Evaluating a new graphical ordinal logit method (GOLDminer) in the diagnosis of myocardial infarction utilizing clinical features and laboratory data.

OBJECTIVE: We used a new graphical ordinal logit method (GOLDminer) to assess a single cardiac troponin T (cTnT) analysis at the time of admission (first generation monoclonal; Roche BMC Corp., Indianapolis, Indiana), the character of chest pain, and electrocardiographic (ECG)findings in predicting the likelihood of acute myocardial infarction (AMI) in patients presenting with suspected myocardial ischemia. The final diagnosis of AMI was based on serial ECG findings and evolution of CKMB isoenzyme levels in conjunction with clinical findings. SUBJECTS: The study population consisted of 293 consecutive patients who presented at a mean of six hours after onset of chest pain or associated symptoms warranting a "rule-out" for AMI assessment to a university-affiliated community hospital. RESULTS: The odds-ratio for an elevated cTnT (> 0. 1 ng/ml) in AMI was 22.2:1. There was an association between typical chest pain and cTnT (chi square = 78.23, p < .0001) and between abnormal ECG findings and cTnT (chi square = 108, p < .0001). The cTnT yielded diagnostic benefit in addition to chest pain characteristics and ECG findings in AMI. We present the odds-ratios for the combined features in GOLDminer plots. CONCLUSION: We demonstrate how the odds-ratios for AMI are obtained after scaling continuous to ordinal the values for a single cTnT determination alone and with other features in patients presenting with chest pain.

Diagnosis of myocardial infarction: integration of serum markers and clinical descriptors using information theory.

OBJECTIVE: We examine the use of information theory applied to a single cardiac troponin T (cTnT) (first generation monoclonal; Boehringer Mannheim Corp., Indianapolis, Indiana) used with the character of chest pain, electrocardiography (ECG) and serial ECG changes in the evaluation of acute myocardial infarction (AMI). We combined a single measure of cTnT (blinded to the investigators) with a creatine kinase MB isoenzyme (CK-MB) measurement to discover the best decision value for this test in a study of 293 consecutive patients presenting to the emergency department with symptoms warranting exclusion of AMI. METHODS: The decision value for determining whether cTnT is positive or negative was determined independently of the final diagnosis by examining the information in the cTnT and CKMB data. Using information theory, an autocorrelation matrix with a one-to-one pairing of the CKMB and troponin T was constructed. The effective information, also known as Kullback entropy, assigned the values for troponin T and for CKMB that have the lowest frequency of misclassification error. The Kullback entropy is determined by subtracting the data entropy from the maximum entropy of the data set in which the information has been destroyed. The assignment of the optimum decision values was made independently of the clinical diagnoses without the construction of a receiver-operator characteristic curve (ROC). The final diagnosis of AMI was independently determined by the clinicians and entered into the medical record. RESULTS: The decision value for cTnT was 0.1 ng/ml as determined by the the information in the data. The method was validated within the same study by mapping the results so obtained into the diagnoses obtained independently by the clinicians using all of the methods at their disposal. The cTnT was different in AMI (n = 60) compared with non-AMI patients (n = 233) (2.08 +/- 0.21 vs. 0.07 +/- 0.10; p < .0001). CONCLUSION: Information theory provides a strong framework and methodology for determining the decision value for cTnT which minimizes misclassification errors at 0.1 ng/ml. The result has a strong correlation with other features in detecting AMI in patients presenting with chest pain.

Outcomes research is the fifth discipline of the fifth generation of managed care: utility of a single troponin-T.

