1 Introduction

The introduction of electronic health records (EHR) has had the biggest information technology impact in health care [1]. The term EHR describes a digital chart including all the medical and treatment history for the patient, which has been designed to be accessed by staff within the organisation which the record belongs to as well as beyond (e.g. General Practitioners) [2, 3]. Numerous EHR benefits have been reported, including better communication and information exchange; decreased errors by medical staff, as well as reduction in time and money [4,5,6]. Healthcare workers also mentioned improved legibility of notes, improved access, use of templates and improved communication and speed [7]. When implemented well, it has the capability to improve safety, perceived job satisfaction, teamwork and overall working conditions of staff involved [8]. However, there are still some apprehensions. Concerns have been raised that electronic notes are more difficult to read and understand due to the excess information included and lack of personal touch, such as ink colour or handwriting, thereby failing to rapidly convey necessary data [9]. Some have also reported difficulty in finding specific notes which was not the case previously [7]. For example, looking for a microbiology consultant note after 10 days of being an in-patient is now more difficult than a clearly labelled Microbiology paper sheet with a familiar handwriting. Strategies to counteract this have been implemented, including various note titles and a search box, but the large number of notes can still make the search difficult. Despite the disadvantages, the transition from paper to EHR has been deemed advantageous to patient care and hospital processes [7].

2 Background

Medical notes should be quick and easy to record, whilst at the same time be structured, retrievable and reliable to assist with clinical decision making [10]. Clinical data and documentation is recorded for four main reasons to support patient care: disease surveillance for public health and research; maintain an accurate record of care provided; payment for the hospital and appropriate remuneration; and for legal defence [11, 12]. The EHR fulfils these requirements; however, it comes with its own drawbacks. The main problems encountered during previous transitions were cognitive overload, hidden information and poor workflow, issues with navigation and lack of trust in the system [13, 14]. Studies have found that these problems resulted in clinician frustration, caused errors in patient care, reduced interaction between patients and clinicians and finally, reduced transition acceptance [13].

A drop in staff morale and clinician burnout has also been mentioned in a number of studies. The exact mechanism which induces burnout due to EHR is not clear, but reports have suggested certain contributing factors, such as increased documentation load [14, 15]; increased time requirement [16]; usability difficulties [1, 11, 17] and fatigue [17, 18]. A recent study uncovered that clinicians who were suffering from burnout reported an unreasonable time spent on admin-related tasks and reported that EHR did not improve efficiency [18]. These clinicians also had reduced personal accomplishment scores, which can be explained by the psychology of motivation [19]. Completion of tasks, which EHR requires, is part of the extrinsic motivation; whereas creativity, autonomy, mastery, sense-making and staying connected are all part of intrinsic motivation, which is what EHR lacks. Due to the intrinsic motivation factors being subconscious and sometimes automatic, they can be overlooked, resulting in increased emotional distress, cognitive workload and lack of sense of fulfilment [19]. This in turn can reduce motivation and cause burnout. Therefore, it’s important to understand the issues surrounding EHR and its implementation, to help clinicians and hospital staff move through the transition smoothly.

So far, seven NHS trusts have made the transition to an EHR system called EPIC (Epic Systems Corporation) – Cambridge University Hospitals, University College London Hospitals, Great Ormond Street, Royal Devon and Exeter, Frimley Health Hospitals and Manchester University Hospitals. However, more NHS trusts are making the arrangements for its implementation. EPIC encompasses numerous other systems used previously into one, to allow the user to gain access to the patient’s notes, past medical records, laboratory results, observations, as well as booking appointments and arranging theatre cases. Additionally, it allows for managerial duties to be carried out, such as reviewing clinic schedules and managing hospital capacity. Patients’ chart, notes and medications can be updated live on the computer or using the corresponding phone/tablet app, ensuring prompt sharing of information. Transition to this system removes the need for the use of paper in the hospital, and ensures all information about patients is accessible under one log-in.

The aim of this project was to understand how the transition from paper to EHR affects staff morale and if this depends on gender, age or job role. This, in turn, encouraged feedback and allowed staff opinion to be heard.

3 Methods

A large NHS Trust moved from paper records to an EHR system EPIC (Epic Systems Corporation) on 10th October 2020, and it was the third trust to undergo the change in the UK at that time. A questionnaire was devised after a patient and public involvement consultation with members of staff (Appendix 7.). Following this, a local research and development approval was granted by the local NHS Trust. The Kubler-Ross change curve was used to monitor staff morale in order to provide quantitative data using the main explanations of differing motivation points (shock, denial/disbelief, frustration, depression, experimenting, feeling more positive and completely happy) [20].

