Although there is a large literature examining head trauma in general, several areas remain understudied. Notably, little is known about symptom expression over the course of a day for adolescents recovering from concussion. Furthermore, intra-individual symptom variability has not been well characterized. This pilot study examined the feasibility of a momentary data-gathering method, as well as the sensitivity of the assessment to the subtle and dynamic changes in symptoms of concussion. Six adolescents, three of whom suffered a concussion and three non-injured controls, provided symptom ratings five times per day for 5 days. This ecological momentary assessment (EMA) was conducted on a personal digital assistant to capture variability in symptom reports while in the natural environment. Preliminary results indicated that the EMA method showed great promise as a research tool in natural settings (e.g., school and home). Adolescents were able to comply with all tasks with little interference in their daily activities. Students with concussion showed generally higher symptom ratings across physical, cognitive, and affective domains, and temporal and diurnal patterns for symptoms emerged. Implications for future research and patient care are discussed.
Mild traumatic brain injury (MTBI) or concussion represents about 75% of all TBI. Most persons with MTBI will become asymptomatic in 7–10 days after concussion (Yeates & Taylor, 2005). The prevalence of persistent symptoms varies across studies and has been estimated to range from 5% to 78% (Alves, Macchiochi, & Barth, 1993; Binder, Rohling, & Larrabee, 1997; Iverson, 2005), suggesting that the aftereffects of concussion are variable and subject to interpretation.
Although there is controversy surrounding the course of MTBI, there is general consensus that a variety of physical, cognitive, and behavioral problems can occur in the days and weeks following the injury (Yeates et al., 1999). These problems are usually referred to as post-concussive symptoms (PCS) and include headache, fatigue, dizziness, mental fatigue, attention and memory problems, irritability, and impulsivity.
Most studies on concussion assess patients at one point in time. Of the handful of prospective studies, students with concussion have been examined at several points post-injury. In general, findings suggest that youngsters with MTBI experience more symptoms than sibling and orthopedic controls (Ponsford, Willmott, & Rothwell, 1999; Yeates et al., 1999). Such studies tend to find that symptoms are most apparent within a week of the concussion, persist in some patients for weeks, months, or longer, sometimes worsen within the first few months, and seem to vary greatly among patients (Mittenberg & Strauman, 2000; Mittenberg, Wittner, & Miller, 1997; Yeates et al., 1999).
Although the effects of concussion may be dynamic, resolving, and sensitive to environmental demands, the assessment of post-injury symptoms has typically been static. We have many studies that examine a group of concussed patients on a symptom checklist or neuropsychological battery at one or two points post-injury. Yet, we have no studies that examine patients over the course of a day for days at a time. Thus, research has not captured the moment-to-moment fluctuations in symptom expression that result from changing environmental demands (i.e., exams in school, physical exertion, social demands). For example, students with concussion may experience more severe headaches or greater fatigue as the school day wears on, or they may be more affected by a bad night's sleep or a stressful social situation, or perhaps they have more difficulties with concentration and memory at certain times of the day.
In order to better understand the dynamic aftereffects of concussion, we need to assess patients throughout the day in their natural environments. One way to do this is to employ a new research method called ecological momentary assessment (EMA). It is a method of data collection that requires the recording of self-reported data on people's experiences as they go about their everyday lives (Smythe & Stone, 2003). Participants are asked to record the occurrence and intensity of the target symptoms on a personal digital assistant (PDA) when they are signaled by the device at certain points throughout the day. Such measurement allows for the close monitoring of symptom fluctuations across time within a patient as well as individual differences between patients. The feasibility and applicability of this approach as a method of measuring post-concussion symptoms still need to be determined.
In this pilot study, we monitored the symptom ratings of adolescents with and without concussion over a period of 1 week using the EMA method. Specifically, six adolescents were asked to report their symptoms five times throughout their day via a Palm Pilot PDA. As a new methodology in this area, we were interested in: (a) the feasibility of the EMA technique in a school setting (defined as compliance with the signaling protocol) and (b) sensitivity of the system in capturing the dynamic nature of symptom fluctuations in the natural setting.
Materials and Methods
Six adolescents ranging in age from 14 to 17 years were recruited for this pilot study. Three adolescents (two girls and a boy) had a concussion within the past 6 months (average time post-injury = 117 days; range 78–165 days) and were still symptomatic and followed in a Concussion Management Clinic in a central New York hospital. They were compared with three adolescents (two girls and a boy) who had no history of concussion and no current physical or psychological illness. All students attended high schools in the same county in central New York and were similar in age, socioeconomic status, race, and education.
