The effect of telephone counselling on reducing post-traumatic symptoms after mild traumatic brain injury: A randomised trial
- K R Bell1,
- J M Hoffman1,
- N R Temkin1,2,3,
- J M Powell1,
- R T Fraser1,4,
- P C Esselman1,
- J K Barber2,
- S Dikmen1,2
- 1Department of Rehabilitation Medicine, University of Washington, Seattle, WA, USA
- 2Department of Neurological Surgery, University of Washington, Seattle, WA, USA
- 3Department of Biostatistics, University of Washington, Seattle, WA, USA
- 4Department of Neurology, University of Washington, Seattle, WA, USA
- Kathleen R Bell, University of Washington, Department of Rehabilitation Medicine, Box 356490, Seattle, WA 98195, USA;
- Received 7 December 2007
- Revised 11 April 2008
- Accepted 16 April 2008
- Published Online First 9 May 2008
Background: Mild traumatic brain injury (MTBI) is a significant public health problem affecting approximately 1 million people annually in the USA. A total of 10–15% of individuals are estimated to have persistent post-traumatic symptoms. This study aimed to determine whether focused, scheduled telephone counselling during the first 3 months after MTBI decreases symptoms and improves functioning at 6 months.
Methods: This was a two-group, parallel, randomised clinical trial with the outcome assessed by blinded examiner at 6 months after injury. 366 of 389 eligible subjects aged 16 years or older with MTBI were enrolled in the emergency department, with an 85% follow-up completion rate. Five telephone calls were completed, individualised for patient concerns and scripted to address education, reassurance and reactivation. Two composites were analysed, one relating to post-traumatic symptoms that developed or worsened after injury and their impact on functioning, the other related to general health status.
Results: The telephone counselling group had a significantly better outcome for symptoms (6.6 difference in adjusted mean symptom score, 95% confidence interval (CI) 1.2 to 12.0), but no difference in general health outcome (1.5 difference in adjusted mean functional score, 95% CI 2.2 to 5.2). A smaller proportion of the treatment group had each individual symptom (except anxiety) at assessment. Similarly, fewer of the treatment group had daily functioning negatively impacted by symptoms with the largest differences in work, leisure activities, memory and concentration and financial independence.
Conclusions: Telephone counselling, focusing on symptom management, was successful in reducing chronic symptoms after MTBI.
Trial registration number: ClinicalTrials.gov, #NCT00483444
Concussion or mild traumatic brain injury (MTBI) is a significant public health problem affecting approximately 1 million people annually.1 Total lifetime costs due to MTBI occurring in 2000 (ie, medical costs and productivity losses) are estimated to be US$ 4.1 billion. Although the majority of individuals with MTBI are expected to recover within weeks and months of injury, 10–15% are estimated to have persistent post-traumatic complaints with associated disruption of daily activities, often characterised as “post-concussion syndrome”.2 3 Although this is a small fraction of the cases with MTBI, the actual numbers involved are large owing to the high incidence rate of MTBI. Patients with persisting post-traumatic symptoms present numerous challenges to the healthcare system, including a high consumption of medical services, lost productivity, emotional distress and reduced quality of life.4
Although there is controversy regarding factors underlying these persisting symptoms (ie, neurological vs. psychological), a number of studies have attempted to intervene early after injury in an effort to minimise persisting symptoms and improve function. Some studies have been conducted that suggest beneficial effects for individuals receiving special treatments compared to standard treatment. In one such study, persons with MTBI enrolled within 48 hours after injury who met with a clinician 1 week later and received an information booklet about MTBI, along with suggestions for coping strategies, reported fewer symptoms and less stress 3 months after injury.5 This study had a 63% follow-up rate for those receiving the intervention; the rate for the control group was not reported. In another study with a small sample size, individuals received an education booklet and met with a therapist to review the information. Individuals in the treatment group had fewer symptoms and of less duration at 6 months compared to the control group.6
Other studies have shown no difference between interventions and control. For example, one study found that a single in-person educational session conducted within 3 weeks of injury focusing on education, reassurance and coping strategies was as effective as weeks of more intensive treatment at 3-4 months and 1 year, with both groups demonstrating improvement.7 8 Other studies have found similar results with in-person educational interventions, but are hampered by poor follow-up and lack of randomisation.9 10 In addition, all of these studies required the subject to return to the healthcare facility for the intervention. None of the studies describe in detail the participants excluded from the studies, leaving the possibility for bias in the treated populations or poor generalisability of the results.
