DISCUSSION
The main finding of this study was recognising the importance of optimisation following, and even before, implementation of EHR systems. We found there were overflowing requests largely related to increasing efficiency of EHR after implementation. This need to make EHR more efficient and usable is real, as evidenced by failing design and usability of implemented EHR systems.17–20 Improvement does not necessarily follow implementation, contrary to an assumption that a smooth go-live will automatically make clinicians’ jobs easier and subsequently improve clinical outcomes.10,21–24 To some extent, there are immediate benefits after implementation, but in practice, it takes an effort of cultivation to ensure such promised results are actually realised.9–12,25,26 Optimisation is a hallmark of successful implementation as McAlearney et al.2 discovered. From the evaluation standpoint, in accordance with Hadji et al.,27 this refinement of the system by optimisation becomes a major determinant of user satisfaction.
The diffusion of innovation theory by Rogers28 is commonly used as a framework to study adoption of innovations related to information technology. According to this theory, there are four main determinants of success of an innovation: communication channels, the attributes of the innovation, the characteristics of the adopters and the social system. In our findings, we see that some respondents would be considered ‘early adopters’ of the optimisation approach (the innovation in this context); an organisation had started the optimisation process even before their EHR implementation phase. The EHR system was the largest financial investment this organisation had made, and ROI was a significant driver in adopting the optimisation approach as early as possible. The ‘relative advantage’ of the optimisation, namely the degree to which an organisation perceives benefits by adopting the innovation, is one of the user-perceived qualities that define an innovation and can affect adoption.28 In this case, the expectation of and commitment to anticipated benefits led to early adoption of optimisation. As Moore and Benbasat29 point out, the more an innovation can address the needs and expectations of potential adopters, the greater the likelihood for adoption. Our findings demonstrate that organisational commitment to maximising benefits from the implementation of an EMR may indeed affect the timing and extent of optimisation efforts.
Second, the present study highlighted the importance of dedicating resources solely for optimisation, validating Cooley et al.30 request not to underestimate the resources necessary to support computerised physician order entry after implementation. Dedicated resources were the biggest facilitator and the second biggest barrier to optimisation. The participant organisations that devoted resources exclusively to optimisation had seen great returns. Cost of dedicating resources was relatively small in comparison to a typical huge investment in implementation. It was largely cost of human resources – a team of about five experts. After go-live, organisations tend to be occupied with maintenance requests, ‘putting out fires all the time’, thereby consuming most of their limited resources (P9). These types of requests are usually urgent because something is broken, but they are not necessarily the most important from an organisational standpoint. Even if there is attention to optimisation, it may not be optimal because in the end, the same people work both in the maintenance and optimisation. The staff may experience divided attention and burnout. In order to move optimisation opportunities to the next level, dedicating resources and staff are required. Optimisation should ideally be separated from maintenance and support of EHR.
Third, this study found there are barriers and facilitators specific to optimisation. Most of the identified barriers and facilitators, for example, resistance to change,31,32 engaging leadership and end users,33,34 importance of informatics professionals33,35 and interdisciplinary committees30 are well documented in many studies.30,32–34,36–41 Despite this overlap, we uncovered additional barriers unique to optimisation such as a bureaucratic process requiring multiple layers of approval for change, poor communication and lack of standardisation. We also appreciated and highlighted those documented success factors in the optimisation perspective, particularly regarding dedicating resources.
Fourth, this study revealed an outcomes-driven approach to clinical decision support. Drawing from the insight of a participant’s experience, clinical decision support should drive actual clinical outcomes (P5). Conventionally, clinical decision support consists of triggering many alerts and sometimes hard-stops, which interrupt the workflow of clinicians and cause alert fatigue. Mandated hard-stops dictate what clinicians should do, leaving them feeling less autonomous and dissatisfied. The ultimate goal is to ensure that clinical care is delivered at the right time, not just reminding clinicians of doing it. Actual clinical intervention, not a reminder, is the closest proxy to a desired clinical outcome. Busy and hardworking clinicians are drowning in a flood of useless alerts. Alerts should be refined and wisely integrated into the clinician’s workflow to drive actual intervention.