OBJECTIVE: The fifth generation of managed care is disease management. Diseases have measurable risk in providing laboratory and medical services. The link between managing services and managing risk can be aided by leveraging the laboratory. We wish to remodel laboratory services to fit the needs of the use, thereby using the laboratory for competitive advantage by redesigning a desired output using a formal structured process. Outcomes research is the systems framework for the remodeling process through the link of laboratory output to clinical and financial outcomes. A process redesign model connects the use of laboratory tests to improved medical services by leveraging resources to achieve measurable improvement over current results. This view of outcomes research seeks both competitive advantage and measurable improvements in quality. METHODOLOGY: This approach is illustrated by the patient presenting with chest pain (CP). A majority of the patients rule out for acute myocardial infarction (AMI), including patients with indigestion, shortness of breath, and other clinical findings. This is the basis for an emergency department (ED) CP observation unit to reduce coronary care unit admission rates. When the Goldman algorithm for discharging low-risk patients with CP from the ED using only clinical features and electrocardiographic findings proved difficult to implement, we turned to measuring the diagnostic efficiency of a new cardiac marker to replace the evolutionary changes in creatine kinase (CK) isoenzyme MB. The physicians making the decision were blinded to the results of the study. We fitted the expected characteristics of the test to the expected results for our program. The test was done on the presenting specimen of 293 evaluable patients with a median of 6.5 hours from the time of onset of CP to the time the specimen was drawn. The result was compared with the evolutionary pattern of CK-MB. RESULTS: The sensitivity of the test at presentation to the ED was 85% compared with < 50% for the presenting CK-MB, the false negative results taken earlier than 3 hours or 10 days after the onset of symptoms. Troponin-T effectively identifies non Q-wave AMI much earlier than the CK-MB. This study led to a prospective randomized clinical trial to demonstrate an improved medical and financial benefit from an early rule in or rule out of severe coronary artery ischemia. CONCLUSION: The study supports our hypothesis that the laboratory can systematically redesign its technology strategy and participate in the construction of a clinical pathway for the discharge from ED or admitting decisions with a test 98% sensitive for identifying patients with serious coronary ischemia by 3.5 hours after the onset of symptoms.

The Bridgeport Hospital experience with autologous transfusion in orthopedic surgery.

The transfusion records of 341 orthopedic patients who donated blood preoperatively in the years 1992 and 1993 were audited to review the transfusion practices associated with the surgical procedures. The study sample underwent 182 total knee (TKA), 123 total hip (THA) arthroplasties, and 33 laminectomies with fusion (LAM) and 3 without. Data used were age, gender, predonation hemoglobin concentration (Hbd), initial (Hbi) and final (Hbf) hemoglobin concentration, surgical procedure, surgical blood order (SBO), and estimated blood loss (EBL). We analyzed for means and associations and differences between covariates. The means of EBL (mL), transfused units, donated units, Hbi, (g/dL), Hbd, and Hbf (g/dL) for the most common procedures were: TKA--272, 1.1, 2.1, 10.4, 13.9, and 10.1; THA--951, 2.3, 2.4, 9.4, 13.8, and 9.9; and LAM--589, 1.5, 2.2, 12.0, 14.6, and 11.2. Phlebotomies for procedures with minimal blood loss, as for total knee arthroplasties, result in wastage. Autodonation under such circumstances takes patients to an unnecessary low hemoglobin concentration prior to either retransfusing the blood taken or discarding part of it. The number of preoperative autologous units donated can be reduced if predonation hemoglobin concentration is > 15 g/dL and expected blood loss is not > 2 g.

Strategic considerations in clinical laboratory management: a laboratory leadership role in clinical pathways. Establishing the laboratory's direct contribution to the institution's performance.

This clinical laboratory serves a unique, though incompletely recognized, function in medical care as the source, moderator, and conduit of vital information supporting the detection, confirmation, and monitoring of disease states. To carry out that function effectively, the laboratory must actively participate in developing clinical pathways based on medical programs. Clinical pathways define the operational and quality requirements for the services involved in meeting the medical program objectives. Because of the laboratory's unique position in the acquisition and flow of information, its contribution must continue as a planned process--based on accurate assessment of value, cost, and resource inputs--rather than in a discrete, event-driven manner. Two examples of clinical pathway development, partly given by the laboratory, illustrate its application in nutritional screening and monitoring for metabolic support and in emergency department admitting decisions on myocardial damage at Bridgeport Hospital (Bridgeport, CT). The paper suggest two activities, point-of-care testing and screening for fetal lung development, that could benefit considerably from clinical pathway information shared across the experiences of several institutions.

