Study design and setting

An online cross-sectional survey was developed using adapted scales on knowledge and awareness of dementia, WTP questions from an unrelated study by Harapan and Anwar,^{Reference Harapan, Anwar, Bustamam, Radiansyah, Angraini and Fasli15} and professional, patient and public contributions, with topic domains mapping onto the TFA. Stakeholder input was gathered as part of a series of workshops regarding DMT service readiness. A survey draft was piloted, and all feedback was incorporated into a final version. The survey was anonymous and took approximately 10–15 min to complete. No financial compensation was provided to respondents. The survey was open from 23 May 2023 to 9 June 2023. Ethical approval for the study was obtained from the HPM/CGH REC Trinity College Dublin on 31 March 2023.

Participants

The survey among the general public was conducted online via the Dynata research company. It was also scripted using Google Forms and emailed to 15 national and local healthcare professional leads specialising in neurocognitive disorders in Ireland. These leads then disseminated the survey link to their extensive networks, reaching over 250 professionals. This network included clinicians in neurology, psychiatry of later life, and medicine for the elderly; nurses, allied health professionals, dementia advisors and medical social workers.

To reach out to the members of the general public in Ireland, Dynata utilised a router for sampling in line with industry standards for online research. Participants were contacted via email invitations, the Dynata app or their membership page. Upon interacting with the platform, participants were offered a variety of surveys to take part in, based on known data about them and their responses to real-time targeting questions.

Participants were provided with the survey details, enabling them to make an informed choice based on their preferences and available time. Although samples from opt-in panels such as these may be subject to unknown selection biases related to people's unobserved characteristics, the use of this data collection method has grown considerably in recent years due to its relative cost-effectiveness, with no indication of inferior results.^{Reference Johnston, Boyle, Adamowicz, Bennett, Brouwer and Cameron16}

Survey structure

The survey comprised a structured questionnaire divided into seven sections, including demographics, disease characteristics, a case study, personal considerations, societal considerations, WTP, and knowledge and understanding of Alzheimer's disease prevention and treatment.

The structure of Ireland's healthcare system, which combines public and private provisions, warrants the importance of assessing the public's WTP for emerging DMTs, reflecting the direct financial implications for the majority of the population not fully covered by public healthcare or private insurance. Access to publicly funded services is cost-free for about 40% of the Irish population. For the rest of the population, access to public healthcare services is available but comes with co-payments attached. Around 40% of residents opt for private health insurance to supplement their healthcare needs.

Study size

The sample size was guided by practical considerations such as the survey company's reach, the accessible participant pool in Ireland and the funding available, aiming for a balance between statistical power and unnecessary use of resources. Thus, 1000 was chosen as the sample size to yield a margin of error of 3% (with 95% confidence intervals). This sample size also aligns with prior research conducted in Ireland that employed a survey company.^{Reference Walsh, O'Shea, Pierse, Kennelly, Keogh and Doherty17,Reference Teahan, Walsh, Doherty and O'Shea18} Indeed, the sample is broadly representative of the population over the age of 18 in Ireland based on age, gender, marital status, education and employment status, though we acknowledge it is somewhat overrepresented by younger and more highly educated people (see supplementary Table 1 available at https://doi.org/10.1192/bjp.2024.24).

Quantitative variables

DMT acceptance was captured using a 5-point Likert scale, ranging from ‘extremely unlikely’ to ‘very likely’. For the multivariate regression analysis, a new binary response variable, based on the original acceptance variable, was generated to explore the factors influencing DMT acceptance. This binary variable takes a value of 1 if a respondent was likely or very likely to accept a DMT and a value of 0 otherwise.