The questionnaire was distributed to the same members of Maxillofacial Outpatients Department (OPD) staff on multiple occasions – weekly for the two weeks prior the change (to gain baseline information), weekly for the first month, bi-monthly for the second month and monthly thereafter for a total of 6 months after EHR implementation. On average, 24.5 responses were received on each occasion, amounting to average response rate of 77%. No staff changes occurred during the study period allowing for the data to remain consistent. The difference in responses occurred when staff were on leave. The responses were collected and typed into a Microsoft Excel (2016) document by the lead author, which was then used for analysis. Quantitative responses were analysed using Microsoft Excel (2016), and ‘NVivo Pro’ [21] will used for qualitative responses, which will be analysed separately.

4 Results

The average number of respondents was 24.5. Job roles of staff in OPD include Trust Doctor (TD), Dental Core Trainee (DCT), OPD nurses, Secretaries, Reception staff, Specialty Doctor and Associate Specialists (SAS), Registrars, Consultants, and other (including medical students and managerial staff). The age, gender and job role distribution can be seen in Table 1.

Table 1 The age, gender and job distribution of respondents
Full size table

From job role point of view, just prior to the change, there was a widespread of feelings (Graph 1). As the transition was taking place, the change curve seemed to be shifting gradually to the right side (Graphs 2, 3 and 4). Interestingly, after week 3, no members were in Shock or Denial/Disbelief stages, indicating a quick adaptation to the system. It’s worth mentioning that nursing staff did not seem to be progressing through the stages between week 3 and week 8. This could be because OPD nurses were not given enough computers, therefore they were not able to practice using the system and their feelings towards it remained the same. Additionally, some of their previous duties were now carried out by the clinician, such as making appointments. By Week 24 the majority of staff were feeling more happy, but only a few were completely happy with the system whilst others were still experiencing frustration.

Graph 1
figure 1

Pre-launch feelings towards the transition according to job role (%)

Full size image
Graph 2
figure 2

Week 3 feelings towards the transition according to job role (%)

Full size image
Graph 3
figure 3

Week 12 feelings towards the transition according to job role (%)

Full size image
Graph 4
figure 4

Week 24 feelings towards the transition according to job role (%)

Full size image

Reviewing the data from gender point of view, before the launch, males seemed to be experimenting a lot more than females, but females felt more positive about the change (Graph 5). After the launch there was a clear difference between the genders with females feeling more frustration (Graph 6). Following week 1, no more gender difference was noted, both genders progressing at the same rate comparatively to each other.

Graph 5
figure 5

Pre-launch feelings towards the transition according to gender (%)

Full size image
Graph 6
figure 6

Week 1 feelings towards the transition according to gender (%)

Full size image

Lastly, from an age point of view, there was a very varied response during the pre-launch period (Graph 7). After the launch, interestingly, the youngest group felt shock, and the rest were in the middle stages (Graph 8). Graphs 9 and 10 show further progression and it appears that the 30–39 year group were progressing through the stages most positively and consistently.

Graph 7
figure 7

Pre-launch feelings towards the transition according to age (%)

Full size image
Graph 8
figure 8

Week 1 feelings towards the transition according to age (%)

Full size image
Graph 9
figure 9

Week 3 feelings towards the transition according to age (%)

Full size image
Graph 10
figure 10

Week 24 feelings towards the transition according to age (%)

Full size image

Finally, approximately 80% of staff reported they did not receive sufficient training for this transition (Graph 11). However, over 80% also reported they would not return to paper notes if given the option (Graph 12).

Graph 11
figure 11

Answers to 'Did you receive sufficient training?’ question

Full size image
Graph 12
figure 12

Answers to ‘Would you go back to paper notes?’ question

Full size image

5 Discussion

Numerous reports have been published on the opportunities and challenges associated with EHR [12, 22, 23]. Units having undergone the transition have reported it took around 2 years for the conversion to happen in order to set up the infrastructure and ensure availability of computers and workstations [7], which was also the case for this transition. The survey was carried out during the first six months after the change and highlights the length of time required to adapt to the new system.

As the trust was the third in the UK to undergo the change, the design of user interface was not consulted with potential end users during the development stage. The authors are not aware whether this was carried out in the first two trusts. This was one of the flaws in the implementation and could have potentially resulted in a smoother transition.

Previous studies on this topic have identified eight factors which improve the usability of EHR: consistency and editing capability, navigation, graphic design, system match to real world tasks, software flexibility and degree of user control, user cognition facilitation, learnability of software and software interaction [24]. If these are optimised, the transition will be smooth. However, organisational change is multifactorial and all aspects would be difficult to predict. We have reviewed common downfalls and suggest potential system alterations in the following paragraphs.

