Where is THE medical record?
Markwell D C
The Clinical Information Consultancy, 93 Wantage Road, Reading, RG30 2SN, UK
Introduction
This paper is divided into two parts. The first part is a light-hearted illustration of the potential problems of distributed medical records. The second part is a more serious reflection on the impact of different approaches to electronic sharing of medical record information.
Records, the universe and everything
Carrie Byte sifted through the envelopes on her desk. She had already searched the practice computer without finding anything of interest. Her face was a picture of tired resignation. She was about to see Arthur Dent. She had never met him before, but her senior partner had once told her something about him. She could not remember what it was, but she did recall it was unusual. Without the record she would be at a serious disadvantage.
She lifted the phone and dialled.
"Hello, George. Mr Dent's record doesn't seem to be here."
George apologised and said he would look for it. As Carrie hung up the phone, there was a knock at the door and, almost immediately, in walked a nondescript man.
"Hello, I'm Arthur. I expect that you are Dr Byte."
Carrie agreed that she was, asked what she could do for him, and apologised that his record was missing. He smiled and said that he had been away for a few years.
"Your record is probably with the FHSA", she suggested.
"Perhaps the Vogons mislaid it when they destroyed the planet", he countered.
"What do you mean?" she asked.
"I admit the FHSA is more likely, but never dismiss the improbable."
"Good advice for a doctor - but tell me what I can do for you today", Carrie said, trying to regain control.
The smell of coffee percolated through the door and she wanted to get this consultation finished. After another knock at the door George entered, carrying three thick Lloyd George envelopes.
"I found it, Doctor."
Arthur noticed that Carrie looked more downcast by the size of the records than she had been by their absence.
"Bit of a trilogy, isn't it, Doctor?" he quipped.
She smiled and he continued "Don't panic! I only want to know the results of my last hospital visit. Your partner, Morris Oxford, referred me for an opinion."
Ten minutes later she had waded through letters, printouts from other practice computers and reams of notes and established that the hospital letter was missing. It took ten more minutes for George to find it in Dr Oxford's in-tray.
As Author left the room he turned and said "Just one more thing, Doctor." Her heart dropped.
"I have a friend who has really good ideas about piles of paper. He'd help you find the right information more easily and have time for a coffee break."
"Oh yes", she replied sceptically.
"Yes", he said emphatically. "You must meet him. One o'clock in the Frog and Sparrow, and bring a towel."
He spoke with such authority that she felt obliged to join Arthur and his friend at the appointed time. Arthur introduced Mr Prefect who frowned as he greeted her and said "No towel. You've forgotten the towel. Oh well, I suppose it doesn't matter."
He took from his pocket a small box with the words "Don't panic!" written on it. "This", he said, "is what Arthur was telling you about."
By the end of lunch break she was convinced. A week later her practice had installed the Instant Transfer Computerised Heath Hyper Record. At the heart of this system, known as the ITCH-Hyper Record, was the concept of the improbably distributed record. Few people understood it, but that didn't seem to matter. It promised an end to 'missing record' misery and it really worked. Within three months, every practice, every hospital and every community unit in the country had installed the ITCH-Hyper Record. Clinical information entered anywhere was seamlessly shared by everyone, subject to necessary and appropriate security constraints. The age of information sharing had arrived, dreams of patients outshining paper became reality, the wood was seen, the trees were spared and crocks of gold appeared under every rainbow.
Six months later Carrie started to worry. She was looking at a record which showed a high blood sugar result two days previously. She switched screens to check the advice on her decision support system. She returned to the record and the blood sugar was normal. The first time it happened she assumed she had misread the screen. The next day it happened to another patient's record, and then another. She queried it with the lab. They said there had been some errors in a batch of tests but insisted they were corrected immediately. Then why, she wondered, had she seen the erroneous results nearly a week later? Soon the medical press were full of stories of similar incidents. Ford Prefect explained by saying "The old record is cached and not refreshed until it is re-accessed. It's just a glitch: we'll fix it."
