545 
Dr Ramachandran
Sivakumar
MD, FRCP
Consultant physician, Colchester hospital, Essex, UK
Mr Stephen Wilson
MA(Cantab), MSc
Software development specialist, Ipswich Hospital, Suffolk, UK
Stroke is the leading cause of disability in adults, and stroke patients are treated by professionals from various disciplines. They work as a multidisciplinary team which typically comprises professionals from medical, nursing, physiotherapy, occupational therapy, speech therapy and dietetics backgrounds.
Various outcome measures are widely used. Effective communication with general practitioners (GPs) is of paramount importance to ensure seamless care and this is done through a discharge summary. To ensure that the expected standards are met and to improve the performance, outcome measures are frequently audited internally. Further, these measures are increasingly being closely monitored by commissioners and authorities.
Typically, a discharge summary is dictated by medical professionals and produced as a Word document by a secretary. This results in a significant delay that could have significant repercussions for the patient’s care. Moreover, this often covers the medical aspects only, while the summary of the interventions and progress pertaining to the other disciplines is entirely lacking or is given only a cursory mention.
Outcome measures, when required for the purposes of audit, could often only be obtained by frequently poring through a thick set of old medical records. These records might be in use by other departments and might not be readily available to the stroke team. Such difficulties would often reduce the number of patients analysed in an audit, thereby affecting its quality. Information needed by commissioners and monitoring authorities would have to be collected by this same painstaking process of digging information from old medical records.
Many separate medical specialties
We felt that in medical conditions like stroke, where professionals from various disciplines are involved in care, it is important that all disciplines contribute to a discharge summary document. Such a document would be more comprehensive and be a good source of information for future management. Furthermore, a prospective collection of important outcome measures for all patients would meet the needs of data collection for audit purposes and external monitoring.
We decided to come up with a tool which would enable us to achieve three aims:
- To provide a multidisciplinary discharge summary
- To collect prospectively the outcome measures for audit
- To collect the information needed for data collection.
As we felt it would be difficult to obtain funding for a novel and untested project, we decided to go for a cost-neutral project using whatever expertise existed in-house. The software development branch of Ipswich Hospitals NHS Trust IT Department was able to offer assistance in developing whatever software it was decided would be suitable, and one of the authors (RS) recommended Microsoft Access, as he had some expertise and could contribute to development. The attractiveness of employing Access increased when we considered that the average clinician’s experience with it would maintain the learning curve at a minimum and allow many of those clinicians to compose their own ad hoc queries and reports. Thus, MS Access became the tool of choice for developing the system.
Once we agreed on the tool and were clear about which disciplines needed to have their data included in the discharge summary, we now had to consider the most appropriate design for the application. Our considerations needed to include:
- Restricting access to the application and its data as per information governance requirements
- Defining easily accessible areas of the interface for each discipline to enter its data
- Making data entry as quick and easy as possible to minimise time spent on this
- Developing a clear and concise discharge summary that largely mimicked a standard hospital discharge summary and could be produced at the touch of a button by any qualified user.
With this in mind, we set about the development. We approached the key members of the various disciplines and got input about the required fields needed for their discipline. In particular, we sought suggestions from them on making the data entry as user-friendly and time-saving as possible.
Policing authorisation and access
Although access to the application and its data could be controlled through Access itself (Users and Groups), in this case it was deemed preferable to bypass that and allow our own IT department to police user authorisation by the application of Windows groups to a dedicated share in which the software would reside. The reasons for this were twofold. Firstly, MS Access user groups can be difficult to maintain for non-IT personnel and the IT department would not commit to administering individual systems directly. Secondly, we did not see a need for assigning complex roles among users – read and write categories would be sufficient for our purposes, and the employment of Windows groups would be ideal in such circumstances. An authorised individual’s username and password would be sufficient to open the application.
With the user interface(UI) itself, again there existed two main concerns: to allow the insertion of data to be as quick and painless as possible for the user, and to make the form view of data easy on the eye and readily navigable, since most users tend to look at data via single record form view.
Simple search facility
First, however, we had to provide a mechanism for the user to know whether an individual patient already existed in the database without having to initiate a search. To achieve this, we forced the application to open with a search form containing a single text box and a button. The user enters the patient’s unique identifier into the box and submits it by clicking the button. If the patient exists in the database, the main form opens showing that patient’s demographics and all his discharge summary records; if not, the main form opens with a blank record ready to accept data entry.
