Main findings

The iCST trial was a pragmatic, multicentre, RCT of a complex, individual, carer-led cognitive stimulation intervention. The trial was designed to evaluate the effects of iCST on cognition and quality of life for people with dementia and their family carers. We recruited a total of 356 caregiving dyads, making this study the largest in the current literature on CST-based approaches.

Primary outcomes

For people with dementia, the primary outcomes did not indicate any specific benefit for those allocated to receive iCST, given that there was no clinically significant improvement in cognition or quality of life compared with people with dementia receiving usual care. Carers’ physical and mental health was not significantly different between the intervention and control groups.

Secondary outcomes

We found no evidence that iCST reduced behavioural and psychological symptoms or depressive symptoms for people with dementia. There was also no evidence of change in activities of daily living. Although no effects were observed on most of the secondary outcomes, analyses indicated that people with dementia allocated to receive the intervention reported improvements in relationship quality with their family carer. iCST did not improve secondary outcomes such as carers’ mood, resilience or relationship quality with the person with dementia. Despite no differences in most secondary outcomes, health-related quality of life ratings for family carers allocated to the intervention group improved at the primary end point. This, however, was in contrast to no evidence of improvement on carers’ SF-12 component scores. This discrepancy in findings between the two generic instruments (SF-12 and EQ-5D-3L) may reflect intrinsic differences between these two instruments or differences in terms of each instrument’s sensitivity to change.

This is the first economic analysis of an iCST intervention for people with dementia and their family carers. Although costs from either the health and social care or the societal perspective did not differ substantially between the groups at either follow-up time point, there was a consistent pattern of lower costs in the iCST group over the 26-week period. In terms of the primary outcomes for people with dementia, it appears that iCST is not more cost-effective than TAU from either cost perspective, when we take sampling uncertainty into account. There are no established societal WTP thresholds for improvements in ADAS-Cog, QoL-AD or QCPR. In terms of QALY gain for carers, iCST was more effective than TAU. Taking carers’ costs into account, costs in the iCST group were lower, but not significantly lower, than in the TAU group. Taking sampling uncertainty into account, and assuming no further change in utility or costs in either group for the following 6 months, the probability that iCST is cost-effective was 93% at a societal WTP per QALY of £30,000. Under the same assumptions, we can be confident that iCST is cost-effective at a societal WTP of approximately £47,300 to gain a QALY, and, excluding costs to carers from this calculation, we can be confident that iCST is cost-effective at approximately £84,200 per QALY. The intervention can be considered to be cost-effective in improving unpaid carers’ health-related quality of life only at a societal WTP well above the NICE threshold of £20,000–30,000. However, societal decision-makers may be willing to accept somewhat lower levels of certainty to achieve this outcome. Given the results of the sensitivity analysis, this conclusion is dependent on a relatively low valuation of carer time.

When considering the number of sessions received, as opposed to allocation, some improvements were observed. People with dementia completing more sessions were more likely to experience gains in terms of the caregiving relationship at 26 weeks. Reports of improvements in the caregiving relationship by people with dementia are consistent with previous studies indicating that meaningful activities conducted alongside family carers can preserve and enhance the caregiving relationship.⁷⁴ Improvements were observed for carers completing more sessions with their relative, expressed by a reduction in depressive symptoms. These findings are consistent with previous research, in which a home-based cognitive stimulation programme was associated with lower depressive affects in carers.¹¹

Overall, however, findings are in contrast to previous studies demonstrating that group short-term CST benefits cognition and quality of life for people with dementia,⁶ and maintains quality of life improvements when provided long term.¹⁹ These results also contrast with previous RCTs reporting benefits in cognition for people with dementia associated with home-based individual RO/cognitive stimulation.¹² In relation to effects on cognition, not replicating results of group CST approaches may be attributable to the lack of a group setting when CST sessions are provided. In addition, although we tried to use similar activities to those provided in group CST, it is unlikely that the activities were the same, indicating that differences may relate to differences on group versus individual CST approaches. In relation to effects of quality of life, this is the first study to include a measure of quality of life for a home-based approach for people with dementia. Our results therefore indicate that the benefits in quality of life are more likely to be associated with interventions that combine or use CST approaches within a social setting. Our findings may be of importance in relation to updating the Cochrane review. This may indicate the need for separate analyses of group and individual approaches to cognitive stimulation.