The next problems Carrie noticed were intermittent gaps in records. These were accompanied by ubiquitous "Don't panic!" messages accompanied by icons depicting sporting events. Enquiries to the support desk were met with the patronising response and brief explanation that the tennis racket meant a 'service fault', the rugby ball a 'line out' and the golf club 'open links'. These errors became more common and it was inevitable that before long the system acquired the less flattering nickname "The Glitch".
Just over a year after Carrie's meeting in the Frog and Sparrow, she was the defendant in the first court case in which two incompatible versions of a patient's ITCH-Hyper Record were presented as evidence. Both carried all the marks of authentication as the genuine record for the same patient. The judge asked for explanations. Ford and Arthur attempted to explain about caches, probability, post-dating, modification, repudiation and the importance of towels for galactic hitchhikers. The judge was unimpressed and asked a simple question: "Where is THE medical record for this patient?"
Ford asked for clarification and the judge obliged. "Where are the original records of the clinical information recorded by each of the health care professionals involved in the care of this patient?"
Ford regarded the judge as a child who wanted to open a television to look for the people inside. He made a contemptuous remark and was silenced.
Arthur came to the rescue. "It's like this, I think. It's sort of everywhere. Well, everywhere in the cyberspace defined by thirty thousand clinical systems in the UK. That is, it could be anywhere, but nobody knows where, and of course, nobody really needs to." The judge was silent for half a minute. During those thirty seconds several dolphins spontaneously appeared, unnoticed, in the courtroom. Then the judge spoke very slowly and deliberately: "I need to know and I need to know now. Otherwise, how can I determine which of these pieces of so-called evidence is genuine?" He paused, waiting for a response. As the silence continued, Ford and Arthur vanished, having hitched a lift on a passing Dolphinian space trawler. The silence resumed. The dolphins smiled and vanished, still unnoticed.
Then the judge summed up. "It is claimed that the medical record is everywhere. That it is distributed across what has been called cyberspace. Yet, since we find that in two places it differs, from a legal perspective there are at least two versions. Neither of these can be located and nor can they be separated or distinguished from one another, except by the factual differences in the statements they contain. It follows that I am unable to determine the legal status of either. Therefore, for the purposes of this court, I rule that this patient's medical record is located nowhere. Since something that is nowhere does not exist, I further rule that it is inadmissible in this court. This being the case, can anyone present any evidence that will help this court to reach a judgment?" There was another pause while the judge and others noticed the absence of the expert witnesses. He smiled and added: "I presume that Messrs. Ford and Dent are now distributed in a manner similar to their records. Is there a shred of evidence to assist us with this case?"
Carrie passed forward a page torn from her diary on which she had made a few rough scribbles. Her counsel rose. "If it please your honour, I have here a piece of paper written on by Dr Byte at the time. It appears to corroborate her account." The judge, having studied the paper, passed it to the plaintiff's legal team and asked if they offered any further evidence. After a hasty conversation with their client, they indicated that no more evidence would be offered and the case was dropped.
The profession was not always the beneficiary of this legal precedent. Patients who made their own rough notes immediately after consultations won otherwise absurd cases against doctors unable to cite admissible medical records. Within a few weeks most GPs were reopening their filing cabinets, brushing the dust off their pens and relearning elaborate hieroglyphics. The ancient art of doctors' writing was reborn. Years later, when Carrie retired, she cleared out her desk and found a perfect glass replica of a 3.5" computer disk. She didn't know where it came from, but it reminded her of the heady days of her youth when she really believed paperless practice had arrived.
Facts in the fiction
The tale above is, of course, pure fiction, but dreams of a distributed medical record are not. The idea of enabling all those who may care for a particular patient to share access to a common distributed record is at first sight a world-beater. The advent of the World Wide Web means that many of us are familiar with a system in which boundaries between systems seem irrelevant. With adequate security provisions such a system might provide a patient's medical record in its seamless entirety. We can imagine moving effortlessly between our GP records, laboratory results and specialist notes. Could a distributed record replace all those structured messages that it is such a strain to develop and implement?
