagnosing depression in primary care
How should depression be diagnosed?
What is the role of screening tools?
Which patients should be referred?
How should depression be diagnosed?
What is the role of screening tools?
Which patients should be referred?
Increasing the recognition of depression and improving the accuracy with which it is diagnosed are two important strategies for improving outcomes in depressed patients which have been included within the Quality and Outcomes Framework (QOF).
Evidence from observational studies suggests that depression has a poor prognosis. The WHO multicentre naturalistic study found that 60% of depressed patients treated with antidepressants and almost half of those with unrecognised depression still had a diagnosis of mental illness at one-year follow-up.1 More recent studies have reported that between one- and two-thirds of primary care patients had not fully recovered from major depression 12 months later.2
This is not so surprising, given that only a small minority of patients receive effective treatment. Approximately 40% of patients do not seek medical help,3 and of those who do, between 30 and 50% are not recognised as being depressed,4 usually as a result of somatic presentations. Patients may not be offered the appropriate treatment if strict diagnostic criteria are not applied, and under half of those receiving treatment will complete a minimal treatment course (? 2 months' antidepressants or 4 psychotherapy sessions).5
Hence, the principal strategies for improving outcomes in depressed patients are:
• Screening for depression in high-risk patients
• Increased awareness of somatic presentations of depression
• Use of a validated measure of depression severity to improve targeting of treatment
• Use of a care management programme to improve adherence to treatment.
Most patients with depression complain of physical symptoms, such as fatigue, headache, back pain or palpitations, rather than depressed mood. An international study found that, overall, 69% of depressed patients reported only somatic symptoms, the figure for the UK was 60%.6
It is also recognised that there is a significant overlap between functional somatic syndromes and depression. Fifty per cent of depressed patients were found to have at least three unexplained somatic symptoms,6 and patients with chronic functional somatic symptoms had high levels of psychiatric morbidity when compared with controls (OR = 2.4).7
Recognition of depression can be particularly difficult in children, elderly patients and patients with chronic disease. It is not known how many children with headaches, abdominal pain or fatigue have unrecognised depression, and they may also present with non-specific behavioural problems such as irritability and aggression.3 Elderly patients often deny feeling depressed and may present with agitation, sleep and/or appetite disturbance, loss of energy and other somatic complaints. In patients with chronic disease, it can be difficult to decide whether somatic symptoms are caused by depression or the medical condition itself.8
Diagnostic criteria and severity measures
There is a continuum of severity from normal sadness to severe depression, so that any distinction between significant depression and subthreshold symptoms is to some extent arbitrary. Nevertheless, there is evidence that the threshold of diagnosis of DSM-IV major depression (see table 1, attached) marks a clinically important boundary above which antidepressants become more effective than placebo. We therefore need to apply strict diagnostic criteria to ensure effective targeting of treatment to those patients who are likely to benefit. DSM-IV is recommended in preference to ICD-10 as it is a better guide to the treatment threshold.4
The British Association of Psychotherapy (BAP) guidelines define four grades of severity, based on the number of symptoms and the degree of functional impairment.4
In contrast, the NICE guideline3 is based on ICD-10 criteria which categorise depression into mild, moderate or severe according to the number and severity of symptoms. Mild depression (4 symptoms) falls entirely within the BAP category of subthreshold depression. NICE recommends a number of treatment options in addition to structured follow-up, including exercise and brief psychological interventions. The definition of moderate depression (5-6 symptoms) includes some with subthreshold depression. Not all patients within this category, therefore, will gain any more than placebo benefit from the NICE recommendation of antidepressant treatment. For patients with severe depression (7 or more symptoms), combined treatment with antidepressants and CBT is recommended
The PHQ-9 is based on the DSM-IV criteria and is therefore a useful guide to severity. It is not a gold standard diagnostic tool, however, and is not a substitute for clinical judgement. Moreover, it counts symptoms if they are present on more than half the days whereas the definition of major depression requires them to be present on nearly every day. A validation study found that more than two-thirds of the patients with major depression had a score of ?15,9 and it is now recognised that a cut-off score of ? 10 is too low10 and may encourage, rather than discourage, excessive prescribing.
