Abstract
Introduction:
Cognitive complaints are common late effects in patients with cancer, and no standard treatment exists. Recent studies with several patient populations have indicated that there is potential to improve working memory (WM) via web-based WM training. However, the feasibility of including web-based WM training as part of inpatient cancer rehabilitation, in combination with unprompted home-based training, has not been studied. The aim of this study was to assess the feasibility of including web-based WM training (using Cogmed QM) during inpatient rehabilitation and its subsequent unprompted completion in a home-based setting.
Methods:
Patients with cancer who self-reported cognitive complaints were instructed to complete 25 Cogmed QM sessions during their 3-week inpatient multidisciplinary cancer rehabilitation and subsequently at home after discharge from rehabilitation. The feasibility was determined by assessing the study recruitment, adherence to the WM training, improvements in training tasks (compliance measure) and patient experiences by individual interviews.
Results:
Twenty-nine (27 women) of 32 eligible patients (89.6%) started WM training, 1 declined participation and 2 patients withdrew before WM training started. Twenty-six of 29 (89.6%) participants adhered to the intervention during rehabilitation, while 19 of 29 (65.5%) also adhered to the subsequent unprompted home-based intervention. All participants who completed the Cogmed QM sessions demonstrated improvements in the training tasks, as defined by the Cogmed Improvement Index (MD = 24.05, SD = 9.38, range 2-44, P < .011). Interview data suggested that practical limitations, including a lack of time, technical difficulties, difficulties finding a suitable disturbance-free environment and low motivation were barriers to completing the training at home.
Conclusion:
The findings show that it is feasible to include web-based WM training during inpatient multidisciplinary rehabilitation for adult patients with cancer with cognitive complaints. However, patient adherence to unprompted web-based WM training after discharge from rehabilitation was not optimal. Thus, future studies should consider the barriers to adherence and the need for supervision and social support to reinforce home-based training.
Introduction
Cognitive impairments are among the most frequently reported symptoms in patients with cancer.1,2 Longitudinal neuropsychological assessment research indicates that up to 75% of patients with tumors outside the central nervous system (non-CNS tumors) experience cognitive impairments during treatment, and the impairments are still present in up to 35% of the patients many years following completion of cancer treatment. 1 A variety of factors may contribute to cognitive impairment in cancer patients, including chemotherapy and new generations of hormonal therapies, immunotherapy and radiotherapy or a combination of these.3,4 In addition, cognitive impairments are influenced by a other factors, such as fatigue, distress, psychosocial components, genetic polymorphism and metabolic disturbances.4,5
Cognitive functioning is a critical clinical outcome measure due to its impact on the daily functioning and quality of life in patients with cancer. 6 Patients with non-CNS tumors report a variety of cognitive symptoms, such as impairment in attention, executive functions, short-term and working memory (WM) and information processing speed.3,4,7,8 Symptoms may occur temporarily during or immediately after chemotherapy, but can also change more gradually over time after treatment.1,3,9 Furthermore, it has been shown that cognitive impairments can persist for 20 years following treatment. 10 Many patients with CNS tumors are already experiencing cognitive deficits at first presentation of clinical symptoms, and they tend to report several cognitive symptoms, including impairment in language, executive function and memory attention. 11 The cognitive impairments may result from the tumor itself or from the anti-cancer treatment (eg, surgery, chemotherapy and radiotherapy).11,12
There is limited high-quality evidence to reduce cognitive impairments in patients with cancer. 4 Pharmacological therapies, such as modenafil and methylphenidate, have not proven successful and are associated with side effects. 13 However, non-pharmacological interventions such as cognitive function training are promising alternatives. 14 Cognitive function training focuses on improving or maintaining cognitive function by focusing on the structured practice of cognitive tasks. The training has demonstrated positive results in cancer patients in objective cognitive functioning, especially in the domains memory and verbal learning and executive function, letter fluency and processing speed. 15 In addition, some studies have reported positive results for subjective cognitive functions. 15 It has also shown promise in patients with CNS tumors. 12 For example, an early cognitive function training program applied post-surgery to patients with primary brain tumors was found to improve objective cognitive functioning after 2 weeks.15,16
Computerized cognitive training (CCT) uses software programs to improve cognitive function and restore impaired skills through repetitive standardized problem-oriented tasks that target specific cognitive domains. 17 The advantage of CCT interventions is that their difficulty level can be adjusted to match the patient’s capacity level and they can be accessed relatively easily in a home-based setting. 18 For patients with cancer, CCT has been shown to not only improve performance in objective cognitive tests that measure trained domains of cognitive function, but also to have transfer effects; improvements in non-targeted cognitive domains and self-reported assessment of cognitive functioning have been reported.18,19 For patients with CNS tumors, CCT has been shown to improve performance in objective cognitive tests that measure attention, memory and verbal fluency. 20 Thus, it represents a promising potential alternative intervention for the improvement of cognitive function.
