| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
|
|
|
|||||||
|
|||||||||
Electronic Letters to:
|
|
Electronic letters published:
![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
The term "diagnostic accuracy" may be misleading
The meaning of "diagnostic"
Timing of testing and the use of serial testing need also to be reported adequately
Please, do not forget consumers of health care services
|
|
|||
|
Joseph Watine, Eur Clin Chem Hôpital de Rodez
Send letter to journal:
j.watine{at}ch-rodez.fr Joseph Watine
|
The definition of diagnostic accuracy proposed by Bruns et al. [1] may be misleading. According to common sense, diagnostic accuracy refers to diagnosis. In the case of prognosis, common sense would lead most laboratory medicine specialists to use the term prognostic accuracy instead of diagnostic accuracy [2, 3]. Would it not be better to use different definitions for prognostic accuracy, diagnostic accuracy, etc. 1) Bruns DE, Huth EJ, Magid E, Young DS. Toward a checklist for reporting of studies of diagnostic accuracy of medical tests. Clin Chem 2000; 46:893-895. 2) Watine J. Prognostic evaluation of primary non-small cell lung carcinoma patients using biological fluid variables. A systematic review. Scand J Clin Lab Invest 2000; 60:259-273. 3) Laboratory variables as additional staging parameters in patients with small-cell lung carcinoma. A systematic review. Clin Chem Lab Med 1999; 37:931-938. Joseph WATINE Laboratoire de biologie polyvalente Hôpital Général 12027 Rodez Cedex 9, France Tel: +33.(0)5.65.75.15.60 Fax: +33.(0)5.65.75.19.73 Email: j.watine@ch-rodez.fr |
|||
|
|
|||
|
David Bruns Department of Pathology
Send letter to journal:
dbruns{at}virginia.edu David Bruns
|
Thank you for your comment. As pointed out in the paper, the authors, too, struggled with the use of the term "diagnostic accuracy." In fact, relatively few test results (other than microscopic examinations of tissue) are truly "diagnostic" in the sense that physicians often use the word. "Accuracy" may be confused with analytical accuracy. The rationale for the use of the term was discussed in the paper and will not be repeated here. It may be useful to point out that a common example of a "diagnostic" system is a weather forcast -- clearly a prognostic use. The term "diagnostic accuracy" is well-entrenched, and the authors of the paper pointed out that they could not find a better one. Needless to say, the title of the paper would have been very long indeed if the single term were replaced by a list of separate terms to describe each of the possible uses of diagnostic systems, e.g., prognosis, monitoring, screening, diagnosis, risk assessment, etc. |
|||
|
|
|||
|
Jacques Massé, Medical biochemist Dept. of medical biology, Centre Hospitalier Universitaire de Québec
Send letter to journal:
jacques.masse{at}bm.ulaval.ca Jacques Massé
|
The performance of a diagnostic test can vary with time. For example, cardiac markers have diagnostic "windows" and testing too early or too late can affect the sensitivity (and thus accuracy) of the test. Similarly, some tests are often repeated over various periods of time to confirm or exclude the diagnostic. If the time of testing is not planified before the study (only possible for prospective studies), the authors should provide the rationale for excluding or including a result on the base that it is in or outside the study testing window. Biases can be introduced if there are some pronostic or therapeutic reasons for doing or not doing the test at the expected time (this is another form of verification bias). If various testing windows have been tried, the authors should indicate which have been tested, and not only report the period for which maximum performance was obtained, since this may induce an overestimation of the real performance of the test(this is analogous to doing multiple statistical comparisons in search of a significant result). If peak values are studied, the authors should report on the importance of the number of repetitions and the adequate interval between two samples. Similarly, if the diagnostic test is reported as a rate of increase or decrease, the authors should also report the minimal number of samples needed for a valid estimate and the appropriate timing of the samples. |
|||
|
|
|||
|
Joseph Watine, PH, Eur Clin Chem, AIHP, AAHU Hôpital de Rodez, France
Send letter to journal:
j.watine{at}ch-rodez.fr Joseph Watine
|
