Knowledge Management

Where is the wisdom we have lost in knowledge? Where is the knowledge we have lost in information?

—T.S. Eliot 1934

Finding the best available answer to a specific clinical question is like finding a needle in a haystack. Essential information is mixed with a vast amount of less credible “factoids” and opinions, and it is a daunting task to sort the wheat from the chaff. Yet, that is what clinicians need to do. Critical reading is only as good as the information found.

Knowledge management is the effective and efficient organization and use of knowledge. This was a difficult task in the days of print media only. Fortunately, knowledge management has become a great deal easier in the era of electronic information. There are more and better studies on a broad range of clinical questions, widely available access to research results, and efficient ways to rapidly sort articles by topic and scientific strength. These opportunities followed the widespread availability of computers, the World Wide Web, and electronic information for clinical purposes.

Finding information may seem to be a low priority for clinicians still in training. They are surrounded by information, far more than they can handle comfortably, and they have countless experts to help them decide what they should take seriously and what they should disregard. However, developing one’s own plan for managing knowledge becomes crucial later on, whether in practice or academe.

Even with recent developments, effective and efficient knowledge management is a challenging task. In this chapter, we review modern approaches to clinical knowledge management. We will discuss four basic tasks: looking up information, keeping up with new developments in your field, remaining connected to medicine as a profession, and helping patients find good health information themselves.

BASIC PRINCIPLES

Several aspects of knowledge management cut across all activities.

Do It Yourself or Delegate?

Clinicians must first ask themselves, “Will I find and judge clinical research results for myself or delegate this task to someone else?” The answer is both. Clinicians should be capable of finding and critiquing information on their own; it is a basic skill in clinical medicine. But as a practical matter, it is not possible to go it alone for all of one’s information needs. There are just too many questions in a day and too little time to answer them on one’s own. Therefore, clinicians must find trustworthy agents to help them manage knowledge.

Which Medium?

One can obtain information via a rich array of media. They range from printed books and journals to digital information on the Web accessed through stationary and handheld platforms. There are audiotapes, videotapes, and more. The information is neither more nor less sound because of how it happens to come to you. Validity depends on authors, reviewers, and editors, not the medium. However, the availability of various media, with complementary advantages and disadvantages, makes it easier to find ones that match every user’s preferences.

A modern knowledge management plan should be based on electronic information on the Internet. The information base for clinical medicine is changing too fast for print media alone to be sufficient. For example, clinically important discoveries in antiviral therapy for HIV, innovative scanning technologies, and state-of-the-science cancer chemotherapy emerge from year to year, even month to month. The Internet can keep pace with such rapid change and also complement, but not replace, traditional sources.

Grading Information

Grading makes it possible for clinicians to grasp the basic value of information in seconds. Usually, the quality of the evidence (confidence in estimates of effects) and strength of recommendations are graded separately. Table 14.1 shows an example of one widely used grading scheme called GRADE, similar in principle to other approaches in general use. This grading is for interventions; grading of other kinds of information is less well developed. Notice that recommendations are based on the strength of the research evidence, depend on the balance of benefits and harms, and vary in how forcefully and widely the intervention should be offered to patients. Although criteria for grading are explicit, assigning grades still depends partly on judgment.

Misleading Reports of Research Findings

Until now, we have acted as if the only threats to the validity of published clinical research stem from the difficulties of applying good scientific principles to the study of human illness. That is, validity is about the management of bias and chance. Unfortunately, there are other threats to the validity of research results, related to the investigators themselves and the social, political, and economic environment in which they work. We are referring to the all-too-human tendency to report the results of research according to one’s own stake in the results.

Conflict of interest exists when investigators’ private interests compete with their responsibilities to be unbiased investigators. There are many possible competing interests:

Financial conflict of interest: When personal or family income is related to research results (this conflict is usually considered the most powerful and most difficult to detect)
Personal relationships: Supporting friends and putting down rivals
Intellectual passion: Being for one’s own ideas and against competing ones
Institutional loyalties: Putting the interests of one’s own school, company, or organization above others
Career advancement: Investigators get more academic credit for publishing interesting results in elite journals.

