Elsevier

General Hospital Psychiatry

Volume 21, Issue 2, March–April 1999, Pages 87-96
General Hospital Psychiatry

Pshychiatry and Primary Care
Diagnosis of depression by primary care physicians versus a structured diagnostic interview: Understanding discordance

https://doi.org/10.1016/S0163-8343(98)00077-2Get rights and content

Abstract

In this paper, false-negative and false-positive cases of depressive illness are examined, differentiating levels of disagreement between a primary care physician’s diagnosis and a standardized research diagnosis. Two stratified random samples of primary care patients in Seattle, USA (N = 373) and Groningen, The Netherlands (N = 340) were examined with the Composite International Diagnostic Interview-Primary Health Care Version (CIDI-PHC). Physician’s severity ratings and diagnosis of psychological disorder were obtained. Three levels of disagreement between physician and CIDI diagnosis were distinguished: 1) complete disagreement about the presence of psychiatric symptoms (true false-negative and true false-positive patients); 2) disagreement over severity of recognized psychological illness (underestimated or overestimated); and 3) disagreement over the specific psychiatric diagnosis among those given a diagnosis (misdiagnosed or given another CIDI diagnosis). All three levels of disagreement were common. Only 27% of the false-negative cases were due to complete disagreement (true false-negatives), and 55% of the false-positives were due to complete disagreement (true false-positives). The true false-negative patients were younger, more often employed, rated their own health more favorably, visited their doctor for a somatic complaint and made fewer visits than the underestimated, misdiagnosed, and concordant positive patients. Complete disagreement in depressive diagnoses between the primary care physician and research interview is not as frequent as indicated by an undifferentiated false-negative/false-positive analysis. Differentiating levels of disagreement does more justice to diagnostic practice in primary care and provides guidance on how to improve the diagnostic accuracy of primary care physicians.

Introduction

Guidelines have been recently promulgated in the USA, the Netherlands, and in other countries to improve quality of care for depressive illness 1, 2, 3. There is growing evidence that management according to these guidelines improves clinical outcomes among primary care patients 4, 5, 6, 7, 8, 9, 10. Accurate diagnostic evaluation of depressive illness is a prerequisite for implementation of these evidence-based treatment guidelines.

Many studies of the accuracy of diagnostic assessment in primary care have focused on recognition of undifferentiated psychological illness rather than the accuracy of the diagnosis of the depressive syndrome 11, 12, 13, 14, 15, 16. Recognition of undifferentiated psychological illness is a first step toward accurate diagnosis, but it lacks sufficient specificity to guide treatment. The accuracy of primary care physicians’ diagnosis of current depressive illness is examined in this paper.

Typically, the accuracy of primary care physicians’ diagnosis, in relation to the gold standard of a research diagnosis, is assessed in terms of false or true negatives and positives. Two complicating factors need to be considered. First, patients with recent onset of depressive symptoms or patients who are recovering may not meet criteria for a depressive disorder when examined with a research diagnostic interview. Their physician may identify them as relapsing or recovering from a depressive episode. Second, there is a high rate of comorbidity of depressive disorder with other psychiatric disorders 17, 18. When primary care physicians do not diagnosis a depressive disorder, they may identify a co-morbid psychiatric illness such as panic disorder, generalized anxiety disorder, or alcohol abuse. This situation needs to be differentiated from those in which no psychological illness is diagnosed at all. Differentiating levels of disagreement may help us understand better why physicians’ diagnoses and research diagnoses disagree.

The data we report were collected in Seattle, USA, and Groningen, The Netherlands, as part of the multicenter collaborative study on Psychological Problems in General Health Care of the World Health Organization [19]. This paper is the result of collaborative efforts of these two centers to develop a differentiated approach for comparing diagnoses by primary care physicians and a standardized research interview. Three levels of disagreement are identified: 1) complete disagreement about the presence of psychiatric symptoms (called true false-negatives and true false-positives); 2) disagreement about the severity of the psychiatric problems (cases with underestimated severity and overestimated severity); and 3) disagreement about what psychiatric diagnosis to assign (cases who were misdiagnosed and cases given another CIDI diagnosis).

We then compared these groups in terms of factors that prior research has found to be associated with recognition 11, 20, 21, 22, 23, 24 (symptom severity, psychiatric history, disability, patient’s health perception, reason for encounter, duration since the last visit, and demographic characteristics). The aim of these analyses is to better understand the reasons for diagnostic disagreement and its significance for clinical practice.

Section snippets

Setting

In Seattle, the study subjects were enrolled from three primary care centers (50 physicians) of Group Health Cooperative of Puget Sound (GHC), a staff model health maintenance organization. Study clinics were selected to represent the range of income and education in GHC’s Seattle area population.

In Groningen, the study was carried out in six primary care practices (11 physicians). The selected practices were typical of the Dutch primary care system and included one solo (1 physician) and five

Study samples

In Seattle, 1962 (93%) of 2110 patients who were asked to fill out the GHQ-12 completed this questionnaire. Among the 608 patients sampled for second-stage baseline assessment, 373 (61%) completed the interview. In Groningen, 1271 (96%) of the 1320 patients approached completed the GHQ-12 and 340 (69%) of the 493 sampled patients completed the second-stage baseline assessment. In Seattle as well as in Groningen, second-stage respondents did not differ from nonrespondents with respect to sex and

Discussion

In this paper we differentiated reasons for disagreement between the primary care physician’s diagnosis of depressive illness and a research psychiatric interview diagnosis of depression. We found that complete disagreement about the presence of psychiatric symptoms contributed somewhat more than one-third of the discordance. Disagreement about severity or specific diagnosis contributed a little less than one-third each.

