TEST CHARACTERISTICS AND DECISION RULES

This study aimed to examine the reliability and clinical decision validities of the Taipei City Developmental Checklist for Preschoolers, 2nd version (the Taipei II, which was filled out by parents) and the screening procedures conducted in the medical setting.

Methodology research and case control study designs were adopted. A total of 310 dyads consisting of children who were developing typically and 196 dyads of children with developmental delays and age 5.5 to 35.5 months were recruited for validity test. Among them, 165 mothers filled out the questionnaire twice within 1 week to examine the test–retest reliability of the total score and individual items. Validity indexes of the single cutoff strategy and multiple cutoff strategies were analyzed. With two cutoff point strategies, the likelihood ratios (LR) of the three test results, positive, neutral, and negative, were calculated.

The test–retest reliabilities of the total scores of the seven checklists of the Taipei II (r_s = 0.54–0.89, p<0.05) and their individual items (agreement 92% to 100%) were acceptable, except for the 30-month checklist and three individual items. The positive LR (LR+) and negative LR (LR-) of the single cutoff strategy were acceptable with most LR+ more than 2, and all LR- less than 0.5. Most of the diagnostic odds ratios of single cutoff strategies were less than 50 and they did not meet the acceptable criteria. When multiple cutoff points were used, all of the LRs with positive test results were equal to infinity that met SpPin criteria, and all of the LRs with negative test results less than 0.5 had at least a small but important diagnostic impact.

Taipei II with multiple cutoff points could give more useful clinical information than using a single cutoff point. The multiple likelihood ratios of Taipei II for children older than 3 years and in different cultural backgrounds need further study.

The purpose of this study was to develop a Cancer Pain Prognostic Scale (CPPS) which could predict the likelihood of pain relief within 2 weeks for cancer patients with moderate to severe pain. Seventy-four (74) consecutive patients who presented with cancer-related pain were managed in accordance with the guidelines for pain management developed by the United States Agency for Health Care Policy and Research (AHCPR). Patients were followed weekly using the Brief Pain Inventory (BPI), and medications were recorded weekly for 3 weeks. Baseline scores from the Functional Assessment of Cancer Therapy (FACT-G), Mental Health Inventory (MHI), Karnofsky Performance Status (KPS), and Memorial Symptom Assessment Scale Short Form (MSAS-SF) at initial interview served as explanatory variables in a logistic regression model. Pain relief ≥ 80% at the end of weeks 1 and 2 were used as outcomes in this model. From this analysis, we developed a predictive formula, the CPPS, which includes the worst pain severity, FACT-G emotional well being, daily opioid dose, and pain characteristics. The rule yields a numerical score that ranges from 0–17. Higher scores correspond to a higher probability of good pain relief. The CPPS has the potential to rapidly identify patients with poor pain prognosis. It can be used as a research tool to characterize pain in cancer patients.

Recent experiments have produced a linked data set of clinical and kinematic responses for human subjects exposed to controlled low-speed rear-end collisions. The purpose of this paper was to examine this paired data set and determine whether the presence or absence of clinical symptoms could be predicted from the peak linear and angular kinematic response of the head and neck. The data were generated using 42 male and female human subjects seated normally in the front passenger seat of a stationary vehicle struck from behind to produce vehicle speed changes of 4 and 8 km/h. Pre- and post-test clinical examinations documented the presence, severity and duration of whiplash-associated disorders (WAD). Logistic regression and backward elimination of independent variables were used to develop the prediction model. The analysis yielded a 16 parameter model that was significantly related (odds ratio=21.2; P=0.0069) to the presence or absence of transient whiplash symptoms. The model correctly predicted symptom presence in 13 of 23 tests (sensitivity 57%) and symptom absence in 49 of 52 tests (specificity 94%) in a population of 75 with a symptom prevalence of 31%. The model’s positive predictive value was 81% and its negative predictive value was 83%. Despite statistical significance, the model did not discriminate between the presence and absence of symptoms in all tests, and indicated that factors other than the selected peak kinematic responses influenced symptom production.