Telephone contact has also been used in interventions with mild brain injury populations as part of larger studies. In one study with persons with varying degrees of injury severity, results showed no differences between individuals receiving both telephone and in-person treatment versus standard care. A second follow-up study at the same location enrolling only hospitalised persons with TBI noted improvement in symptoms at 6 months.11 12 The telephone contact in these studies was used as an adjunct to other types of contact and was not assessed separately. A recent study used a similar methodology of providing specialised treatment in person or by phone, but only for individuals with complaints of post-concussive symptoms 2–8 weeks post-injury. They found improvement at 1 year for those with symptoms who received intervention compared to the control group, which received standard care.13
In this article, we report the results of a single-centre, prospective, randomised trial that compared standard care with an intervention consisting of a series of telephone contacts over 12 weeks aimed at ameliorating post-traumatic symptoms at 6 months. To our knowledge, this is the first study to use only telephone counselling as an intervention to influence outcome of MTBI. In addition, enrollment in the study and baseline assessment were performed within hours after injury, which should improve accurate identification of subjects and reduce the probability of recall bias. Our hypothesis was that individualised education and recommendations for symptom management and physical remobilisation would minimise the occurrence of chronic symptoms.
Subjects in this study were recruited in the emergency departments (EDs) of Harborview Medical Center, a level I trauma centre, and the University of Washington Medical Center. Study entrance criteria included: admission to the ED within 48 hours of injury; likely circumstances for MTBI (ie, motor vehicle crash or fall); ED Glasgow Coma Scale score of 13–15;14 documented, self-reported or witnessed loss of consciousness (LOC) of ⩽30 minutes, a period of impaired consciousness (eg, confusion) of ⩽24 hours or post-traumatic amnesia of ⩽24 hours; sufficient knowledge of the English language to participate in the intervention calls; and permanent home address and phone number. These criteria for diagnosis of MTBI were based on the recommended conceptual definition of the Centers for Disease Control (CDC).4
Study exclusion criteria included: age <16 years; intracranial abnormality on CT; admission to the Intensive Care Unit; serious non-extremity injuries (eg, organ lacerations, pelvic injuries); extremity injury(ies) more severe than a single limb closed fracture; current or progressive neurological disease (eg, multiple sclerosis, stroke, brain tumour) or end-stage terminal disease; evidence of recent or current major psychiatric illness; injury received during sexual assault; and hospitalisation ⩾2 nights for head injury in the past year. Persons with evidence of serious or long-standing alcohol abuse (ie, ED blood alcohol level ⩾250 mg/dL or self-report of alcohol consumption on waking up in the morning) and/or current dependence on illegal or prescription drugs were excluded, as were prisoners or those in custody. This study was approved by the Institutional Review Board of the University of Washington.
Study personnel conducted the subject recruitment, consenting and baseline assessment in the ED, with the exception of some cases admitted to the hospital for treatment of non-excluding injuries or observation in which the baseline assessment was completed following transfer to a medical unit. ED activity was monitored for the first month of the study and analysed for peak activity periods. We staffed the ED during these peak periods for recruitment purposes (including weekend days and evenings, as well as weekday afternoons and evenings). After preliminary determination of eligibility from ED admission logs and patient screening, study personnel administered the Galveston Orientation and Amnesia Test (GOAT) to assess the subject’s level of orientation and capability to give informed consent.15 A score of ⩾75 was considered a sufficient level for consent purposes. The GOAT was followed by a brief structured interview to determine eligibility for the study. The interview questions included: asking the patient to recap the injury events to determine the presence and length of amnesia for events before or after the injury (ie, retrograde (RA) or post-traumatic (PTA) amnesia); and questions regarding the presence and length of a LOC or period of confusion. After consent was obtained, baseline assessments were administered, followed by randomisation. Randomisation was stratified on gender, race (non-Hispanic White vs. other) and age (<40 vs. ⩾40 years) and blocked. The study coordinator entered identifying and stratification information into a password-protected website running an Access program that selected and returned the next assignment from a set of concealed, computer-generated lists compiled by one of the study biostatisticians (JKB). Notification of group assignment was done in person or by mail if the subject had been discharged. If a subject was deemed ineligible following randomisation when the full extent of the injuries became known, the person was notified by mail and were considered lost to follow-up.
