Approximately 46.8 million people are living with dementia worldwide and it is expected that this figure will double every 20 years [1]. The number of people affected by dementia will probably reach 74.7 million in 2030 and 131.5 million in 2050 [2]. Alzheimer's disease (AD) is the most common type of dementia, representing between 60% and 80% of dementia cases [3]. The costs of this disease are very high now and will become even higher in the future, and will impact severely, directly, and indirectly on health systems and patients’ families [4]. No therapy has been found to stop dementia progression, but any potential treatment should be administered as early as possible. Therefore, it is crucial to provide a cognitive assessment of elderly patients as early as possible and continue monitoring it [5-7]. Information and communication technologies have been employed in dementia and they are particularly helpful [8], allowing physicians, psychologists, and nurses to reach patients without displacing them [9,10]. Several studies have evaluated the use of telemedicine for dementia disorders, but doubts still exist about the reliability of cognitive tests applied by videoconference [11,12].

In this field, telemedicine shows benefits and limitations. Among the benefits, there is the possibility to include patients who prefer to stay at home rather than going to the hospital or clinics [13] and for their caregivers to not be alone in providing better care. Costs and the displacement of patients can be reduced, more frequent monitoring can be assured, and waiting lists and hospital staff work are lowered and the caregiver time is preserved [14]. These benefits justify why cognitive assessment via videoconferencing is potentially useful and valid in dementia patients.

However, telemedicine also has some risks: older adults are frequently resistant to use new technologies and their use requires necessary skills. The security of sensitive data, the identification of the evaluator, and the quality of data transmission may represent further limitations [15].

The main concern is the reliability of data obtained by telemedicine. The assessment of AD requires the evaluation of cognitive functions, which is done through specific tests. These include the Mini Mental State Examination (MMSE), the most used for clinical purposes, and the Alzheimer’s Disease Assessment Scale cognitive subscale (ADAS-cog), the most used for measuring the effect of treatments. The feasibility of the MMSE by videoconference has been investigated by several authors [16-23] (Table 1). One of these studies was done in an Italian population. No differences were reported between the scores obtained when the MMSE was administered face-to-face versus videoconference, except in one study in which it was found for 40% of the patients the videoconference MMSE was two points lower than the face-to-face MMSE [21].

The majority of previous studies have considered mild to moderate AD outpatients, living in different contexts, more often in rural areas. All studies have used the traditional 30-item MMSE, except one which has used a 28-item MMSE [17]. In summary, MMSE studies in videoconference have not yet evaluated the follow-up of AD patients. No study, to our knowledge, has evaluated the reliability and feasibility of the ADAS-cog test by videoconference modality. Moreover, the feasibility of MMSE has not been fully evaluated. On this basis, we wanted to assess if MMSE and ADAS-cog are reliable at follow-up.

On this basis, our aims were (1) to evaluate if the videoconference administration of MMSE and ADAS-cog were comparable to the face-to-face administration, and (2) to assess the acceptance of patients and caregivers of the videoconference modality. Our study was focused on these aspects and not on the diagnosis of AD that in the majority of cases requires more extensive and articulated diagnostic procedures.

A total of 28 AD patients (8 male, 20 female) with a mean baseline MMSE of 19.6 (SD 3.0), ADL mean 3.1 (SD 1.0), and IADL mean 2.0 (SD 0.8) were evaluated by administering the MMSE and ADAS-cog in both a face-to-face and videoconference modality. All participants reached the end of the study period of 2 years.

Baseline and follow-up MMSE scores obtained by face-to-face or videoconference modality are summarized in Table 2. As shown, no significant differences were noticeable between face-to-face or videoconference testings. The same was true for baseline and follow-up ADAS-cog scores, shown in Table 2. The time required for performing the tests was also comparable, with mean 37 (SD 8) minutes in the face-to-face modality and mean 38 (SD 10) minutes in the videoconferencing modality.

Data derived from the MMSE and ADAS-cog tests were also analyzed separately, dividing patients into three groups according to their MMSE scores at baseline recorded at the enrollment (eg, severely impaired, moderately impaired, and slightly impaired). As shown in Table 3, at baseline and follow-up, the two modalities did not influence the MMSE scores of the first and second groups of patients. Patients in the third group, who had a lower MMSE, obtained lower scores by the videoconference modality compared to the face-to-face modality. The same differences were observed in the ADAS-cog test (Table 4) with patients in the third group having higher scores by videoconference than in the face-to-face modality.

The acceptance of videoconference examination documented by a short questionnaire was quite high, with both patients and caregivers appreciating this modality. In particular, the preference for the web modality versus coming to the hospital was very high for both patients (mean 3.3, SD 1.5) and caregivers (mean 4.3, SD 1.5; Table 5).