Treatment implementation is an important parameter when evaluating psychosocial interventions. In order to ensure that iCST was ‘delivered as intended’, we used a treatment protocol that specified all components of the intervention in detail, in order to ensure that the intervention was delivered as planned. Each unblinded researcher received training in iCST, and there were frequent opportunities for researchers supporting carers in the delivery of iCST to receive supervision and feedback. Furthermore, there were a number of challenges associated with implementing iCST, particularly with regard to carers fitting iCST into a busy timetable, the lack of ‘stimulation’ for some participants and difficulties reported by carers in engaging in the intervention with their relative. Compliance will be, therefore, an important component to consider in order to optimise similar psychosocial interventions for people with dementia and their carers. At this stage, it is not clear if there are any specific characteristics that would identify the very low compliance group who completed no sessions at all, but we plan to look at further analyses on this topic in due course.

In line with our projections in relation to sample size, a total of 81% of the sample completed the 13-week assessments and 71% completed the 26-week assessments. The most common reason for not being available to complete follow-up appointments was reporting problems with engaging in iCST, indicating that, although the intervention may have a high uptake, it still may be difficult or not suitable for some carers and people with dementia. Although most carers were able to engage in iCST with their relative, there were frequent reports by carers of struggling to engage with their relative in the sessions and, often, this was a common reason of loss to follow-up. A total of 22% of dyads allocated to receive iCST did not complete any sessions and 13% completed fewer than 10. A threshold of completing over 38 sessions was set on the basis of number of sessions completed by the whole of the sample, as indicative of compliance to the intervention. We did not find any evidence of differential attrition, and analyses exploring differences between completers versus non-completers did not suggest any differences between the two groups in terms of baseline characteristics.

In terms of cognition and quality of life, our findings do not provide support for the use of home-based cognitive stimulation programmes for people with dementia, which is contrary to previous work on home-based memory rehabilitation¹⁰ and RO for people with dementia.¹² It is unclear which factors could account for the differential efficacy. It is likely that in some studies participants are highly selected, resulting in different patient groups recruited and, therefore, different levels of dementia severity and cognitive function. For example, in our study most of the sample had mild dementia in comparison to previous studies. Other factors could reflect differences between studies in terms of power or chance variation, given that previous studies are of varying quality with small samples overall.² Importantly, although treatment compliance is not reported in detail across studies, the ‘dose’ received in each of the studies may be an important determinant of efficacy.

Despite generally negative findings, people with dementia reported better relationship quality with their carer, indicating that individual cognitive stimulation interventions have the potential to improve inter-relationship outcomes for people with dementia. Relationship quality, rated by the person with dementia, has been shown to be an important contributor to quality of life for people with dementia and is the cornerstone of relationship-centred care.⁷⁵ We also found that health-related quality of life for family carers significantly increased, indicating that carer involvement in cognitive-based interventions may increase carer well-being. This may be related to the fact that, although iCST was developed largely as a home-based, carer-led, individual cognitive stimulation approach, the intervention incorporated additional psychoeducational elements such as communication, opportunities to increase pleasant events for both carers and people with dementia, which are components less likely to be incorporated in group CST-based approaches. This is consistent with the findings of the qualitative study where carers reported that iCST provided opportunities to understand dementia, its impact on communication and confidence for the person with dementia, and to increase pleasant activities both for themselves and their relative. Interventions that therefore target communication between people with dementia and their carers may reduce the strain on the caregiving relationship and may potentially improve general well-being outcomes for both.

A further variable of interest is unblinded researchers’ level of expertise and experience, which is an important factor to consider in the development of psychosocial interventions. Consultation groups with unblinded researchers indicated that they were generally well received by carers and people with dementia, although dyads differed widely in their level of engagement with the intervention. The level of expertise was judged as sufficient, and training and background in dementia care was considered to be important in supporting carers in delivering the sessions.

Implications for health care

It was expected that iCST may be beneficial for people with dementia in terms of cognition; however, this was based on smaller studies using both cognitive stimulation and RO techniques and using less well-defined methods. However, it is notable that participants in this study had better cognition (mean baseline MMSE score of 21) than those in the original group CST study (mean baseline MMSE score of 14), which was also limited to people with dementia scoring between 10 and 24 on the MMSE. This suggests that in the current study there was less scope for improvement, and some participants may have had cognitive function that was too high to benefit from iCST.