The study was a randomised two-group design with a blinded examiner performing the outcome assessment at 6 months after injury. The control group received the usual ED standard of care for MTBI consisting of a patient instruction handout and standard outpatient treatment, if prescribed. The treatment group received scheduled telephone contacts over the first 3 months after injury, along with the standard patient instruction handout, a wallet card with the study’s toll-free telephone number, and the CDC booklet “Facts about Concussion and Brain Injury and Where to Get Help”, in addition to usual care.16 The study was initiated with a third group that consisted of persons receiving only the patient education handout, the toll-free telephone number card and the CDC booklet. This arm was discontinued 6 months after enrollment was begun when the number of cases enrolled was lower than expected. The cases were folded into the usual care group.
The scheduled telephone contact intervention was designed to facilitate recovery from symptoms associated with MTBI and resumption of normal activities. It consisted of an initial phone call within 2 days of injury and four follow-up calls at 2, 4, 8 and 12 weeks after injury. The calls were subject-centred, with the goal of providing information and reassurance on the general course of recovery from MTBI and assisting people in developing a plan for managing specific symptoms related to the MTBI.
The research coordinators all had bachelor or masters degrees in psychology, sociology or anthropology, and experience in interviewing or telephone crisis management. They were trained in motivational interviewing and self-management techniques17–20 to assist in their interactions over the phone, and were provided with education on MTBI symptoms and the typical course of recovery. The calls ranged from brief check-ins lasting less than a minute to more problem-focused calls lasting up to 33 minutes, averaging 8 minutes. Phone calls were recorded and a random subset was reviewed by a physician and a psychologist who supervised the study. Adherence to study protocol was reviewed and calls were also used as an educational tool during group supervision. Group supervision of the research coordinators occurred on a bi-monthly basis and individual coordinator supervision occurred on a weekly basis and as-needed for individual questions/issues as they arose. Emergency back-up by the physician and psychologist was available at any time.
The content of the calls followed a general script that included review of previously noted problems and their current status, new problems or issues, and a prompt to review physical, cognitive, emotional and activity status. The coordinators used techniques of problem clarification and reflective listening to help the subject to identify and prioritise symptoms related to the MTBI. They suggested appropriate management strategies and modelled problem-solving behaviours by prioritising issues and setting short-term goals. In addition, they reinforced positive behaviour aimed at physical mobilisation and provided encouragement for addressing ongoing difficulties. Subjects were advised to seek appropriate medical providers and treatment if their situation warranted. Educational material relating to specific symptoms was mailed when indicated following a phone call.
The baseline assessment at the time of injury consisted of questions relating to demographics (age, education, gender), major role activities (work, school, retired or homemaker), and prior medical history.