Working memory is an important target of CCT, and refers to the small amount of information that can be held in mind and used in the execution of cognitive tasks. 21 WM is commonly viewed as playing a central role in nearly all cognitive abilities such as memory, selective attention, learning and emotional regulation, and it facilitates planning, reasoning and problem solving.22,23 WM abilities are essential in several goal-oriented processes and even small differences in WM capacity may have considerable consequences in various domains including academic and job performance and mental health.24-26 These domains are all clinical important outcomes to address in patients with cancer, and therefore WM training is highly relevant and should be considered incorporated and evaluated in cognitive function training within this patient group.
CCT is labor intensive, and patients must adhere to the program to receive a sufficient amount of training. In fact, several studies have shown that patients with cancer adhere to CCT to varying degrees, 18 and both self-motivation and external support are important factors for sustained training.18,27 As such, it may be beneficial to integrate computerized cognitive training into the multidisciplinary programs offered by rehabilitation clinics so that patients have access to assistance and support during the training. Another advantage of this integration is that several treatments important for cognition could be combined (eg, exercise, psychological support).28,29
Cogmed QM (Pearson Education, Inc.) is an adaptive web-based WM program to increase the processing capacity of WM. 30 Cogmed QM has the largest effect sizes in WM improvements following training relative to other programs. 31 Cogmed QM has been researched in a wide variety of populations and has demonstrated benefits in populations without cancer. 32 Moreover, the feasibility of Cogmed QM has been shown in several patient groups such as Fragile X syndrome, Huntington disease and multiple sclerosis.33-35 and Cogmed QM appears to be well suited for clinical practice for several patient groups. However, past feasibility studies were not conducted in a multidisciplinary rehabilitation setting, where several other clinical interventions are offered as part of the rehabilitation program for adult cancer patients. A pertinent issue is that web-based WM training protocols may last longer than inpatient cancer rehabilitation programs in different health care systems.36,37 In such cases, patients may need to complete the program at home following discharge. The aim of the study was to assess the feasibility of including web-based WM training for adult patients with cancer during an inpatient rehabilitation program and at home after discharge from rehabilitation.
Materials and Methods
Participants and Recruitment
The study participants were recruited from patients participating in an inpatient multidisciplinary cancer rehabilitation program at Unicare Røros Rehabilitation (Røros, Norway) from January 2017 to July 2018. All participants were over 18 years old. The patients were referred for rehabilitation by their general practitioner or the responsible physician at the hospital. All patients completed their primary cancer treatment before initiating their rehabilitation program. All patients referred for a 3-week inpatient rehabilitation program were screened for eligibility by a study physician at the rehabilitation unit. Patients who expressed interest, met the inclusion criteria were asked to participate.
Inclusion Criteria
Inclusion criteria were based on patient-reported outcomes on memory complaints on the self-administered 11-item Fatigue Questionnaire (FQ), 38 and reported subjective cognitive complaints to the study physician were asked to participate. The FQ measures physical and mental fatigue, and items were answered using a 4-point Likert-type scale that ranged from asymptomatic to maximum symptomatology. The cut-off for inclusion based on the FQ was a score of 3 or more on 4 items addressing cognitive complaints (ie, difficulties concentrating, slips of the tongue, difficulty finding the right word and memory problems). Additional inclusion criteria were having a cancer diagnosis, included age ≥18 years, in stable medical condition and Norwegian-language fluency.
Exclusion Criteria
Exclusion criteria were: history of neurological disorder with cognitive symptoms (eg, Alzheimer’s disease, Parkinson’s disease, multiple sclerosis or seizure disorder) as per self-report or as documented in the medical record; history of mental retardation, learning disorder, psychiatric disorders of psychosis, bipolar disorder, schizophrenia, substance use disorder(s), or uncontrolled depression as documented in the medical record; history of traumatic brain injury with ≥a 30-minute loss of consciousness or cognitive sequelae as per self-report or as documented in the medical record; history of stroke as per self-report or as documented in the medical record; evidence of cancer recurrence at time of inclusion; a hearing or visual deficit that impairs the ability to use the software; significant cognitive or psychiatric disturbance sufficient, in the investigator’s judgment, to preclude providing informed consent; self-report of learning disabilities; substance addiction; previous participation in a cognitive training program; visual impairments, such as uncorrected vision or color blindness; uncorrected hearing impairments; and inability to use a mouse or computer keys to navigate around the computer screen. Written informed consent was obtained from all participants who met the inclusion criteria and were willing to participate at their first appointment, prior to initiation of any study procedure. Participants were informed about the time and logistical demands of the WM training.
