Learning from errors

An observation of the results of cross-validation revealed that all but one of the compounds in the dataset had been modeled pretty well. The last (31st) compound behaved weirdly. When we looked at its chemical structure, we saw that it was the only compound in the dataset which contained a fluorine atom. What would happen if we removed the compound from the dataset The quahty ofleaming became essentially improved. It is sufficient to say that the cross-vahdation coefficient in-CTeased from 0.82 to 0.92, while the error decreased from 0.65 to 0.44. Another learning method, the Kohonen s Self-Organizing Map, also failed to classify this 31st compound correctly. Hence, we had to conclude that the compound containing a fluorine atom was an obvious outlier of the dataset. 

Log D predictions are more difficult as most approaches rely on the combination of estimated log P and estimated pK. Obviously, this can lead to error accumulation and errors of 2 log units or more can be found. Some algorithms, however, are designed to learn from experimental data so that the predictions improve over time. 

A complication arises. We learn from considerations of multiple regression analysis that when two (or more) variables are correlated, the standard error of both variables is increased over what would be obtained if equivalent but uncorrelated variables are used. This is discussed by Daniel and Wood (see p. 55 in [9]), who show that the variance of the estimates of coefficients (their standard errors) is increased by a factor of... 

It is also salutary to note figure 2, which reminds us that agreement and correctness are not always linked. [This figure is from the on-line dBase of particle properties http //pdg.lbl.gov.] Systematic errors always exist, and may be much larger in amplitude than expected. In general, deducing from uncertain data that a model is acceptable is not useful scientific progress. One learns from the failure of models, not from their successes. 

All the objectives presented in Box 11.1 can be described as quality improvement in the structure and process to support improvement in each patient s health outcome. This support the patient medication care process presented in Fig. 3.1. Various problems and tools and models for improvement have been described in this book. Another approach to prevent medication errors and to improve care is to be open and continuously learn from mistakes. The basis for this is not to punish health care providers who make errors, as this may lead to less reporting of errors. In Britain, the government has taken steps away from this blame-culture (Wise 2001). In a declaration it is stated that honest failure should not be responded to primarily by blame and retribution, but by learning and by a drive to reduce risk for future patients. 

As we increase our knowledge and hope to learn from our past errors, it becomes important that we as a professional group servicing society continue our efforts to (1) minimize and reduce the uncontrolled and accidental release of petroleum hydrocarbons and organics into the environment, (2) improve our understanding of the overall behavior of these compounds in the subsurface, and the health risks associated with their presence, and (3) continue to develop sound strategies for the recycling, remediation, and restoration of impacted soil, water, and air. 

Psychoactive plants have been a part of human life since our beginning. Our first experiences with them probably came from foraging among plants for food. Through trial-and-error learning and behavioral reinforcement, experience has shaped our use of herbal drugs. Similarly,... 

The extension of methods for the prediction and prevention of harm (including audit and learning from errors of the past) and the investigation of patients concerns. 

The preparation for physical chemistry laboratory is a little harder to judge. Today s students seem less experienced in working with their hands than students of a few decades ago, except in the area of video games. I have used an initial experiment in which the students were required to carry out three simple measurements after assembling simple apparatuses, a technique that I learned from Ed Bair at Indiana University. Close observation of the students as they carry out such an experiment gives some information about their aptitude. This experiment also provided the opportunity to discuss data reduction, error analysis and report writing at the beginning of the course. 

The ability to innovate also involves learning from past mistakes, not just one s own. Mistakes and errors in practices can be both costly and dangerous but can be prevented from occurring successively if their causes are determined. This can be difficult because a near-miss scenario can be seen either as an infrequent event or as an averted disaster. The first reaction to this is usually to continue without changing current practices, allowing for similar mistakes to occur. Learning requires all levels of an organization to participate and create channels of communication to innovate effectively, as the inability to share experience denies new opportunities [13,29-40]. 

Multidisciplinary educational programs should be developed for health care personnel about medication error prevention. Because many errors happen when procedures are not followed, this is one area on which to focus through newsletters and in-service training. It also is important for pharmacy staff not just to focus on their own internal errors but also to look at other pharmacies errors and methods of prevention and to learn from them. Organizations such as the ISMP, USP, and many others provide ongoing features to facilitate these reviews in publications such as Hospital Pharmacy, Pharmacy Today, U.S. Pharmacist, and Pharmacy and Therapeutics or newsletters that report on current medication safety issues and offer recommendations for changes. 

Accidents are very rare relative to the number of near accidents and human errors. Fortunate as it may seem, this poses a real problem for complex systems with a high catastrophy potential (nuclear power plants, chemical plants, commercial aviation) few accidents means few cases to analyse and hardly any feedback to learn from. This leads to the undesirable situation of ad-hoc corrective measures after each single accident, because the database is far too small to generate statistically sensible preventi ve measures. 

There certainly is an art to developing a therapeutic relationship. The experiences of numerous practitioners functioning within various pharmaceutical care demonstration projects, as well as those who have built practices through trial and error, provide us with insights regarding some of the basic dos and don ts for establishing a therapeutic relationship. Some of the lessons learned from caring for thousands of patients are discussed in the section on the patient as a holistic system. 

The Medication Error Reporting Program (MERP) is a voluntary program administered by the U.S. Pharmacopeia (USP) in conjxmction with the ISMP. This confidential reporting system improves patient safety by alerting practitioners and the industry to potential or actual problems. Practitioners are asked to report errors and near misses to this program so that others learn from errors and prevent similar errors in the future. 

Learning from experienQ --04e.your owni -" experience of good and bad leetufers to - shape-your performance- S6me of the -more common errors include ... 

---