This Week in BMJ | Editor's Choice | Press releases | Advertisement details

Letters Saturday 22 March 1997

What clinical information do doctors need?

IT supports clinical decision making

Editor,
Richard Smith paints a challenging scene for clinical information systems.1 For many years, Wirral Hospital Trust's information technology strategy has been to support clinical decision making. Wirral is one of the two national pilot sites for the electronic patient record. Junior doctors use the system daily, and requesting pathology and radiology is done through the computer, as is inpatient prescribing.

Two approaches have been adopted. Firstly, we provide information to the clinician when tests are ordered. We have adopted many of the Royal College of Radiology's guidelines and these have been shown to reduce the number of requests for radiology. Using the computerised pharmacy system, we have altered prescribing behaviour and stabilised the drug budget despite an increase of 14% in FCE (finished consultant episode) activity. With the Wolfson Institute in Birmingham we are developing a rules based prescribing system to search the patient's record for specific data and inform the doctor on the safety of the prescription.

The second approach is to deliver structured information to the clinician by using the Path.Finder system, a locally developed information system for general practitioners and hospital clinicians. It has been shown to influence both clinical and referral practice.2 It, rather than the Internet, was chosen as the most efficient means of delivering targeted information to the clinician as the information can be structured and condensed for rapid assimilation and yet reflect local cirumstances. The information is often referenced as evidence based or consensus based and has been mostly derived from colleagues on the Wirral, who have been most generous in their support.

The project is now being developed by 10 other trusts, under the auspices of the British Association of Medical Managers. Each trust will share the Wirral set of information, adapt it to reflect local circumstances, and then report back to the database's national library. We have been able to show that a Read code can trigger the correct page of Path.Finder to open. The project also provides the opportunity for patient groups to share specific information about relevant services such as leisure and health facilities, disease and drug information, advice on benefits, and other patient support groups. Further modules, including multimedia clinician education, are currently under development. We believe this project will provide a solution to many of the issues raised by Smith's article.

T D Kennedy
Director

S Magennis
General practitioner

Cathy Harris
Path.Finder coordinator

Clinical Practice Research Unit,
Wirral Hospital Trust,
Wirral Hospital,
Upton L49 5PE

References

1 Smith R. What clinical information do doctors need? BMJ 1996;313:1062-8. (26 October.)

2 Buchan I E, Kennedy T D. Path.Finder: an interactive clinical information system. Int J Health Care Quality Assurance 1995;8(7):32-5.

Excellent retrieval tools are available in libraries

Editor,
Medical librarians have long understood the problems doctors face in dealing with the questions that arise daily in medical practice. It is perfectly true, as Williamson et al conclude, that "science information management is a critical professional skill that is not adequately taught in undergraduate medical education."1 It is, however, taught in graduate programmes in library and information science.

At this hospital, the doctors who are most often sought out as experts by their colleagues are the ones who come into the library, introduce themselves, and find out what services are available to them. They tell the medical librarian which subject areas are of interest to them, and they take advantage of the library's table of contents service, current awareness database searches, document delivery service, and more. When a question arises in the treatment of a patient, they telephone the library and ask for a literature search.

The medical librarian is responsible for these services and for deciding (with much input from the medical staff) which journal subscriptions to maintain, which books to purchase, and which databases to search. In Richard Smith's description of the characteristics of the ideal "information tool that may transform medicine" the first requirement is that it "must be able to answer highly complex questions and so will have to be connected to a large valid database."2 Not only are medical librarians "connected" online to the databases they use, they have a  clear understanding of how each one is put together, how it is indexed, and how best to retrieve articles on a particular subject.

There also is the human element. The medical librarian gets to know individual patrons, and is able to anticipate their information needs. Often, medical librarians can put a crucially important article in a physician's hands long before he or she would otherwise know of its existence and before it is cited in any database - because the librarian sees it first and knows who will want it. Medical librarians may not be the computers that Negroponte wished for, but they can "know you, learn about your needs, and understand verbal and non-verbal languages."3

I believe that Smith is correct in assuming that there will evolve a "family of tools" to help doctors gather the information they need, but for searching (and sifting through) the medical literature a very good "tool" is already available. Your medical librarian is here to serve you - stop by and introduce yourself.

References 1 Williamson J W, German P S, Weiss R, Skinner E A, Bowes F. Health science information management and continuing education of physicians. A survey of US primary care practitioners and their opinion leaders. Ann Intern Med 1989;110:151-60.

2 Smith R. What clinical information do doctors need? BMJ 1996;313:1062-8. (26 October.)

3 Negroponte N. Being digital. London: Hodder and Stoughton, 1995:92-3.

Sequoia Hospital,
170 Alameda de las Pulgas,
Redwood City,
CA 94062-2799,
USA

Karen W Moody
Medical Librarian

Electronic medical references are being used by practitioners

Editor,
A common conclusion of literature reviews such as Richard Smith's1 is that current information systems are not used because they are not based on users' requirements. In fact, electronic medical reference tools driven by users' needs are being developed and used.

