Populations are aging worldwide [1]. In Canada, the rural population is older and is disproportionally aging relative to the urban population [2]. The incidence of dementia increases with age, and formal dementia services are the least accessible in rural and remote communities where the proportion of older adults is the greatest [2,3]. Long travel distances and transportation difficulties further limit accessibility [3]. Telemedicine, or telehealth, is the remote delivery of health care services, where distance is a critical factor, by means of information and communications technology [1]. Expanding telehealth services has been suggested to reduce disparities in urban and rural health care [4]. Interventions to support the function of older adults with cognitive concerns, including dementia, are needed, and it is essential that these interventions are accessible to all families. Cognitive rehabilitation is a promising but understudied nonpharmacological individualized treatment that has been shown to help individuals with mild cognitive impairment (MCI), early stage dementia due to Alzheimer disease (AD), and vascular dementia set [5] and attain personally important functional goals [6-8]. This study investigated the feasibility and acceptability of delivering cognitive rehabilitation to individuals with subjective cognitive impairment (SCI), MCI, and early stage dementia using telehealth videoconferencing.

Research on telemedicine and dementia has primarily focused on diagnosis (eg, [9,10]), clinical consultation, follow-up appointments [11], and support for family caregivers [12-14]. In comparison, relatively little work has studied interventions for individuals diagnosed with dementia, although research in this area has begun to emerge. For example, Dal Bello-Haas et al [15] demonstrated that videoconferencing is a feasible method to deliver an exercise intervention for rural individuals with dementia due to AD. These examples of successfully delivering remote interventions for individuals with dementia suggest that remote cognitive rehabilitation could be feasible.

Telerehabilitation is “the set of instruments and protocols aimed at providing rehabilitation at a distance” [16]. Telerehabilitation has been used to provide a range of interventions to individuals diagnosed with a number of different disorders. For example, telerehabilitation has been used to treat stroke [17], spinal cord injury [18], traumatic brain injury [19,20], multiple sclerosis [21], and cognitive impairment following intensive care [22]. Diverse use of telehealth includes delivering diagnostic assessments, caregiver support groups, individual and group psychotherapy [23], home exercise programs, clinical consultations, and cognitive rehabilitation using information and communications technology [16].

Clare and colleagues have developed a goal-focused approach to cognitive rehabilitation for individuals with early stage dementia due to AD, or mixed AD, and vascular dementia [7,24]. In this approach, cognitive rehabilitation begins with an assessment, which is followed by collaborative goal setting [24]. Typically, functional goals related to everyday memory problems, practical skills, and activities and concentration are set, and improved function in these areas has been reported in multiple studies (eg, [25-31]). Generally, these collaborative goals are addressed in weekly 1-hour sessions using empirically supported techniques such as spaced retrieval, cuing and fading, errorless learning, and external memory aids [24]. Briefly, errorless learning involves training without allowing trial-and-error guessing; cues and supports are given to maximize the probability of guessing the correct response instead [24]. Numerous methods facilitate errorless learning, which includes cuing, gradual fading of cues, and spaced retrieval where recollection occurs over increasing time intervals [24]. Although promising, research evaluating cognitive rehabilitation for individuals with early stage dementia is still emerging [32].

The majority of literature on remotely-delivered cognitive rehabilitation focuses on interventions with individuals who have a traumatic brain injury (TBI). Early research suggests that remotely-delivered rehabilitation for individuals who have sustained TBI is feasible (eg, [33,34]). For example, Tam and colleagues [35] reported a series of 3 case studies where individuals with TBI participated in cognitive rehabilitation using customized software. This software combined videoconferencing with screen-sharing, and participants completed computer-based activities that targeted word recognition, semantic memory (ie, memory for factual knowledge), and prospective memory (ie, memory to perform something in the future). In other work, Bergquist and colleagues [33] opted to use technology to remotely teach participants who had had a severe TBI to use a calendar as a compensatory memory strategy. They adapted Sohlberg and Mateer’s [36] calendar training procedure to an instant messenger format and also taught participants to use a personal diary [33]. Both interventions led to increased use of compensatory strategies and improved mood [33]. Finally, in an approach that is more similar to the type of cognitive rehabilitation reported here, Bourgeois and colleagues [37] had participants with chronic TBI identify three everyday memory problems (ie, forgetting appointments, forgetting day planner at home, and losing items), and they provided either an errorless learning approach, spaced retrieval, or memory strategy instructions over the telephone. Individuals in the spaced retrieval group made greater gains in their target goals than those given strategy instructions, and both groups improved their everyday memory functioning. These studies suggest that traditional, in-person cognitive rehabilitation strategies can be delivered by videoconferencing, instant messaging, or telephone.