The main form interface was designed so that a patient’s demographic details remain permanently in view while all discharge data are entered for each of the five disciplines involved. Thus, an inexperienced user can always be confident that the data they are committing corresponds to the relevant patient. With regard to the insertion of discharge data, the most efficient approach seemed to be to provide a tabbed sheet containing six pages, one each for the five disciplines and one for reports.
Of the five disciplines, some were required to include considerably more data than others, so there was a danger that one or two pages would be too ‘busy’ and thereby confusing and difficult to navigate. To deal with this potential problem, we determined that the more complex pages should contain clearly delineated areas that would comprise the details of a given category, such as risk factors and investigations. Different coloured outlines demarcated the various category areas so that one set was clearly distinguishable from another. The tabbed pages’ approach to the interface produced the added benefit of allowing the whole UI to be visible on a screen without the need for scrolling either horizontally or vertically. For, although we had no qualms about requiring scrolling if necessary, to some users the absence of such a need is seen as an advantage and a time-saver.
Time-saving checkboxes
Since one of the main objectives of the application was to shorten the time spent in producing a summary for the GP, we clearly needed to render the process of inputting data as quick and easy as possible. Therefore, attention was given to providing time-saving methods such as checkboxes (tick boxes) and combo boxes (dropdown boxes) where at all possible. This, of course, proves beneficial not only in terms of time saved, but also for data quality and the more selection fields that can be incorporated the better.
An approximate calculation of selection fields to free text fields reveals a ratio of slightly better than one to one (almost 53% selection fields). In addition, where the opportunity arose, automatic arithmetical calculations (such as Barthel score) were incorporated into the system so that users would not be burdened with the responsibility of calculating the score for themselves.
Again, this also benefits data quality, and the quality of data is of paramount importance for all audits, both internal and external. Contrary to usual choice, we preferred checkboxes over combo boxes for many fields with yes/no options, such as risk factors. User could just check the ones with yes, thereby saving time and we programmed our discharge summary to show up the checked ones only.
Automatic discharge summary
Having considered the provision of an effective interface for data input, we now turned attention to the outputting of that data in the form of a discharge summary to be sent to the patient’s GP. The automatic provision of such a document was seen as one of the main benefits of the whole system. Its advantages would be evident not only in terms of the amount of time saved in its production but also in terms of its accuracy.
Also, the report was to be a pre-set report that could be summoned for any patient at the click of a button on the ‘reports’ tabbed page by any user at their own convenience. Prior to the existence of the system, secretaries had typed a summary for the individual patient, each of which might take as long as 45 minutes to complete. So, assuming an average of 20 transient ischemic attack (TIA) discharges and ten stroke discharges per week, the time saving made possible by the system amounted to saving the salary of a band 4 clerical worker on a half-time equivalent contract, a benefit to the NHS Ttrust of some £10,000 per annum.
The layout of the summary itself more or less replicated a typical hospital discharge summary and so provided the GP with what he expected of such a document. However, judicious arranging of the report’s fields enabled us to separate the medical details, in the form of a first-page letter to the GP, from other clinicians’ additional data, such as those of physiotherapists and dieticians. Some fine-tuning of the presentation was required before all parties appeared satisfied with the arrangement, but, on the whole, very little change was requested or required. In fact, the one-click-summary paradigm proved sufficiently popular to prompt clinicians in some of the disciplines to establish their own dedicated reports.
The ‘reports’ tabbed page is a repository for buttons that run macros to call up, in preview or print mode, reports set up by clinicians for their own purposes. But the system’s primary functions still remain the effective capturing of multidisciplinary data that provide accurate evidence for internal and external audits and supply information in a well-defined format to assist the patient’s GP when the patient is discharged from hospital. Besides achieving all the objectives, it proved to be a useful resource for clinicians for answering any patient-related queries without any delay due to non-availability of old medical records.
This project won first prize in the software category in the competition organised by NHS Innovations East.
In summary, by using our medical and technical backgrounds, we collaborated effectively and developed a cost-neutral project, which doubles up as an electronic discharge summary and data collection tool.
