The caregiver training program consisted of applying the cognitive stimulation model of the CUIDA-2 program. This application included theoretical training made up of three modules: 1) person-centered care, 2) communication strategies, and 3) mediated cognitive stimulation strategies. The training was given in two group sessions of two hours each, plus 50 hours of individual practice, either on the job (in the case of professional caregivers) or in the home (in the case of family caregivers), in both cases supervised by psychologists who were experts in the program. In these individual practice hours, the caregivers were required to keep a weekly log. Here they had to plan in advance the activities that they were going to carry out with the older adult, and once they had taken place, they had to record how they were done and how the older adult had responded.

This screening instrument is widely used for detecting cognitive impairment. The final score ranges from 0 to 35 points and is often used as a global index and method for monitoring the evolution of cognitive functions in cognitive impairment and dementia. It has high internal consistency (α = 0.88), good test-retest reliability (0.64-1.00; p < 0.01) and good interjudge reliability (0.69-1.00; p < 0.01).

In the present study, we used the cognitive assessment scale only, which includes one part on information and orientation and another part on mental ability. The final score ranges from 0 to 23 points and higher scores mean a better outcome. Its test-retest reliability falls between 0.79 and 0.90, and between 0.61 and 0.69, for the information and orientation scale and for the mental ability scale, respectively.

Generic measure of health-related quality of life. The individual rates his or her own state of health, first rating levels of severity by dimensions, and afterward a more general assessment using a 20-centimeter visual analog scale (VAS) that goes from 0 to 100. Scores range from 0 to 2 on each of the test scales, and higher scores mean a worse result.

Evaluates the person's functional ability based on 10 items answered by the caregiver. Scores range from 0 to 100, with 0 being completely dependent and 100 completely independent. Its internal consistency presents an alpha coefficient between 0.86 and 0.92 and interjudge reliability between 0.84 and 0.97.