Different diagnostic practices of the primary care physicians in the two

Acknowledgements

This study was financially supported by grant MH47765 from the National Institute of Mental Health, Bethesda, Md (Seattle), Delagrange Laboratories and Synthélabo Pharmacie, Paris, France, Grants 900-571-036 and 940-20-802 from the Dutch Organization for Scientific Research, Medical Sciences, KWAZO-program, and by SGO, the Promotion Program Health Research (Groningen).

References (42)

  • J Miranda et al.

    Intervention for minor depression in primary care patients

    Psychosom Med

    (1994)
  • L.M Mynors-Wallis et al.

    Randomised controlled trial comparing problem-solving treatment with amitriptyline and placebo for marjor depression in primary care

    BMJ

    (1995)
  • W Katon et al.

    Collaborative management to achieve treatment guidelines. Impact on depression in primary care

    JAMA

    (1995)
  • H.C Schulberg et al.

    Treating major depression in primary care practice. Eight-month clinical outcomes

    Arch Gen Psychiatry

    (1996)
  • D Skuse et al.

    Screening for psychiatric disorder in general practice

    Psych Med

    (1984)
  • D Goldberg et al.

    Screening for psychiatric illness in general practicethe general practitioner versus the screening questionnaire

    J R Coll Gen Pract

    (1987)
  • M VonKorff et al.

    Anxiety and depression in a primary care clinic. Comparison of Diagnostic Interview Schedule, general health questionnaire, and practitioner assessments

    Arch Gen Psychiatry

    (1987)
  • J Ormel et al.

    Recognition, management and outcome of psychological disorders in primary carea naturalistic follow-up study

    Psychol Med

    (1990)
  • L Borgquist et al.

    Perceived health and high consumers of carea study of mental health problems in a Swedish primary health care district

    Psychol Med

    (1993)
  • W.W.K Zung et al.

    The comorbidity of anxiety and depression in general medical patientsa longitudinal study

    J Clin Psychiatry

    (1990)
  • Cited by (67)

    • Different answers to different questions: Exploring clinical decision making by general practitioners and psychiatrists about depressed patients

      2014, General Hospital Psychiatry
      Citation Excerpt :

      The contrast between primary and specialized care will further increase due to the introduction of a new structure of health care in 2014 in which a more explicit and more strict distinction between general care, basic mental health care and specialized health care will be introduced. Although the three perspectives on differences between GPs and psychiatrists do not exclude each other, they stem from three different research traditions, apparently without much interaction or mutual influence [19]. The gold standard-approach relies for a large part on outcomes of efficacy trials in combination with output from epidemiological surveys, the second “input” vision more often uses results from effectiveness studies in the field of primary care with special focus on the need- and demand characteristics of the patients [20], and the third information processing perspective is mostly based on qualitative studies of clinical reasoning [21].

    • Depression, neighborhood deprivation and risk of type 2 diabetes

      2013, Health and Place
      Citation Excerpt :

      All cases of major depression and type 2 diabetes were clinically identified, which is not synonymous with true prevalence (particularly for depression). Although major depression is under-detected in primary care (Harman et al., 2006), in the past decade there has been a push to increase detection in Sweden general practice (Hansson et al., 2008) and the discordance between primary care and clinical interviews is generally one of specific diagnosis (e.g., major depression vs. dysthymia) rather than whether or not psychopathology is present at all (Tiemens et al., 1999). Also, the quality and validity of electronic medical records from primary care in Sweden is high (Nilsson et al., 2003; Grimsmo et al., 2001), and the relative availability of primary and psychiatric care in Sweden should increase the detection of major depression relative to places such as the United States.

    • Do GPs' medical records demonstrate a good recognition of depression? A new perspective on case extraction

      2011, Journal of Affective Disorders
      Citation Excerpt :

      These estimates may not reflect the actual diagnostic process in primary care. Depressed patients might be labeled by their GP as having other psychiatric problems, or the registration code could have been limited to a symptom code (‘feeling depressed’), to psychosocial problems or to a prescription of an antidepressant (Tiemens et al., 1999; Volkers et al., 2005). Other previous studies have used a simple questionnaire or form to assess whether GPs are able to diagnose depression (Ormel et al., 1990; Tiemens et al., 1999; Balestrieri et al., 2007; Menchetti et al., 2009).

    • Recognition and subsequent treatment of patients with sub-threshold symptoms of depression in primary care

      2011, Journal of Affective Disorders
      Citation Excerpt :

      The failure of GPs to spot depressive disorders in patients that have been independently assessed as depressed using research tools (false negatives) has been much investigated (Freeling et al., 1985; Kirmayer et al., 1993; Tylee et al., 1993; Coyne et al., 1995; Simon and VonKorff, 1995; Schwenk et al., 1996; Callahan et al., 1997; Garrard et al., 1998; Borowsky et al., 2000; Bertakis et al., 2001; Aragones et al., 2004; Volkers et al., 2004; Diminic-Lisica et al., 2005; Pfaff and Almeida, 2005; Ani et al., 2008). Less attention has been paid to those patients who do not meet criteria for depressive disorders formally, but who are diagnosed as depressed by general practitioners (false positives) (Tiemens et al., 1999; Klinkman et al., 1998; Aragones et al., 2006; Berardi et al., 2005; Gerber et al., 1989; Wittchen et al., 2001). Additionally, it has been shown that non-psychiatric physicians, including GPs, are much better at correctly ruling out depressive disorders (demonstrating high specificity) than appropriately recognising them (demonstrating lower sensitivity) (Cepoiu et al., 2008; Mitchell et al., 2009).

    View all citing articles on Scopus
    View full text