Remotely-delivered cognitive rehabilitation has also been demonstrated for persons with dementia. Joltin et al [38] used the telephone to train spaced retrieval, a memory intervention, to help 3 women previously diagnosed with dementia recall target information. The goals addressed using spaced retrieval were set in collaboration with family caregivers, staff at the assisted living facility, and the individual diagnosed with dementia [39]. Two participants set the goal to recall what time to take their medications, and one participant set the goals to recall her grandson’s names, her room number, and the year [39]. The first participant (Mini-Mental State Exam, MMSE=17) did not always answer the telephone when the researchers called to provide spaced retrieval training and after 4 sessions, she was still unable to recall the times to take her medication for longer than 2 min [39]. The second participant (MMSE=17) was able to pick up a prompt card listing the times she needed to take her medications across a 5-min interval at the conclusion of the intervention. The third participant (MMSE=13) achieved all three of her goals (grandchildren’s names, room number, and year) and was able to recall the target information across 3 sessions [39]. The authors concluded that it is feasible to modify spaced retrieval for remote delivery. To date, no research has explored videoconferencing for remote delivery of cognitive rehabilitation.

Telerehabilitation is a developing field with the promise of increasing the accessibility of specialized interventions such as cognitive rehabilitation. To date, remotely-delivered cognitive rehabilitation for persons with dementia has not been systematically studied. Interventions that are included in cognitive rehabilitation, such as spaced retrieval, have been applied in a telerehabilitation format, suggesting that this may be an acceptable and feasible approach to increasing the accessibility of cognitive rehabilitation for dementia for persons residing in rural and remote areas. The purpose of this study was to investigate the acceptability and feasibility of delivering cognitive rehabilitation to individuals diagnosed with dementia due to AD using telehealth videoconferencing.

This small-scale randomized control trial used a combined single-case and between-subjects design. Random assignment to in-person versus videoconferencing conditions occurred by random number generator. Random assignment to condition occurred before recruitment (conditions were determined before the study commenced and hidden in envelopes) and was not stratified by diagnosis or by rehabilitation goal. Regardless of randomized condition, all participants received an in-person initial assessment. At the initial in-person assessment participants selected at least two goals for cognitive rehabilitation. The features of a between-subjects design were combined with the features of a multiple-baseline design. Multiple baselines were measured within-subjects, and treatment modalities were compared across participants. After 3 weeks of baseline assessment, Goal 1 was targeted and baseline assessment for Goal 2 continued. After 3 weeks of Goal 1 intervention, Goal 2 was targeted. In this way, both the in-person and telehealth groups were observed repeatedly during the baseline and treatment phases. The repeated observations over the baseline and treatment phases meet the criteria for a multiple-baseline design (across groups) [39]. In single-case experimental design guidelines, three is the minimum number of data points required to establish a baseline, and the minimum number of data points needed in each phase [40].

