Portal de Periódicos da CAPES

Misuse of psychiatric epidemiology

1998; Elsevier BV; Volume: 351; Issue: 9116 Linguagem: Inglês

10.1016/s0140-6736(05)77684-4

1474-547X

Eve Leeman,

Schizophrenia research and treatment

20 years ago, psychiatric epidemiology was still in its infancy. Several investigators had tried to ascertain the true prevalence of psychological disorders in community populations, not just among people receiving psychiatric treatment.1Regier DA Myers JK Kramer M et al.The NIMH Epidemiologic Catchment Area Program.Arch Gen Psychiatry. 1984; 41: 934-941Crossref PubMed Scopus (1032) Google Scholar These studies suggested that mental illness—commonly untreated—was quite prevalent in the USA, but the data lacked uniformity. Results could not be reliably compared across sites since the studies differed in design, objectives, case-funding methods, and even case-definitions. An explicit diagnostic system did not exist, nor did a reliable instrument for large-scale surveys. The 1978 US President's Commission on Mental Health,2The President's Commission on Mental HealthReport to the President from the President's Commission on Mental Health. vol 1. US Government Printing Office, Washington DC1978Google Scholar which aimed to identify research and service gaps, served as a call to action. The development of the National Institute of Mental Health Diagnostic Interview Schedule (DIS),3Robins LN Helzer JE Ratcliff KS Seyfried W Validity of the Diagnostic Interview Schedule version II: DSM-III diagnoses.Psychol Med. 1982; 12: 855-870Crossref PubMed Scopus (395) Google Scholar a case-identification instrument based on the Diagnostic and Statistical Manual of Mental Disorders, 3rd edition (DSM-III), paved the way for the landmark Epidemiologic Catchment Area study (ECA) in 1980–85. Over 18 000 patients in five sites around the USA were surveyed with the DIS by non-clinician interviewers. 15·4% of the US population were estimated to meet criteria for at least one alcohol-related, drug-abuse, or other mental disorder during the month before the interview,4Regier DA Boyd JH Burke JD et al.One month prevalence of mental disorders in the United States.Arch Gen Psychiatry. 1988; 45: 977-986Crossref PubMed Scopus (1473) Google Scholar 20% had met criteria during the previous 12 months, and 32% reported a lifetime history of one or more disorders.5Regier DA Kaelber CT Rae DS et al.Limitations of diagnostic criteria and assessment instruments for mental disorders.Arch Gen Psychiatry. 1998; 55: 109-115Crossref PubMed Scopus (442) Google Scholar The ECA was received with much optimism. It was seen as providing a topographic map from which further studies could be generated and allow epidemiology to be linked with biological, nosological, genetic, familial, and clinical research.6Freedman DX Psychiatric epidemiology counts.Arch Gen Psychiatry. 1984; 41: 931-933Crossref PubMed Scopus (56) Google Scholar Policymakers, armed with data rapidly obtained under strong federal science leadership, could address service needs, comforted by the existence of discrete and definable disorders with total prevalence approximately equivalent to that of hypertension.6Freedman DX Psychiatric epidemiology counts.Arch Gen Psychiatry. 1984; 41: 931-933Crossref PubMed Scopus (56) Google Scholar In the early 1990s the ECA was partly replicated to confirm its numerical findings and to focus more specifically on the relation between comorbid mental and addictive disorders.5Regier DA Kaelber CT Rae DS et al.Limitations of diagnostic criteria and assessment instruments for mental disorders.Arch Gen Psychiatry. 1998; 55: 109-115Crossref PubMed Scopus (442) Google Scholar The National Comorbidity Study (NCS), a single-wave interview of a national probability sample of 8098 respondents, identified many risk factors already detected by the ECA for psychiatric illness, but reported significantly higher prevalence rates. Nearly 50% of all respondents reported at least one lifetime disorder, with almost 50% meeting criteria in the previous year.7Kessler RC McGonagle KA Zhao S et al.Lifetime and 12-month prevalence of DSM III-R psychiatric disorders in the United States.Arch Gen Psychiatry. 1994; 51: 8-19Crossref PubMed Scopus (10862) Google Scholar D A Regier and colleagues5Regier DA Kaelber CT Rae DS et al.Limitations of diagnostic criteria and assessment instruments for mental disorders.Arch Gen Psychiatry. 1998; 55: 109-115Crossref PubMed Scopus (442) Google Scholar have reviewed the current state of the art of psychiatric epidemiology. They focus in particular on the “health policy implications of discrepant and/or high prevalence rates for determining treatment need in the context of managed care definitions of medical necessity”. Warning that the ECA and NCS “do not adequately differentiate between diagnosis and treatment need”, they advise an increase in symptom threshold, and the addition of duration and of criteria for impairment or disability in assessment techniques.5Regier DA Kaelber CT Rae DS et al.Limitations of diagnostic criteria and assessment instruments for mental disorders.Arch Gen Psychiatry. 