Practicing and perfecting the art of medicine demands recognition that uncertainty permeates all clinical decisions. When clinicians make decisions, diagnostic or therapeutic, combining the art and the science is a complicated matter replete with unknowns. Herein, the authors do not propose to address the rivalry between intuition and analytic models of cognition.³⁸ One can recognize the former when a master clinician exhibits outstanding diagnostic acumen. On the other hand, scientific education in medical school and beyond involves analytic cognition. The “evidence-based medicine” movement can be viewed as an attempt to reduce reliance on intuition. Nevertheless, the application of medical knowledge is always done in the context of uncertainty; textbook cases occur only in textbooks. A variety of modes of clinical problem-solving exist, ranging from automatic decisions, in which a clinician makes a decision without consciously considering alternatives or even specific features of the problem, to “rational” decision-making using simplified strategies to take into account probabilities and values of outcomes to formal decision analysis. The potential deficiencies of automatic decisions (nondecisions one might say) are obvious. Because formal decision analyses, even if accepted as a gold standard (which they are not), are beyond practical use for the vast majority of clinicians, simplified processes are followed to combine likelihoods and values. This article discusses one method that clinicians can use to assist them in decisions about which tests to order and how to interpret the results—the bayesian approach. Even when striving to reduce uncertainty and to come up with the best course of action for patients, it is critical to recognize that uncertainty will always be present. In using a quantitative approach, one can only become more certain about the probability of the outcome. Such an approach cannot guarantee the desired outcome.

In approaching a case, the clinician makes a series of inferences about the nature of a patient's condition and proceeds toward a working diagnosis along with some alternatives—the generation of a differential diagnosis.18, 36, 42 Although one could create a possibilistic differential, that is, a listing of all the possible causes of the patient's complaint, experienced clinicians generate a differential diagnosis that is a combination of probabalistic (i.e., considering first those disorders that are more likely), prognostic (i.e., considering first those disorders that are more serious if missed), and pragmatic (i.e., considering first those disorders most responsive to treatment). The clinician then refines the diagnostic hypotheses, sometimes using clues from the history and physical examination and often with the use of laboratory or radiologic tests. In doing so, the clinician tries to reduce the inherent uncertainty so that the most appropriate course of treatment can be prescribed.³ In the Bayesian model, the physician updates his or her belief in a hypothesis with each new item of information, with different weights given to the new information depending on its operating characteristics. In this article, the authors use a bayesian approach to address the question, “How can a clinician use laboratory tests to approach the diagnostic dilemma in an individual patient when a mass is incidentally found in the adrenal, pituitary, or thyroid glands?” The data presented are derived from individual reports, recognizing that a more rigorous approach would involve a systematic search and systematic critical review of the literature. In the case of diagnostic tests, even a systematic search would be insufficient because of the variability in test performance from laboratory to laboratory. Because of these limitations, the examples herein should be considered just that—examples.

The first step in the Bayesian approach is to understand the concept of probability. A probability is an expression of likelihood and thus represents an opinion of the relative frequency with which an event is likely to occur. When making a diagnosis, probability is a numerical expression of the clinician's uncertainty about a diagnosis. The use of numerical wording is urged because expresasing a clinical opinion in subjective terms such as “likely” and “possible” is fraught with imprecision and misunderstanding.⁷ Certainty that a disease is present is assigned a probability of 1, certainty that a disease is not present is assigned a probability of 0, and the clinician's opinion of the disease being present or absent usually falls somewhere in between. Probabilities are derived from different data sources that vary in their reliability and application to a given patient, such as the clinicians's experience (remembered cases), research studies, and population-based epidemiologic studies; therefore, a certain ambiguity is inherent in a given probability, and the confidence with which one can rely on a given probability depends to a large extent on the data underlying it.

