Persons interested fill out an online form that allows the identification of those who meet the pre-established requirements before attending an individualized face-to-face screening visit.

For the online pre-screening procedure interested participants access a web-based formulary [Research Electronic Data Capture (REDCap)-based; Harris et al., 2008, 2018] to collect basic sociodemographic data, a standardized questionnaire that allows to estimate a CAIDE dementia risk score and the presence of exclusion criteria (Table 1). Prior to the fulfillment of the on-line form (Supplementary Table S2), participants receive detailed information on the study objectives, future activities, and evaluations.

Participants with an estimated CAIDE dementia risk score ≥ six and no reported exclusion criteria are invited for an in-person selection evaluation visit and full written information is provided. At the beginning of the interview, information regarding the study details is provided to all participants. After reading the informed consent form and solving all possible questions and doubts, participants sign an informed consent. Subsequently, past medical history is recorded to rule out exclusion criteria. Demographic and anthropometry data together with cholesterol determinations are collected and a current CAIDE dementia risk score is obtained. Brief cognitive testing is then administered and CCI scores are accounted for to complete inclusion/exclusion criteria accomplishment. The variables obtained at the screening visit are depicted in Table 1.

In accordance with the WWFINGERS network data harmonization recommendations, cognitive performance change is the primary outcome. It is assessed using the modified Neuropsychological Test Battery (mNTB) (Harrison et al., 2007), previously described for in our pilot study (Tainta et al., 2024a) and in Supplementary Table S3. It is recorded by trained neuropsychologists who are blinded to the group assignment at baseline evaluation and at the 12-month and 24-month visits. The mNTB Cognitive Global score is obtained. Composite cognitive z-scores are calculated based on the results of all single tests. The higher the score, the better the performance. Changes in global cognition are assessed by differences between Global Cognitive composite z-scores at baseline and 24-month visits.

Cognitive Domains – (cognitive performance II): mNTB executive function, Memory, and Processing speed scores are obtained. Apart from the mNTB scores, perception and construction abilities function composites are also assessed using the Judgement Line Orientation, Rey Complex Figure copy, and 15 object test. Differences in pre-and post-intervention scores in mNTB-episodic memory, mNTB-executive function, and mNTB-processing speed z-scores together with perception and construction abilities composites are used to assess changes in specific cognitive domains.

A comprehensive assessment of the socioemotional domain, dietary and culinary habits, physical fitness, and vascular risk factors have been devised to evaluate the impact of the multidomain intervention (Table 2).

Participants provide information on past medical history and current illnesses, which is confirmed through their clinical history obtained from their General Practitioners (GPs) at the beginning of the study period and follow-up visits. Any change in a participant health along the study is recorded as an adverse event. Adverse events are actively identified through structured interviews at every follow-up visit. Current diagnoses, medication, and lab values, including blood count, cholesterol, glucose, renal and liver function, thyroid hormones, B12 vitamin, and folic acid, are documented before randomization.

This in-person visit is conducted by the study neurologists, who review medical history, risk factors, cognitive symptoms, functional status, cognitive performance and perform a general and neurological physical exam. Individuals diagnosed with dementia, as defined by the DSM-IV, are excluded. A diagnosis of Mild Cognitive Impairment (MCI) (Petersen, 2004) is confirmed when applicable. Participants are then randomly assigned to the Group 1 or Group 2 to complete the randomization visit. Participants are then informed of their designated group and receive both verbal and written information and recommendations on active and healthy aging, control of risk factors dietary advice, physical activity, social engagement, and routine suggestions aimed at fostering cognitively and affectively stimulating practices.

Participants assigned to the RHA group undergo follow-up evaluations at 12-month and 24-month marks. Adherence to the lifestyle recommendations provided during the randomization visit is self-directed, with no additional guidance on the activities pursued to follow these recommendations. Participants may connect with other healthcare professionals, including their general practitioners, who are not part of the study team according to their individual health needs.

The MD-Int program represents an update of the GZ pilot study intervention program, introducing a socio-emotional component along with a greater frequency of visits and extended materials within each intervention domain. The program aims to promote healthy brain aging, provide tools for incorporating these routines into daily life, and strengthen the social environment. The program is tailored to each participant needs and abilities throughout the study. Apart from the 12-month and 24-month visits participants in the MD-Int group receive individual interviews with the study physician and physiotherapist every 4 months. In these visits cardiovascular and metabolic parameters (blood pressure, pulse, weight, BMI, hip and waist perimeters, capillary glucose, and cholesterol), Mediterranean diet adherence, physical capacity and activity, active medications and adverse events are reviewed and monitored.

The MD-Int program is administered in small groups of 10–15 participants, with a designated investigator or “group leader” overseeing the various MD-Int activities and intervention staff. Intervention components are listed in Table 4. Cognitive stimulation intervention comprises 16 group sessions in workshops (24 h) conducted by a neuropsychologist and over 100 h of individual at-home cognitive training exercises using the EXERCITA® program that was specifically designed after the pilot study. Nutritional and culinary intervention includes eight nutritional and show cooking workshops facilitated by study nutritionists and professional chefs. Individual recommendations for physical exercise based on the VIVIFRAIL©² multicomponent physical exercise tool. And finally, up to 16 group workshops focused on promoting socio-emotional skills the intervention.