The Cancer Rehabilitation Program
Rehabilitation programs in Norway are organized and offered as outpatient programs at hospitals or rehabilitation clinics located near the hospitals and as inpatient programs at rehabilitation clinics mostly located in rural areas. The current study was conducted at an inpatient rehabilitation clinic located 150 km from the hospital. All patients completed their primary cancer treatment before starting their rehabilitation program. The cancer rehabilitation program was conducted as a 3-week stay at the rehabilitation clinic and included a scheduled program from Monday to Friday. The rehabilitation program consisted of group-based physical exercise, patient education and discussions and needs-based individual consultations. All participants had a consultation with a nurse and a doctor, during which they set individual goals for their rehabilitation based on their self-defined needs. Physical exercise consisting of different indoor and outdoor activities was performed twice a day, with sessions lasting from 60 to 120 minutes. Patient education was delivered in the form of lectures in combination with group discussions and was led by nurses, dieticians, social workers, oncologists/physicians and physiotherapists/sport instructors. In addition, the participants discussed their challenges with their peers during their stay. Details of the rehabilitation program and patient education topics can be found elsewhere. 39
The Cogmed QM Working Memory (WM) Program
The commercial software product Cogmed QM was used for the adaptive WM training. 40 This program was chosen because it is designed to give adaptive feedback, and because it is a well-studied and readily available program that may be used on any computer, laptop or iPad with internet access. It focuses on different sensory modalities (auditory and visual) and includes 12 rotating exercises, presented as games, that address visuospatial WM (eg, remembering the location of previously highlighted boxes and rotating exercises) and/or verbal-numeric WM (eg, remembering digits and repeating them backwards). A description of the different Cogmed QM exercises can be found at www.cogmed.com. The standard Cogmed QM administration protocol requires users to complete 8 daily tasks lasting a total of 40 to 50 minutes, 5 days a week for 5 consecutive weeks. All participants started at the same level in the first training session, beginning with 2 items. To progressively increase WM capacity, the Cogmed QM program uses an adaptive algorithm that automatically adjusts the difficulty to an increasingly challenging level. The task difficulty level adapts to the capacity level of the user in real time and is altered by changing the number of items that need to be processed by the WM. The program was designed such that each participant could obtain 60% correct responses on each task in every session. In each session, participants started at the level they achieved in the previous session. 30
Procedure
After their inclusion was confirmed, each participant met with a health care coordinator who was trained in administering the Cogmed QM program to be familiarized with Cogmed QM and to discuss the WM training expectations and schedule during their rehabilitation. Participants practiced on computers stationed in their rooms, which were made available by the rehabilitation center. Participants were instructed to complete 25 sessions of WM training within a period of 6 weeks on weekdays. An extra week was added to the Cogmed QM standard administration protocol of 5 weeks; it was presumed that participants would require extra time to complete the program due to moving from the rehabilitation center to their homes. Participants were instructed to practice every weekday during and after the rehabilitation program until the termination of the WM training. The period from the first day of the WM training until the day before departure from the rehabilitation clinic was defined as the available WM training days during rehabilitation.
During the rehabilitation program, participants were given feedback about their WM training progress. The feedback was related to motivational aspects and their performance in terms of progression, as assessed using the Cogmed Improvement Index. Short breaks during WM training, at the participants’ discretion, were permitted and encouraged throughout the session. This was done to ensure and optimize the participants’ training performance.
Participants were instructed to continue the WM training immediately after returning home following discharge from the rehabilitation clinic. To avoid significant prompting, the health care personnel did not initiate any training-related contact with the patients through this part of the study. However, participants could receive support by telephone if there were any technical problems or other training-related issues. At home, participants used their own internet-connected computer, laptop, or iPad.
