That’s a good question!
Critical Clinical Inquiry
That’s a good question, and good questions deserve a good answer. What is a good question? Many people say that there are no bad questions. But that does not logically imply that all questions are good. There can be “medium” questions. And in certain contexts, I believe there can be bad questions. It’s the context that is important.
In the movie, Finding Forrester, one of the lead character’s William Forrester (played by Sean Connery) tells the other lead character Jamal (played by Rob Brown) that a question he just asked was “Not a soup question.” To Forrester, a question Jamal had just asked about why Forrester was making soup the way he did, was a good question because it provided Jamal with information that Jamal could use, information he could improve his life with in some small way. The question that Jamal asked that was “Not a soup question” - that is not good based on Forrester’s view point - because it was about Forrester’s personal life and did not (according to Forrester) fulfill the criteria of providing Jamal with information that was needed. This is a very pragmatic view point on the value of questions. If you’ve seen the movie you understand why this entire conversation simply foreshadows the demise and implications of Forrester’s overly pragmatic worldview such that even something that is “Not a soup question” can have inherent value in the context of personal relationships.
In the context of a teacher - student interaction and relationship, such as in class, in conversation, or in a clinic, both pragmatic (soup questions) and non pragmatic (non soup questions) are important. With a teacher - student interaction learning is always occurring in both directions. From teacher to student; from student to teacher. Learning occurs from the interaction and communication itself. In such a context there are no bad questions because it is vitally important for the teacher to hear the question. The question is always informative to the teacher. The question itself communicates the insights and understanding of the student. It fulfills that role regardless of what the question is about - even the question that displays no understanding, or an incorrect understanding, of the concept being taught is useful for the teacher. In the teacher - student interaction context - there are no bad questions. That still does not mean all questions are good questions in all contexts.
This first post in the category Critical Clinical Inquiry is on questions because questions are the beginning of Critical Clinical Inquiry. All such inquiry, indeed all inquiry, starts with questions. The iterative process of questioning is fundamental to the inquiry.
Rather than starting the series on Clinical Critical Inquiry with a broad theoretical, abstract or universal discussion on questions, I’m going to approach it in a clinical way (a practical way). The clinical way of approaching inquiry tends to be talking about specifics, particulars. The clinician - at home with the situation full of all its detail and particulars and variability and multitudes of interactions and plentitudes of possible causal paths - is, thankfully, comfortable with the utter confusion of the particular situation. For those that have never seen, or heard me discuss my 2005 letter to the editor of the PT Journal about what we can do to try to reconcile clinical research with clinical practice - my quote that has framed me for now 17 years is:
The simple study of the parts in isolation—reductionism—is the modus operandi of the scientific method, attempting to isolate sources of variation. Clinicians, however, are faced with all sources of variation at the same time and must deal constantly with the full burden of the complex system. They and their patient management become immersed, interconnected, and part of the complex system.
Since writing that I have come to recognize - thanks to my friend and Systems Scientist / Theorist Bog Sniezek - that the “complexity” of a complex system is epistemological, not ontological, but that’s a conversation for another day. The simple point here is that clinicians are, by necessity, comfortable with particular situations. My goal as a theorist, and their goal as clinicians constantly learning from experience, is to extract from those particulars that which can be a universal principle and applied to other particulars (i.e. inductive / statistical inference).
Therefore, let’s talk about a particular type of question.
“Does X work?” - ok, give me a little flexibility to be abstract, but no more abstract than you dealt with in 7th grade algebra.
“Does X work?”, where X is any “intervention” that a physical therapist might do (consider doing, part of the scope of practice, at any level of detail you feel inclined to make it - could be exercise; could be strength exercise; could be 10 reps and 3 sets; could be 10 reps for 1 set and attempting 10 reps but working to failure for the remaining 2 sets. Point is - make X, the intervention, as detailed as you want.)
“Does X work?” Could also be “Is X Effective” - I’ve been asked numerous times lots of different “Does X work” questions - Does estim work? Does ultra sound work? Does dry needling work? Does stretching work? Does cupping work? Does airway clearance work? Does X work? Does exercise work? They are all bad questions (unless they are from a student, because then I can teach them about PICO questions).