All subjects were called at 6 months post-injury for the outcome assessment by an examiner blinded to their treatment assignment. We used a battery of measures to assess post-traumatic symptoms and general health status to determine treatment effects. They included the Head Injury Symptom Checklist,21 the Short Form Health Survey-12 (SF-12),22 Modified Perceived Quality of Life (PQOL),23 Patient Health Questionnaire (PHQ)-Depression and Panic/Anxiety,24 and questions on change in major role performance and participation in community activities. The Head Injury Symptom Checklist consists of 16 symptoms typically associated with MTBI (ie, headaches, fatigue, dizziness, blurred vision, trouble concentrating, bothered by noise, bothered by light, irritability, lose temper easily, memory difficulty, anxiety, sleep trouble, balance, sexual difficulties, coordination, sense of taste). The subject was asked if the symptoms had been present during the preceding 2 weeks and, if present, whether they had experienced any of the symptoms before the injury. Symptoms were counted if they were new following the brain injury or worse now than pre-injury. Subjects were also asked whether their symptoms had interfered in the following 12 functional areas: self-care, walking, travel in the community, work, school, home management, leisure and recreation, social interactions, memory and concentration, standard of living, financial independence, and ability to manage feelings.
This study was designed to evaluate the effect on symptoms and general health of scheduled telephone counselling compared to usual care at 6 months post-injury. The study was originally planned to have three groups of 200 participants each to give 80% power to detect a 16% difference in the symptom-free rate among the groups accounting for multiple testing. After protocol amendment to include two groups rather than three, we recalculated that with the 312 cases with available data at 6 months, the study had 72% power to detect a 16% difference. After the decision to amend the protocol, the randomisation procedure was modified to yield approximately even-sized groups.
The study had two primary outcomes composites chosen prior to the start of the study, each based on several measures. Our intervention, although structured in its approach, was individualised to address each subject’s concerns, symptoms and questions. Because of this individual focus, we felt that any single measure chosen as a primary outcome would be unlikely to demonstrate change. Therefore, we chose composite outcome statistics for each of the two co-primary outcomes. The composite outcome used is the average across measures of the rank (expressed as a percent of participants scoring worse) of the patient on each measure.25
The post-traumatic symptom composite was based on the number of new or worse post-traumatic symptoms and the number of functional areas impacted by the symptoms (Head Injury Symptom Checklist and 12 functional areas listed in Measures section). The general health composite was based on assessments in five areas: physical functioning (SF-12 Physical Score), satisfaction with life (PQOL), emotional well-being (PHQ-Depression, PHQ-Anxiety/Panic, PHQ-Anxiety/Other, SF-12 Mental Score), major role (days missed from work, school or household work, and a score indicating if the work/school situation was better), and community integration (number of days the person participated in each of six activities—informal recreation, organised recreation, volunteer work, informal courses, exercise programme and religious activities—outside the home in the preceding week). Each area received similar weight in the general health composite. For both composites, scores can range from 0 to 100, with a mean of 50, and higher scores indicating more favourable outcome.
The two groups were compared on baseline characteristics using Mann–Whitney tests for ordinal or continuous variables, and Chi-squared tests for unordered categorical variables. The effect of treatment was assessed by linear regression on the post-traumatic symptom and general health composites controlling for the three stratification factors (gender, race/ethnicity and age) and baseline factors that were imbalanced despite the randomisation. A two-sided significance level of 0.025 was used for each test to account for the two primary outcomes. Similar analyses were run for four subgroups consisting of: 1) non-Hispanic Whites; 2) minority race or ethnicity; 3) males; and 4) females to see if the treatment effect was similar in each of these groups. The percent of participants endorsing each symptom and area of impact was assessed by Chi-squared tests with confidence intervals on the difference in the rates.26 All analyses were intent to treat.
Figure 1 shows the participant flow through the study. All baseline measures were completed while the subject was still in the ED or within the first 24 hours of admission to the hospital. Few people refused to be assessed for eligibility or to participate in the randomised study if eligible. Seventy-nine percent of subjects randomised to receive telephone intervention participated in at least 4 of the 5 scheduled calls and 86% were assessed for outcome. Five cases were randomised in error and were not followed for outcome. Four had injuries severe enough to meet the exclusion criteria and one had no permanent address. Patients were recruited between 2/23/04 and 2/10/06 with follow-up extending to 9/5/06. No emergency situations requiring immediate back-up by the physician or psychologist were identified during the study.
The groups were well-balanced with respect to demographic and injury characteristics (table 1). The group assigned to receive the telephone intervention had more people with injury to other body systems besides the head and more with Glasgow Coma Scale score of 15 in the ED.