Feasibility Criteria and WM Training Outcomes
The feasibility was determined by assessing the study recruitment, accrual, attrition, adherence to the WM training during the multidisciplinary rehabilitation, training tolerance, patient experiences and Cogmed Improvement Index, which is a built-in compliance measure. The feasibility measures were adapted from guidelines for reporting non-randomized feasibility studies. 41 The Cogmed QM program automatically records training information, such as date, duration, breaks and performance. These parameters were used to determine adherence and tolerance. Accrual was defined as the number of included patients compared to the number of invited patients. The attrition rate was derived by dividing the number of withdrawn participants by the number originally included. The number of participants lost during the trial was recorded (drop-out) and calculated as a percentage of the total sample size.
The adherence rate is shown as a percentage and was calculated as the number of completed WM training sessions divided by the maximum possible WM training sessions. Two adherence measures were assessed, and the a priori adherence threshold was guided by previous research that evaluated the effects of the Cogmed QM program.40,42 The primary measure was adherence to WM training during the rehabilitation program, which was defined as completion of ≥ 80% of the possible WM training sessions. The second measure was adherence to the combination of WM training during the rehabilitation program and consecutive home-based training after discharge from the rehabilitation clinic, which was defined as completion of ≥ 80% of the 25 sessions (the entire training protocol) within 6 weeks.
WM training tolerance was evaluated by examining the ratio of time spent on active training compared to time spent on breaks, during sessions. WM training tolerance was defined according to the Cogmed QM guidelines, which suggest a 2:1 ratio of active training time to breaks per session. An individual who exhibited a ratio ≥ 2:1 was described as “tolerating the training well.” WM training engagement was evaluated by calculating mean active time per day and mean pause time per day.
The Cogmed Improvement Index, a built-in compliance measure, was calculated automatically by the program. The Cogmed Improvement Index score is based on 2 training indices: the Start Index (the score on the third day of training) and the Max Index (the best score obtained during training). The Cogmed Improvement Index is calculated by subtracting the Start Index from the Max Index. According to Gray et al successful progress is represented by a score of ≥17. 43 Several general performance parameters were included: number of completed sessions during rehabilitation and during home training, mean active training duration per day and mean pause duration per day.
Evaluation Questions
The participants were asked additional questions during and after the WM training to evaluate the usability of the Cogmed QM program. The questionnaire that covered the participant’s subjective experience of the training comprised the following questions: (1) Overall, did you find it easy to use the WM program? (2) Did you manage to get enough time to train during the rehabilitation stay? (3) Was WM training useful? (4) How was the program useful or not useful to you? (5) Did you manage to get enough time to train at home? (6) How did you experience using the WM program at home?
Other Outcome Measures
Sociodemographic characteristics were collected using a proprietary questionnaire developed by the clinic. Education level was assessed using 3 response alternatives: elementary school, secondary school or university or equivalent. Marital status was assessed using 2 response alternatives: married/cohabitating or living alone/widow(er). Employment was assessed using 3 response alternatives: full-time work, part-time work or social benefits.
Statistical Analysis
First, descriptive statistics of the quantitative data were calculated, including the sociodemographic characteristics of the sample. Between-group comparisons were conducted using Fischer’s exact test for all nominal variables. Independent sample t-tests or Mann–Whitney U tests were used for the scale variables.
Next, the feasibility was examined by computing descriptive statistics for the WM training feasibility measures. The statistical analysis was performed using the paired Student’s t-test for changes in the Cogmed Improvement Index and Pearson’s correlation test for correlation analysis of the Cogmed Improvement Index over the training days. All statistical analyses were performed using the IBM SPSS Statistics 22.0 software program (IBM Corporation, Armonk, New York). P-values < .05 were considered statistically significant. All evaluation answers were transcribed verbatim. The first author reviewed all the transcribed interviews and met with the research assistants who conducted the interviews to validate the answers. For reliability and validity checking, peer debriefing with the last author was conducted.
Ethics
The study was approved by the Regional Committee for Medical and Health Research Ethics in Central Norway (REK); 2016/1928-1 REK Midt. Appropriate written informed consent was obtained.
Results
Participants, Recruitment, and Attrition
A total of 277 patients with cancer were screened for eligibility, and 245 patients were ineligible due to not meeting the inclusion criterion of presenting self-reported cognitive complaints (Figure 1). A total of 32 patients met the inclusion criteria and were invited to participate. The accrual rate was 96.9%; one patient declined to participate and explained that “cognitive intervention did not appeal to the need during rehabilitation.” Two patients withdrew from the study before the first WM training session, resulting in an attrition rate (drop-out before starting WM training) of 6.5%. The 2 drop-out patients explained that they “did not have the motivation and energy to conduct WM-training during an extensive rehabilitation program.” The study sample of 31 participants consisted of 28 women and 3 men, and the mean age was 50.7 (SD = 10.8) years. The mean time from diagnosis to the start of the rehabilitation program was 36.1 (SD = 56.3) months. Table 1 presents the clinical and demographic information of the included participants.
