Cognitive aids for people with early stage dementia versus treatment as usual (Dementia Early Stage Cognitive Aids New Trial (DESCANT)): study protocol for a randomised controlled trial

TAU comprises help from memory clinic staff, post diagnostic counselling and advice, and specialist follow up as appropriate. Participants in the comparator group also receive a general guide to dementia for patients and carers [13].

Index people with dementia must be aged 50 years or more; cared for by a trial memory clinic or equivalent; within one year of their first attendance at that clinic for dementia; physically able to engage with the intervention (as judged from clinical records); clinically able to do so, as judged by a responsible clinician and have dementia of mild to moderate severity. At baseline, they must live in their own home or share a home with a relative. They must have an identified carer, defined as the primary person who feels responsible for and supports them. This could be a family member, a close friend or a neighbour.

Sites will be recruited via professional and research networks, including the UK NIHR Clinical Research Network (CRN). Participating Trusts will identify memory clinics for inclusion in the study. Each site will confirm their capacity to undertake the trial by completing the Health Research Authority Statement of Activities as a formal agreement with the Sponsor.

Clinical staff in participating Trusts are briefed about the trial and provided with participant information sheets (PISs) for potential participants. They will complete a screening schedule to check eligibility, give the PIS to participants, and seek oral consent to refer them to the DESCANT team. Recruitment, documents and processes were designed following guidance from the PPCRG on language and format, to ensure that those with cognitive impairment are fully informed and engaged in the decision to take part.

Clinical staff within agencies will introduce the study to participants, provide them with PISs and obtain their consent to contact by a member of the research team. Before conducting the interview, researchers explain the study to participants and obtain their formal consent. Guided by the Mental Capacity Act 2005 [16], it is judged that people with mild to moderate dementia approached in this way have capacity to provide informed consent, given sufficient time to decide.

Participants may withdraw from the trial at any time for any reason. If they withdraw, a reason will be recorded unless they choose not to provide one. PIs may also withdraw people with dementia from the trial if they feel it is no longer in their best interest to continue. Unless participants request otherwise, data collected before withdrawal will be retained for analysis. Each site maintains a log of people who satisfy the inclusion criteria but are not recruited, including demographic details and reasons for not consenting to participate if given.

Interviewers, who are masked to the allocated group, interview participants face to face at home at baseline and 13 and 26 weeks after first interview. Data quality will be promoted in two ways: interviewers undertake online training (focussing on completing outcome measures) and sites check all documents to minimise missing data.

The Trial Manager (TM) coordinates recruitment from all sites and forwards details to STU through their email-based randomisation service. After baseline interviews, the unmasked Trial Data Manager (TDM) at STU oversees stratification of people with dementia by Trust, age, gender, whether living with a carer and time since first attendance at memory clinic or equivalent. Allocation between intervention and comparator groups using dynamic software randomises participants in real time, thus preventing subversion while ensuring (stochastic) balance between the two groups [35].

The person with dementia receives a letter specifying their allocated group and reminding them what this entails. Those in the comparator group receive a general guide to dementia [13]. The intervention group receive confirmation of arrangements for an initial visit by the DSP after about 2 weeks.

Cost-effectiveness analysis will be undertaken to evaluate the full cost of this intervention and whether it yields value for money. National unit costs for specific items of service will be used to estimate the costs of services used. Primary analysis will estimate quality-adjusted life years (QALYs) from the EQ-5D-5 L using the value set recommended by the National Institute for Health and Care Excellence (NICE) at the time of the analysis [45]. Current value sets include the EQ-5D-5 L specific utility weights [46] and the crosswalk algorithm to map the EQ-5D-5 L survey to the EQ-5D-3 L value set [47]. QALYs gained between baseline and follow-up interviews (13 and 26 weeks) will be estimated as the number of weeks multiplied by the utility of observed survival. Covariates that affect costs or outcomes from relevant published economic evaluations will be identified [48]. Adjusted costs and QALYs will be bootstrapped to estimate pairs of net costs and QALYs. These will then be used to estimate the probability that the intervention may be cost-effective. Net benefit will be estimated by valuing net QALYs to reflect decision makers’ willingness to pay (WTP) for one QALY. Recent decisions by the NICE suggest that WTP is at most £20 k [49, 50]. Cost-effectiveness planes and cost-effectiveness acceptability curves (CEACs) will be plotted to summarize uncertainty associated with incremental cost-effectiveness ratios by the probability that the intervention in question is cost-effective [51].

This economic model will be varied to explore uncertainty arising from study design decisions by using higher and lower sets of unit costs; an earlier estimate of utility from the EQ-5D-5 L [47]; BADLS, the primary outcome measure, or the DEM-QoL to measure health benefit; different approaches to missing data (for example by analysing only complete cases) and predicting costs and QALYs over longer time horizons (for example 5 and 10 years).

This analysis will use the framework in Fig. 2. Again, continuous variables will be reported by means and standard deviations and categorical variables by counts. Both standardised and study-specific measures will need analysis of covariance to identify predictive factors.

Thematic and narrative analysis of interview transcripts will be undertaken to elicit the experiences of people with dementia and their carers. At least two researchers will analyse each transcript. Once all issues have been identified, an iterative process will be used to identify the major themes and processes through conceptual abstraction. Achievement of data saturation will be informed by a focus group of interviewers [52].