Concussion was defined as a traumatically induced physiological disruption of brain function, as manifested by a least one of the following: (a) any period of loss of consciousness; (b) any loss of memory for events immediately before or after the accident; (c) any alteration in mental state at the time of the accident (e.g., feeling dazed, disoriented, or confused); and (d) focal neurological deficit(s) that may or may not be transient, including posttraumatic amnesia less than 24 hr, an initial Glasgow Coma Scale of 13–15 (if available), and loss of consciousness of <30 min. No participants had a premorbid neurological disorder, previous head injury requiring medical care, mental retardation, or history of severe psychiatric disorder (criteria drawn from the American Congress of Rehabilitation Medicine, 1993).
Three of the students in this study sustained a sports-related concussion. None were involved in litigation. Each of the students was taking medication: Welbutrin for mood, Elavil for headaches and sleep, and Celexa for labile and irritable mood. It should be noted that these students were symptomatic 2–6 months post-concussion and not representative of the concussion population.
A parent of the students with concussion completed a background questionnaire that covered previous medical history, issues related to the head injury, past and current academic functioning, social and interpersonal relations, past and present personality characteristics, and demographic variables. Parent data were used to verify the presence of the head injury and describe the students' past and current functioning.
Each student with concussion was administered the ImPACT Trauma Version computerized neuropsychological battery (Lovell, 2004) that has been shown to be sensitive to the subtle neurocognitive effects of concussion, even in high school athletes (Schatz, Pardini, Lovell, Collins, & Podell, 2005). This battery consists of five tasks that assess five factors (visual memory, verbal memory, reaction time, motor, and impulse control). Students also completed the Post-Concussion (symptom report) Scale (Lovell et al., 2006, provide normative data for adolescents). The time of testing was approximately 20 min. This testing was employed to provide objective evidence of possible neurocognitive deficiencies.
The device used to assess symptom reports throughout the school day was a Palm Pilot 100. Palm Pilots are battery-operated organizers that can be carried easily in pockets or purses. Information is entered on the Palm Pilot touch screen using an attached stylus pen. Appointment alarms were set to signal subjects when to record information and to remind subjects if they failed to complete scheduled ratings (five beeps per day). The questions on the Palm Pilot provided information as to the context (classroom, lunch home) at each time of testing. Next, students were queried with a Symptom Severity Scale (SSS). This consisted of 13 symptoms (e.g., headache, fatigue, irritability) that were rated in severity within the past hour as none = 0 to severe = 6. The SSS was followed by another rating scale, the Functional Status Scale (FSS). The FSS is a measure of functional impairment and it consisted of 15 items (e.g., memory problems are affecting my work; it is hard to finish what I have started) that were rated from 0 = no problem to 6 = severe problem. Both the SSS and FSS were developed for this study and designed for easy use on the PDA.
A focus group consisting of three adolescents with MTBI, three parents and two teachers was held to ascertain personal experiences following concussion. The information derived from this 2-hr session was used to inform the questions on the PDA and the procedure for data collection.
Three additional students with concussion were accompanied by a parent to an initial session at a hospital concussion management clinic. The parent completed consent form, demographic questionnaire, and post-concussion symptom checklist. The student completed an assent form and the ImPACT neuropsychological battery. The session ended with training on the Palm Pilot and instructions for 1 week of use. Control students did not complete the neuropsychological battery, but they did complete the assent form and training on the Palm Pilot.
All students were instructed to record data (SSS, FSS, and questionnaire) on the Palm Pilot for five consecutive days (Monday–Friday) at five time intervals: 9–10 a.m., 11 a.m.–12 p.m., 2–3 p.m., 5–6 p.m., and 8–9 p.m. Each student had school permission to collect data during the school day. Field testing allowed us to set an alerting tone that the student could hear yet would not distract surrounding students. Students were prompted by a beep from the Palm Pilot at the appointed time. Students had up to 50 min to complete the data recording that took approximately 3–5 min. This required them to read a prompt and check a box with their response or rating. Every 5 min they received an alert beep until they responded to all questions on the Palm Pilot.
Verification of Status
Scores from the ImPACT neurocognitive battery for all concussed students were sent to the ImPACT laboratory in Pittsburgh where they were compared with normative data for same age and sex peers (M.R. Lovell, personal communication, January 2007). Two of the test profiles were identified by Dr Lovell as non-normal or comparable to the clinical (concussion) norms. One student's profile was considered borderline, suggesting possible mild and/or resolving deficits. A second analysis of concussion status was based on the symptom reports provided by parents. A symptom severity score on a post-concussion checklist showed a dramatic difference in reported symptoms for the concussed versus typical students, validating their symptomatic status.