To assess the WTP for a DMT, we generated a variable to capture the participants’ highest WTP, following the method outlined by Harapan and Anwar.^{Reference Harapan, Anwar, Bustamam, Radiansyah, Angraini and Fasli15} Specifically, a list of DMT price ranges was presented to respondents (€1–4999; €5000–9999; €10 000–15 000; €15 001–20 000; €20 001–25 000). For each price range, respondents were asked how likely they were to be willing to pay, using a 5-point Likert scale ranging from very unlikely to very likely. If respondents indicated they were likely or very likely to be willing to pay, they were coded as willing to pay. To determine the amount that individuals would be willing to pay, the median of each interval was calculated. If the participant was willing to pay the lowest price (€1–4999), they were assigned the interval of that price range. The price was then increased until the participant was no longer willing to pay, i.e. they answered unlikely or very unlikely. Thus, WTP was defined as the median of the highest accepted price, i.e. the highest price the participants said they were still very likely or likely willing to pay.

The survey included a number of explanatory variables of relevance to analysing DMT acceptance and WTP. Firstly, perceived knowledge of Alzheimer's disease or dementia was captured using three self-report questions relating to the prevention, treatment and progression of Alzheimer's disease. Respondents were asked to indicate their understanding of these three domains on a 5-point Likert scale. If they indicated that they were knowledgeable or very knowledgeable about any one of these domains, they were coded as being knowledgeable regarding Alzheimer's disease. In addition, respondents were asked to indicate their understanding of available Alzheimer's disease treatments on a 5-point scale. A binary response variable was generated, whereby respondents were coded as ‘understands available Alzheimer's disease treatments’ if they specified that they were knowledgeable or very knowledgeable. Finally, a range of attitudinal questions were included in the survey, which are also included as controls in the regression analysis. In particular, respondents were asked the following: the degree to which they think finding a cure for Alzheimer's disease is a priority; whether DMT should be available to everyone regardless of stage; and whether people with Alzheimer's disease should get the best treatment, even if the government says it is not cost-effective. The remaining explanatory variables are aligned directly to the survey. Full definitions and sample descriptive statistics for all variables are presented in Tables 1 and 2.

Table 1 Sample characteristics (general public)

Table 2 Sample characteristics (healthcare professionals)

Statistical methods

A series of descriptive and multivariate regression analyses were undertaken. Descriptive statistics are presented for all demographic and attitudinal variables. In our multivariate analysis, DMT acceptance for individual i is modelled as a function of a vector of personal characteristics $( {X_I^P } )$ and a vector of awareness and knowledge variables $({X_I^K})$ such that:

(1)$$DMT\,Acceptance_i = f( {X_I^P , \;\,X_I^K , \;\,{\varepsilon}_i} )$$

Here, DMT Acceptance_i is an indicator variable taking a value of one if an individual is willing to accept a DMT and a value of zero otherwise, and ${\rm \varepsilon }_i$ is a stochastic error term. The variables included in $X_I^P$ are age, gender, marital status, education, employment status, private health insurance status, region and whether the individual is a healthcare professional. The variables included in $X_I^K$ include whether the respondent is a care partner of someone with Alzheimer's disease, whether they know someone with Alzheimer's disease and whether they have a family history of Alzheimer's disease, as well as the two indicator variables described above capturing knowledge of Alzheimer's disease and understanding of available Alzheimer's disease treatments. In addition, three indicator variables were included, which capture respondents’ attitudes towards Alzheimer's disease, details of which are provided above. Given the binary nature of the dependent variable, a logit model was estimated, and average marginal effects are presented.

In addition, a multivariate regression analysis was undertaken to explore the factors associated with WTP for a DMT. A generalised linear model (GLM) was employed with a log-link and Poisson distribution. The distribution was informed by a modified Park test; the link function was informed by a Pearson correlation test, a Pregibon link test and a modified Hosmer and Lemeshow test. The same explanatory variables were included as detailed above, along with the ordered variable capturing DMT acceptance. Missing data were minimal and not a factor in the analysis. Statistical significance was explored at the 0.05 level, and model fit by the log-likelihood, Akaike information criterion (AIC) and Bayesian information criterion (BIC), and all analyses were performed using Stata 17 for Windows.¹⁹

Content analysis was performed for open-text answers relating to factors that participants would consider when deciding whether to accept a new DMT for Alzheimer's disease.