DDIs can cause adverse drug event in ICU population. DDIA system have indicated their potential to decrease DDIs. However, override rate of DDIA in ICU setting is high(12). Therefore, the primary purpose of this study was to design, development and evaluation of serious DDIASs in the ICU. To our knowledge, this is the first study to develop and implement a DDIAS in the CPOE in the ICU of an Iranian hospital. DDIAS interface contents included how the physician reacts to alerts that recommend ordering or canceling current medication against keeping or discontinuing existing order, or overriding drug alerts (order current drug and keep existing order). This study demonstrated that the acceptance rate of triggered DDIAS rates is high and the usability level of the DDIASs is in reasonable level.
In this study, knowledge-based included severe DDIs for ICU setting. In contrast, some studies showed that knowledge-based of routine DDIASs contains any potential DDI without intensity leveling (32). Our results are consistent with Paterno et al. study (33), which showed that classification of DDIs by severity increases compliance rate (33). Also, Knight et al. indicated that removing moderate severity DDIAs decreases the override rate, albeit they remained high (34). There are several potential reasons for the low rate of overridden DDIAs in our study; Firstly, triggered DDIAs at the exact point and time of prescribing of the order by the physician in CPOE system. Secondly, actionable recommendations when the physician encounters the serious DDIAs. Third, an ICU administration and team interested in using health information technology in point of care (35). Fourth, the already working structured CPOE system that has been implemented since 2015 in the ICU of Nemazee general teaching hospital. Therefore, health care providers are familiar with CPOE, and all medication orders were entered electronically by the physicians. Fifth, in this study we used Node.js and MongoDB for implement DDIAS that improved the speed of DDIA system. Sixth, grading DDIAs may improve the acceptance of this system (36). Seventh, in this system development process, we considered five right (the right information, to the right person, in the right CDS intervention format, through the right channel, at the right time in workflow) of CDSS (37). Eighth, our system was its integration into a homegrown CPOE and its especial design for the ICU setting. In contrast, some studies indicated that overridden rate of DDIAs remains high (range from 40.1% to 96.2 %) (12, 14–16, 38–43). However, most physicians actions over the overridden recommendations were appropriate (39). Even though a small fraction of serious DDIAs were overridden in this study, some overriding of DDIAs are acceptable in the ICU because patients are closely observed (13, 44). The significance of these findings is that the serious DDIAs and actionable recommendations contribute to reduce the overridden rate of DDIA in the ICU (14, 45). The clinicians should not follow DDIAs recommendations without considering the total risk-benefit balance for a DDIs. Here, severe DDIAs have been stored in an electronic database. However, without this system, DDIs collection data could be incomplete. Therefore, a DDIA system is imperative for decreasing drug error events. The results of this study could be considered as a solution to reduce the rate of overriding of DDIAs.
The usability of the DDIASs was at a reasonable level. Therefore, users like to use safe and usable warning systems in the field of DDIs. The results of our study are in line with Kavumas et al. (46), which showed the high usability levels of EMR systems implemented in sub-Saharan Africa. On the contrary, some reports showed lower usability rates of the EMR (47, 48). An important implication of these findings is that end-user’s participation in designing the DDIASs interface may play an essential role in improving the usability of this system. Also, in this system, consistency medication terminology, visual cues for severity rate of DDIs, and minimal text was used that may impact on improving on usability of DDIAS(49). Nevertheless, the usability of DDIASs is not excellent. Clinical information systems(CIS) have some unintended consequences, including workload, ignoring the role of the clinical pharmacists, alter in the authority structure, and overdependence on the technology(50, 51).
The most important limitation of this study was the sample size. This system brings up a novel approach to DDI that could not be evaluated by our team in a large population as there is no mandating rule for use of DDIAs in the health systems in Iran. In this study, adverse drug events due to override-triggered DDIs were not investigated because of the constraining time and source of the project. We also must consider that our study could have been affected by location bias and suffered from a lack of randomization design.