Flowchart of participation enrollment, inclusion and involvement.
Participants Sociodemographic and Medical Characteristics.
Abbreviations: n, number of patients; SD, standard deviation.
Adherence, Tolerance, and WM Training Outcomes
Adherence
During the rehabilitation program, 26 of 29 (89.6%) participants adhered to the WM training, while 19 (65.5%) participants adhered to the subsequent home-based training after discharge from the rehabilitation clinic. The 19 participants who adhered to the WM training during the rehabilitation program and at home completed a mean of 24.27 (SD = 1.56) out of 25 Cogmed QM sessions, of which 48.4% (SD = 9.8) were performed during the rehabilitation program, and the remaining sessions were performed at home.
The 3 participants who did not adhere to the WM training during the rehabilitation program all completed one Cogmed QM session. Their non-adherence was related to a lack of training motivation. In addition to the 3 participants who did not adhere to the WM training during rehabilitation, 7 participants did not adhere to the home-based training; in total, these participants completed a mean of 10.8 (SD = 2.8) Cogmed QM sessions. Eight participants did not start the home-based WM training after discharge from the rehabilitation clinic, while 2 participants each completed 5 and 6 WM training sessions at home, respectively. No participant used the Cogmed QM program for more than 6 weeks.
The participants who did not adhere to the home-based WM training provided the following reasons for their non-adherence: everyday life being too busy and stressful to continue training (4 participants), technical problems with the internet connection (one participant), problems with finding a disturbance-free environment (one participant) and low general motivation (one participant). No significant differences were found between adherent and non-adherent participants concerning age, duration from diagnosis to start of rehabilitation, education level, employment status and marital status (Table 2).
Characteristics of Adherent and Non-Adherent Participants.
Abbreviations: N, number of patients; SD, standard deviation.
Tolerance and engagement
Both adherent and non-adherent participants who attempted the WM training were tolerant to the training as defined by a ratio of at most ≥2:1 time spent on active training versus breaks. Participants who adhered to the Cogmed QM program spent a mean of 39.84 (SD = 6.82) total min on each session and a mean of 37.95 (SD = 5.73) active min, which excluded breaks and inter-trial pauses. Non-adherent participants who |completed >1 Cogmed QM session (n = 7) spent a mean of 37.40 (SD = 7.00) total min on each session and a mean of 36.50 (SD = 6.57) active min. Neither the mean total WM training time per session (P = .122) nor the mean active WM training time per session (P = .219) differed between the adherent and non-adherent participants.
WM training outcomes
All participants who adhered to the WM training during the rehabilitation program and at home showed improvement in the Cogmed QM tasks, as defined by the Cogmed Improvement Index (a summary variable generated by the Cogmed QM software to indicate the extent of change in a training task; mean difference = 24.05, SD = 9.38, range 2-44, P = .011). Figure 2 shows a scatterplot of the average Cogmed Improvement Index scores and training day, and a positive and significant correlation (r = .926, P < .001) is observed.
Mean Cogmed Improvement Index across training period in adherent participants. N = 19.
Participants who adhered to the WM training during the rehabilitation program showed improvement in the Cogmed Improvement Index; mean difference = 13.0, SD = 9.86, range 3 to 34, P < .001)
Summary of evaluation questions
Twenty-three of the 29 participants responded to the evaluation questions (Table 3). The majority (91.3%) reported that the Cogmed QM program was easy to use (Table 3). Furthermore, just over half (56.5%) thought they had enough time to complete the WM training during the rehabilitation program. Overall, just under half (48.7%) found the Cogmed QM program to be useful, ~39% were uncertain about whether it was useful and ~13% did not find it to be useful. Among the 11participants who found the program useful 3 reported improved memory functions, 3 better concentration and 2 improvements in attention. Participants who were uncertain about its usefulness described a lack of perceived effects or uncertainty about potential effects, and they even reported that the WM training induced distress (Table 4). More than half (55.0%) of the participants reported that they had enough time to complete the WM training at home, while the remaining reported that they did not (Table 3). Among the 12 participants who were uncertain about whether the program was useful or did not find it useful 7 (58.3%) participants reported that they had enough time to complete the WM training at home. The increase in the Cogmed Improvement Index between those who were uncertain about whether the program was useful or did not find it useful (mean difference = 20.8, SD = 9.83) and those who found the program useful (mean difference = 21.8, SD = 12.0) was not different (P = .725). The participants reported mixed experiences regarding the home-based WM training following the rehabilitation program. All the participants who had positive experiences (n = 9) reported that it was easy to use the program at home. The participants who had negative experiences (n = 13) reported issues such as technical, emotional and motivational problems, as well as difficulties with finding a suitable, disturbance-free environment to conduct the training (Table 4). No participant reported any adverse events with the use of the Cogmed QM program.
