Smart@home – supporting safety and mobility of elderly and care dependent people in their own homes through the use of technical assistance systems and conventional mobility supporting tools: a cross-sectional survey

The present study was conducted from February to March 2018 in a district with an above-average number of elderly people. As part of a cross-sectional survey, tenants at the age of 65 to 90+, some of them living independently, others in need of care (Table 1), were asked to fill in a questionnaire (Supplementary file 1). In their apartments, individual smart home solutions and other conventional mobility supporting aids such as grab handles, service sockets and balcony elevations for easy entry and exit were previously installed between one and six months before study participation. The survey focused primarily on the aspects of the need for care, communication/social contacts and security (Table 2). The questionnaire was pre-tested for comprehensibility and legibility by researchers and nurses in Geriatrics, and modified accordingly.

The written survey was conducted by means of questionnaires with an enclosed stamped envelope. The survey included all tenants aged 65 to 90+ years of a municipal social housing association, in 27 of whose apartments smart home solutions and/or conventional mobility supporting aids to support mobility had been installed. The apartments are located in a large city near the city center and are available for socially disadvantaged persons with lower incomes. As a control group, a random selection of 100 additional tenants with the same age structure was available, in whose apartments no smart home solutions or conventional mobility aids had been installed so far. In order to ensure the highest possible response rate, the tenants had the opportunity to complete the questionnaire at home. The tenants could either return the completed questionnaire directly by mail in a stamped envelope or in a sealed envelope to the housing association.

Tenants with installed smart home and conventional solutions were provided with information on the study and the course of the investigation at an information event. Afterwards they had the opportunity to fill in the 2-page questionnaire (Supplementary file 1) directly on site or to take it home and return it afterwards. Tenants of the control group received a letter from the research group by internal mail from their housing association with explanations of the survey and a 1-page questionnaire with a stamped envelope. The data collection was carried out according to the medical-ethical principles of the Ethics Committee of the State of Berlin.

The questionnaire was developed by the authors based on literature review and usability criteria developed in further research projects on care-dependent geriatric residents and patients with the use of new technologies and consisted of three parts in total. In the first part, demographic data was collected. In the second part of the questionnaire, the participants were asked to give an assessment of various aspects of their individual housing situation in order to be able to compare the intervention and control group with regard to need for care, communication/social contacts and security in combination with smart home solutions (Table 2). The respondents were able to indicate whether and to what extent the statements on the questionnaire were applicable using a 5-level scale (1 = “fully agree”; 2 = “rather agree”; 3 = “partly agree”; 4 = “rather not agree”; 5 = “not agree”). In the third part of the questionnaire, tenants with built-in solutions were asked about their satisfaction with the smart home solutions which had been installed as well as other built-in conventional mobility supporting aids (Table 3 offers a complete list of sensors and aids installed). In order to assess user satisfaction with the installed solutions, the participants in the intervention group were given a scale of 1 (very good) to 6 (unsatisfactory).

Data were screened prior to analysis for any anomalies (e.g., missing data, outliers, and nonnormality). Data were analyzed using SPSS Statistics for Windows, Version 24 (Statistical Package for the Social Sciences, Chicago, Illinois). Data analysis was primarily descriptive. Scales used were interpreted pseudo metrically resulting in an analysis with mean values (standard deviations), since the use of medians did not allow for sufficient discrimination. Statistical significance of the contingency table (Table 2) was tested by Pearson’s chi-square test.

The cost-benefit analysis focused on quality, price and performance. Price performance ratio and the quality of the installed solutions – smart home solutions and conventional mobility supporting aids –were analyzed and evaluated accordingly. In determining the best price performance, the price of the installed solutions was compared to tenants’ evaluation of the installed solution, while quality performance was determined exclusively on the basis of tenants’ evaluation. Prices of the smart home solutions and conventional mobility supporting aids are displayed separately.

The price and quality assessment of the installed solutions (Table 4) was based on the individual prices for each solution, as given by the manufacturer, in combination with the perceived utility evaluation (quality) with grades ranging from 1 (very good) to 6 (unsatisfactory) as rated by the individual tenants. In a first step, ratings on the conventional mobility supporting aids were looked at separately from the smart home solutions, and in a second step, they were compared with each other.

The response rate was above average in both groups surveyed. A total of 37 of the 46 tenants with installed smart home solutions and/or conventional mobility supporting aids submitted the completed questionnaire, which corresponds to a response rate of 80.4%. In the control group, 64 questionnaires were returned from a total of 100 questionnaires sent out, which corresponds to a response rate of 64%.