1998; 55: 109-115Crossref PubMed Scopus (442) Google Scholar Improvement in standardisation of epidemiological tools and the ability to distinguish measured diagnoses from clinical need is a laudable goal. The policy implications are more disturbing. The shift in national mood from one of progressive outreach to cost-effectiveness in health care casts an ominous shadow on what began as an inspired search for truth. Scientific shortcomings in the predictive value of these surveys are now being raised as reasons for limiting treatment in the managed-care era. With illness categories no longer clearly defined and data not replicable, policymakers see a “bottomless pit” of psychiatric illness, fearing that the US talk-show-driven culture will bankrupt the health-care system in a mad rush for treatment. The discrepancies in ECA and NCS findings are not surprising. From the outset, the NCS investigators predicted that they would find higher prevalence rates,8Blazer DG Kessler RC McGonagle KA Swartz MS The prevalence and distribution of major depression in a national community sample: The National Comorbidity Survey.Am J Psychiatry. 1994; 151: 979-986PubMed Google Scholar and Regier himself wrote in 1984 that the results of the ECA were to be seen as “method-dependent” approximations.1Regier DA Myers JK Kramer M et al.The NIMH Epidemiologic Catchment Area Program.Arch Gen Psychiatry. 1984; 41: 934-941Crossref PubMed Scopus (1032) Google Scholar Differences in study tools, such as the earlier placement and greater number of stem screening questions in the NCS than in the ECA, can be expected to yield higher positive responses because order effects are quite substantial.8Blazer DG Kessler RC McGonagle KA Swartz MS The prevalence and distribution of major depression in a national community sample: The National Comorbidity Survey.Am J Psychiatry. 1994; 151: 979-986PubMed Google Scholar People with psychiatric disorders are under-represented in cross-sectional community surveys. The ECA did not address this issue; the NCS provided a financial incentive to reinterview initial non-respondents, thereby increasing the number of ill interviewees.8Blazer DG Kessler RC McGonagle KA Swartz MS The prevalence and distribution of major depression in a national community sample: The National Comorbidity Survey.Am J Psychiatry. 1994; 151: 979-986PubMed Google Scholar Moreover, the bulk of the data was collected in single interviews by non-clinicians.9Eaton WW Jozer CE Von Korff M et al.The design of the Epidemiologic Catchment Area Surveys.Arch Gen Psychiatry. 1984; 41: 942-948Crossref PubMed Scopus (220) Google Scholar Psychiatrists would rarely expect a diagnosis made in one interview to be accurate over time. Even diagnoses made during lengthy inpatient stays often change during outpatient follow-up. And those are diagnoses made in acutely ill people seeking help. Responder bias is likely to be higher among participants in community research who have to recall distant symptoms. Regier and colleagues state that “modifications in both diagnostic criteria and assessment instruments have revealed their sensitivity to seemingly small changes and their possible limitations in defining the need for mental health services.”5Regier DA Kaelber CT Rae DS et al.Limitations of diagnostic criteria and assessment instruments for mental disorders.Arch Gen Psychiatry. 1998; 55: 109-115Crossref PubMed Scopus (442) Google Scholar How relevant to assessment of need is psychiatric prevalence data anyway? Robert Spitzer, an architect of DSM-I, says not very, although many would disagree. He postulates that “mental disorders, like physical disorders, vary in their severity and associated functional impairment. No one is interested in the prevalence of any physical disorder, so why the interest in prevalence rates for any mental disorder?”10Spitzer RL Diagnosis and need for treatment are not the same.Arch Gen Psychiatry. 1998; 55: 20Crossref Scopus (121) Google Scholar Prevalence of mental illness certainly does not correlate with treatment demand, as shown by a particularly significant finding of the NCS that Regier does not address. The vast majority of people with psychiatric disorders do not get treatment.7Kessler RC McGonagle KA Zhao S et al.Lifetime and 12-month prevalence of DSM III-R psychiatric disorders in the United States.Arch Gen Psychiatry. 1994; 51: 8-19Crossref PubMed Scopus (10862) Google Scholar Only four of every ten respondents with a lifetime history of at least one DSM-IIIR disorder obtain professional help. Among those reporting a 12-month disorder, only one in five receive any form of treatment, only one in nine are seen in the mental-health sector, and only one in 25 are treated in substance-abuse facilities.7Kessler RC McGonagle KA Zhao S et al.Lifetime and 12-month prevalence of DSM III-R psychiatric disorders in the United States.Arch Gen Psychiatry. 1994; 51: 8-19Crossref PubMed Scopus (10862) Google Scholar That concerns about the economic burden of seemingly ever-increasing psychiatric illness outweigh distress about the barriers to adequate care is a reflection of our times. Scientists may argue about the best way to further refine psychiatric epidemiology. But lack of concordance in community samples should not give policymakers and insurance companies permission to refuse psychiatric care to ill patients who do ask for help.