Summary of Close-Ended Evaluation Questions.
Abbreviation: N, number of patients.
Summary of Evaluation Questions.
Abbreviations: n, number of patients.
Discussion
To our knowledge, this is the first study to explore the feasibility of web-based WM training in a group of patients with different types of cancer with memory complaints during multidisciplinary rehabilitation and at home after discharge from rehabilitation. The results show that most of the participants (89%) adhered to the web-based WM training during their inpatient multidisciplinary rehabilitation and that fewer participant (65%) adhered to the subsequent unsupervised home-based training. Thus, the findings suggest that web-based WM training is feasible during inpatient multidisciplinary rehabilitation and that continuing unsupervised training at home after discharge from rehabilitation is not ideal.
The high recruitment rate (97%) in the current study is comparable to that in 2 studies that examined cognitive function training in a rehabilitation setting.44,45 The participant who declined to take part reported “that cognitive intervention did not appeal to the need during rehabilitation.” In addition, 2 participants withdrew from the study before the first WM training session, both reporting that they “did not have the motivation and energy to start training during an extensive rehabilitation program.” Thus, it seems that the motives of these 3 subjects for not participating were connected to a perception of a lack of need for the intervention and/or a sense of being overwhelmed by the demands of long-term rehabilitation.
To date, no systematic review has examined web-based WM training adherence in a comparable population. However, several studies on supervised WM training have provided insights. For example, web-based WM training adherence similar to that found here has been reported in several other studies, including for patients participating in vocational rehabilitation, 30 patients with fragile X syndrome 46 and older adults. 47 A few studies have included adults with cancer. One study reported that 26% of included patients with breast cancer never started a 3-month web-based cognitive function training program and that 42% of those who started the training dropped out within 3 months. 48 The second study was a large, randomized study evaluating a web-based cognitive function training program in 242 adult cancer survivors. In that study, 14% of the patients in the intervention group did not start the training, and the average total training time for the other patients in this group was 25.06 hours of the recommended 40 hours. Only 27% of patients completed the program within the recommended 15-week timeframe. 45
As discussed in these and other studies, both patients’ expectations of the proposed treatment and their motivations are important for treatment adherence. It is worth mentioning that other factors are also involved, such as depression, socioeconomic background, side effects of the medication and morbidity resulting from the disease. 49 In the present study, the 3 participants who did not adhere to the WM training during rehabilitation and the 2 participants who withdrew before the WM training commenced all reported a lack of training motivation. However, it is not known why these participants lacked motivation, and further studies should assess individual differences in motivational style when analyzing adherence and developing selection criteria for web-based WM training for patients with cancer.
While most participants (91%) reported that it was easy to use the Cogmed QM program and adhered to the WM training during rehabilitation, only a slight majority of participants (~ 56%) thought that they had sufficient time for the WM training during rehabilitation. In addition, a few participants reported that using the program was stressful. This was surprising, since the protocol was designed to allow for flexibility in training schedules and the participants could continuously adapt their WM training to accommodate participation in the rehabilitation program. Although the Cogmed QM program allows for flexibility, patients were expected to complete a total of 25 sessions (+/-40 minutes per session) within a period of 6 weeks. It may be the case that including WM training in a comprehensive rehabilitation program in which patients are engaged in and have to prioritize other activities is suboptimal. Clinicians should consider whether undergoing cognitive function training in a particular phase of the disease may be too burdensome with respect to fulfilment of other needs, recovery and the impact of medical treatment and treatment transitions. An alternative approach would be to investigate the feasibility of supervised Cogmed QM-based WM training at home after the completion of rehabilitation, which may allow patients to better concentrate on the WM training.