Sample size calculations were based on the FINGER trial (Ngandu et al., 2015) and GZ pilot study data (Tainta et al., 2024a). Accepting a significance level of 5% and a statistical power of 90% in a one-sided test, 584 subjects would be necessary for each group, with a total of 1,096 participants, to detect a statistically significant difference between the two proportions of participants who decline in performance on the mNTB after 2 years, 49% for the RHA group and 32% for the MI group. The sample size has been calculated to allow a predicted 20% dropout rate according to the GZ study data. In the same way, considering the same level of significance, statistical power, and dropout rate, in order to detect a decline in performance on the mNTB score after 2 years, considering data from the GOIZ ZAINDU Pilot study, we calculated that a sample size of 802 randomized participants would be sufficient to detect by a two tailed Student t test, a difference between groups of 0.10 units in global mNTB score change, considering a 0.33 standard deviation.

Randomization list is created to follow a 1:1 ratio stratified by age, sex, and presence of mild cognitive impairment (MCI). To minimize bias, the allocation of participants to each group will be concealed through the REDCap randomization procedure, previous to the recruitment process begins. A member of the research team who will not carry out primary outcome assessments will enter participant details into the randomization system and inform relevant parties of the outcome of group allocation. The study participants are distributed in 4 waves of approximately 250 subjects, in which age, sex, and cognitive status are equally similarly distributed. As the distribution of those variables would be unknown at the time the randomization list is created, the resulting groups could have minimum size differences.

As with any other intervention studies on lifestyle habits, CITA Go On has limitations regarding the maintenance of masking of the group assigned to each participant. Nevertheless, participants are said in the randomization visit to be referred to Group 1 in the case of RHA and Group 2 in the case of MD-Int assignment. Therefore, participants are not actively told which group, intervention or control, they belong to. Blinded neuropsychologists perform the primary and secondary cognitive assessment, and participants are specifically requested to avoid discussing the intervention with the research staff and other participants. Primary outcome raters did not participate in any of the cognitive and socio-emotional intervention activities and have no access to randomization information in the database.

Data is collected in a computerized database using an electronic Clinical Record Form (eCRF) record system based on REDCap (Harris et al., 2008, 2018). REDCap is a secure web-based software platform designed to support data capture for research studies.

First, the randomness of the samples will be verified to ensure that the baseline characteristics of the two groups do not show statistically significant differences. Univariate analysis will be conducted to identify sociodemographic and diagnostic variables that could differ between the two groups selected randomly before the study.

Continuous variables will be analyzed using Student’s t-test or the nonparametric Mann–Whitney test to compare medians when they do not follow a normal distribution. Categorical variables will be analyzed using the Chi-square or Fisher’s exact test for contrast in each case. Confounding variables will be included in the adjustment for all subsequent analyses. Logistic regression analysis will be performed to the control group and the MD-Int group. Composite scores will be calculated as Z-scores standardized to the Baseline mean and standard deviation (SD), with higher scores indicating better performance. The decline will be defined as a decrease in z-scores between Baseline and 24-month visit assessments (Z24month - Zbaseline < 0). Those covariables considered to influence changes in cognition will be introduced as model variables.

Third, mixed models will be used to study the trajectories of the different scores (NTBm) in the control and intervention groups. Mixed models explicitly account for the correlations between repeated measurements within each patient. The factors assumed to have the same effect across many patients are called fixed effects, and those likely to vary substantially from patient to patient are called random effects. In these mixed models, the interaction between the group variable (control vs. intervention) and the time will be included in all final analyses because the study’s final objective is to determine the effect of the intervention on effectiveness along time, adjusted for other variables that could influence them. For statistical analysis, the programs SPSS® (version 23) and R (version 4.2.3 or superior) will be used with a significance level of 95%.

A simulation model will be developed to represent the natural history of dementia to assess the cost-utility of the CITA GO-ON Study intervention (Mar et al., 2020). The model will compare two cloned cohorts that had slight differences in progression to dementia according to whether they received the multidomain intervention. It will calculate costs in Euros (€) in each cohort and utility in quality-adjusted life years (QALYs). Finally, the incremental cost-utility ratio (ICUR) will be estimated by dividing the incremental cost (Ca-Cb) by the incremental utility (Ua-Ub). The time horizon will be the lifetime of the patient. A 3% discount rate will be applied to costs and effectiveness. As the main economic impact of dementia is related to the care required as patients become dependent, a societal perspective will be used. Therefore, we will consider direct and indirect costs (Lopez-Bastida et al., 2006), Data for model parameterization will be mainly obtained from the CITA GO-ON study, although some parameters to build the model will be taken from the literature. The NTBm and functional scales (A-IADL-Q) will be used to obtain the model parameters that determine the progression of dementia from the pre-MCI stage to more advanced stages of the disease according to A-IADL-Q cut-off points (MCI < 60, mild dementia < 50 and moderate dementia < 40) (Dubbelman et al., 2022), die from the disease, or die from other causes. Mixed models for repeated measures allow the progression of the scales to be reproduced over time and will be used to simulate the trajectories of real patients (Mar et al., 2020). The model will reflect the improvement introduced by the effectiveness of the intervention in patients to the extent that those treated with multimodal intervention delay their evolution to dementia and improve their health-related quality of life score. We have constructed a preliminary version of the model using a synthetic database. The result has been cross-validated with other nine models within the International Pharmaco-Economic Collaboration on Alzheimer’s Disease (IPECAD) collaboration (Handels et al., 2025).