The evidence for this development is more likely to be held by commercial organisations than found in publications. "Mentor", for example, is an immediate use electronic medical reference for primary care team members and junior hospital doctors (jointly developed by Egton Medical Information Systems and Oxford University Press). Evidence based medicine and best practice are incorporated in 2200 succinct, peer reviewed articles which are regularly - sometimes immediately - updated electronically. "Mentor" is linked to patients' records, patient information, protocols, Read codes, and a drug database in more than 2100 general practices serving about 22% of the British population.

Subjective comment on the value of each interaction, requests made to the computer, and the information viewed can be gathered electronically. New material can be reviewed by users, enabling system validation and personal education.2

Recent analysis of 93 481 "Mentor" interactions showed that about a third of general practitioners used the system once per working day (G J Brooks, primary health care specialists meeting, Cambridge, 1996). A quarter of these interactions took place during the consultation. Two thirds of assessment scores indicated that clinical problems were resolved or performance enhanced through using the system. The figures show real usage and perceived value of a passive information system.

My experience, in developing "Mentor", is that some doctors will use familiar text based references and others will use electronic sources if they are faster and more up to date. Ease of use and robustness of software are also critical determinants of source chosen.

"Mentor's" development cycle includes continuous investigation of users' needs and resultant refinement of content, software, and software interconnection. The scope and speed of change render standardised objective methods of measuring performance inappropriate and might, as the system is commercial, account for our inability to attract funded independent assessment with publication of results.

Enormous potential exists for providing clinicians with layered electronic medical references that link local resources available for immediate use with access to more remote information. Smith's vision can be realised in Britain only if academics, the government, publishers, educators, and system suppliers pragmatically collaborate to prevent a further "missed opportunity."

27 Monckton Road,
Gosport,
Hampshire PO12 2BG

Gordon Brooks*
Medical systems designer and developer

*Dr Brooks is an employee of Egton Medical Information Systems.

References

1 Smith R. What clinical information do doctors need? BMJ 1996;313:1062-8. (26 October.)

2 Westerman C, Brooks GJ, Longmore JM. Information overload. BMJ 1993;307:679.

Information lines run by doctors are useful

Editor,
Richard Smith1 clearly points out the main issues about information needs of doctors. Studies show that doctors need information during patients' visits and that they prefer to get answers from colleagues.

To address these problems some Italian physicians set up a private medical information service, "Doctorline," accessible five days a week through a toll free telephone number. It uses computerised databases on CD ROM (Medline, Micromedex-CCIS, Embase), books, serials, bulletins, international formularies, and its own files and is staffed by suitably trained doctors from different specialties.

Since 1991, Doctorline was available to 52,180 Italian physicians, selected according to specialty and provided with an identification card by pharmaceutical companies who sponsored the service (without interfering on the scientific side).

In this period 60,653 calls were received (nearly 12,000 calls a year; 51 per day and 3.6 per physician), of which 24,868 (41%) concerned clinical problems; 10,918 (18%) pharmacological issues; 9,098 (15%) requests for full text articles; 5,458 (9%) specialised centers, congresses, and legislation; 6,065 (10%) service activities and diagnostic and clinical instrumentation; and 4,246 (7%) were follow up calls. General practitioners had the highest call rate (16,840 calls, 28%), followed by cardiologists (10,815; 18%), orthopaedic specialists or rheumatologists (8,615, 14%), gastroenterologists (7,960, 13%), and urologists (5,446, 9%). Dermatologists (3,304), gynaecologists (3,416), clinical pharmacists (1,629), internists (1,364), and chest physicians (1,264) each made less than 5% of calls. General practitioners needed more information on drugs than did other doctors.

The service is used mainly during surgery hours, and 40% of questions concern clinical problems; 20% of questions are related to drugs. Doctors who called Doctorline felt comfortable discussing clinical issues directly with a colleague.

It is difficult to provide doctors with scientifically based answers to "questions generated in consultations"; we have no data regarding if and how these answers can "lead to better patient outcomes or better doctors." We think it is useful to develop independent services, in which trained physicians use electronic information technologies to provide scientifically based answers.

Alessandro Nobili
Senior researcher

Associazione per lo Sviluppo della ricerca in FarmacoEpidemiologia,
Viale Certosa 148,
20156 Milan,
Italy

Gebru Frewini
Doctorline medical staff

A V E Rossetti Doctorline medical staff

Doctorline,
c/o Medical Economics Italia,
Piazza Esquilino 5,
20148 Milan

References

1 Smith R. What clinical information do doctors need? BMJ 1996;313:1062-8. (26 October.)

Few doctors are expert at evaluating information

Editor,
In the inaugural article in the Information in Practice series,1 we were especially pleased to see that Richard Smith used our "usefulness equation"2 to answer clinicians' information needs. We would like to expand on our perspective by commenting on the low usefulness of some information sources commonly used by doctors.

