The following is a conversation between Devashish Tiwari, Assistant Professor in the Department of Physical Therapy at Simmons University, and Taylor Eubanks, graduate student in the Gender and Cultural Studies program. Professor Tiwari discusses his current research on pediatric dizziness in post-concussion children and adolescents using IRT (item response theory).
What is your current research project?
Devashish Tiwari: To begin, concussion is something that has been on the rise of late, especially in the pediatric population. We have seen an increase in involvement of childhood and team sports, like football, baseball, wrestling and so on. Until recently we have not noticed nor paid attention to pediatric concussions.
The peculiar thing about studying pediatric concussions is we cannot consider them in the same way we consider adult concussion. The expectations after injury should be considered according to the child’s perspective while examining a child. We cannot take a measure or a questionnaire that is designed and validated for the adult population and use it in a pediatric population. This necessitates the creation of a measure that is appropriate for a pediatric population.
So this project is targeted towards assessing one of the post-concussion symptoms, dizziness, because previous studies have suggested that dizziness is one of the most prevalent symptoms after a concussion. Additionally, dizziness presents longer and it can track or delay the recovery of a child after their concussion. It has been identified that if a child has prolonged dizziness, this child may struggle in school, especially in terms of participation in class and both academic and physical education.
How are you assessing post-concussion injury?
Devashish Tiwari: To accurately assess, there are two routes: one is the patient reported, or questionnaire, way. The second is "objective testing" where we use multiple instruments or scales.
"It is really important for us to understand what is valuable or important to the child in order to create a questionnaire that will accurately assess their injury."
My focus is on one of the patient reported measures. It is really important for us to understand what is valuable or important to the child in order to create a questionnaire that will accurately assess their injury. To do that, we looked at one of the existing questionnaires used in the assessment of dizziness in children.
Whenever we are using measurements, the instrument measuring should be valid, and it should present questions in such a way that the difficulty is evenly distributed across questions—just like when we take an exam. If the questions are all too hard, then we get a poor score that is not necessarily an accurate representation of ability. On the other hand, if the measure is too easy we get a ceiling effect which is also not an accurate measure of ability.
To address and create a well balanced questionnaire, we need to first examine if the present or existing questionnaire is suitable. I have been focusing on this measure and I’m trying to assess all the items in the current questionnaire: are they really representing the construct—should each question even be part of the questionnaire—and how is the difficulty level distributed? This is achieved by something called item response theory (IRT).
Your research tables display categorization changes in the measure or values of assessment. What was the process in making those changes?
Devashish Tiwari: That's a great question. In the original measure three domains were identified, and the questionnaire had a list of 25 questions. The questions were distributed—nine items, nine items, and seven items—among those three identified domains. The domains were physical, functional and emotional.
Concussion is something that affects multiple parts of an individual. Not only the physical and functional, it also affects the emotional status and participation of the child in their everyday community, so we really want to be sure whether the questionnaire is complete or if there is something hidden within the questionnaire that is not being identified from the original three domains.
For instance, the functional domain category on the original scale has items not exclusively representing one particular domain, but that actually represents multiple domains. So we separated some of those questions into subdivisions and created two new categories: walking or mobility and community participation.
"If we standardize documentation in rehabilitation sciences, we can very strongly improve the effectiveness of physical therapy interventions."
If we can target the domain that is most affected, let’s say if it the emotional subscale, I may refer my patient to a psychologist, a neuropsychologist, or psychiatrist because [a physical therapist] would not be the best health care provider to address that domain. The direction of my treatment can be built more specifically around the needs of the individual patient based on the subscale involved.
How did you realize the original questionnaire needed to be reassessed?
Devashish Tiwari: Let me say first of all, concussion and I, we have a very old relationship. As a child and as a young adult I’ve had four concussions, all due to different reasons. At that point in time—I’m talking about 15 or 20 years ago when nobody knew about the seriousness of concussion—I was not treated for any of these. Since those injuries, I’ve had it in my mind that I need to draw people's attention to this problem.
Now my area of research is outcome measures, and I really like to measure improvements. That is to say, if I'm treating somebody, to know that my patient is really improving I need to have a good measure to assess them. The original questionnaire on dizziness is the first measure that I came across and, to my and to my research team’s knowledge, it is the only existing measure to assess dizziness impact in children and adolescents.
As we spoke before about the validity component of the questionnaire, there was no established validity of this measure. So in one of my previous papers we did a validity test, and we found some inconsistencies in the structure and content of the measure. We knew we needed to explore it further and conduct a more detailed analysis, which entailed the utilization of item response theory. We wanted to identify the exact inconsistencies, including any problems in how that questionnaire comes out as a whole.
You open your article with the statistic that since 2007, the incidence of childhood concussions have gone up 60%. Why do you think that is?
Devashish Tiwari: There are a few reasons, in my mind, for that rise. One reason is that in our culture, kids are increasingly participating in many more physical activities, like sports, and so on.
However, on the flip side, the second reason is our iPhone age, or smartphone age, where kids have started utilizing less time overall in physical activity and spending more time on their phone. These kids then go out in the field and are more prone to concussion because prolonged FLEXED position of the neck brings about weakness of the neck muscles. Imagine when this kid puts on a helmet and goes out in the field, they are like a bobblehead—the neck is not strong enough to hold the weight of the head with that helmet on, and once the kid sustains an impact, the intensity of impact increases significantly.
"The peculiar thing about studying pediatric concussions is we cannot consider them in the same way we consider adult concussion."
A final reason is that people are growing more aware of concussions in children, so the reporting and diagnosis has increased, contributing to an overall rise in the number of concussions. When we were younger and got a concussion, you sat down and rested for a few minutes, but then you got back out there again.
Why does this matter?
Devashish Tiwari: If we take our child to a health care provider, our intention is that the child should show improvements with the treatment and we should understand that this treatment is effective. We need to have very age appropriate very specific measures. If you look at medicine, for that matter, dosage of medication is based on the child's height, age, and weight. It is not based on the measure for adults.
Interventions and rehab which are applicable to adults may not be applicable for the pediatric population, and how we know what works for pediatrics is to look into the measures which are designed specifically for the pediatric population.
If we standardize documentation in rehabilitation sciences, we can very strongly improve the effectiveness of physical therapy interventions, and this will lead to changes in healthcare policies related to rehabilitation. In the long run, that's my vision.
Edited for length and clarity.