The University of Saskatchewan Behavioural Ethics Review Board provided ethical approval. Participants were recruited through community-based organizations and a hospital-based geriatric assessment program. Initially, the researchers hoped to recruit participants solely from clinical settings, but low enrollment led us to expand the recruitment strategy and inclusion criteria. With the expanded criteria, individuals with SCI and no diagnosis, MCI, early stage dementia due to AD, or mixed AD and vascular dementia, were all invited to contact us if they were interested in participating in the study. Diagnosis was self-reported (ie, participants reported that they had received a diagnosis of dementia due to AD from their neurologist, reported no diagnosis), but all self-reported diagnoses were consistent with the clinical interview, neuropsychological tests, and questionnaires administered in the assessment for the study. Prior to enrolling in the study, participants completed a brief clinical interview where cognitive rehabilitation was reviewed, informed consent was obtained (from participant and caregiver wherever appropriate), and a Mini-Mental State Exam (MMSE; [41]) was administered. All individuals were encouraged to participate with a family member or friend, but this was not mandatory.

Two sets of measures were used in this study: pre-post measures and weekly measures. A set of measures was administered in-person to all participants before and after the intervention. Second, weekly observational measures and measures of goal performance and satisfaction were collected using the medium of in-person versus telehealth as per the random assignment. The measures were selected to be similar to those used by Clare and colleagues in their 2010 randomized control trial [7].

Cognitive rehabilitation followed the procedures outlined by Clare [24] in her manual Neuropsychological Rehabilitation and People With Dementia. This manual emphasizes individualized, person-centered goal setting. One participant set goals related to mood and sleep. Here, cognitive behavioural strategies were used to treat insomnia [57] and low mood [58,59]. All of the interviews, assessments, and interventions were completed by a senior doctoral student in clinical psychology (RLB) and supervised by a neuropsychologist (MEO).

Eight individuals were recruited to participate in this study, participants were immediately randomly assigned to in-person versus telehealth, and the initial assessment was completed in-person. Two discontinued the study following the initial assessment, and these two happened to be randomly assigned to the telehealth condition. In one case, the family member support person reported that she and the participant did not have time to participate. In the other case, the family member support person reported that initial assessment had been distressing for the participant, and following a family discussion it had been decided that participating in the study was likely to be more distressing than helpful. Demographic and descriptive data for the six individuals who participated are presented in Tables 1 and 2.

Participants each collaboratively set between one and five goals for cognitive rehabilitation. Table 3 lists the specific goals and the cognitive rehabilitation strategies used to address them. The study was designed to deliver 8 sessions of cognitive rehabilitation after 3 sessions of the baseline phase. All 3 participants randomly assigned to the in-person intervention completed 8 sessions. In the telehealth group, one individual completed 8 sessions, one individual (Ms D) completed 7 sessions, and one individual (Ms F) completed 6 sessions. Ms D reported that she had decided to go on vacation, and therefore the researchers decided to cancel the final telehealth rehabilitation session and complete posttreatment assessment. Ms F only had one goal for cognitive rehabilitation, and she felt it had been accomplished after 6 telehealth cognitive rehabilitation sessions.

The primary outcome measure was goal performance as measured by the COPM. Figures 1 to 6 display the COPM scores across the baseline and intervention phases for each of the 6 participants.

The pretreatment and posttreatment scores for the secondary outcome measures are presented in Table 2. To facilitate comparisons reliable change indices (RCI) were provided where they were available in the literature. When RCI were not available, standard error of the difference (SE_D) or minimum clinically important differences (MCID) were provided in Table 2.

There were few changes in the secondary measures that exceeded these estimates of change, and only changes greater than the SE_D, RCI, or MCID are reported below.

The research journal was used to reflect on the process of conducting this study, to document any challenges and successes that may not have been fully captured by the quantitative measures, and to document modifications that were made to deliver the intervention to the individuals in the telehealth group. The codes that were generated were organized into two major themes: “relationship and therapeutic alliance” and “method and technique.” Text pertaining to how the researcher (RLB) felt working with the participant, comments the participant made regarding comfort, or how they felt in the session were coded in the “relationship and therapeutic alliance category.” “Engagement” (interest in the intervention and attendance), “connection and enjoyment” (partnership with participants and fun during the sessions), and “responsibility” (researcher’s sense of personal accountability) were coded as subthemes. Text pertaining to study design, measurement, or comparisons between in-person and telehealth treatment were coded in the method and technique theme. “Adjustment to telehealth” and “challenges of measurement” were the themes within “method and technique.” The findings of the thematic analysis are summarized in Table 4, and characteristic examples of text from the journal are presented in the table.