Conflict of interest exists in relation to a specific topic, not in general, and regardless of whether it has actually changed behavior.

How is conflict of interest expressed? Scientific misconduct—fraud, fabrication, and plagiarism— are extreme examples. Less extreme is selective reporting of research results, either by not reporting unwelcome results (publication bias) or reporting results according to whether they seem to be the “right” ones. Industry sponsors of research can sometimes block publication or alter how results are reported. To create a public record of whether this has occurred, randomized controlled trials are now registered on publically available Web sites before data collection, making to possible to follow-up on whether the results were published when expected and whether the reported endpoints were the same as when the trial began (1).

More subtle and more difficult to detect are efforts to “spin” results by the way they are described, for example, by implying that a very low P value means the results are clinically important or by describing effects as “large” when most of us would think they were not (2). All of us depend on peer reviewers and editors to limit the worst of this kind of editorializing in scientific articles.

We mention these somewhat sordid influences on the information clinicians (and their patients) depend on because they are, in some situations, every bit as real and important as the well-informed application of confidence intervals and control of confounding, the usual domain of clinical epidemiology. Research and its interpretation are human endeavors and will, therefore, always be tinged, to some extent, with self-serving results. There are ongoing efforts to limit bias related to conflicts of interest, mainly by insisting on full disclosure but also by excluding people with obvious conflicts of interest from peer review of manuscripts and grants, authorship of review articles and editorials, and from guidelines panels.

Table 14.1 Grading Recommendations for Treatment According to the Quality of Evidence (Confidence in Estimate of Effect, A-C) and Strength of Recommendation (1, 2) with Implications. Based on GRADE Guidelines

Grade of Recommendation	Clarity of Risk/Benefit	Quality of Supporting Evidence	Implications
1A. Strong recommendation, high-quality evidence	Benefits clearly outweigh risks and burdens, or vice versa	Consistent evidence from well-performed randomized controlled trials, or overwhelming evidence in some other form. Further research is unlikely to change confidence in the estimates of benefits and risks	Strong recommendations apply to most patients in most circumstances without reservation. Clinicians should follow a strong recommendation unless there is a clear and compelling rationale for an alternative approach.
1B. Strong recommendation, moderate-quality evidence	Benefits clearly outweigh risks and burdens, or vice versa	Evidence from randomized controlled trials with important limitations (inconsistent results, methodologic flaws, or imprecision), or very strong evidence of some other research design. Further research (if performed) is likely to change our confidence in the estimates of benefits and risk	Strong recommendation that applies to most patients. Clinicians should follow a strong recommendation unless there is a clear and compelling rationale for an alternative approach.
1C. Strong recommendation, lowquality evidence	Benefits appear to outweigh risk and burdens, or vice versa	Evidence from observational studies, unsystematic clinical experience, or randomized controlled trials with serious flaws. Any estimate of effect is uncertain.	Strong recommendation that applies to most patients. Some of the evidence base supporting the recommendation is of low quality.
2A. Weak recommendation, high-quality evidence	Benefits closely balanced with risks and burdens	Consistent evidence from wellperformed randomized controlled trials or overwhelming evidence of some other form. Further research is unlikely to change our confidence in the estimates of benefits and risks.	Weak recommendation. Best action may differ depending on circumstances or patient or societal values
2B. Weak recommendation, moderate-quality evidence	Benefits closely balanced with risks and burdens, with some uncertainty in the estimates of benefits, risks, and burdens	Evidence from randomized controlled trials with important limitations (inconsistent results, methodologic flaws or imprecision), or very strong evidence from some other research design. Further research (if performed) is likely to change confidence in estimates of benefits and risks.	Weak recommendation. Alternative approaches likely to be better for some patients under some circumstances.
2C. Weak recommendation, low-quality evidence	Uncertainty in the estimates of benefits, risks, and burdens; benefits may be closely balanced with risks and burdens	Evidence from observational studies, unsystematic clinical experience, or randomized controlled trials with serious flaws. Any estimate of effect is uncertain.	Very weak recommendation. Other alternatives may be equally reasonable.
Adapted from Guyatt GH, Oxman AD, Vist GE, et al. for the GRADE Working Group. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 2008;336:924-926.

LOOKING UP ANSWERS TO CLINICAL QUESTIONS

Clinicians need to be able to look up answers to questions that arise during the care of their patients. They need this for things they do not know but also to check facts they think they know but might not, because the information base for patient care is always changing.

It is best to get answers to questions just at the time and place where they arise during the care of patients. This has been called the point of care and the associated learning just-in-time-learning. Answers can then be used to guide clinical decision making for the patient at hand. Also, what is learned is more likely to be retained than information encountered out of context in a classroom, lecture hall, book, or journal, apart from the need to know for a specific patient. In any case, postponing the answering of questions to a later time too often means they do not get answered at all.

For just-in-time learning to happen, several conditions must be in place (Table 14.2). Most patient care settings are time-pressured, so the answer must come quickly. As an office pediatrician pointed out, “If I added just 1 to 2 extra minutes to each patient visit, I would get home an hour later at the end of the day!” What clinicians need is not an answer but the best available answer, given the state of knowledge at the time. They need information that corresponds as closely as possible to the specific clinical situation their patient is in; if the patient is elderly and has several diseases, the research information should be about elderly patients with comorbidities. Clinicians need information sources that move with them as they travel from their office to home (where they take night and weekend call) and to hospitals and nursing homes.

When all this happens, and it certainly can, the results are extraordinarily powerful.

Example

A patient sees you because he will be traveling to Ghana and wants advice on malaria prophylaxis. You are aware that the malaria parasite’s susceptibility to antimalarial drugs varies across the globe and that it is continually changing. The Centers for Disease Control and Prevention have a Web site (http://www.cdc.gov) with current information for travelers to all parts of the world. Using the computer in your clinic, you quickly find out which prophylactic drug this patient should take and for how long before, during, and after the trip. You are also reminded that he should have a booster dose of polio vaccine and be vaccinated for hepatitis A and B, typhoid, and yellow fever. The site lists clinics where these vaccines are available. The site also shows that northern Ghana, where your patient will be visiting, is in the “meningitis belt,” so he should also be vaccinated against meningococcal disease. The information you are relying on is an up-to-date synthesis of the world’s best advice and is available to you within seconds.

Table 14.2 Conditions in Which Information Is Available at the Point of Care

Condition	Rationale
Rapid access	The information must be available within minutes for it to fit into the busy workflow of most patient care settings.
Current	Because the best information base for clinical decisions is continually changing, the information usually needs to be electronic (as a practical matter, on the Internet).
Tailored to the specific question	Clinicians need information that matches as closely as possible the actual situation of their individual patient.
Sorted by scientific strength	There is a vast amount of information for almost any clinical question but only a small proportion of it is scientifically strong and clinical relevant.
Available in clinical situations	Clinicians cannot leave their place of work to look up answers; they must find it right where they work.

Solutions

Clinical Colleagues

A network of colleagues with various and complementary expertise is a time-honored way of getting point of care information. Many clinicians have identified
local opinion leaders for this purpose. Of course, those opinion leaders must have their own sources of information, presumably more than just other colleagues.

Electronic Textbooks

Textbooks, even libraries, are on the Internet and made available to clinicians by their medical schools, health systems, and professional societies. For example, UpToDate (http://www.uptodate.com) is an electronic information resource for clinicians, the product of thousands of physician-authors and editors covering 9,000 topics in the equivalent of 90,000 printed pages (if it were ever printed)^†. Information is continually updated, peer reviewed, searchable and linked to abstracts of the original research, and recommendations are graded. UpToDate is available at the point of care throughout the world wherever the Internet can be accessed by computers or mobile platforms.