“Does X work?” Is a bad question - no matter how detailed X is if it is not a “PICO” question. And X cannot be a PICO question without an understanding of the “population” (or patient - more on this in a moment); the intervention (obviously Does X work? Includes the intervention); the comparison and the outcome.
The population: If you’re asking “Does X Work” you are not asking “Did X work?” As in some recent experience where you attempted X and it worked on a patient, or you have good reason to believe that it worked. Believing it worked means believing it was the cause of whatever outcome you observed or the patient reported to you. Causality in the singular case is one of the benefits of the dispositional view of causality - one of the features of the “Cause Health” project that I find most appealing. I do think that our adjudication about X working in a case requires some set of criteria for consideration that I could consider in another post. For example, the time frame. If you do X on Monday, and the patient comes back next Monday and is feeling better it is much more difficult to infer that the single cause was what happened last Monday. Certainly not impossible, just needs further justification. Compared to someone having a big change immediately from an intervention (like a manual intervention such as airway clearance, or dry needling or mobilization of a joint) and having an immediate objective change in oxygen saturation, or muscle tone and movement, or range of motion; or a subjective change in breathlessness or pain. The temporal proximity of an intervention and its outcome are certainly considerations in the causal inference of the singular. For now, I’m switching back to “Does X work?” - as opposed to “Did X work?”.
“Does X work?” Is a question about both the past being being projected to the future. The question is asked by someone that wants to know whether they should “Do X” or not. If it works, they’ll do it. If it does not work, they won’t do it.
“Does X work?” Requires statistical inference - identifying a characteristic from observing particulars and saying it is a characteristic of a sample, and then from a sample to a population. And we’re more confident when we have more observations in a sample; and we’re more confident when we have more samples that are consistent (i.e. meta-analysis with homogeneity of effect sizes). Why the population - because if you ask “Does X work?” - you do not intend to use X on the same people that were in the samples that were tested. You do intend to use X on someone (a person) that is reasonably associated with (part of) the population that we can answer the question and we can only do that if we know the population. Populations can be defined with varying degrees of detail. It’s set theory. There’s a universal (all inclusive) population (set) - all living humans for example. As soon as you add one characteristic - like age categories or sex or the present of a disease - you’ve made subsets of the population. To use a phrase made population in the PSU DPT program by Dr. Kelly Legacy “It depends” can often by answered by the population component of “Does X work?” Does X work when P is Y? Does estim work when the patient is congruous with the population that includes muscle atrophy due to disuse?
The outcome: Based on where that last example ended I feel compelled to jump to “Outcome” of our PICO approach to making “Does X work?” a good question. The outcome. When you ask if something works you need to have an outcome in mind and the question about whether it works is always dependent on that outcome. Does estim work when the patient is congruous with the population that includes muscle atrophy due to disuse for the outcome of improving muscle mass, or strength, or health, or blood flow, or all of these? Is it biologically plausible that these have a particular temporal order (of course, causality as we know it occurs across time).
“Does X work to achieve O when P is Y?”
The final component - C is the comparison. The comparison has a research and a practical interpretation. The research interpretation is that all of the research that has attempted to answer the question - “Does X work to achieve O when P is Y?” - has done so with a comparison group. The group may not exist (single group observational studies) but in such cases the group is implied - it is implied (not always correctly) that the group getting the intervention would NOT have had the effect that they had if they had not had the intervention. Because this implication is often wrong, we much prefer evidence that has an actual comparison group that truly does nothing (whether a sham or not), or that does something different - an alternative intervention. I cannot say this enough - the comparison used in research studies is very important (there will be an entire post on the comparison because this post is approaching my personal limit of 2000 words per post).
We are left with:
“Does X work to achieve O when P is Y as compared to C?”
That is a good question about the effectiveness of a clinical intervention. Please consider asking all future “Does X work?” Questions in that format.
So, if I say "that works for me" that may or may not be true. On another note: I've never been quite sure how to handle implied questions -- my wife is really good at those.