The diagnostic approach to minimizing uncertainty requires four steps. First, the clinician starts with an estimate of probability based on initial clinical impressions. This starting point is dubbed the “pretest” or a priori probability and is a number between 0 and 1 that expresses the likelihood of disease. For example, a clinician who sees a large population of patients with diabetes mellitus may conclude that a 55-year-old patient with polyuria, polydipsia, weight loss, and fatigue has a probability of 0.70 of having diabetes mellitus. In other words, if there were 100 such individuals, 70 would have diabetes. Second, the clinician determines the probability at which he or she is compelled to treat. This probability is dubbed the “threshold” probability (often denoted as p*). If the pretest probability is greater than the threshold probability, the clinician may opt to treat; if it is less than the threshold probability, the clinician may opt not to treat. In the previous example, the physician will conclude that the benefits of administering insulin immediately outweigh the risks of insulin when the probability of diabetes mellitus is 0.60 or greater. Third, if the clinician is not comfortable enough about the presence or absence of disease, he or she may chose to order further tests with the objective of getting closer to certainty. If the physician in the previous example recognizes that there is a finite chance of other diseases presenting similarly, he or she will order a fingerstick blood glucose level. The final step involves taking information gained from the test and using it to update the pretest probability. This updated or “posttest” probability may now cross the treatment threshold probability and add more certainty to guide therapy. If the physician finds a normal blood glucose level, he or she will adjust the probability of diabetes mellitus below 0.60, thereby saving the patient from insulin therapy. Additionally, the updated or posttest probability can serve as a new pretest probability for the next step in hypothesis testing. In the example, the physician may remain very concerned about the possibility of diabetes mellitus (the probability is now about 0.40) and may perform a serum blood glucose test to rule out diabetes mellitus as the cause of the patient's symptoms. Figure 1 shows this process graphically.

The following sections discuss each of the four steps in depth and use them to detail a reasonable diagnostic approach in the evaluation of pituitary, adrenal, and thyroid incidentalomas.

The ELISA for detection of anti-phosphatidylethanolamine antibodies in patients with lupus anticoagulant was studied. The influence of microtitre plate types, antigen concentration, antibody dilution, and buffer system was evaluated. Human albumin in the blocking buffer gave greater numbers of positive plasma samples than bovine serum. Comparison of simultaneous plasma and serum samples showed only minor variation in binding characteristics.

The desmopressin (DDAVP) test has been proposed to discriminate Cushing's disease (CD) from pseudo-Cushing states (PC); however, current information on its value is scarce and contradictory.

The aim of the study was to assess the ability of the DDAVP test in distinguishing between these conditions, with emphasis on subjects with mild hypercortisolism.

We conducted a retrospective/prospective study at the Division of Endocrinology, Polytechnic University of Marche, Ancona, Italy.

The study included 52 subjects with CD, 28 with PC, and 31 control subjects (CT).

We performed the DDAVP test and standard diagnostic procedures for the diagnosis of Cushing's syndrome.

The diagnosis/exclusion of CD was measured.

Interpretation of the DDAVP test based on percentage and absolute increment of cortisol and ACTH did not afford acceptable values of both sensitivity (SE) and specificity (SP). CD diagnosis based on simultaneous positivity for basal serum cortisol greater than 331 nmol/liter and absolute ACTH increment greater than 4 pmol/liter and its exclusion in subjects negative for one or both measures yielded an SE of 90.3% and an SP of 91.5%. The approach was also highly effective in distinguishing PC from: 1) CD with moderate values of urinary free cortisol (SE, 86.9%; SP, 92.8%); 2) CD with moderate values of serum cortisol after dexamethasone suppression (SE, 86.6%; SP, 92.8%); and 3) CD with moderate values of midnight serum cortisol (SE, 100%; SP, 92.8%).

Interpretation of the DDAVP test through a combination of parameters allowed effective discrimination of CD from PC, even in subjects with mild hypercortisolism.