Current Issues

Even with the fast promotion of the electronic health records (EHRs), the old fashioned mouse and keyboard data entry still gives a problem to most of the clinicians. The alternative to nursing notes production is speech recognition (SR) technology that is subject to scrutiny regarding its accuracy in comparison with handwrittendocumentation that does not always correspond well to it. The study is an interventional study that compared and reviewed the spelling errors in nursing admission notes that were created using online, offline and a traditional handwriting process of SR. A sample of 35 nurses was involved in two separate groups, where they recorded patient admission notes using each of the three methods in a randomized order, with at least a onemonth interval in between the alternation of the methods. Tailored SR vocabularies were made to minimise terms that are not understood. The number of errors was established both prior and subsequent to correction by the user. Findings revealed that online SR had an accuracy of 96.4 whereas offline SR was a little bit better. The initial errors per report were higher in both SR methods as compared to handwriting (online: 6.76, offline: 4.56). Online and offline error rates of a corrected set plummeted (by 94.75 per cent and 97.20 per cent). Online SR brought the greatest amount of incorrect reports. Though both of the evaluated systems were comparatively precise, they did make a bigger number of errors as opposed to handwriting, which shows that a system should be improved and optimized. Documentation errors were tremendously brought down by user review and correction.