There were several areas of interest with regard to feasibility of the new EMA methodology for students with concussion. First, training students to use the Palm Pilot device was remarkably easy. Students required one demonstration and set of verbal instructions to learn to use the system. There were no problems with breakage, battery depletion, incomplete responding, or lost data. Second, the system was manageable in a school setting, not disruptive to others, or annoying to the students using it. Third, the compliance rate for data collection was relatively high. Across the three concussed students of a possible 75 EMA data entries, there were 70 recordings, for a compliance rate of 93.3%. A debriefing interview with the three concussed students indicated that the data collection went smoothly and did not present a problem (at least at this intensity of data capture and for only 1 week).
Given the exploratory nature of the study and small number of participants, the data collected by EMA are presented for illustration purposes. Data from the SSS included up to five ratings per day per student for 5 days. Viewed globally, the concussed students reported a higher level of symptom severity on all 12 symptoms (mean rating 4.79; 0 = none, 6 = severe) than controls (M = 2.96) over the 5 days. Although some of the average differences were small to moderate (0.3 SD), other symptoms yielded relatively large differences (2–3 SD; headache, concentration, irritability, and confusion).
We were interested not only in the mean symptom ratings, but also the pattern of symptoms across the day. To illustrate the patterns, we selected three symptoms, one from each of the major categories: physical headache, cognitive concentration, and emotional irritability. There were mean score differences shown via t-tests between groups for each of these symptoms (p < .01). Graphs of the severity ratings are presented for each of these symptoms across times of the day collapsed across 5 days of data collection (Figs. 1–3). Inspection of the graphs illustrates that students with concussion experience greater severity of symptoms, even though controls also incur mild levels of these common symptoms. Also, the graphs show variability across individuals and times of the day. In general, symptoms tend to get worse as the day wears on, with increased severity by the end of the school day (2–3 p.m.). Interestingly, students with concussion tend to maintain this level of symptom severity into the evening (8 p.m.), whereas controls tend to be relatively symptom-free once out of school.
In addition to symptom severity, we also used the EMA to assess possible impairment from concussion. Students completed the Functional Severity Scale five times per day for five days. Data from the FSS revealed group differences that indicated greater impairment in the concussed group (M = 5.11, SD = 2.36) than controls (M = 3.34, SD = 1.96), t(110) = 5.03, p < .0001; d = 0.86.
The purpose of this study was to pilot a novel data collection system (EMA) that would capture the dynamic nature of perceived symptomatology in adolescents with concussion. We were particularly interested in feasibility of the method as well as its sensitivity to symptom changes throughout the day. Findings from this exploratory study, although preliminary and small in scale, were favorable. Our observations and feedback from the students indicated that the EMA is a viable methodology for collecting frequent, repeatable, and meaningful data about a person's well-being or functioning. In fact, this seems to be an ideal research tool for monitoring the dynamic and sometimes subtle aftereffects of concussion, even in students who are attending school. Our students demonstrated that this is a feasible and acceptable methodology that may be more useful than the conventional research study requiring an individually administered lengthy test battery. The EMA clearly has potential as a research tool in this field and may also be helpful in clinical management.
We realize the limitations of a small pilot study comprised of adolescents enrolled in a concussion management program. These three students were studied because they were still symptomatic 10–24 weeks after injury. They represent a select set of the MTBI population, yet a set that seems to have management needs as they adjust to school and social life. We realize that these students may not be representative and their data not generalizable to most concussed adolescents. However, it is the methodology in this study, not the sample or the dependent measures, that was of greatest interest.
We certainly have evidence that the students with concussion were different from controls, two dramatically and one slightly. We see these differences reflected not only in the baseline measures from parents and ImPACT, but also in the ongoing severity ratings and functional impairment ratings. These findings are consistent with the aftereffects noted by Lovell and colleagues in their studies of adolescents with concussion (Collins et al., 2002; Lovell et al., 2003). However, it appears that the EMA gives us more information than the static measures. It allows us to track changes over time by individuals as well as by group. With a larger sample, we would be able to use EMA data to examine relationships between variables, and even determine if certain antecedent conditions (e.g., poor night sleep, exam stress, physical exertion) are related to changes in symptoms and performance. By better understanding the context and situations that exacerbate symptoms, we might be able to design interventions that are time and activity sensitive.
It seems noteworthy that these symptoms tended to become worse over the course of the day and persisted into the night. It has been suggested that exertion, whether physical or mental, may exacerbate symptoms (Landre, Poppe, Davis, Schmaus, & Hobbs, 2006), and our preliminary data seem to support that supposition. Some of our patients in interviews and focus groups have described it by saying “my brain hurts.” They clearly are worn down by the end of the day and even the demands of school can overwhelm them. Although fatigue and headaches are physical indicators of their impairment, we also hear them report problems sustaining attention and concentration and also becoming more irritable and moody. Symptom changes and aggravation are reported regularly by students with concussion. In both the research and clinical domains, ecological validity may be increased by capturing changes in post-concussion symptoms across different times and contexts.
Conflict of Interest