Fifth, this study provided insights on defining EHR optimisation. Participants recognised optimisation as an ongoing process improving EHR systems. We noted there were two different approaches to optimisation. One was user-driven. Starting from issues or requests, it continually brought refinement and enhancement to the system, thus, making it more useful and efficient. The second was organisation-driven; it was an organisational project, not an IT project. It was truly about optimising processes, practices, workflows and performance of the organisation by leveraging the implemented EHR system. It was an intentional commitment to realise actual benefits of the EHR implementation. This finding is comparable to a conceptual understanding of EHR optimisation noted in the literature in the absence of any direct definition.3,42 Blavin et al.42 recognise EHR optimisation as ‘continually modifying technology for optimal use as better able to use technology to meet an organisation’s performance goals’ in a comprehensive literature review. Blavin et al.42 acknowledge optimisation as one of four distinct stages of the EHR implementation process, which are ‘planning and vendor selection’ (acquisition), ‘workflow and software design’ (implementation), ‘training and user support’ (implementation) and ‘optimisation and modification’. Notably, Blavin et al.42 discovered a need to examine EHR optimisation with emphasis on outcomes. Drawing from the findings from the study and literature, optimisation of EHR should be defined as an ongoing commitment with dedicated resources to improve the EHR system and realise its benefits to the fullest by achieving measurable outcomes both after and before implementation.
Lastly, this study validated findings of prior studies. There was dissatisfaction among clinicians following EHR implementation. The decreased productivity, burden of documentation, sensory and cognitive overload and increased time required to get things done in EHR systems contributed collectively to clinicians’ widespread dissatisfaction and frustration over EHR systems.5,43–45 The reduced productivity was consistent with previous findings.46,47
The study findings present numerous implications. Recognising a strong demand to improve efficiency of EHR systems as evidenced by surging requests, the healthcare organisations planning on, in the midst of implementation or post go-live should allocate adequate resources to optimisation. Despite limited resources, even small healthcare organisations could still make an intentional commitment to optimisation within their capacity. Ideally, a dedicated optimisation team should be established. The organisations should also cultivate an optimisation-friendly environment by eliminating or mitigating the barriers to optimisation while promoting the facilitators. Additionally, during the design and implementation phase, they should ensure more robust engagement of clinicians (e.g. usability test, seeking clinicians’ iterated feedback on EHR system design) to incorporate the clinicians’ optimal workflows and insights into development of EHR system which will result in more refined system. Finally, an evaluation system measuring actual outcomes of optimisation and implementation should be established and employed. Making the EHR system ‘go-live’ is just the beginning. Real success of EHR implementation or optimisation must be measured by its realised benefits.
This study has limitations to consider. First, the collected data may not fully represent the participant organisations’ perspectives, because most organisations were represented by a single participant. Additionally, a 1-hour interview may not be enough time to capture the full experience of a large organisation. However, strong efforts were made to mitigate this risk by referring to publicly available information (e.g. HIMSS Davies Award applications) for validation of findings. Second, a single researcher completed the study interviews, transcriptions and large part of coding and analyses due to inadequate resources, thus, possibly introducing a personal bias and violating the best practice in qualitative research. In order to overcome this limitation, a rigorous study methodology was designed and followed and there was regular frequent discussion among the research team throughout the study. During the coding and data analysis, other team members reviewed the coding, refined development of the complete coding guide and data analysis framework and validated data analysis and interpretation. The team’s detailed involvement is documented in Authors’ Contributions. Third, the sample size is small (N = 15) with a low response rate (20 respondents out of 997 contacted individuals). Although we could recruit more participants with follow-up invitations or phone calls, we intentionally did not pursue it due to practical considerations and also, considering the qualitative nature of the study and the fact that data saturation was reached. Expanding sample size was not feasible, for this study was carried out under very limited resources. Lastly, the research team did not include patients and caregivers who would be an ultimate recipient of EHR optimisation benefits.
Future research should focus on measuring the impact of optimisation on clinical outcomes. This study was not designed to identify a concrete connection between optimisation efforts and actual clinical outcomes for the patient. This is one area that needs improvement and further study.12