Name:
What is the Question? Gordon Guyatt, MD, MSc, discusses "Chapter 3: What Is the Question?" from the Users' Guides to the Medical Literature.
Description:
What is the Question? Gordon Guyatt, MD, MSc, discusses "Chapter 3: What Is the Question?" from the Users' Guides to the Medical Literature.
Thumbnail URL:
https://cadmoremediastorage.blob.core.windows.net/e1cf9522-2db8-4c5e-899d-08b1d6fcfec6/thumbnails/e1cf9522-2db8-4c5e-899d-08b1d6fcfec6.jpg?sv=2019-02-02&sr=c&sig=KTIgBRiINOewZSFuaSaJFrmYCTEsu2ENukU2wWYCZUY%3D&st=2025-05-10T08%3A59%3A11Z&se=2025-05-10T13%3A04%3A11Z&sp=r
Duration:
T00H09M04S
Embed URL:
https://stream.cadmore.media/player/e1cf9522-2db8-4c5e-899d-08b1d6fcfec6
Content URL:
https://cadmoreoriginalmedia.blob.core.windows.net/e1cf9522-2db8-4c5e-899d-08b1d6fcfec6/6830455.mp3?sv=2019-02-02&sr=c&sig=wYeAqEYsTErPcu0737Rl02z3DyhE41oKW5%2B7Jz7BJjo%3D&st=2025-05-10T08%3A59%3A11Z&se=2025-05-10T11%3A04%3A11Z&sp=r
Upload Date:
2022-02-28T00:00:00.0000000
Transcript:
Language: EN.
Segment:0 .
>> I'm Joan Stephenson, editor of JAMA's Medical News and Perspectives section. Today, this JAMAevidence podcast will focus on chapter three of the User's Guide to the Medical Literature, which deals with formulating a structured clinical question. Our guest expert is Dr. Gordon Guyatt, the lead author of this section in the guide. Dr. Guyatt, why don't you introduce yourself to our listeners? >> My name is Gordon Guyatt. I am a professor of clinical epidemiology and of medicine at McMaster University. And I do my clinical work as a hospitalist, and I'm involved in teaching and research.
>> Why is the section of the User's Guide to the Medical Literature the most important one? >> This section provides the structure that clinicians need for the whole process of using the medical literature to improve patient care. It starts with defining the question. And people often neglect this. They often, when they're teaching it, they go straight to an article. But in fact, defining the clinical question is surprisingly challenging, and it's certainly very important.
The other part of this chapter that makes it very important is that it tells clinicians once they have defined the clinical question, what sort of study they should look for to answer that question, what sort of study design would be appropriate. >> What are the three ways to use the medical literature? >> Three ways to use the medical literature are first of all the way that might be seen as most traditional, which is browsing. So historically, clinicians used to be subscribing to large numbers of medical journals, which would gradually pile up in the corner of their offices, and they might pick up one occasionally and leaf through.
And after the pile got big enough, they would throw them all out and wait for another set to accumulate. So browsing was not particularly efficient. And nowadays, if one does browse in the mode just to see what's new, that's the idea of browsing, one does it ideally by looking at secondary journals for internal medicine, for example, ACP Journal Club. A second way of using the medical literature, however, is in a problem-solving mode. You have a patient.
You have a clinical question, and you are not sure what to do, and you use the literature. You find the right information, the best information available, to solve the problem. The third issue in using -- ways of using the literature is defining whether you are looking for background information or foreground. Background information is what the medical student and early trainee tends to do more. What exactly is diabetes insipidus? What are the appropriate management strategies are available?
Potential management strategies available? How do I monitor my patient and so on? Foreground questions are those of the seasoned clinician. What is -- is treatment A better than treatment B for this particular condition? What exactly are the magnitude of the effect and the downsides? And does it apply to my particular patient? >> Dr. Guyatt, what are the five types of clinical questions? >> We define five types of clinical questions, which include first of all treatment questions, best addressed by randomized trials.
Is this intervention -- and it could be a screening intervention or a traditional drug or surgical intervention. What -- is intervention A better than a comparator or not as good as a comparator? A second issue is issues of harm, typically addressed by observational studies. Is this particular exposure going to be harmful to my patient? The third is differential diagnosis. A patient is presenting with a particular clinical presentation.
What are the possible diagnoses? And how likely are each of them? Then there is diagnostic tests. What is the value of this particular diagnostic test? What are its measurement properties? How much will I gain? How should I interpret the results? And finally, prognosis in a patient presenting with a particular clinical presentation, what is the likely outcome?
Are they likely to do well or not so well? What is the outcome over a short period of time? What is the outcome over many years? >> And how do you frame a clinical question? >> The framing of a clinical question, we have a little memory aid that clinicians can use, and it's called PICO. And it has to do with the patients, the interventions, the comparator, and the outcome. So who are the patients? What exactly is the intervention you're considering?
What are you comparing it to? And what are -- ideally identifying all patient important outcomes that might be relevant. That is very applicable to treatment and harm. For issues of diagnosis and prognosis, the not so catchy memory aid would be PEO, which is the patients, the exposure, which is the diagnostic test or in the case of prognosis time. And then, for diagnosis, the outcome would be the gold standard to which the test is compared, the exposure being the test, the outcome being the gold standard.
And in prognosis it is morbidity and mortality as you follow patients over time. >> Dr. Guyatt, can you give us an example of an unstructured clinical question, and then show us how to change it into a searchable question? >> All right. Let us take a middle-aged man who presents to the emergency room with loss of consciousness. The very unstructured question would be what do I do now? The more structured question might be do I do a very extensive workup?
Or do I do a much more modest workup? And one can actually translate this unstructured question into a number of different structured questions. So the question, for each of these questions, the patient is a middle-aged individual presenting with loss of consciousness. The differential diagnosis question, the exposure would be a full workup, including long-term follow up to see what the ultimate diagnoses are.
And the outcome would be what those diagnoses are. And if the diagnoses were benign, or most of them were benign, one might choose a limited workup. If there were many patients who had diagnoses with serious implications that require immediate management, one might choose a more extensive workup. One could frame it as a question of prognosis to patients at the same. This time, the exposure is time, and the outcome is whether morbidity and death.
So if patients in natural history studies presenting with loss of consciousness, transient loss of consciousness do well, one would consider a more narrow workup; if they do badly, a more extensive one. We could frame it as a question of diagnosis, where the exposure is an electroencephalogram, and the outcome is the gold standard of seizures or no seizures, again, probably from long-term follow up. And finally, one could frame it as a treatment question where the intervention is the intensive investigation.
The comparator is the more limited investigation, and the outcome is death or morbidity. >> Is there anything else that JAMAevidence users should know about clinical questions? >> One of the most important things perhaps to remember is just how important understanding the structured clinical question is. And when I run workshops and teaching sessions, I often tell people that if all they remember is the PICO, that will be a lot, and that will be very helpful.
One of the ways that it is very important is that clinicians tend to forget about looking at all patient important outcomes. And investigators sometimes fail to measure them, particularly the harm outcomes. Also, by noting the patient important outcomes in advance, it alerts you to the fact that sometimes all you have is substitute or surrogate endpoints that are measured in the trials, which lead to much less confidence in applying an intervention to a patient.
And finally, the PICO is very useful in defining the terms for a search and finding the best evidence from the literature. >> Thank you, Dr. Guyatt, for this overview of formulating clinical questions. For more information, JAMAevidence subscribers can consult chapter three of the User's Guide to the Medical Literature, which was coauthored by Dr. Guyatt. This has been Joan Stephenson of JAMA interviewing Dr. Gordon Guyatt about the clinical questions for JAMAevidence.