Most of the evidence on effects on cognition for cognitive stimulation is based on group approaches, so future research will need to focus on understanding the mechanisms that are more likely to be associated with the reported effects and differential effects of outcomes between group versus individual approaches. There was evidence in this trial that some people with dementia and their family carers will not be able to engage successfully in iCST, as 34% of the sample allocated to receive the intervention completed 0 or fewer than 10 sessions. Our qualitative findings indicated that, although most people with dementia and carers enjoyed the sessions, there were also a few comments that the iCST sessions were not challenging enough, indicating that the type of carer-led, cognitive-based intervention may be key in terms of producing a therapeutic effect. Despite overall negative findings, improvements on the carer–patient relationship and carers’ health-related quality of life suggest that iCST may have a key role in improving communication for people with dementia and their carers.

Limitations

Participants dropping out will have introduced bias if they had a different response to the intervention or TAU conditions compared with those that completed the trial. However, there was no evidence from demographic variables or baseline outcome scores that those who did complete the study were different from those who did not.

Although significant efforts were made to obtain outcome data, the trial may have been underpowered to detect significant differences for the primary outcome measure owing to the attrition rate and low levels of compliance in the overall number of sessions completed. Nevertheless, this is the largest RCT of a cognitive stimulation intervention in which carers lead the sessions and which shows no effect on cognition and quality of life for people with dementia compared with usual care.

An important limitation is that the observed differences of improvement in the caregiving relationship for people with dementia and health-related quality of life for carers may be attributable to incidental findings, driven by the multiple comparisons tested, and may, therefore, be attributable to chance. Despite people with dementia and carers expressing interest in the intervention, compliance was low overall, indicating that cognitive stimulation interventions delivered by carers may not be the ideal mode of delivery for many, thereby limiting wider applicability and generalisability of this approach, and indicating that better methods of monitoring and support for adherence are needed. Identifying subgroups of caregiving dyads that are more likely to benefit from this intervention is likely to be an important aim for future research.

Recommendations for future research

The null effect reported in this study leaves unanswered the question of whether or not carer-led cognitive stimulation interventions are effective. Despite the appeal of home-based programmes led by carers, cognitive stimulation approaches may be better provided on a group basis unless further evidence becomes available.

1. If carer-led interventions are to be further pursued, future research should identify which factors are more likely to make the intervention most successful and adaptable to the needs of people with dementia. As feedback from people with dementia in the qualitative interviews indicated that some sessions were not stimulating enough, future studies should consider that people earlier in the disease trajectory may have different cognitive stimulation needs from those with moderate dementia.
2. Given that less than half of the iCST group completed at least two sessions per week this reduced the power of the study to identify potential differences with the control group and indicates limitations in relation to the applicability of the intervention. Future work is needed to investigate the characteristics of caregiving dyads that are most likely to adhere to and benefit from carer-led individual cognitive stimulation interventions and methods to improve adherence. The involvement of paid domiciliary care workers or volunteer befrienders in delivering sessions should be explored further, as this may increase adherence by placing less responsibility on family carers.
3. Given that carer-led, cognitive-based approaches are relatively new treatments, research designs that address their efficacy are more relevant than designs addressing mechanisms at this point. Therefore, comparisons of group versus individual approaches are likely to be premature at this stage.
4. It is important that future research and different research groups evaluate further the effects of individual cognitive stimulation interventions. This work will help to ensure the reliability and robustness of the effects reported in this trial, in relation to benefits of relationship quality for people with dementia and health-related quality of life for carers.

Conclusions

The evidence from this trial suggests that taking part in iCST sessions does not result in improvements on cognition or quality of life for people with dementia. There was no evidence of improvements for carers’ mental and physical health. There was no evidence that iCST improved secondary outcomes for people with dementia, such as activities of daily living, mood or behavioural and psychological symptoms. Analyses indicated that iCST did not confer any benefit for carers’ mood, resilience or relationship quality with the person with dementia. Although people with dementia receiving iCST reported better relationship quality with their carer (in itself, an important component of quality of life for people with dementia) and carers reported improved health-related quality of life, these findings need to be interpreted with caution given limitations owing to multiple testing. Despite efforts to minimise loss to follow-up, the study may have still been underpowered to detect significant differences for the primary outcome. Given that iCST did not achieve the expected change, it is unlikely to lead to clinical benefit in improving cognition and quality of life for people with dementia.