For most doctors, it would be inconceivable to manage common long term disorders (such as diabetes or asthma) or serious risk factors (like hypertension or hypercholesterolaemia) without using standard clinical, pathology or other investigative parameters. The data from standard clinical measures have underpinned the drive towards outcomes-based healthcare and have long been recognised as the optimal response to "uninformed patients, skeptical payers, frustrated physicians and besieged health care executives".1

Recent Commonwealth Department of Health and Ageing measures provide incentives for general practitioners to improve the management of patients with complex, chronic or relapsing disorders, including asthma, diabetes and mental disorders.2 Doctors do not need to be convinced of the need to monitor peak flow in asthma or blood glucose (or other metabolic parameters) in diabetes. These simple measures do not try to describe the aetiology of the disease, the breadth of clinical manifestations or the patient's experience of illness. However, as proxy measures of clinical status, they can be used reliably over time and be gauged by different care providers operating in different care settings. Furthermore, the significance of the results can be easily communicated to the patient. Such measures are also used to monitor the response to treatment over time or responses to different types of interventions. Hence, such outcome data, in combination with other clinical and patient factors, are a crucial part of high-quality clinical practice.

By contrast, few patients treated for depressive or anxiety disorders by GPs or psychiatrists will have any systematic measure of their ongoing clinical status or outcome documented.3 This deficit is confined largely to doctors, as clinical psychologists consider repeated systematic measurement an essential clinical tool.4 This deficiency in clinical record keeping by medical professionals persists despite the availability of valid and reliable outcome measures.

Depression and anxiety are relapsing or chronic conditions5 and are often comorbid with other medical conditions in primary care. Patients will be managed mainly by primary care providers6 and the time frame for such care is years (not weeks or months). Consequently, we propose that all medical providers need to incorporate standard outcome measures for depression or anxiety into the management plans of patients with depression or anxiety. The other, clinical reasons why such a practice should be pursued are detailed in Box 1.

Although not the prime concern for the treating clinician, health services planning and other research benefits rely on the collation of clinical data in combination with other practitioner and organisational information. The evidence base for high-quality mental health practice in primary care is limited and urgently needs relevant longitudinal data.9 These non-clinical outcomes (Box 2) should also be of concern to the wider medical profession.

Although there are no specific laboratory-based diagnostic tests for depression or anxiety, clinical diagnoses can be made reliable by determining whether a sufficient number and type of clinical symptoms are present. The number of diagnostic categories that are relevant to primary care is limited and easily incorporated within standard checklist formats.9 There are specialised neuroimaging and neuropsychology strategies that may emerge as useful clinical markers of risk factors for illness, severity of illness, or predictors of course or response to treatment; however, none are yet suitable for application in everyday clinical practice.

Consequently, our current focus is more on the use of appropriate outcome measures rather than diagnostic measures. We now have a wide range of instruments that rate relevant psychological symptoms or resultant disability. For people with depression or anxiety, it is necessary to recommend the measurement of symptoms and disability, because some people minimise their symptoms simply by avoiding stressful situations. For example, a housebound person with agoraphobia, or a socially withdrawn person with chronic depression, may show only mild elevations on current symptom measures but reveal severe disability on other measures. Similarly, while people with comorbid physical and mental disorders can show elevated symptom scores, inspection of the mental and physical scores on disability measures can be helpful in prioritising treatment. Importantly, the common general disability measures (eg, Short Form Health Survey [SF-12]14 and Brief Disability Questionnaire15) can be used across the broad spectrum of general medical practice.

To encourage widespread application in general medical settings, measures need to have a number of key characteristics. These include brevity, low cost (preferably being in the public domain), self-report rather than clinician-administered format, results that are readily communicated to the patient, and demonstrated responsiveness to change in clinical status. Further, symptom-based instruments should be focused on the common symptoms of depression, anxiety or related somatic constructs.9 Such instruments also lend themselves to automation and electronic record keeping. While illness-specific measures are commonly used in specialist practice (eg, for major depression, panic disorder, social anxiety or obsessive compulsive disorder only), they are far less attractive in general medical settings.

Given these considerations, we recommend a small series of instruments for use in general medical settings. For the assessment of common symptoms of depression and anxiety, we recommend the Kessler Psychological Distress Scale (K10),16 the Somatic and Psychological HEalth REport (SPHERE),8 or the Hospital Anxiety and Depression Scale (HADS).17 The K10 is a 10-item instrument that was developed internationally and has been used in the Australian National Survey of Mental Health and Wellbeing. Scores range from 10 to 50, with scores above 30 being highly predictive of a depressive or anxiety disorder. The sensitivity and specificity of the K10 (ie, its ability to predict diagnosis) is superior to that of the 30-item General Health Questionnaire (GHQ).18 In the national mental health survey, it correlated 0.5 with the GHQ, –0.6 with the SF-12 disability measure, and 0.3 with the number of consultations for a mental health problem in the previous year.19 It measures the relevant constructs, is simple to use and requires little expertise to score and interpret the results.

The SPHERE has a 34-item version that was developed to rate the psychological and somatic symptoms reported by people with common mental disorders in primary and other medical care settings. The utility of the shorter 12-item version for use in primary care was demonstrated as part of SPHERE: a national depression project and was based on more than 46 000 consultations in Australian general practice.8 The instrument is best used to differentiate two levels (and three types) of common mental disorders (patients reporting both characteristic psychological and somatic symptoms [Level 1, Type 1], and patients reporting either psychological symptoms [Level 2, Type 2] or somatic symptoms [Level 2, Type 3]). These levels predict not only formal psychiatric diagnoses, but also disability and doctor- and patient-perceived need for mental healthcare.8

The HADS is a 14-item instrument that was developed to rate the severity of specific depressive and anxiety symptoms in patients with comorbid medical disorders. Hence, it avoids rating those common somatic symptoms that accompany many mental disorders. While this narrower focus is advantageous for assessing changes in depressive and anxiety symptoms, it may mean that it is less useful in patients with the more mixed psychological states that are commonly seen in primary care.

For rating disability, we recommend the SF-12.14 The 36-item Medical Outcomes Study Short Form Health Survey (SF-36) was developed as a measure of disability or functioning and is applicable to people with any illness, physical or mental. It is now available as a 12-item version (SF-12) that correlates 0.95 with the results of the parent SF-36 instrument. It has three sections: a preliminary question about self-perceived health status; eight questions about the extent to which current health status limits activities; and three questions about feelings. Two scores are generated: a mental and a physical component score. The scoring method is complex and is arranged so that scores on the physical and mental components are independent. The SF-12 has excellent psychometric characteristics, is sensitive to change, and is probably the de facto world standard for measuring outcome of treatment for both physical and mental disorders.

For physicians who work predominantly in academic, specialist or administrative settings, the arguments for routine outcome measurement are obvious.3,20 In association with the move to manage the common psychological disorders like depression and anxiety more effectively in primary care, there is now an urgent need to promote the clinical utility of standard outcome measures to GPs. The historical and professional resistance to the use of such measures has hampered the delivery of standardised and effective treatment in primary care settings. As new treatments are developed and as governments move to support major service innovations in primary mental healthcare, we need to measure routinely whether "improved" treatments and services are actually changing the lives of the individual patients who present for treatment.