People assigned to receive scheduled telephone calls had significantly better outcome with respect to the post-traumatic symptoms composite (p = 0.016; table 2).
There was no effect on the general health composite, which was the other primary outcome. Fewer people in the treatment group had each individual symptom except anxiety in the 2 weeks prior to assessment, with the largest differences in the reporting of new or worse sleep trouble, fatigue, memory difficulties, dizziness, trouble concentrating, blurred vision and sexual difficulties (table 3). Similarly, fewer people assigned to the treatment group had areas of functioning impacted by symptoms, with the largest differences in negative impact on memory and concentration, work, leisure activities and financial independence. Secondary analyses of subgroups based on gender and race/ethnicity suggest that males had more improvement in overall symptoms than females, although the observed effect in females was also slightly positive (see table 4).
Persons with acute MTBI who received 4–5 telephone counselling sessions aimed at education, symptom management and encouragement to resume everyday activities over 12 weeks after injury demonstrated fewer symptoms and less effect of symptoms on everyday functioning at 6 months after injury. This type of intervention reduces the development of chronic post-traumatic symptoms and their disruptive effects on everyday functioning, the so called “post-concussive syndrome”. Virtually all symptoms were endorsed less in the treatment group, with the greatest effect in the areas of fatigue and sleep. Males appeared to benefit more from the intervention. There were no differences in treatment outcomes based on race or ethnicity.
Interestingly, although there was a difference between the groups on the symptom composite (presence of new or worsened symptoms and perceived effect of symptoms on function), there was no difference between the groups on the general health composite, which included measures of health and emotional status, and productive activity. The intervention itself specifically targeted symptom management and so, perhaps, it is not surprising that the most robust gain was an improvement in the occurrence or worsening of symptoms as opposed to other more general health and activity measures. Another potential reason for this finding includes the nature of the measures. The head injury symptom checklist asks the patients to indicate whether they have symptoms and, if they do, whether the symptoms are new or have worsened since the injury. Here, individuals are asked to make a judgment regarding pre-post injury change (within-individual change) and the individuals may be the best assessors of this after taking into account the impact of other factors in their lives. In contrast, almost all of the general health measures are based on what the individual is doing or feeling at that point in time regardless of cause and the comparison is between the treated and the control groups. Within-group variability is likely to be much larger than within-individual variability, masking potential benefits of the treatment. In fact, several findings give hints in that direction. On comparing the employment situation change from pre-injury to 6 months post-injury, a within-individual comparison, 6% of those in the intervention group reported a worsening job situation, whereas 17% of the control group reported worsening; both groups had 8% reporting an improved job situation with the remaining subjects reporting no change (p<0.10, Mann–Whitney test). Similarly, on looking at days missed from work, school or household responsibilities due to the injury, there was a trend towards shorter time away for those in the treatment group (p<0.10 Mann–Whitney test; 53% vs. 41% back in a week or less). These differences are not significant, but are compatible with the findings on the impact of symptoms on role activities.
The positive finding on the measure requiring the patient to report change as a function of the injury, but not on those assessing level of functioning or performance or mood at that specific point in time compared to controls, is important and deserves some comment. Patients often complain of fatigue and problems with memory and concentration, although they have returned to their prior roles and responsibilities. Functional imaging studies performed in persons with MTBI, including those with sports injuries, indicate increased activation (representing greater effort) while performing on neuropsychological measures in spite of comparable actual performance compared to non-injured controls.27–29 It is unfortunate that subjective reports of fatigue and cognitive inefficiency are often dismissed because of a lack of “objective” evidence and the possible effect of other factors (eg, litigation) on subjective reporting. Yet, subjective report may be the most sensitive measure available given the absence of baseline measure on persons injured.