Amniotic fluid polarization of fluorescence and lecithin/sphingomyelin ratio decision criteria assessed.

A negative finding of amniotic fluid (AF) phosphatidyl glycerol (PG) does not eliminate the need for determining the lecithin/sphingomyelin ratio (LSR). We use a novel approach to classify fetal lung maturity (FLM) data, and to validate the fluorescence polarization (FP) surfactant assay (Abbott), which replaces the PG assay and reduces the frequency of repeat LSR. This method finds the values (decision points) of these tests that allow for classifying the data with least errors. These tests best identify the risk of respiratory distress syndrome (RDS) from fetal lung immaturity. We find the decision values for tests by exploring the data for information content and optimize their selection using group-based reference. We previously defined normal reference as the maximum entropy set with no information. The uncertainty resolved by information provided in the data allows formation of syndromic classes. This is greatest at the values for the variables (decision-points) associated with the greatest decrease in entropy. Decision-values found for PF, EGA, PG, LSR that classify amniotic fluids into the mature and not-mature classes are in agreement with the results of ROC analysis. We validate the replacement of PG by the PF method. We also find a level of FP below which LSR might be required to resolve uncertainty and above which the FP indicates maturity. We confirm the ability to evaluate fetal maturity methods using information analysis.

Cholecystokinin-stimulated monocytes produce inflammatory cytokines and eicosanoids.

OBJECTIVES: Plasma cholecystokinin increases with enteral feeding. Cholecystokinin increases intracellular calcium in lymphocytes/monocytes and is a lymphocyte co-mitogen. We hypothesize that decreased cholecystokinin production with "bowel rest" and parenteral nutrition may be beneficial in inflammatory bowel disease by down-regulating gut immune/inflammatory mechanisms. The majority of cells observed in mucosa of inflammatory bowel disease are monocytes and neutrophils. Cholecystokinin effect was therefore measured on monocyte production of proinflammatory mediators (tumor necrosis factor alpha, interleukin-1 beta, interleukin-6) and neutrophil chemotaxins/activators (interleukin-8, granulocyte-macrophage colony stimulating factor, and leukotriene B4). METHODS: Peripheral blood monocytes (0.5 x 10(6)) from healthy donors in 1 mL of RPMI 1640 plus 5% fetal calf serum were cultured for 24 h in 5% CO2 at 37 degrees C with 5 micrograms/mL endotoxin, 1 x 10(-7) M cholecystokinin, or no agonist. Supernatants were analyzed by ELISA for cytokines and leukotriene B4. RESULTS: Endotoxin-stimulated monocytes produced 1130 pg/mL tumor necrosis factor versus 81 pg/mL for cholecystokinin, 612 pg/mL interleukin-1 versus 10 pg/mL, 694 pg/mL interleukin-6 versus 30 pg/mL, 4531 pg/mL of interleukin-8 versus 3848 pg/mL, 21 pg/mL granulocyte-macrophage colony stimulating factor versus 9 pg/mL, and 21 pg/mL leukotriene B4 versus 12 pg/mL. Controls produced no cytokines/eicosanoids (N = 8, p < 0.001). CONCLUSION: Cholecystokinin increase with enteral feeding may up-regulate gut immune response. Cholecystokinin suppression with parenteral alimentation may decrease inflammatory mediator production.

Bridgeport Hospital autologous blood donation experience from 1992 to 1996.

