Too often clinical questions are couched in terms specific to a particular patient. Asking “What test should I order for this 28-year-old woman with chest pain?” is appropriate when talking to a human consultant but not when searching a medical reference. As phrased, the question requires that you know the patient and the particulars of their case.
You are much more likely to get a helpful answer from a medical reference or a medical librarian if you ask a well-formed, more generalizable question. For example, the above question could be rephrased as: “What is the best test to rule out myocardial infarction in this person with chest pain and a low likelihood of disease?”
PICO
“PICO” is a useful acronym that describes the elements of a well-formed clinical question about therapy. The letters stand for:
Population
Intervention
Comparison
Outcome
Let’s consider each element in turn. Population is really a description of the group to which your patient belongs, and may include age, gender, race, ethnicity, and stage of disease. The description should be specific enough to be helpful, but not overly specific. You are unlikely to find studies of “Asian 63 year old women” (too specific); describing the population as “Post-menopausal women” is much more likely to be helpful.
Intervention is a description of the test or treatment that you are considering. Comparison is the alternative. Not all questions need a comparison, but clinical questions typically involve a choice between two options, even if the alternative is only watchful waiting.
Finally, the outcome should generally be something that not only matters to you, but matters to the patient. You may well want to involve the patient in defining the outcome. In making a decision about anti-depressant drug therapy, some patients may wish be more concerned about dry mouth and fatigue, while others are more concerned about orgasmic dysfunction. The only way to know is to ask.
Case #1
Jean Birkenstock presents with a complaint of fatigue and loss of interest in her usual activities. She denies any suicidal ideation, and has a normal history and physical examination. After diagnosing her with minor depression, you are preparing to write a prescription for an Serotonin Specific Reuptake Inhibitor (SSRI). Your patient asks, what about St. John’s Wort? You wonder: “Is St. John’s Wort a reasonable choice for this patient?”
Population: Adults with minor depression
Intervention: St. John’s Wort
Comparison: An SSRI
Outcome: relief of symptoms
Putting it all together: “In adults with minor depression, is St. John’s Wort or an SSRI more effective at relieving symptoms?”
Case #2
Jimmy Coryza is a 14 year old with seasonal allergies. You wonder: should I prescribe a nasal corticosteroid or a non-sedating anti-histamine?
Population: Patients with seasonal allergies (if you find data for adolescents or children, so much the better, but don’t box yourself in at the outset)
Intervention: nasal corticosteroid
Comparison: non-sedating antihistamine
Outcome: relief of symptoms (other important outcomes could include sedation and cost, or you could look at the symptom that is most troublesome to the patient. Coryza, for example.)
Putting it all together: “In patients with seasonal allergies, is an inhaled corticosteroid or a non-sedating antihistamine better at relieving coryza?”
Case #3
Hack Fredrickson has been coughing for 6 weeks. He is a heavy smoker, and you are considering skipping the chest x-ray and going straight to a CT scan.
Population: Patients with cough
Intervention: chest x-ray
Comparison: chest CT
Outcome: positive and negative predictive value
Putting it all together: “In patients with cough, does chest x-ray or chest CT have a better positive and negative predictive value?”
Comment: This is a tricky one. While we want more accurate tests, that really isn’t the most important thing. The most important thing is whether the increased accuracy improves patient outcomes. Another way to phrase this question is: “In patients with cough, does initial chest x-ray or initial chest CT result in better patient outcomes?” While a better question, in most cases you won’t find evidence that answers this type of question.
Case #4 (therapeutic)
Aiken Jones has had back pain for several days now, after a weekend of chopping wood at his cabin. You wonder whether a muscle relaxant would be helpful.
Now, put it all together in a single, focused question.
Case #5 (diagnostic)
Achilles Smith has had persistent heel pain. He wonders if he might have a heel spur, and says "Aren't you going to order an x-ray, doc?"
Now, put it all together in a single, focused question.