The serum digoxin test and digoxin toxicity: a Bayesian approach to decision making

The clinician may often be uncertain about the presence of digoxin toxicity. This uncertainty is particularly important when the clinician must make initial therapeutic decisions about continuing or discontinuing digoxin. We describe a method that helps to clarify the role of the serum digoxin test in decreasing the uncertainty surrounding the diagnosis and treatment of toxicity. The relation between the test and toxicity was first determined in our patient population. An approach to the interpretation of the test based on the likelihood ratio was then developed by combining our data with selected data from the literature. The relation between the pretest risk of toxicity (the estimated risk of toxicity in the population under investigation before the test result is known) and the predictive value of the test was established. This relation was also used to analyze the importance of the degree of elevation of the test. The appropriate threshold probability for institution of treatment of toxicity was then determined by an interview technique. The test was able to make the patient's probability of toxicity cross the threshold probability for treatment of toxicity for an intermediate range of pretest risk. Our analysis suggests that the serum digoxin test may have a critical effect on therapeutic decisions and can be best considered as contributing to the spectrum of risk.