While participant adherence to the web-based WM training during rehabilitation was good, it was less ideal in the home-based setting after discharge from rehabilitation 65% of the participants met the adherence criterion. The unprompted home-based training was part of a pragmatic study design and different from the standard coaching based Cogmed QM protocol, which recommends active guidance during the entire training protocol. The study was designed in this way so that it corresponded to the specialist health care system in Norway; after discharge from rehabilitation, patients often return home without further interaction with the rehabilitation clinic. 37 Thus, these results may be of special relevance to countries with similar types of rehabilitation services (eg, Germany) 50 and to other settings in which there are limited resources for supervised web-based WM training.
Few studies have investigated adherence to unprompted home-based computerized cognitive training, and an extensive literature search did not reveal any such studies conducted in patients with cancer. The adherence rate found in this study (65%) is comparable to those of home-based WM training in neurological surgery patients 51 and patients with multiple sclerosis, 52 while lower than in a varied group of patients with several neurological diseases. 53 Home-based training is likely to be more sensitive to individual differences in motivational factors than training conducted in a laboratory setting or institution, 54 which may directly influence the amount of effort allocated to training. The non-adherent participants reported high performance requirements, lack of time in professional and everyday life, lack of motivation and technical problems as reasons for discontinuing the training in the home-based setting. This is in line with other studies that have shown that both computer difficulties and similar time- and environment-related personal factors are associated with reduced adherence to online psychological interventions. 27 It is reasonable to conclude that the change in the conditions (from the inpatient clinic to home) required the participants to establish a new training routine that suited their daily life and that this somehow became too demanding or perhaps not prioritized.
The absence of active support may have resulted in lower adherence in participants experiencing barriers, even if they had experienced guided WM training during rehabilitation and could phone or email the rehabilitation center at any time. Active support can provide an avenue to discuss challenges and offer suggestions for progress. Users are advised to complete 25 Cogmed QM sessions over 5 weeks with active coaching, and it cannot be ruled out that using the Cogmed QM program in an irregular fashion may diminish its useability and effectiveness. It has been discussed in the limited scientific literature on this topic that more intensive cognitive function training should presumably produce larger effects on WM performance. One study showed that the number of Cogmed QM sessions was related to improvement in a digit span task in healthy subjects. 55 Nevertheless, future studies should consider discharge planning with the development of a sustainable plan for home-based WM training and the inclusion of periodic follow-up between patients and health care professionals to avoid isolation and reinforce training.
In the present study, participants who adhered did not differ from those who did not adhere in terms of age, duration from diagnosis to start of rehabilitation, education level, employment status and marital status. The results for education level are in line with those of a 2016 review by Beatty and Binnion which concluded that education level was not associated with adherence to online psychological interventions, while the influence of marital status and age on adherence was mixed and the direction of the relationships was unclear. 27 Since treatment adherence has clear clinical implications, with poor adherence limiting exposure to the required treatment, there is still a need to identify the medical, sociodemographic, motivational and technical determinants of adherence. 27
All participants were tolerant to training according to the Cogmed QM guidelines. This indicated that patients were able to complete the Cogmed QM sessions within the recommended time limit for each training session. Since no participants spent a disproportionate amount of time on breaks, they may not have struggled to complete the training when they first started a training session. The fact that non-adherent participants were also tolerant to training may suggest that when they first initiated a training session, they were motivated to complete it. The feedback the participants received from the health care personnel during the rehabilitation and WM training phase may have served as a motivation to continue the training and potentially reduced training frustration.
The aim of this study was to determine whether adult patients with cancer could engage in web-based WM training during rehabilitation and in home-based settings. A positive outcome would permit future evaluation of the effectiveness of web-based WM training. The results of this study do not allow any conclusions to be drawn about the transfer effects of web-based WM training. However, the difference between the Max Index and Start Index was significant at ~24 units and is comparable to Cogmed QM normative data with a mean Cogmed Improvement Index of 29 (normal range 15-49) for individuals aged 18 to 65 years. 30 The data show a strong positive correlation and incremental progress in performance over the entire training period, which suggests that the adherent participants improved in the Cogmed QM tasks and were not randomly clicking. Although the Cogmed Improvement Index results suggest that there were some effects, according to the program-based performance standards, the index does not necessarily reflect the final effect of the training, because the Max Index is not necessarily from the last session. The score may better represent the capacity for improvement across the training sessions. 56 It should also be mentioned that the Cogmed Improvement Index value could have been influenced by other components of the rehabilitation process, in addition to the WM training. Thus, future studies that examine the effects of adaptive web-based WM training during rehabilitation should also control for other mediating factors, such as exercise.