GPs vary considerably in their ability to recognise depressive illness, but educational programmes do not appear to improve their performance.4 NICE has recommended the alternative strategy of case-finding through the use of screening questionnaires in high-risk patients.3 Approximately 20% of patients become depressed following a myocardial infarction and the prevalence of depression in patients with diabetes is twice that of the general population.11 Depression increases the mortality rate in both CHD and diabetes, and this can be mostly explained by poor health behaviour, especially physical inactivity.10,12
There is no good evidence, however, that case-finding is effective. Observational studies suggest that patients whose depression is recognised fare no better than those whose depression is unrecognised.2 A systematic review found that routine screening and feedback of scores did not increase the recognition of depression. Selective feedback of high scores did improve recognition but did not increase the intervention rate.13
There are three major difficulties. First, the positive predictive value (PPV) of a positive screen is too low to be of any clinical value. The point prevalence for major depression in adults aged 18-65 is 2.1% in the UK.10 In patients with diabetes, this will be increased to 4.2%. This gives a PPV for the two-question screen (see box 1, above) and HADS of only 0.11 and 0.22 respectively.4 Second, the case yield is disproportionately low compared with the time expended carrying out the screening programme. Prior to the introduction of the QOF, we screened a sample of our CHD patients using the HADS questionnaire. Of 93 patients screened, only four were identified as having unrecognised depression. Third, the evidence from the US Preventive Services Task Force suggests that screening is only beneficial if it is linked to a managed care programme.14
Tackling the QOF depression indicators
The main difficulty with indicator 1 (see table 3, attached) is the low PPV of the two-question screen. In our practice we have therefore set up a three-step screening process. Those patients who answer ‘yes' to either of the two questions are asked to complete a HADS questionnaire. Patients with a HADS depression score ?8 are interviewed by our practice nurse, who has experience of providing shared care for depressed patients and therefore has the necessary skills to filter out some of the false positives. If she feels the patient may be depressed, she arranges a GP appointment.
An audit of our screening programme in the first year of the QOF indicated that, although the three-step screening process appeared to work well (31% of those identified by the two-question screen were given an appointment with their GP), the intervention rate for those patients who were referred on was very disappointing (23%). We are therefore considering whether patients should be referred to the mental health lead rather than their usual GP, as part of a care management programme.
The main purpose of indicator 2 is to improve the targeting of antidepressant treatment. It is important that the questionnaire score should inform the decision whether or not to prescribe and I spread the diagnostic process over two consultations to facilitate this.
The latest QOF guidance recommends that we should consider intervention if the PHQ-9 score is ?12.10 Diagnosis of depression is complex and incorporates assessment of cognition, functional impairment, previous history, duration and illness trajectory as well as the number and severity of symptoms. It is a mistake, as in the ICSI guideline, to impose rigid treatment guidelines based solely on the PHQ-9 score.15 If the score is ? 15, it is likely that the clinical suspicion of major depression is correct, but a score between 10 and 14 is borderline, and we need to think carefully before initiating antidepressant treatment.
Patients should be referred to psychiatric services if any of the following criteria are met:
• high suicide risk
• psychotic major depression
• major depression in bipolar affective disorder
• two or more attempts to treat a patient with medication have failed, or resulted in an insufficient response.3, 4
Assessing suicide risk
It is essential to make and document an assessment of suicide risk for every patient with depression. This should incorporate a review of:
• long-term risk factors (e.g. family history of suicidal acts)
• short-term risk factors (e.g. lack of social support)
• precipitating factors (e.g. recent loss)
• protective factors (e.g. responsibility for children).16
It is important to bear in mind that risk factors differ in importance between men and women.17 We should always ask patients directly about suicidal ideas and intent.3
There appears to be little evidence that screening for depression leads to improved outcomes.4 Furthermore, depressed CHD patients may benefit more from exercise programmes than from antidepressants.12 We need to increase the effectiveness with which we manage patients with recognised depression. Improved targeting of treatment through the use of severity measures will help, but we are most likely to improve outcomes by using case management programmes to identify and address problems with adherence or response to treatment.18 It is a welcome move that GPs will now be rewarded for actively following up their depressed patients.10Table 1 Table 3 Box 1 Key points Author
Dr Phillip Bland
BA BM BCh MRCGP
GP with an interest in mental health, Dalton-in-Furness