One of the key concerns from staff members was the information overload. The screen tends to be busy and includes objects which the user don’t use, causing cognitive strain [25]. A recent study identified that in some cases the user of EHR has to spend over half their time performing cognitive tasks, resulting in fatigue which can lead to mistakes [26]. It has been suggested that in order to increase efficiency the system needs to have less content, which can be achieved by having fewer dialog windows showing the same information, and reduced number of screens [25, 26]. Additionally, only ‘important’ boxes should be present and the non-important options should be hidden but accessible if necessary [25]. This way the number of steps to achieve an outcome would be reduced, improving the efficiency of the system and reducing cognitive overload. Other suggestions to help with this issue are colour coding the main menu to provide recognition-based assistance [26] and making order status information more visible to reduce the reliance on memory of already completed tasks and orders made [27]. The system should also recognise that sometimes multiple orders are placed based on the same information. To address this, EHR should inherit the information from previous orders to avoid having to re-enter the same information [28]. Lastly, to help the clinician, the system could offer potential disease states or a set of differential diagnoses based on the options selected and the problem description. This way the cognitive load can be reduced even further [1].

Another major issue in EHR systems is the placement of orders. Some studies reported improper medication dosing requests, whilst others published reports about patient overdose due to system errors [29,30,31]. The system can suggest the dosing of the medication, but it’s easy to press a button with different dosing, allowing for human-error missteps. Therefore a suggestion has been made to present drug interactions and conflicts more clearly, and highlight the most commonly used dosing regime [1]. As well as that, a wrong test can be ordered for a patient, or the wrong patient selected for a test. Because of this, some EHR systems have included an additional step for order verification, which includes patient identifiable information and their photograph, significantly reducing the frequency of incidents [24].

One of the key factors to take into consideration when undertaking the transition to EHR is improved presentation of information, which in turn allows faster interpretation of data and encourages more accurate diagnoses [32]. This study identified some flaws which could have been picked up if the users were involved in the design and development of the system, such as having to approve all orders twice, once for activating the order and again on behalf of the consultant in charge, even when the consultant in charge is ordering the test. Additionally, previous reports noted that knowing one part of the software should lead to knowing other parts as well [33]. This means one user action should always result in the same system action, allowing for less learning and minimising errors. However, in some EHR systems, a number of actions can result in the same outcome. For example, there are a number of ways to check blood results, which leads to confusion. Ideally, the interface should anticipate the next necessary step, preventing possible mistakes [34], or even providing a shortcut [33, 35]. This is because some users might not know what the next step is, or what to do when they get there [28]. For example, in the currently used system, in order to book a case for theatre, a ‘clinically safe date’ is a required field after indicating the urgency of the case (< 24 h, < 1 month, < 3 months, > 3 months), but no explanation is given on how to answer this, resulting in data entry errors.

This study identified that six months after the transition staff were still not fully happy with the system. Previously published studies have shown that more than 75% of staff report mis-trusting the system [14]. The suggestion is that training and support are extremely important to ensure the terminology and content are appropriately explained. This study identified that over 80% of staff did not feel they had sufficient training, which was one of the major flaws and should be addressed in future transitions. The training for this transition was brief because the change was happening during the Covid-19 pandemic, causing further distress to already tired members of staff. This could have contributed to the reduced response and satisfaction levels with the system. Data on staff sick leave due to anxiety/stress/depression, acquired employing the Freedom of Information Act 2000, highlighted 2.2% of total trust staff were off sick in September 2020, compared to 2.4% in October 2020 (month when EPIC was launched). The levels went back to normal 2.21% in March 2021, 6 months after the launch. Additionally, lack of feedback from the system can cause frustration as users can be unsure of where the data went after the order is complete, causing further mistrust [27]. This can be rectified by having a clear status information board which is visible to the user. An additional issue to consider surrounds the navigation of the system. Some studies report navigation to be not intuitive and users unsure of where to look for information, causing excessive clicking and resulting in frustration and mistrust [34, 36]. Again, adequate training and support should be offered to the users to resolve this issue.

The authors acknowledge several study limitations. Firstly, the study was carried out in a single hospital department to achieve a high response rate and ensure continuity and consistency of staff throughout the study period. The findings were discussed during hospital-wide Clinical Governance meetings and staff from other departments expressed they felt the same; although it would be useful for future studies to include wider hospital staff. Additionally, the Kubler-Ross change curve might not accurately represent the change of this scale and its impact. Nevertheless, the results identified issues with the transition and training which affected staff morale and should be addressed during future transitions. Furthermore, the study was conducted over the first six months of the transition. As can be seen in the results from the last month, most of the staff were not yet completely happy with the system. Therefore, it is recommended that further studies initiate a longer follow-up period to identify a point when staff become proficient and happy with the new EHR. Notwithstanding the limitations, the findings of this study can act as a pilot to confirm the need for future hospital-wide studies.

6 Conclusion

This study has identified that six months into the transition from paper to EHR, the users are still not completely happy with EHR system. This aligns with other published reports which call for an increased focus on clinicians and users during the design of EHR system [37]. One of the major flaws of the implementation of EHR to this trust was the of lack of training reported by 80% of staff, which should be addressed in future transitions to improve system acceptance and user satisfaction. However, the majority of staff are feeling more positive about digitisation and 80% would not want to return to paper notes. The authors would like to offer the suggestions made in the discussion section to other hospitals undergoing the change in the future.