A Probabilistic Model for Reducing Medication Errors

Abstract

BackgroundMedication errors are common, life threatening, costly but preventable. Information technology and automated systems are highly efficient for preventing medication errors and therefore widely employed in hospital settings. The aim of this study was to construct a probabilistic model that can reduce medication errors by identifying uncommon or rare associations between medications and diseases.Methods and Finding(s)Association rules of mining techniques are utilized for 103.5 million prescriptions from Taiwan’s National Health Insurance database. The dataset included 204.5 million diagnoses with ICD9-CM codes and 347.7 million medications by using ATC codes. Disease-Medication (DM) and Medication-Medication (MM) associations were computed by their co-occurrence and associations’ strength were measured by the interestingness or lift values which were being referred as Q values. The DMQs and MMQs were used to develop the AOP model to predict the appropriateness of a given prescription. Validation of this model was done by comparing the results of evaluation performed by the AOP model and verified by human experts. The results showed 96% accuracy for appropriate and 45% accuracy for inappropriate prescriptions, with a sensitivity and specificity of 75.9% and 89.5%, respectively.ConclusionsWe successfully developed the AOP model as an efficient tool for automatic identification of uncommon or rare associations between disease-medication and medication-medication in prescriptions. The AOP model helps to reduce medication errors by alerting physicians, improving the patients’ safety and the overall quality of care.

An automated technique to identify potential inappropriate traditional Chinese medicine (TCM) prescriptions

“…After validating the AOP model, the results showed 54.66% sensitivity, 86.98% specificity, 97.08% PPV, and 19.48% NPV by using threshold that was equal to 1. Because the pharmacology and indication of TCM are not as clear as for western medicine, the outcome of detecting TCM may be not as good as detecting for western medicine where the NPV was 43.1% …”

“…Because the pharmacology and indication of TCM are not as clear as for western medicine, the outcome of detecting TCM may be not as good as detecting for western medicine where the NPV was 43.1%. 18 There was also one reason that might explain the reason of low NPV. If there was one TCM that could not be explained by disease or the other TCM, that particular prescription would be identified as potential inappropriate prescription by the AOP model.…”

“…The AOP model was developed and presented in our previous study, which was able to identify the appropriate and inappropriate prescriptions by following these rules: The number of positive DTCMs and positive TCMMs should be greater than or equal to the number of TCMs. All diagnoses should have at least one positive DTCM. Each TCM should have at least one positive DTCM or TCMM. …”

“…The AOP model was developed and presented in our previous study, 18 which was able to identify the appropriate and inappropriate prescriptions by following these rules: It is expressed mathematically as in the following discussion. where n is the number of diagnoses, m is the number of TCMs, Q D i ÀTCM j is the Q-value of disease and TCM, Q TCM j ÀTCM k is the Q-value between different TCMs within the same prescription, and α is the threshold of the model.…”

“…Nowadays, there is already much research going on by using data mining techniques in the healthcare or drug safety research . Some studies claimed a good performance for identifying the association between western medication, laboratory data, and problems by analyzing structured electronic health records data . To our knowledge, we could not find any research study that has been done to identify potential inappropriate TCM prescriptions.…”