The tale of Carrie Byte suggests some problems that may arise if, in our search for improved communication, we forget the importance of control, ownership and origin of data. Present paper and computerised systems implicitly provide physical tokens of control and ownership. Furthermore, even in the worst case they provide circumstantial evidence about the origin and originality of the information they contain. The story also highlights the general risks associated with seductive technical solutions that are poorly understood by users and the communities that they serve.
Reasons for sharing information
Medical records serve many different purposes2. There are also many different reasons for sharing GP medical record information:
The information that needs to be shared varies according to the purpose. For current paper-based GP-to-GP transfers, the entire record is passed. In other cases, a summary relevant to the underlying purpose is sent. A further consideration is the need for structured information. A simple textual structure may be easy to read and stored for later reference. However, information that the receiving system will use in analysis and decision support, must be structured and represented in such a way as to enable efficient access by data retrieval tools.
The timeliness and availability of shared information must also be considered. If the shared information is to be used in the care of patients, it must be rapidly available when the patient is treated and it should be as up-to-date as possible. Information that is required most frequently should be instantly available in the most common situations in which it is required. Delays of a minute or two may be acceptable in less common situations (for example, for temporary residents).
Methods of sharing information
There are several methods by which medical record information can be shared. These are outlined in the following sections:
Each patient has a single medical record. This is stored in one place - hopefully with a backup somewhere else. All those needing to access a patient record use this central resource. Security restrictions are possible but depend upon a secure and accepted way of distributing and updating access permission to different record elements. In theory it would be possible for these records to be held centrally, in regional centres, in a hospital, heath authority or in the patient's registered general practice.
Central storage would be especially vulnerable to attacks on patient privacy. More importantly the central resource and the intervening network would need to support simultaneous access by more than a hundred thousand health care professionals.
Regional or local storage would require a directory system pointing to the current location of a particular patient's record. It would also require a mechanism to support the transfer of records when patients moved from area to area or practice to practice. A practice-based single record would be unlikely to suit hospital specialists and I would not expect many GPs to welcome a hospital-based single record.
The idea of the distributed medical record is simple. Records are made and stored at different points in an interconnected network. These records can all be accessed because they are linked to one another, or because a central directory service indicates where records about each patient are held. When a health care professional needs to see a patient's record, the distributed parts are found and merged to form a single record. This may be a summary with links to more detailed information held on other systems. The main problems with this approach are guaranteeing adequate security while delivering acceptable performance. If the required information needs to be collected from many clinics and practices, this becomes an impossible feat. There are ways of delivering a virtual distributed record. In this case, the distributed information is fetched from other systems in advance rather than when the patient is seen. This is similar to the system of caching used by many Internet service providers. However, caching or mirroring of information produces problems. Records may not be refreshed, so the users may inadvertently be looking at out-of-date information. Worse still, patient records are frequently added to or updated by different people. Over time, inconsistencies may develop as users of two or more systems may modify the same information without reference to one another.
Currently, patients may have separate paper-based medical records in general practice and in several different hospitals. Communication occurs between the holders of these records using forms or letters. Most of these communications are initiated and controlled by the sender of the information. In some situations, a copy of all or part of a record may be requested and provided to another record holder. In general practice the entire record is passed when a patient registers with a different practice. It is understandable that some people see the existence of multiple records as undesirable. It can cause confusion in interpretation of statistics and occasionally unnecessary investigations or treatments may be undertaken because relevant information is in another record. However, while there is room for improved communication, separate records do have advantages. Health care professionals can keep records that they need in a form that is appropriate to their activity. This does not force other professionals to use the same structures and expressions. They can control the information that is passed on to others to protect the confidentiality of the patient. They can simplify the task of the recipient by summarising the relevant information and removing extraneous detail. EDI communications between users of different records offer benefits without disrupting underlying clinical processes and responsibilities. The application of this approach to referrals and investigation requests is well understood4. There are already signs of change in GP-to-GP communication. Many practices retain archives of former patients' computerised records and electronic records transfer will increase this trend. It seems likely that an authenticated archived record will be retained when patient record information is forwarded to the new practice5. This will reduce the complexity of the transfer without undermining the medicolegal status of the original records.
