Errors In Healthcare Interventions

Drawing correct conclusions regarding the safety and effectiveness of healthcare interventions requires data to be as accurate as possible. Moreover, systematic errors in data can lead to biased results and incorrect allocation of resources (Adler-Milstein and Jha 2013).
There is little evidence that provides guidance for decision makers on how to qualify data sources and evaluate their appropriateness for use in analysis. However, there is recognition of the need to provide guidance on qualifying data sources, which is evident in the 2020 strategic plan of the International Society for Pharmacoeconomics and Outcomes Research (ISPOR). Therefore, it is expected that in the next few years more guidance on this topic will be available for decision

makers.
3.3.2. How can decision makers assess accuracy of data linkages?
Assessing the success of data linkage is a difficult task.

A gold-standard method allows for the comparison of resulting data with an independent source. Given the lack of such method for evaluating data linkage quality, several indirect methods are available for this purpose. These methods, however, can be flawed. The available methods provide sensitivity and specificity estimates that range between 74% - 98% and 99% - 100%, respectively (Bohensky, Jolley et al. 2010). Decision makers and stakeholders should be aware of the possibility of data linkage errors and the difficulty of assessing linkage success.
3.4. Erroneous information.
3.4.1. Data entry errors
Another challenge to electronic healthcare data is erroneous information. One of the sources of errors in healthcare datasets is data entry errors. In healthcare systems, most data is entered by practitioners and staff, which makes the data prone to human error. Data entry errors can result from direct mistakes in transferring the data to computers, or from issues related to computer use (De Lusignan, Liaw et al. 2011). For example, using different computer softwares leads to inconsistent data entry practices and scoring systems (De Lusignan, Liaw et al. 2011).
For these reasons, some measures are suggested in the literature to improve the quality of data entry (De Lusignan, Liaw et al. 2011). Moreover, it is suggested that electronic healthcare data should be validated before its use in research (Bayley, Belnap et al. 2013).
3.4.2. Data linkage

errors
Linking data related to the same individual from different sources provides integrated and more complete data about patients. However, several types of errors can take place during the linkage process, which leads to erroneous data and limited utility of studies’ results. These errors include failing to link data related to the same person or mistakenly linking unrelated data (Bohensky, Jolley et al. 2010).