The research journal reflects the importance of building rapport and an alliance to carry out the cognitive rehabilitation. Initial codes were organized into the minor themes of “engagement,” “enjoyment and connection,” and “responsibility.” It is notable that these themes were similar irrespective of treatment modality (in-person vs telehealth).

The theme of “method and technique” is comprised of journal entries that comment on the adaptation of cognitive rehabilitation to telehealth videoconferencing. Participants adjusted easily and quickly to working through telehealth videoconferencing. Journal entries were organized into the subtheme of “different but not worse.” Participants commented that although they might have preferred to meet in-person, the telehealth sessions ran smoothly. As a clinician, the researcher (RLB) noted challenges because of not being able to physically interact with materials. For example, the researcher was unable to pick up a day timer and read through what the participant had written. Therefore, the researcher had to cue participants to read out written notes or from worksheets. Initial codes in the research journal were organized into the minor theme “greater reliance on verbal description.” Lastly, journal entries comment on working to adjust and modify goals and sessions to make the intervention measureable and adhere to the multiple baseline design. Initial codes were organized into the subtheme “measurement challenges.”

The results of this study cautiously suggest cognitive rehabilitation can be adapted to telehealth videoconferencing for older adults with subjective and objective memory impairment. The study also adds to the growing body of literature that suggests goal-oriented cognitive rehabilitation delivered in-person is a promising nonpharmacological intervention for older adults with SCI, MCI, and early stage dementia due to AD. For participants who completed the initial assessment and baseline sessions, participation was excellent with all in the in-person group completing 8 out of 8 sessions. Although two in the telehealth group completed fewer than 8 sessions, one terminated because of the goal having been attained and the other terminated because of scheduling conflict with vacation. Both groups demonstrated high completion rates, however the lower rate of session completion for the telehealth group may suggest that telehealth-delivered treatment is less acceptable to participants or something about this modality of treatment (such as the virtual nature of the interpersonal connection, or added challenge of describing steps and materials verbally rather than physically interacting or handing something in to be read) delivery made completion of the sessions less motivating. Despite this caveat, the high completion rates over 12 weeks (3 baseline and 8 intervention sessions) suggest participating through either delivery modality was acceptable to participants. The themes from the research journal also support this conclusion; although some participants assigned to telehealth were initially apprehensive or even disappointed to be assigned to the telehealth condition, as sessions progressed the theme “different but not worse” as a description of videoconferencing-delivered sessions emerged from the research journal entries. These qualitative data revealing apprehension about telehealth could have interacted with the decision to withdraw made by the 2 participants initially assigned to telehealth, but who withdrew after the initial assessment and before the intervention. In order of theme of “different but not worse” to be created, an individual needed to first interact through telefhealth to discover that expectancies regarding apprehension of technology were unfounded.

Importantly, the results suggest participants’ goal performance improved across both treatment delivery modalities. Of the 15 goals set in this study, performance on only two goals (sleep set by Participant C, and concentration set by Participant D) did not improve by 2 or more points on the COPM. Participants C and D were both assigned to the telehealth group, so this raises the possibility that telehealth may reduce the efficacy of cognitive rehabilitation for older adults with SCI. It may also be the case that these goals are less amenable to cognitive rehabilitation. Improved sleep, in particular, is not a typical goal for cognitive rehabilitation, however, improving sleep and managing daytime sleepiness were both reported as goals set in Clare and colleagues’ [26] study (goal attainment was not reported goal by goal). In the sleep intervention literature more generally, CBTi is an effective treatment, demonstrates efficacy that is similar to pharmacological interventions with better long-term outcomes, and has been recommended as a standard treatment for insomnia [68]. Importantly, a full course of CBTi, which is typically between 6 and 8 sessions was not delivered here (Ms C participated in 5 sessions that focused on her goal to improve her sleep). Overall, the results of this study suggest that it is worthwhile to pursue adapting cognitive rehabilitation to telehealth videoconferencing. This is consistent with previous research that has explored remotely-delivered cognitive rehabilitation (ie, [37]) as well as remotely-delivered psychotherapy (ie, [14,23]).