The safety of the blood supply, an issue in the 1970s and 1980s, created an increased need to screen the blood supply for HIV-1 and hepatitis C virus infections. The possibility exists that other contamination could again affect the blood supply. This has resulted in the increased use of strategies to minimize the transfusion of allogeneic blood, such as autologous blood predeposit for elective surgical procedures. Many studies indicate, however, that autologous blood donation is overutilized so that half of the blood withdrawn for autologous use is discarded. Cost-effectiveness studies have indicated that autologous blood donation has little benefit compared with many medical procedures, from which one might conclude that the procedure could be eliminated. Alternatively, the benefit could be improved by reducing the wastage of autologous donated blood. This wastage must occur only because of a premise that autologous blood is obtained to ensure avoidance of a homologous transfusion. This results in an amount of blood withdrawn that is more than is used in an uncomplicated procedure. We examined the transfusion requirements in surgical procedures for which there is autologous blood donation to establish the optimum amount of blood to be taken based on expected blood use. The transfusion records of 493 patients who donated blood preoperatively (340 orthopedic, 69 urological and 83 gynecological, in the years 1992 and 1993) were audited to determine the characteristics of the transfusion practices associated with the surgical procedures. The study sample underwent 182 total knee and 123 total hip arthroplasties, 33 laminectomies with fusion and three without, 83 hysterectomies and myomectomies, 59 radical retropubic prostatectomies and 10 nephrectomies and lymph node resections. Data used for evaluation were age, sex, units donated and transfused, predonation hemoglobin concentration, initial and final hemoglobin concentration, surgical procedure and surgical blood loss. The study suggests that autologous predeposit is not indicated for hysterectomies because of the low likelihood of transfusion. Even when a transfusion is likely according to the surgical blood order schedule, predonation is greater than actual use. Use of predonation hemoglobin could facilitate better efficiency of use for procedures where use is anticipated, thereby significantly reducing a wastage near 50 percent.

Financial implications of malnutrition.

PEM or the possibility of developing PEM occurs in 30% to 50% of hospitalized patients, the frequency determined by the criteria used in its assessment and the case mix of patients in the hospital population. This condition exists independently of other medical conditions and results from preadmission or postadmission failure to meet nutrient requirements with associated loss of body weight and function, as well as impaired immunity. PEM also frequently arises in patients with a chronic condition and decreased functional reserve when a superimposed acute metabolic stress leads to accelerated nutrient depletion. Whether preexisting or not, PEM increases morbidity and mortality along with LOS and may be associated with complications such as pneumonia, sepsis, operative site infection, delayed wound healing, or decubitus ulcers. The cost of these complications and an extended LOS is a significant financial burden and a controllable medical liability for hospitals. Other costs include identifying patients at risk of PEM, providing nutrition support, not to mention treating any of its complications (mechanical, metabolic, and so forth). A proper analysis of the financial implications of late or untreated PEM versus nutrition support must therefore take into account not only the costs of complications or extended LOS due to the delay or failure to provide nutrition support but also the costs associated with this intervention itself. In this review, we described a model for examining the financial implications of malnutrition and nutritional therapy.

Effect of nutrition status and other factors on length of hospital stay after major gastrointestinal surgery.

Various factors may prolong postoperative recovery and the length of stay (LOS) in the hospital. In a retrospective study of 245 adult patients who had no oral intake for 5 days after major gastrointestinal surgery, we used a correlation matrix to describe the population and determine the effects of the following factors on LOS: malnutrition, complication status, stress level, type of surgery, pathology, period of inadequate nutrient intake, and use of nutritional support. LOS was markedly prolonged in malnourished patients compared with those who were not (23.5 +/- 16.5 vs. 16.5 +/- 10.7 days, means +/- SD, p < 0.001). Patients were then grouped into those who had nutritional support and those who had not, and a nutrition classification was derived by examining the uncertainty (entropy) in the data matrix that allowed separation of the population into distinct groups. Nutrition and complication status and days without oral nutrient intake were discriminative. Analysis of variance and multivariate studies were also used to determine whether the presence of malnutrition, complications, both together, or neither could predict LOS and to determine the confounding effect of nutritional support on LOS. A significantly extended LOS persisted for patients with malnutrition or complications and was most prolonged for those with both, but patients who received nutritional support had a greater LOS than those who did not. In addition to the effects of nutritional support, malnutrition, and complication status, LOS correlated with the duration of the postoperative period without oral nutrient intake. We therefore recommend systematic and early nutritional intervention for selected gastrointestinal surgical patients.