A slim majority of respondents to the evaluation questions acknowledged improvements in their memory following WM training, and these participants reported improvements in attention, memory and concentration. Similar observations of subjective WM changes with the use of Cogmed QM have been reported34,57; however, subjective improvement must be interpreted with caution, as it may reflect participants’ expectations of training effects. 58 Other studies have demonstrated positive effects of other web-based cognitive function training programs in adults with cancer44,45 and of the web-based Cogmed QR program in children with cancer. 59 There is now a need for randomized controlled studies to determine the effects of the Cogmed QM program on the cognitive performance and the symptoms of cognitive impairments in adults with cancer.
Limitations
The interpretation of this study’s results must be tempered with an understanding of its limitations. This was a single-center study, which may impose some restrictions on the generalizability of the findings. Of the 277 patients initially assessed, 32 (11.5%) fulfilled the inclusion criterion for cognitive complaints. This number is lower than the higher rates of cognitive complaints reported in the literature, 1 and may suggest that patients participating in inpatient rehabilitation may not be representative of the population of cancer survivors who experience cognitive complaints. After a patient has been referred for cancer rehabilitation, he or she must decide whether to attend. This is a potential point of self-selection. 60 Whether the patients had knowledge of the inpatient rehabilitation program and felt the need for rehabilitation or whether the physicians advised them to attend is not known. Furthermore, the rehabilitation program emphasized physical training, and it cannot be ruled out that the rehabilitation program attracted patients with less memory complaints than the broader cancer patient population. It should be noted that the majority of the study sample consisted of middle-aged female patients with breast cancer, but patients with other cancer diagnoses were also represented. This could be considered a limitation in a large-scale trial assessing the effects of WM training and thus should be of specific concern when planning a randomized controlled trial.
Furthermore, the investigation utilized a non-randomized feasibility study design and lacked baseline neuropsychological screening. In the absence of a randomized control group, it is difficult to determine the influence of the WM training on adherence to other rehabilitation components and the extent to which improvement in the Cogmed Improvement Index was due to training as opposed to induction bias. The lack of baseline neuropsychological screening eliminated the possibility of investigating the relationship between WM performance and adherence to web-based WM training. For effect studies, it has been recommended to include a neuropsychological assessment at baseline to identify cognitive strengths and limitations and to identify the relationships between the neuropsychological findings and the medical condition. 61
This feasibility study was conducted in a clinical setting and used selection criteria for cognitive complaints based on the FQ 38 in combination with a clinical evaluation. A similar selection approach was used in a study of patients with chronic fatigue syndrome and myalgic encephalopathy to identify memory and concentration complaints. 62 The FQ is a widely used instrument for assessing mental and physical fatigue in both clinical and nonclinical settings in patients with cancer. 63 Although there is no instrument designed specifically to measure CRCI, 1 it cannot be ruled out that the mental fatigue component of the FQ taps into different experiences for cognitive function than those described for CRCI in other studies. Therefore, future studies should include validated neuropsychological tests to determine the effects at various levels of cognitive function 61 and to potentially investigate the relationship between the results of neuropsychological tests and questionnaires identifying symptoms of cognitive complaints. Given these limitations, it is important that future research will endeavor to extend our findings to different cancer patient populations, include better representation of cognitive complaints and ensure that motivational measures are administered at the outset of the study.
Conclusions
This study provides important evidence for the feasibility of including web-based WM training for cancer survivors with cognitive complaints during cancer rehabilitation and at home after discharge from rehabilitation. The findings suggest that it is feasible to include web-based WM training for patients with cancer who participate in multidisciplinary rehabilitation. Patient adherence to unprompted web-based WM training after discharge from rehabilitation was not optimal in this study and implies that home-based intervention is not feasible for all participants in its current form. Future studies should consider barriers to adherence, periodic communication with trained personnel and supportive strategies to promote adherence to training. This study has highlighted important aspects of both the study and the intervention design that require consideration, and these include the cognitive impairment criteria, logistical considerations (eg, lack of time to complete the intervention) and need for discharge planning to ensure adherence to home-based WM training. Further feasibility studies could investigate web-based WM training in the home setting following completion of inpatient rehabilitation and with professional assistance to provide structure, motivation and support throughout the WM training period.
Footnotes
Acknowledgements
The authors extend special thanks to the cancer rehabilitation team at Unicare Røros, Norway, who helped to conduct the study.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
