Intended for healthcare professionals


Recognition of depression and anxiety in primary care

BMJ 1999; 318 doi: (Published 05 June 1999) Cite this as: BMJ 1999;318:1558

Patients' attributional style is important factor

  1. Andrés Herrán, Associate professor. (herran{at},
  2. José Luis Vázquez-Barquero, Head professor,
  3. Graham Dunn, Professor of biomedical statistics.
  1. Clinical and Social Psychiatry Research Unit, Department of Psychiatry, University Hospital “Marqués de Valdecilla,” University of Cantabria, Santander 39008, Spain
  2. School of Epidemiology and Health Sciences, University of Manchester, Stopford Building, Manchester M13 9PT
  3. Belfast City Hospital, Belfast BT9 7AB

    EDITOR—Kessler et al found that doctors detected psychiatric illness in less than half of patients scoring highly on the general health questionnaire (85% of patients with a normalising attributional style and 38% with a psychologising style were not detected).1 These data are in accordance with the work that we did in four Spanish primary care centres. Using the general health questionnaire-28 in the first part of the study and a SCAN interview 2 3 in the second, we found similar figures of non-recognition of psychiatric illness4 and the same relevance of somatisation to lower rates of recognition of mental illness by general practitioners.5

    In her commentary on the paper Heath doubts that scoring highly on the general health questionnaire could be equated with having a treatable disorder. We agree with her that the general health questionnaire is a screening questionnaire, not a diagnostic tool, and that doctors should not talk of depression and anxiety just because patients scored highly on the questionnaire. But Kessler et al's findings are relevant. Questionnaires such as the general health questionnaire provide an approximation of the rate of well defined psychiatric illness in primary care. The diagnosis and treatment of psychiatric illness by general practitioners are usually based on a suspicion about mental illness rather than on diagnostic criteria according to current nosology.

    In an attempt to analyse this issue we repeated the analysis of our data only for those with a definite affective or anxiety disorder. In our published study we collected data on patterns of symptom presentation (“attributional style”—physical, psychologising, and mixed symptoms). Among the (unweighted) 72 patients with these diagnoses as defined in the international classification of diseases, 10th revision (18 depression, 12 dysthymia, 2 cyclothymia, 3 panic disorder, 31 generalised anxiety disorder, 2 obsessive-compulsive disorder, 4 phobic anxiety), only 11 (23%) of the 47 with somatising attributions were recognised by their general practitioner as having affective or anxiety disorder. Equivalent figures were 93% (14/15) of the psychologising attributions group and 90% (9/10) of the “mixed” group.

    We also performed a logistic regression, with recognition of psychiatric caseness by the general practitioner as a dependent variable; symptom attribution and marital status were included as significant variables influencing recognition by the general practitioner. Patients with psychologising and mixed style attributions were more likely to be detected by general practitioners (table), as were married and previously married patients. Thus our results support the relevance of attributional style in patients with well defined depressive and anxiety disorders, for whom effective treatments exist.

    Variables included in logistic regression with dependent variable “general practitioner's recognition of mental illness” for 72 patients with diagnosis of depression or anxiety disorder according to ICD-10 (international classification of diseases, 10th revision)

    View this table:


    General health questionnaire alone is not sufficient for making psychiatric diagnosis

    1. Tim Johnston, Specialist registrar in general adult psychiatry. (tjohnston{at},
    2. Peter Talbot, Specialist registrar in general adult psychiatry.
    1. Clinical and Social Psychiatry Research Unit, Department of Psychiatry, University Hospital “Marqués de Valdecilla,” University of Cantabria, Santander 39008, Spain
    2. School of Epidemiology and Health Sciences, University of Manchester, Stopford Building, Manchester M13 9PT
    3. Belfast City Hospital, Belfast BT9 7AB

      EDITOR—In their study on depression and anxiety in primary care Kessler et al aimed at examining reasons why depression fails to be detected in general practice.1 Several fundamental methodological flaws in the study, however, mean that their conclusions are irrelevant.

      The authors incorrectly use the general health questionnaire as a diagnostic instrument when in fact it can only indicate caseness. A further structured psychiatric interview is always necessary to make a psychiatric diagnosis. This error accounts for the 52% of patients considered by the researchers to have measurable depression and anxiety.

      The general health questionnaire used is able to screen only for psychological disorder in general. The larger 30-item general health questionnaire, when it has been examined by factor analysis, has been found to contain three other factors in addition to anxiety and depression—difficulty in coping, feelings of incompetence, and social dysfunction.2 With the authors' methodology, readers remain ignorant of the actual number of patients with a diagnosis of depression and anxiety. As a result, to compare the authors' rate against attributional style is flawed.

      This study seems to be suggesting that all of life's major problems should be labelled as depression or anxiety. Heath's analysis in her commentary on the paper is pertinent when she refers to “the medicalisation of human distress.”1 This indeed is a contemporaneous trend, which some would have us believe has no end point. As she rightly indicates, however, “normalisers” could be seen as showing a healthy cognitive attributional style in which normal feelings are not medicalised or psychologised. By avoiding such medicalisation the individual is likely to adapt to life's difficulties instead of receiving inappropriate drug treatment or psychotherapy.

      The study's hypothesis therefore remains untested. This is disappointing in view of the importance of the subject matter. The detection of depression and anxiety and depression will be aided by the use of screening instruments but only if they are used within their natural limits. Personally we think that most general practitioners do a difficult job well. Any improvement in detection is likely to come about from research into closer liaison of general practitioners with their psychiatric colleagues. Certainly it will not happen if attempts are made to undermine psychiatric skills with inappropriately used screening instruments.


      View Abstract