As stressed in their paper, one of the aims that Bruns et al.’s have in mind is “making the reports (on diagnostic accuracy) more useful for systematic reviews” [1]. Indeed, it is currently believed that systematic reviewing may help to satisfy one of the most legitimate expectations of the consumers of health care services: having medical practice based on scientific evidence, regarding not only effectiveness, and safeness, but also cost-effectiveness [2-4]. Laboratory medicine is part of medicine. A term that is understood by laboratory medicine specialists should be understood by all medical professionals (and conversely). In addition, we, medical professionals, must try to be intelligible to as many consumers of health care services as possible, and not only to a certain professional “elite”. It is therefore very important, when we “create” professional terms, to make them as intelligible to all as possible [5]. I am quite convinced that it would be more intelligible to most health care consumers (this also includes the health care professionals) to read this: “accuracy” of a medical test refers to its ability to provide accurate information about screening, diagnosis, prognosis, treatment, monitoring, risk of disease, and other medical issues. or even: “medical accuracy” refers to the ability of a medical test to provide accurate information about screening, diagnosis, prognosis, treatment, monitoring, risk of disease, and other medical issues. than this: “diagnostic accuracy” refers to the ability of a medical test to provide accurate information about diagnosis, prognosis, risk of disease, and other clinical issues [1]. My proposed term for “medical accuracy” (or alternatively “accuracy” alone) would only be a heading below which it would be possible to use sub -headings such as “diagnostic accuracy”, “prognostic accuracy”, “screening accuracy”, etc. According to Bruns et al. [1], the term “diagnostic accuracy” is “well-entrenched” in the biomedical literature and a list of separate terms would be too long. Such an opinion might well be not only subjective (what do “well-entrenched” and “too long” mean?), but also wrong. For example, the Cochrane Collaboration Center, has a Committee on Systematic Reviewing of Sreening and Diagnostic Tests [6]. The existence of this committee shows that some well known authors may consider that different terms are indicated for different situations (i.e. in this case “screening” is clearly distinguished from “diagnostic”). The Medline indexers also distinguish between diagnosis and prognosis [7]. Besides, as suggested by Bruns et al. [1], the heading “diagnostic accuracy” may have some problems (e.g. being confused with analytical accuracy [1]). It would therefore perhaps be better to try to “trench out” this “well-entrenched” heading before it becomes still more “well-entrenched”. After all, it is quite usual, in the scientific field, to see such “well-entrenched” definitions change. Indeed, it is already very well understood by most people that some medical decisions (or opinions) may be more accurate than others, regarding not only diagnosis, but also all other sorts of medical issues. For example, evidence-based medicine would lead most health care consumers to believe that a patient suffering from colorectal liver metastases would better consult an excellent medico-surgical team than to try to benefit from psychological support alone [5, 7, 8]. Such an excellent medico- surgical team would use various more or less accurate medical tests (including clinical, radiological, and pathological ones) to assess if the patient needs to be resected (or not) and/or to be given (or not) anticancer chemotherapy, etc. David Bruns and/or some of his colleagues might of course have various thoughtful objections to the use of “medical accuracy” (or alternatively “accuracy” alone) as main heading instead of “diagnostic accuracy” in their proposed definition [1]. In this case, may I suggest that they could express such objections in a new e-response to the present e-letter? Let us hope that in such an e-response, they will bear in mind that we are not at all dealing with “weather forecasts” but with medical tests [see David Bruns’ first e-response to my first e-letter] (this last sentence aims at being slightly humorous). Thank you very much in advance. [1] Bruns DE, Huth EJ, Magid E, Young DS. Toward a checklist for reporting of studies of diagnostic accuracy of medical tests. Clin Chem 2000; 46:893-895. [2] Ohlsson A. Systematic reviews - Theory and practice. Scand J Clin Lab Invest Suppl 1994; 219:25-32. [3] Oosterhuis WP, Niessen RWLM, Bossuyt PMM. The Science of Systematic Reviewing Studies of Diagnostic Tests. Clin Chem Lab Med 2000; 38:577-588. [4] Sandberg S, Oosterhuis W, Freedman D, Kawai T. Systematic reviewing in laboratory medicine. Position paper from the IFCC committee on systematic reviewing in laboratory medicine. J Int Fed Clin Chem 1997; 9:154-155. [5] Watine J, Borgstein J. Evidence-based illiteracy or illiterate evidence. Lancet 2000; 356:684. [6]http://som.flinders.edu.au/FUSA/COCHRANE/cochrane/sadtdoc1.htm [7] Hunt DL, Jaeschke R, McKibbon KA for the Evidence-Based Medicine Working Group. Users' guides to the medical literature: XXI. Using electronic health information resources in evidence-based practice. JAMA 2000; 283:1875-1879. [8] Bismuth H, Adam R, Levi F, Farabos C, Waechter F, Castaing D, Majno P, Engerran L. Resection of nonresectable liver metastases from colorectal cancer after neoadjuvant chemotherapy. Ann Surg 1996; 224: 509- 520; discussion 520-522. |
|||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
eLetters: