Adults learn differently than children. While children have their curriculum
and agenda set for them, adults typically take a much more active role in
determining their learning needs (13).
An attorney may decide to learn
Spanish because she has many Spanish clients; a researcher may learn to use a
statistics program because their research project requires it; or a homeowner
may learn how to drywall because they want to finish their basement.
Applying this to physicians suggests that we learn best when learning (1) is in
the context of patient care, (2) answers our questions,
(3) is directly
applicable to our work, and (4) does not take too much time. Thus, asking and
answering the clinical questions that arise at the point of care is central to
our continuing education as physicians. Continuing medical education has not
been terribly successful at improving patient outcomes, largely because it is
disconnected from patient care.(11) Physician
practice changes in haphazard ways, often not driven by the best available
evidence. In fact, the best predictor of physician prescribing behavior
regarding antihypertensive medications is the year of graduation from medical
school!(13)
Physicians who want to be successful lifelong learners therefore need highly developed critical reflection skills. Applying these skills to their practice and generating clinical questions at the point of care are central to physician education and lifelong learning. To take an evidence-based approach to practice, these questions must be answered using the most valid, relevant information available.
What
Questions Do Clinicians Ask at the Point of Care?
When asked at the end of a half-day of patient care to recall how many questions
they had related to patient care, physicians reported 1 question for every 4
patients.[2] Direct observation of primary care physicians, however,
has shown that they generate an average of 2 questions for every 3 patient
encounters.[3] A physician seeing 25 patients per day will generate
approximately 15 questions. These and similar studies have shown there is
tremendous variety to the questions physicians ask, and these questions are
often complex and patient specific. Approximately 33 percent relate to
treatment, 25 percent to diagnosis, and 15 percent to pharmacotherapeutics.[3-5]
When questions are pursued and answered, more than one half of the answers come
from textbooks and human sources, including both office partners and
consultants.[5] The Physician’s Desk Reference (PDR) is
perhaps the most commonly named source of answers. Electronic sources of
information are rarely used.[3,4]
Unfortunately, approximately two thirds of the clinical questions generated at the point of care go unanswered.[3] Are these questions important? One study took the unanswered questions and gave them to medical librarians. The authors then gave the answers to the physicians who had asked them and found that approximately one half of the answers would have had a direct impact on patient care.[6]
Why do we not answer more of these questions? Limitations include a lack of convenient access to reference materials at the point of care, the time needed to search for information, and the challenge of formulating an answerable question.[4] Two characteristics that predict whether physicians will seek and find an answer to a clinical question are the urgency of the problem and their confidence that they will find an answer.[7]
For example, consider the physician who wants to know how to prescribe famciclovir for herpes zoster. Although he or she can be confident of finding a dosage recommendation in the PDR, this reference will not answer questions about the medication’s effectiveness. Information about the number of patients this physician would have to treat to prevent a case of postherpetic neuralgia might be found in a randomized trial, but access to that information is unlikely at the point of care, and the question remains unanswered. Thus, the information a patient and physician might want so they can decide whether the medication is worth paying for and taking is not available.
A useful way to think about clinical questions is by the type of information needed.[8] For example, a relatively large randomized controlled trial has shown that patching corneal abrasions only increases discomfort and healing time.[9] A family physician who is unaware of this outcome and who continues to patch corneal abrasions has an unrecognized information need, because his or her patients would benefit from a change in practice. When that same family physician asks this clinical question: “I wonder whether there is any evidence that patching corneal abrasions improves outcomes that my patients and I care about?” he or she has recognized an information need and asked a clinical question. When the physician asks a colleague, he or she has begun to pursue this information need. Searching the Journal of Family Practice POEMs (patient-oriented evidence that matters) Web site (http://www.medicalinforetriever.com) using the term “corneal abrasion” will locate an article that answers that clinical question; the information need is now satisfied. Finally, the information must be implemented in the physician’s practice to affect patient outcomes.
This pathway is not linear. Rather, it is a cycle, because medical science is dynamic rather than static, and new information is constantly becoming available. Yesterday’s satisfied and implemented information need is tomorrow’s unrecognized need. For example, a new study has shown convincingly that topical nonsteroidal anti-inflammatory drops reduce pain and speed healing in corneal abrasion.[10] If use of nonsteroidal anti-inflammatory drugs was not the standard practice for this physician, the physician has an unrecognized need and begins the process anew. In the next section, we will discuss this cycle, the Information Pathway.
In addition to the type of need, there are several other ways to classify physician information needs. Woolf and Benson[5] describe several other characteristics of an information need: the type of information (eg, diagnostic, prognostic, therapeutic), the organ system, and the source of information used to answer the question. Osheroff et al[8] describe the generalizability of an information need as whether it can be satisfied by generalizable sources of information, such as original research and the PDR, or whether the information need is specific to a particular patient. These different ways of classifying information needs (including clinical questions) are summarized in Table 1.