In some specialties, patients have paper-based patient-held records. The possibilities of using computer-readable cards for this purpose have been extensively explored in the past6,7 and work on this is continuing in many countries8,9. Patient-held records provide a means of giving the patient control over clinical communications. They may be particularly valuable when patients seek treatment outside their home area. The cards may either contain clinical information or may contain pointers and access keys that allow the holder to access a secure record over a network. Relying on the health card as the only patient medical record is unsatisfactory. The patient may lose the card and thus be without a record. Population-based studies and preventative care will be impossible unless the relevant information is accessible without the card. Finally, unless health care professionals have access to patient records, they may be unable to defend themselves against claims for negligence. However as part of a combined approach, a card held by the patient can make a significant contribution to rational information sharing.
It makes sense to exploit different methods of information-sharing to meet different needs:
Conclusion
Consideration of any innovation should focus on user requirements rather than the cleverness of a fashionable new technology. There are many different requirements for clinical communication and we should be surprised and sceptical about cure-all solutions. Health care professionals treating the same patient need to share information, but this requirement should be met by methods that reflect the ways in which they work. A truly distributed patient record introduces complexity and risks. Furthermore, other methods of communication may be more effective in meeting user requirements.
It is worth remembering that too much information is as dangerous as too little. Systems that provide selective views of the information may help to remove some of the confusing noise but it may be better to build on tried and tested approaches. The ability of health professionals to summarise key issues of relevance to one another is a valuable skill and cannot be replaced by access to each other's records. Very few consultants welcome the inclusion of comprehensive patient record printouts in referrals from enthusiastic computerised GPs.
So where is THE medical record? There is no single answer to this question but there are answers that are more satisfactory than "everywhere in cyberspace". Patients will continue to have several different medical records. The location, ownership and responsibility for each record should be clear. Furthermore, the fact that information in the record from external sources may now be out of date should be as obvious as a yellowing, crumbled letter in a Lloyd George envelope.
The records held by a practice will, of course, include information communicated from the holders of other records about the same patient. Subject to appropriate controls a practice may also have limited remote access to records held by others caring for the same patient. Members of a primary health care team may share access to a common patient record or summary. However, ownership and responsibility for the integrity of each record must be clearly defined.
When contention between different electronic records occurs - and I'm sure it will - there should be straightforward answers to the question, "Where are the original records of the clinical information recorded by each of the health care professionals involved in the care of this patient?" Otherwise, there is a risk that one day, like Carrie Byte, all we will have left will be a perfect glass disk inscribed "Goodbye and thanks for all the chips."
References
1 Markwell D. Fear of Flowing. In Teasdale S (ed.). Proceedings of the Annual Conference of the Primary Health Care Specialist Group. PHCSG, Worcester 1995:36-42
2 Williams J. Report on the Clinical Computing Special Interest Group of the Primary Health Care Specialist Group of the British Computer Society. Journal of Informatics in Primary Care, 1993 (Apr):11-13
3 Perry A. MIQUEST - rolling out and rolling on. In Teasdale S (ed.). Proceedings of the Annual Conference of the Primary Health Care Specialist Group. PHCSG, Worcester 1995:115-121
4 Pill S. NHS Standard messages for clinical EDI: the real thing at last. In Richards B (ed.). Proceedings of Health Computing 1995. BJHC Books, Weybridge, 1995:262-269
5 NHS Executive. Computerised Patient Records in General Practice: Guidelines for Good Practice (consultation draft). March 1995
6 Hopkins R. The Exeter Care Card Project. In de Glanville H, Roberts J (eds.). Proceedings of Health Computing 1990. BJHC Books, Weybridge, 1990:236-242
7 Markwell D. Patient Held Records in Communications Strategy. In de Glanville H, Roberts J (eds.). Proceedings of Health Computing 1990. BJHC Books, Weybridge, 1990:136-142
8 Pernice A, Doare H, Rienhoff O (eds.). Healthcare Card Systems. IOS Press, Amsterdam, 1995
9 Cirre P, Gallagher M. CardLink Project France and Ireland. In Köhler CO, Rienhoff O, Schaeffer OP (eds.). Proceedings Health Cards '95. IOS Amsterdam 1995:107-109