Our study used only telephone counselling to achieve symptom reduction in contrast to previous MTBI intervention studies that relied on face-to-face education and counselling or a combination of face-to-face with telephone contact.5–7 11–13 An important finding of this study relates specifically to the advantages of a telehealth approach, as underscored in the review by Liss et al (2002)30–that is, a) increased access to rehabilitation for underserved populations, b) increased efficiency in the delivery of rehabilitation and health services, particularly rural access, and c) lower cost of services without sacrificing the quality of care. Paraprofessionals can be trained, as was done in this study, to conduct this type of telephone intervention. Although a common perception in mental health is that visual contact is a requirement for intervention, this perception has not been shared by consumers with disabilities and their families and is not supported by this study’s findings.31
This study achieved a very good follow-up rate for all subjects in comparison to many studies with subjects recruited in an ED. We feel that two factors contribute to this rate. First, the research coordinators who enrolled the subjects were scrupulous in obtaining multiple means of contacting the subjects afterwards—including home and mobile telephone numbers, names and numbers of relatives, and even work contacts, and updated these contacts at every telephone contact. Second, the outcome examiners were selected for having good interpersonal skills and were expected to use all available information and to obtain high follow-up rates.
There are also some limitations to the present study. There were some differences between the intervention and control groups, with more individuals in the intervention group having a GCS of 15 and having associated injuries. However, regression models adjusted for these differences in groups, along with age, gender and race. It is difficult to know the degree to which the subjects enrolled are representative of patients with MTBI in general. Most of our enrollment occurred at a single Level I trauma centre. We did not screen every case or attempt a sampling method to ensure representative samples within this trauma centre, and we eliminated certain classes of subjects whom we could not accurately assess for the occurrence of MTBI or for whom we were not able to implement the intervention (eg, no telephone). However, within these constraints of subject selection, the trial was randomised. Therefore, in spite of limitations of representativeness of the subjects, the results are internally valid in suggesting treatment efficacy for subjects such as those included in our study. In order to establish generalisation to the entire population of MTBI, a multi-centre effectiveness study would be needed.
We did alter the study protocol during the course of the study, changing from the original plan of three groups to two groups. Because identification and enrollment were less than we had originally predicted, the group receiving an educational pamphlet at the time of enrollment and baseline assessment were folded into the usual care group. Because of the extra educational material afforded to these subjects, it is likely that the effect of the telephone intervention would be slightly underestimated, rather than overestimated. We cannot identify the specific elements of the intervention that made the difference in outcome (eg, content of the intervention, number of contacts, proactive contact by personnel). Further studies are needed to identify the active ingredients of this type of intervention.
With these caveats, this study represents the largest randomised intervention for MTBI to date, with a sufficiently large sample to detect important differences and excellent follow-up. The results demonstrate that telephone counselling with a focus on symptom management and encouraging an early return to everyday activities was successful in reducing the development of chronic symptoms after MTBI.
We would like to acknowledge Darwin Hoop, Nicholas Vanneman, Nadya Zawaideh and April Sather for their excellent work in data collection; and Drs Michael Copass, MD, Medical Director Emergency Services of Harborview Medical Center, and Carin Olson, MD, Emergency Services, University of Washington Medical Center.
Author Contributions: KRB had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study Concept and Design: KRB, SD, PCE, NRT, JMP, JMH, RTF.
Data Acquisition: KRB, JMP, JMH, JKB, NRT.
Analysis and Interpretation of Data: KRBG, JMH, SD, NRT, JMP, RTF, JKB.
Drafting of Manuscript: KRB, JMH, SD, NRT, JMP, RTF.
Critical Revision of Manuscript for Important Intellectual content: KRB, JMH, SD, NRT, RTF, PCE, JMP, JKB.
Statistical Analysis: JKB, JMH, NRT, JKB, SD.
Obtained Funding: JKB, SD, NRT, PCE, JMP, JMH, RTF.
Administrative, Technical or Material Support: KRB, NRT, SD, PCE, RTF.
Study Supervision: KRB, JMH, JMP, NRT.
Competing interests: None.
Funding: This research was funded by the Centers for Disease Control, Award No. R49/CCR023226-03. The Centers for Disease Control provided funding for the research but had no control over the design, data collection, analysis or interpretation of the data.