Clinicians often turn to "expert based" sources (colleagues, continuing medical education lectures, textbooks, and standard journal reviews) for new information. As the usefulness formula states:

Usefulness of medical information
=(relevance x validity)/work to access

These sources are potentially useful because the "work" factor to access the information is low, but the validity and relevance of the information that they provide may be in doubt.

Most doctors are good at diagnosing disease and performing procedures because of their accumulated clinical experience, but because critical appraisal has been added only recently to the medical school curriculum, few doctors are expert at evaluating the primary literature - performing a "validity" assessment. As a result, expert based therapeutic recommendations often rely only on clinical experience rather than on a critical evaluation of the available evidence.

More importantly, the "relevance" of the information that these sources provide also may be in question. Clinical evidence can be categorised as either patient oriented or disease oriented.3 "Patient oriented evidence that matters" considers outcomes that patients would care about (morbidity, mortality, quality of life) and that would "matter" because the interventions should change the way clinicians practice. Disease oriented evidence is the large amount of intermediate or surrogate endpoint information that makes up the knowledge base of almost all practising clinicians. Reliance on disease oriented information is the main reason, therefore, that most information found in expert based systems is not relevant, and why it is likely not to be useful to either clinicians or their patients.

David C Slawson,
Associate professor

UVA-HSC,
Department of Family Medicine,
Box 414, Charlottesville,
VA 22908,
USA

Allen F Shaughnessy,
Director of research

Harrisburg Family Practice Residency,
PO Box 8500,
Harrisburg,
PA 17105-8700

References 1 Smith R. What clinical information do doctors need? BMJ 1996;313:1062-8. (26 October.)

2 Shaughnessy A F, Slawson A F, Bennett J H. Becoming an information master: a guidebook to the medical information jungle. J Fam Pract 1994;39:489-99.

3 Slawson D C, Shaughnessy A F, Bennett J H. Becoming a medical information master: feeling good about not knowing everything. J Fam Pract 1994;38:505-13.

Hospital libraries provide crucial information

Editor,
Richard Smith1 has unfortunately missed a key article, the Rochester study.2 Doctors were asked to request some information related to a current clinical case and then to evaluate its impact on the care of their patients. Of the 208 doctors participating in the survey, 80% said that, as a result of the information provided by the hospital library, they probably or definitely handled some aspect of patient care differently than they would have otherwise done. Changes in several areas of care were reported: diagnosis (29%), choice of tests (51%), choice of drugs (45%), reduced length of stay (19%), and advice given to the patient (72%). The doctors also said that the information provided by the library helped them to avoid the following: hospital admission (12%), patient mortality (19%), hospital acquired infection (8%), surgery (21%), and additional tests or procedures (49%). In general, the doctors rated the information provided by the library more highly than that provided by other information sources such as diagnostic imaging, laboratory tests, and discussions with colleagues.

The Rochester study confirmed earlier studies that information provided by hospital libraries is perceived by doctors as having an important impact on clinical decision making. With the advent of evidence based medicine the importance of getting the right piece of information into the hands of the right doctor at the right time, for quality patient care, cannot be overemphasised.

Royal College of Surgeons in Ireland,
123 St Stephen's Green,
Dublin 2,
Republic of Ireland

References

1 Smith R. What clinical informaion do doctors need? BMJ 1996;313:1062-8. (26 October.)

2 Marshall J. The impact of the hospital library on clinical decision making: the Rochester study. Bull Med Lib Assoc 1992;80:169-78.

Citing old research may mislead readers

Editor,
Richard Smith has broken a fundamental rule by not citing the original authors from whom he drew information.1 Instead, he quoted from an author2 who cited the three original papers.3-5 This suggests that he has not read the original articles. Moreover, it may have misled readers. In citing this paper from 1995, Smith implied that this research on costs and time spent on information handling took place in the 1990s - but the three papers cited date from 1966, 1970, and 1973. I find it hard to believe that nothing has changed in costs as well as time spent on information handling over the past 30 years. In this way, Smith has contributed to his own statement that "some of the information in doctors' heads is out of date and wrong."

Anita Verhoeven,
Research librarian

University Library,
PO Box 559,
9700 AN Groningen,
Netherlands

References 1 Smith R. What clinical information do doctors need? BMJ 1996;313:1062-8.

2 Hersch W R, Lunin L F. Perspectives on medical informatics: information technology in health care. Introduction and overview. J Am Soc Inf Sci 1995;46:726-8.

3 Jydstrup R A, Gross M J. Cost of information handling in hospitals. Health Serv Res 1966;1:235-71.

4 Mamlin Jo, Baker D H. Combined time-motion and work sampling study in a general medicine clinic. Med Care 1973;11:449-56.

5 Richart R H. Evaluation of a medical data system. Comput Biomed Res 1970;3:415-25.