The importance of establishing a strong therapeutic relationship was a major theme that emerged from the research journal. This aspect of cognitive rehabilitation has perhaps not been emphasized enough in the literature, and clinicians who are providing the intervention (whether or not they have been trained as psychotherapists) may benefit from attending to the research on the common factors of psychotherapy (see [69] for a recent summary of the common factors literature based on meta-analyses). This is not to suggest that the therapeutic relationship has been ignored in the cognitive rehabilitation literature, but to highlight the importance of not emphasizing technique (ie, errorless learning and spaced retrieval) at the expense of developing an alliance. One imagines that telehealth videoconferencing could impact developing an alliance, however, this was not noted in the research journal and psychotherapy noninferiority trials (ie, [23]), and other videoconferenced work [14] detail how the therapeutic relation can be established and maintained remotely. Future researchers might consider adding a formal measure of alliance to their protocols.

In carrying out this study, the researchers learned a number of things that may be helpful for future researchers. First, the researchers were surprised by how challenging it was to recruit research participants with MCI or early-stage AD. Those recruited and retained in the study were highly motivated and engaged, which is a self-selection bias. This recruitment challenge and the way in which participants were randomly assigned to the in-person or telehealth videoconferencing limits the conclusion the researchers can draw about delivering cognitive rehabilitation through videoconferencing to individuals with MCI or dementia due to AD (the 3 telehealth participants were individuals with subjective memory impairment). The researchers were also surprised to find that the majority of the participants in this study opted to participate without a support person. This was either because no support person was available (Ms C), because a support person was not interested in participating (Mr B and Ms E), or because it was decided that the support person was not needed (Ms D and Ms F). Only Ms A’s husband accompanied her to every session. This is noteworthy because Ms A was the only participant with a diagnosis of dementia due to AD. Previous research (ie, [8]) has recommended that a support person always be included in the intervention. The results of this study suggest that for individuals with SCI a support person is not necessary, but for individuals with dementia due to AD or MCI, the researchers continue to recommend a support person.

The researchers also came to reconsider the experimental design. The multiple baseline design was chosen to be able to infer that any improvements in COPM scores were due to cognitive rehabilitation interventions rather than common therapeutic factors such as establishing a positive relationship with the researcher delivering the treatment. As the study progressed, it became apparent that skills being taught for one goal carried over to other goals, and in some cases (Participant D in particular), participants spoke explicitly about generalizing strategies from a goal that was being trained to a goal that was not being trained. Of course, this is excellent for that individual, but it does limit the usefulness of within-person multiple baselines for cognitive rehabilitation. Therefore, this type of experimental design is not recommended, at least not for similar goals. Furthermore, in this study the researchers chose to rely on visual inspection to examine these data. This has the advantage of highlighting strong effects, which are more likely to be functionally relevant. The limitation of this approach is that subtle trends such as serial dependency are not readily observable using visual inspection, visual inspection is unreliable when effects are not large, and for these reasons, statistical methods of analyzing single-case data have been increasingly studied and used [60].

The findings presented in this study support developing goal-oriented cognitive rehabilitation delivered both in-person and expanding the accessibility of this intervention by adapting it to videoconferencing. Further research is needed to replicate the results presented here. Additionally, this data does not fully explore the adaptation of cognitive rehabilitation to videoconferencing for individuals with cognitive impairments consistent with MCI or dementia due to AD. Given the increasing prevalence of cognitive impairment in late life in both urban and rural areas, interventions aimed at supporting the personally relevant functional goals of these individuals are clearly needed.