Application of software design principles and debugging methods to an analgesia prescription reduces risk of severe injury from medical use of opioids.

A prescription is a health-care program implemented by a physician or other qualified practitioner in the form of instructions that govern the plan of care for an individual patient. Although the algorithmic nature of prescriptions is axiomatic, this insight has not been applied systematically to medication safety. We used software design principles and debugging methods to create a "Patient-oriented Prescription for Analgesia" (POPA), assessed the rate and extent of adoption of POPA by physicians, and conducted a statistical process control clinical trial and a subsidiary cohort analysis to evaluate whether POPA would reduce the rate of severe and fatal opioid-associated adverse drug events (ADEs). We conducted the study in a population of 153,260 hospitalized adults, 50,576 (33%) of whom received parenteral opioids. Hospitalwide, the use of POPA increased to 62% of opioid prescriptions (diffusion half-life = 98 days), while opioid-associated severe/fatal ADEs fell from an initial peak of seven per month to zero per month during the final 6 months (P < 0.0016) of the study. In the nested orthopedics subcohort, the use of POPA increased the practice of recording pain scores (94% vs. 72%, P < 0.00001) and the use of adjuvant analgesics (95% vs. 40%, P < 0.00001) and resulted in fewer opioid-associated severe ADEs than routine patient-controlled analgesia (PCA) (0% vs. 2.7%, number needed to treat (NNT) = 35, P < 0.015). The widespread diffusion of POPA was associated with a substantial hospitalwide decline in opioid-associated severe/fatal ADEs.

Predicting in-hospital mortality of patients receiving cardiopulmonary resuscitation: unit-weighted MultiODA for binary data.

Widely used, linear classification functions may be derived on the basis of several different statistical paradigms. Since, regardless of choice, analysis generally yields an equation with unwieldy intercept and attribute coefficients, researchers often construct simpler scoring schemes based on unit weights. Accordingly, for applications involving entirely binary data, we discuss a simple procedure for obtaining unit-weighted (that is, we restrict attribute coefficients to the values of 0, 1 or -1) MultiODA functions that explicitly maximize classification accuracy in the training sample. We illustrate this with an application involving prediction of in-hospital mortality of patients receiving cardiopulmonary resuscitation. In training analysis of 88 patients, unit-weighted MultiODA outperformed prior scoring schemes and logistic regression analysis. Unit-weighted MultiODA also yielded superior hold-out (cross-generalizability) validity for an independent sample of 26 patients.

Predicting in-hospital mortality of patients with AIDS-related Pneumocystis carinii pneumonia: an example of hierarchically optimal classification tree analysis.

A non-linear statistical classification methodology known as hierarchically optimal classification tree analysis (CTA) offers promise of outperforming linear alternatives. We present the first example of CTA in medicine, for an application that uses four attributes (age, body mass index, alveolar-arterial oxygen gradient, prior history of AIDS) to predict in-hospital mortality for a sample of 1339 patients with AIDS-related Pneumocystis carinii pneumonia. We also illustrate use of a hold-out (cross-generalizability) sample that suggests a three-attribute model (alveolar-arterial oxygen gradient, total lymphocyte count, private insurance). Both CTA models achieved approximately 25 per cent of the total possible theoretical improvement beyond chance in classification accuracy (for comparative purposes, logistic regression and regression-tree models obtained in prior research returned less than 10 per cent). We illustrate how one can use CTA models to construct staging systems for assessing severity of illness, and for identifying important directions for future research. In the context of the example, we discuss additional advantages of CTA versus linear alternatives that include treatment of missing data, control of experimentwise type I error, richness of the substantive findings, and ease of use.

Application of multivariable optimal discriminant analysis in general internal medicine.

OBJECTIVE: To illustrate the use of multivariable optimal discriminant analysis (MultiODA). DESIGN: Data from four previously published studies were reanalyzed using MultiODA. The original analysis was Fisher's linear discriminant analysis (FLDA) for two studies and logistic regression analysis (LRA) for two studies. MEASUREMENTS AND MAIN RESULTS: In Study 1, FLDA achieved an overall percentage accuracy in classification (PAC) for the training sample of 69.9%, compared with 73.5% for MultiODA. In Study 2, the LRA model required three attributes to achieve a 76.1% overall PAC for the training sample and a 79.4% overall PAC for the hold-out sample. Using only two attributes, the MultiODA model achieved similar values. In Study 3, the FLDA model achieved an overall PAC of 82.5%, compared with 87.5% for the MultiODA model. In Study 4, MultiODA identified a two-attribute model that achieved a 93.3% overall training PAC, when an LRA model could not be developed. CONCLUSIONS: MultiODA identified: a superior training model (Study 1); a more parsimonious model that achieved superior overall training and identical hold-out PAC (Study 2); a model that achieved a higher hold-out PAC (Study 3); and a two-attribute model that achieved a relatively high PAC when a multivariable LRA model could not be obtained (Study 4). These findings suggest that MultiODA has the potential to improve the accuracy of predictions made in general internal medicine research.

Heart rate variability and susceptibility for sudden cardiac death: an example of multivariable optimal discriminant analysis.

The statistical classification problem motivates the search for an analytical procedure capable of classifying observations accurately into one of two or more groups on the basis of information with respect to one or more attributes, and constitutes a fundamental challenge for all scientific disciplines. Although there are many classification methodologies, only optimal discriminant analysis (ODA) explicitly guarantees that the discriminant classifier will maximize classification accuracy in the training sample. This paper presents the first example of multivariable ODA (MultiODA) in medicine, for an application in which we employ three attributes (age and two measures of heart rate variability) to predict susceptibility to sudden cardiac death for a sample of 45 patients. MultiODA outperformed logistic regression analysis on every classification performance index (overall accuracy, sensitivity, specificity, and positive and negative predictive values). In fact, the worst performance result achieved by MultiODA (in total sample or leave-one-out validity analysis) exceeded the best performance achieved by logistic regression analysis. We conclude that ODA offers promise as a methodology capable of improving the classification performance achieved by medical researchers, and that clearly merits investigation in future research.

Androgyny predicts empathy for trainees in medicine.

Data from 65 medical students and residents support the hypothesis that scores on a measure of androgyny are predictive of those on an index of empathy but relatively modest predictive accuracy was observed for sympathetic responses. Further exploration is suggested.