In this study, we evaluated the reliability and feasibility of the MMSE and ADAS-cog tests by videoconference in mild to moderate AD outpatients. Two blinded raters administered the tests by face-to-face and videoconference modalities and compared the results. The MMSE results were the same in the two modalities. Only in nine patients with more pronounced cognitive deficits (MMSE<17) did the videoconference modality overestimate the impairment.

The results from the MMSE are consistent with findings reported by other studies [16-23]. Most studies evaluated patients affected by dementia or MCI, except in one study [18] that examined the reliability of MMSE administration via remote administration in a group of 72 patients with diabetes. These studies did not find differences between the scores of the MMSE administered face-to-face versus videoconference or telephone interviews [16-23]. In contrast to a larger previous study by another group [17], we used the traditional 30-item MMSE, which includes the writing and drawing tasks.

For the ADAS-cog test, this is the first study, to our knowledge, using a videoconference modality. In our investigation, we observed that the face-to-face or videoconference modalities did not influence the ADAS-cog scores of the first and second group of patients (MMSE scores 21-24 and 18-20, respectively), whereas patients in the third group, who had a lower MMSE (15-17), obtained more severe (higher) scores by videoconference compared to the face-to-face modality. This is probably due to the difficulty of understanding the meaning of specific questions. These findings suggest using the videoconference modality with patients with mild to moderate AD and excluding those with a moderate to severe AD.

Videoconferencing has been shown to be reliable not only in AD, but to assess the cognitive functions in other pathologies. Interesting results were provided for psychiatric patients [32-33], patients with Parkinson disease [34], and stroke patients [35], as well as in older adults [36]. Furthermore, videoconferencing has been found useful to evaluate the clinical status [37] and rehabilitation [38] of patients affected by different neurological disorders. All the preceding studies confirm that videoconference instructions are clear enough. This also was observed in our sample of patients, who reported that for saving time they preferred the teleconference modality than coming to the hospital. That is one good reason for supporting the videoconference approach. The patients and their caregivers were informed that videoconference evaluation ensured high levels of security and encrypted communication. From responses to the questionnaire on acceptance, they did not have concerns about the privacy of their medical information. The acceptability of the videoconference modality in this study is consistent with previous investigations in AD, which reported that 98% of the patients were satisfied and felt at ease with it [32,39].

Previous studies found that women, older patients, and less educated patients may be less receptive to technology. In contrast, men, younger patients, and those with higher education are more receptive to telemedicine and report less anxiety [40,41]. The link found between education and technology acceptability is worth being considered. Studies have revealed that higher levels of education are associated with higher levels of both computer knowledge and computer interest and lower levels of computer anxiety [42].

The range of settings where videoconferencing can be used is wide and it can represent a useful and effective method for assessing cognitive functions. However, continued validation studies and adaptation of neuropsychological instruments is warranted.

Ethical and legal consequences of assessing cognitive functions via telemedicine deserve to be discussed as well. Caring for patients suffering from dementia poses complex ethical problems because of the nature of dementia and the way that dominant ethical principles apply to its clinical features [43]. These problems include informed consent, the duty to maintain the confidentiality, and privacy of patient examination and records [44-48]. In this study, patients and/or their caregivers were informed about potentiality and the limits of the videoconference approach used and in our sample and none reported any issues regarding privacy. However, the study was performed in a dedicated hospital setting where patients were routinely examined. The option of carrying out remote videoconferences with patients staying at home was not applied; therefore, all ethical implications of home-based cognitive function assessment should be further investigated.

Our study has strengths and limitations. The long duration (24 months) of observation, the well-defined diagnosis, and comparable clinical characteristics of all patients, which make the sample quite homogenous, are strengths of our study. In fact, only sparse studies have evaluated MMSE and ADAS-cog by videoconference for follow-up. The presence of two independent and blinded raters with specific experiences should be also positively considered.

On the other side, we recognize that the number of patients investigated is obviously small because this was a pilot study. We are also aware that we observed patients in a hospital setting with the presence of an assistant. This was done to check the audiovisual system. However, we cannot exclude that the same patients could behave differently if they were at home. These limits should be addressed in a future study with a larger sample group.

Despite these possible limitations, the MMSE and ADAS-cog administration via telemedicine is useful to simplify the assessment of patients and to allow wider participation to clinical trials of people living in remote geographical areas. Further research with a larger sample group and a remote geographical location are required.

In conclusion, videoconferencing can be used to assess patients with mild to moderate dementia, document their cognitive stability or decline, and measure the effects of the treatments.