CRITICALLY APPRAISED PAPER (CAP) FOCUSED QUESTION: For adults who have an acquired brain injury, what is the effect of personal assistant devices on memory, compared with standard occupational therapy? Lannin, N., Carr, B., Allaous, J., Mackenzie, B., Falcon, A., & Tate, R. (2014). A randomized controlled trial of the effectiveness of handheld computers for improving everyday memory functioning in patients with memory impairments after acquired brain injury. Clinical Rehabilitation, 28(5), 470 481. http://dx.doi.org/10.1177/0269215513512216 CLINICAL BOTTOM LINE: This is the first randomized controlled trial to assess the use of a low-cost model of a personal assistant device and its applicability to the improvement of memory among people with acquired brain injury. The study based its design on previous research done with pagers. The study had a sample size of 42 participants: Twenty-one in the intervention group received personal assistant devices and training from an occupational therapist for 8 weeks, and 21 in the control group received standard occupational therapy for 8 weeks. The occupational therapist trained the client to use the personal assistant device, as well as instructed him or her in methods to regain memory function in relation to meaningful goals. Participants in both groups identified their own goals in regard to memory and worked toward completing them in the study. In relation to occupational therapy practice, the results of this study demonstrate that handheld computers may be an effective technique to enhance memory among people with a traumatic brain injury. It is within an occupational therapist s scope of practice to educate clients with brain injuries in the use of evidence-based strategies for memory. This potentially low-cost intervention, in comparison with the cost of caregiver time, could be implemented easily in the clinic. However, more research is needed to strengthen the evidence for this approach in occupational therapy. As for limitations, this randomized controlled trial was only able to blind the assessor of the data obtained from the intervention and quantified memory outcomes with several subjective measurements. RESEARCH OBJECTIVE(S) List study objectives. To determine whether patients with brain injury would benefit more from the use of personal assistant devices in comparison with nonelectronic memory aids 1
DESIGN TYPE AND LEVEL OF EVIDENCE: Level I: randomized controlled trial SAMPLE SELECTION How were subjects selected to participate? Please describe. Participants were selected from a private occupational therapy practice in Queensland, Australia, and the Brain Injury Rehabilitation Unit and Brain Injury Community Rehabilitation Service at the Royal Rehabilitation Centre Sydney, Australia. Inclusion Criteria Participants were invited to be a part of the study if they were older than 17 years of age, were diagnosed with acquired brain injury, had functional memory impairment, had come out of posttraumatic amnesia, and had sufficient hand function to be able to use a personal assistant device. Exclusion Criteria Participants were excluded if they had severe physical disabilities, had no memory impairments, lived outside of metropolitan Sydney, or did not speak English. Exclusion factors labeled as other were not described. SAMPLE CHARACTERISTICS N= (Number of participants taking part in the 42 study) #/ (%) Male 26/(61.9%) #/ (%) Female 16/(38.1%) Ethnicity NR Disease/disability diagnosis Acquired brain injury INTERVENTION(S) AND CONTROL GROUPS Add groups if necessary Group 1: Intervention group Brief description of the intervention The intervention group (participants and their caregivers) was trained to use personal assistant devices as memory aids. The intervention was conducted at the Brain Injury Rehabilitation Unit and Brain Injury Community Rehabilitation Service at the Royal Rehabilitation Centre Sydney, Australia. The participants were given the lowest cost model available during the time period in which in the intervention occurred. The device contained a calendar, alarm, camera, and address book but no telephone (the telephone was determined to be a burden of care before the study implementation). 2
How many participants were in the group? Where did the intervention take place? Who Delivered? How often? For how long? The first session focused on identifying goals that were important to the participants. These were based around meaningful activities, including transportation, pet care, medication use, and shopping for groceries. Neurological occupational therapists conducted five training modules, which featured choosing an appropriate digital device for the client, training on its use for organization and application skills, and training on the awareness of deficits. The therapists allowed participants to take as long as they needed to learn to use a proper personal assistant device, with awareness of their own deficits. A total of eight sessions were conducted, on average, varying depending on how much assistance was needed, with a median of 40 min per session. Sessions were conducted over the course of 8 weeks. The client s caregivers were involved whenever possible. The occupational therapist used an errorless model, emphasizing the importance of not making mistakes. This model allowed the therapist to tailor session time on the basis of the individual participants needs. 21 NR Occupational therapist One session per week, on average Seven total hours on average over all sessions, 40 min per session Group 2: Control group Brief description of the intervention How many participants were in the group? Where did the The control group received standard occupational therapy with nonelectronic memory aids, with the same approach as the intervention group and the same focus on meaningful activities. The memory aids included creating lists, keeping a journal, integrating mnemonics, and using cueing strategies. Sessions were individual and group based. Once the study was completed, the control group was given personal assistant devices. Both participants and caregivers took part in this group as well. Overall, there was an average of nine sessions. The therapists used the errorless model in this group as well. 21 NR 3
