Medical records

A pilot study of patient-generated electronic medical histories in primary care

Description of Instant Medical History

Instant Medical History (IMH) is a computer-based artificial intelligence program, developed for primary care in the USA (1). We used a version with vocabulary modified for use in the UK.

IMH is backed by a battery of 15,000 branching questions and clinical scales, enabling patients to provide a self-directed medical history. This is used to generate a clinical summary and indicate any relevant diagnostic tests, before the consultation takes place.

It has been shown to be a useful tool in the USA, with high patient acceptability (2). A clinician based in a ‘family practice residency’, who had worked with the software producers, reported improved records, reduced computer use in consultations and simplified administrative processes (3).

This application seems to have the potential to enable a more patient-centred approach by freeing up consultation time for more discussion between patient and clinician. However, a thorough search of the literature found no references to a systematic evaluation of its possible value to primary care in the UK.

Aims of the study

We wanted to pilot the use of this program in two of practices, ideally recruiting about 30 patients, so that we could explore its acceptability and feasibility in busy surgeries. If this stage was successful, it might eventually lay the groundwork for a larger, randomised controlled trial. Ultimately we hoped to explore the IMH potential for managing chronic diseases and increasing healthcare access for hard-to-reach groups. An early presentation to the Department of Health indicated that it could fit well with the emerging NHS IT 'spine'.

Preparation

We discussed the pilot with all GPs, nurses and administrative staff to ensure that they would be willing to support it. We then met with the Friends of Abbey Surgery, as patient involvement was clearly vital. This was a very useful and positive occasion. Despite some fears about confidentiality and reduced face-to-face time with their doctors, the committee unanimously agreed to support us in every way they could. The practice managers were also entirely supportive. They organised things to minimise disruption to the everyday work of the surgery and set aside space to ensure privacy for patients taking part.

Several people had volunteered to act as 'buddies' for patients using IMH. They then had a session trying out the software and, with their consent, we taped the subsequent discussion. They reported some technical difficulties: a rather non user-friendly format, a failure to recognise some common conditions and some apparently irrelevant or unanswerable questions. But the volunteers were still enthusiastic. They said there was more time to think than in a face-to-face consultation, that it was usable even by older people who had never touched a computer before and that it asked questions they hadn’t thought of.

The trial

Despite this encouraging start, problems were encountered when we started engaging patients.

It had been agreed that patients should be recruited on arrival rather than when they rang to book an appointment, as this would have been complex and time-consuming. All the reception staff had been briefed about how best to help patients with the information sheets and consent forms. However, it did mean there was no prior knowledge about likely activity levels. The pilot was also limited to patients presenting with a new medical problem, since providing histories for ongoing conditions or routine checks would simply waste patient and clinician time.

During the first three days, only four out of 21 patients were eligible for the study and of these, one was pregnant and another was deemed too ill to take part. Generally there seemed to be little interest amongst patients and it was a time-wasting experience for the volunteers. Eventually, one of the practices had to withdraw for unrelated reasons and we decided that it was best to put the pilot on hold until circumstances changed.

Lessons learnt

Ethics: Completing the ethics application demanded the premature delivery of a very detailed protocol for a very small pilot study. This swallowed a substantial part of the available research time (funded by a small grant from the local research network). Rather than a research pilot, it might have been better to have undertaken the first stage as a piece of low-key, exploratory innovation within a single practice.

Patient involvement: The lay ‘buddy’ system was a very useful way of engaging users in the research activity, but it did mean that the history-giving process was very slow. The volunteers clearly needed much better training in how to help patients translate their symptoms (‘earache’, ‘numbness’, ‘muscle weakness’) into a clinical vocabulary. In the American system a healthcare worker is used for this clinical ‘mediation’ process.

However, the volunteers undoubtedly brought an invaluable user perspective to the project. They would also have made a very useful contribution at the stage of writing up and disseminating the report. On reflection, it might have been better for them to have been directly involved in a hands-on steering group at the beginning, mid and endpoints of the study.

Clinical benefit: In all, the GPs saw about a dozen patients. They felt the system had potential but that it failed to add value in its current form. We belatedly discovered that the software was set at sub-optimal levels and was defaulting to a high level of complexity, not best suited to the generic environment of primary care. We also realised that the report print-outs for GPs failed to show negative responses. This gave an incomplete picture of the overall patient history — a particular problem when reviewing patients with complex conditions. The software needs to be re-tested with volunteer GPs and patients, so that all glitches are ironed out before it is used live again.

Use in practice setting: It is difficult to dovetail pre-consultation history-giving with a surgery’s routine appointment system. An alternative would now be to launch an initiative encouraging patients to use the practice website to generate their own medical histories. These could then be emailed to the GP who could authorise any relevant diagnostic tests before the consultation. This option was not available at the time of the pilot, but could increase access for hard-to-reach groups, as well as immobile or housebound patients.

Conclusion

IMH needs further testing. We had a lot of teething difficulties and do not have enough data at this stage to draw any conclusions about its acceptability and feasibility. However, there are some indications that it may be over-sophisticated for day-to-day general practice in the UK and might be better used as an indepth tool.

It could be useful in managing chronic diseases, such as asthma and COPD or reviewing chronic conditions such as depression, where patients could complete relevant clinical scales before a follow-up consultation. Dovetailing the data with electronic records could simplify information management and audit. It may also have future relevance for practice-based commissioning, but research is needed to demonstrate clear evidence of hard outcomes such as reduced hospital admissions.

Getting new ideas/technologies accepted in general practice needs additional resources, time and energy in an already busy setting. However, there has been a rapid increase in the number of patients, at all ages and stages, who are prepared to use new technologies to access healthcare. Some practices are already using texts and emails to make appointments or monitor symptoms. It may turn out that patients now become the drivers of change, along with some enthusiastic clinicians, if we can just give them a chance to see what they might get from it.

Tess Lomax, Devon Primary Care Trust.

Mark Cullen GP, Abbey Surgery, Tavistock, Devon.

Rupert Jones GP, Peninsula Medical School, Plymouth, Devon.

References

1. Bachman JW. The Patient-Computer Interview: a neglected tool that can aid the clinican. MayoClinic Proceedings Jan 2003; 78(1): 67–8.

2. Pierce B. The use of instant medical history in a rural clinic. Case study of the use of computers in an Arkansas physician’s office. Journal of the Arkansas Medical Society 2000, 96(12).

3. Zelnick CJ. Incorporating patient-entered data in your EMR with instant medical history. 2003.