Free thyroxine estimation for the screening of hyper- and hypothyroidism in an adult population.

Fukazawa, Hiroshi; Sato, Toshiro; Yoshida, Katsumi; Kitaoka, Hirofumi; Nomura, Takashi; Suzuki, Michiko; Kaise, Nobuko; Kaise, Kazuro; Yamamoto, Makiko; Saito, Shintaro

doi:10.1620/tjem.148.411

Cited by 5 publications

(1 citation statement)

References 17 publications

(16 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Physicians rely heavily on blood tests to make diagnoses. Although lab tests are useful in diagnosing thyroid abnormalities, the positive predictive rate can be as low as 15-26% (Fukazawa, Sakurada, & Yoshida, 1986). The high error rates occur because test results characteristic of hypo-and hyperthyroidism are not consistent across patients and are influenced by a variety of factors such as pregnancy, drug interactions, nonthyroidal illnesses, and psychiatric problems (Gavin, 1988;Wilke, 1986;Wong & Steffes, 1984).…”

Section: Thyroid Diagnosis Problemmentioning

confidence: 99%

The Effect of Misclassification Costs on Neural Network Classifiers

Berardi

Zhang

1999

Decision Sciences

View full text Add to dashboard Cite

The potential of neural networks for classification problems has been established by numerous successful applications reported in the literature. One of the major assumptions used in almost all studies is the equal cost consequence of misclassification. With this assumption, minimizing the total number of misclassification errors is the sole objective in developing a neural network classifier. Often this is done simply to ease model development and the selection of classification decision points. However, it is not appropriate for many real situations such as quality assurance, direct marketing, bankruptcy prediction, and medical diagnosis where misclassification costs have unequal consequences for different categories. In this paper, we investigate the issue of unequal misclassification costs in neural network classifiers. Through an application in thyroid disease diagnosis, we find that different cost considerations have significant effects on the classification performance and that appropriate use of cost information can aid in optimal decision making. A cross-validation technique is employed to alleviate the problem of bias in the training set and to examine the robustness of neural network classifiers with regard to sampling variations and cost differences.

show abstract

Section: Thyroid Diagnosis Problemmentioning

confidence: 99%

The Effect of Misclassification Costs on Neural Network Classifiers

Berardi

Zhang

1999

Decision Sciences

View full text Add to dashboard Cite

show abstract

Screening depressives for causative medical illness; the example of thyroid function testing: I. Literature review, meta‐analysis, and hypothesis generation

1993

View full text Add to dashboard Cite

There is ample evidence that thyroid disease causes mood symptoms. However, little is known regarding the prevalence of hypo‐ and hyperthyroidism among mood disordered patients. Such data are necessary in order to develop rational screening strategies for reversible causes of depression. To investigate the prevalence of thyroid disease in mood disordered patients we conducted a comprehensive review of the literature to determine prevalence rate far hypo‐ and hyperthyroidism in three types of samples, focusing on clinical thyroid disease, as contrasted with sub clinical forms of the disorders for reasons discussed in detail in the body of this report. 1) In unselected general population and ambulatory medical clinic samples, prevalence rates for hypo‐ and hyperthyroidism were found to be highly dependent upon age and gender. Somatic thyroid symptoms, but not symptoms of fatigue or depression, increased the yield of thyroid Junction screening in studies which investigated the issue. 2) Two studies were available specifically on fatigued general medical samples, and only one on ambulatory depressives. Rates of hypo‐ and hyperthyroidism did not differ from the meta‐analysis‐derived prevalence rates for general population and medical clinic samples. 3) Analysis of inpatient depressive studies revealed a bimodal distribution of prevalence rates, with the bulk of studies reporting negligible rates of hypo‐ or hyperthyroidism. Sample size was modest in most of the inpatient studies, and no factors could be identified which predicted increased rates of hypo‐ or hyperthyroidism among inpatient depressives. We therefore hypothesize that although hypo‐ and hyperthyroidism produce mood symptoms in a high proportion of affected persons, major depressives may not comprise an enriched sample for hypo‐ or hyperthyroidism. If so, screening guidelines for the general population, rather than routine screening of all depressives, may be the appropriate testing strategy. Limitations of the current data are discussed in light of conceptual and methodological issues in testing strategy development. Depression 1:210‐219 (1993). © 1993 Wiley‐Liss, Inc.

show abstract