intervention take place? Who Delivered? How often? For how long? Neurological occupational therapist One session per week, on average Eight total hours on average over all sessions, 40 min per session Intervention Biases: Circle yes or no and explain, if needed. Contamination Co-intervention Timing Site Caregivers and other family members assisted in the use of the device. The interventions were performed in the brain injury unit consistently. Use of different therapists to provide intervention MEASURES AND OUTCOMES Complete for each measure relevant to occupational therapy: Measure 1: Name/type of measure Goal Attainment Scale (GAS): functional memory failures goal 4
used: What outcome was measured? subscale, use of prompts subscale, and use of memory aids subscale to make the total score of the GAS Functional memory goal performance, specified for each patient. Each patient set a memory failure goal; t score measure for goals set for use of memory aids Is the measure reliable? Is the measure valid? When is the measure Twice, once at pretest and once at posttest used? Measure 2: Name/type of measure used: What outcome was measured? Memory Functioning Questionnaire (MFQ): Frequency of Forgetting subtest, Seriousness of Forgetting subtest Caregiver s assessment of the participant s frequency of forgetting, caregiver s perspective on how many daily items were forgotten Is the measure reliable? Is the measure valid? When is the measure Twice, once at pretest and once at posttest used? Measure 3: Name/type of measure used: What outcome was measured? Memory Compensation Questionnaire How participant used prompts and memory aids Is the measure reliable? Is the measure valid? When is the measure Twice, once at pretest and once at posttest used? Measure 4: Name/type of measure used: What outcome was measured? Rivermead Behavioural Memory Functioning Questionnaire Whether improvement in functional memory was due to improved cognitive performance Is the measure reliable? Is the measure valid? When is the measure Twice, once at pretest and once at posttest 5
used? Measurement Biases Were the evaluators blind to treatment status? Circle yes or no, and if no, explain. Assessors were blind, but occupational therapists were not. Recall or memory bias. Circle yes or no, and if yes, explain. Others (list and explain): Subjective responses to GAS might have brought bias. RESULTS List results of outcomes relevant to answering the focused question Include statistical significance where appropriate (p<0.05) Include effect size if reported Primary Outcomes The GAS Functional Memory Failures Goal t score and the MFQ Frequency of Forgetting subtest (carer report) yielded significant evidence for an increase in memory in the intervention group, in comparison with the control group. The researchers calculated the differences between the two groups (GAS, p =.0001, effect size = 1.6, CI [1.0, 2.2]; MFQ, p =.021, effect size = 12.5, CI [2.0, 22.9]). No significant group differences were found after calculation. Secondary Outcomes Total GAS t score was not significant in increasing memory (p =.165, effect size = 6.3, CI [ 2.7, 15.4]). This might have been due to subjective responses on this questionnaire. Other MFQ subtests showed an insignificant difference between the two therapy groups (Seriousness of Forgetting subtest, p =.455, effect size = 4.9, CI [ 8.3, 18.3]; Mnemonic Usage subtest, p =.301, effect size = 2.7, CI [ 8.1, 2.6]). Another test that had insignificant findings was the Memory Compensation Questionnaire s External Strategies subscale (making shopping lists, writing notes, putting things in their proper place, marking appointments and numbers, etc.; p =.580, effect size = 0.89, CI [ 2.3, 4.1]. Significant outcome measures included the Memory Compensation Questionnaire s Internal Strategies subscale (using memory tricks, taking your time, planning, repeating, linking old 6
and new, making associations, etc.; p =.021, effect size = 4.4, CI [1.6, 8.4]). In addition, the MFQ s Retrospective Memory Functioning subtest showed significant differences in memory loss between the two study groups (p =.042, effect size = 4.3, CI [1.6, 8.4]). No significant group differences were found after calculation. Was this study adequately powered (large enough to show a difference)? Circle yes or no, and if no, explain. Power was calculated for primary outcome measures only; secondary outcome measures might have Type I error. Were appropriate analytic methods used? Circle yes or no, and if no, explain. Were statistics appropriately reported (in written or table format)? Circle yes or no, and if no, explain. Was the percent/number of subjects/participants who dropped out of the study reported? Limitations: What are the overall study limitations? Possible limitations of this study include the fact that the study could not be performed as a double-blind randomized controlled trial; only the assessors of the data could be blinded. Participants and occupational therapists knew which groups they were allocated to. In addition, the subjective responses to several measures of the study might have introduced bias and compromised validity of the study. For instance, external validity may be questionable depending on what the GAS total score was actually measuring; it might have measured recall of specific tasks rather than executive dysfunction. CONCLUSIONS State the authors conclusions that are applicable to answering the evidence-based question. The study demonstrates the clinical relevance of personal assistant devices for reducing 7
memory loss among clients with acquired brain injury. Particularly, this study shows that the use of personal assistant devices was beneficial in decreasing memory failures in comparison with standard occupational therapy pen-and-paper memory training techniques. Neurological occupational therapists created the intervention, which makes the results of the study more clinically relevant. The potential low cost in comparison with burden of care costs will likely make this method of intervention attractive to clinicians and clients alike. The study s 100% follow-up rate strengthens the evidence for this approach in therapy, because it rules out several potential biases. Education on the use of a handheld computer can carry over into the home and community. This work is based on the evidence-based literature review completed by Savannah Colhour, OTS, and Ashley Halle, OTD, OTR/L, Faculty Advisor, University of Southern California. CAP Worksheet adapted from Critical Review Form Quantitative Studies. Copyright 1998 by M. Law, D. Stewart, N. Pollack, L. Letts, J. Bosch, and M. Westmorland, McMaster University. Used with permission. For personal or educational use only. All other uses require permission from AOTA. Contact: www.copyright.com 8