The first part of the questionnaire was used to collect general personal data. Almost three quarters (74.3%) were between 65 and 80 years old, 25.7% were over 80 years old. Among the tenants in the control group, 63 persons indicated their age, 58.7% of whom were between 65 and 80 years old and 41.3% over 80 years old. In both groups, more than three quarters of the respondents were female. With regard to their housing situation, 42.9% of the tenants with installed smart home solutions stated that they were living alone, which was true for a significantly larger proportion of tenants in the control group, namely 61.7% in total. 43.8% of the tenants in the intervention group were in need of care. In the control group, on the other hand, this was true for a significantly lower proportion, namely 13.6% of the tenants surveyed (Table 1). Of the 37 tenants in the intervention group, a total of 15 persons (40.5%) were in need of long-term care, with care degrees 1 and 2 (care degrees (Pflegegrade) in Germany range from 1 = minor impairment of independence or skills to 5 = severe impairments of independence or abilities that are associated with special demands on nursing care) being the most frequently represented with a share of 33.3% each. None of the tenants in this group had the highest degree of care (care degree 5). Among the tenants from the control group, 8 out of 64 people surveyed required nursing care, which corresponded to 12.5% of the residents. Care degree 2 was most frequently represented (50%). Care degrees 4 and 5 did not occur in this group. The most frequent physical complaints were mobility limitations, both in the intervention group (59.5%) and in the control group (43.8%).

The second part of the questionnaire was intended to provide insights into various aspects related to smart home solutions. With regard to the aspects of care, communication/social contacts and safety in connection with smart home solutions, tenants with installed smart home solutions evaluated the given statements in the questionnaire similarly to the control group (mean values ranging from 1.08–3.73 and 1.3–3.33). Significant results were available for aspects of communication/social contacts and safety. For example, the survey showed that both groups were fully satisfied that they had regular contact with a reference person in order to report that they were doing well (Table 2 mean 1.08; n = 37 and mean 1.70; n = 63). In addition, the participants in the study stated that they felt safer in their homes thanks to smart home solutions. This was fully true for tenants from the intervention group (mean 1.55; n = 31), and nearly equally true for tenants from the control group (mean 2.38; n = 26). Finally, the participants of both groups said that it was fully true for them to know whom to inform immediately in the case of an emergency (mean 1.09; n = 35 and mean 1.34; n = 61).

Table 3 shows the mean values of user satisfaction among tenants with installed smart home solutions, displaying the three best rated technical and conventional solutions. On a scale of 1 (very good) to 6 (unsatisfactory), the tenants’ assessments of the smart home solutions and conventional mobility supporting aids installed in their homes were in the range from very good to good. The technical solutions were rated with average mean values between 1.00 and 1.83, while the ratings for conventional mobility supporting aids ranged between mean values from 1.00 to 1.80.

The highest rated component of the installed smart home solutions was the “home emergency call” (mean 1.00), which for example in the event of falling, forwards an emergency call to the corresponding emergency call centre. The “visual doorbell”, which signals the ringing of the doorbell to tenants with impaired hearing both acoustically and visually by means of a wireless bell, was rated second best (mean 1.07). In third place was the “orientation light” (mean 1.15), which activates a glare-free floor lighting at night when leaving the bed. The “tablet”, which is used to operate the automated modules, such as the “light control”, received the lowest rates (mean 1.83). Regarding the installed conventional mobility supporting aids, the item “handles balcony” was rated “very good” (mean 1.00), followed by the “balcony exit/elevation” (mean 1.45) and the components “service socket” and “handles bathroom” (mean 1.50). The lowest rating was given to the “handles toilet” (mean 1.80). Overall, the tenants’ satisfaction with all installed smart home solutions and conventional mobility supporting aids was rated “very good” (mean 1.41).

Three limiting methodological problems have to be considered in the study carried out: First, the small sample size was small, which is due to the fact that smart home solutions were only installed in 27 apartments of the respective housing association. Second, the actual testing phase of the installed smart home solutions and conventional mobility supporting aids by the users was quite short due to the fact that the installation of some of the corresponding solutions was completed four weeks before the start of the survey. A third methodological problem is, as it is often the case in surveys, the social desirability of the response behaviour. Social desirability is promoted, among other things, by the adaptation to structural characteristics, in this case the installation of smart home solutions and conventional mobility supporting aids by the own housing association. However, socially desirable behaviour may also be promoted by the design of the survey instrument. In order to prevent this tendency as far as possible, emphasis was placed on specific questions oriented towards the individual installed solution or aid in order to reduce the willingness and also the possibilities for criticism. An additional methodological precaution against socially desirable response behaviour was the accompanying letter, in which the tenants were informed that they would participate in the survey voluntarily and anonymously.

Moreover, regarding statistical significance, multiple testing should always be considered. However, since the number of chi-square tests in our study is limited, we assumed the type 1 error to be rather small and thus, further statistical tests have not been carried out.