Table 2.

Summary of findings in relation to the evidence categories of the evidence standards framework for digital health technologies (N=16).

Tiers			Findings		Studies, n (%)		Risk of biasa
Tier 1
	Credibility with health and social care professionals	Minimum evidence standards show that relevant social care professionals were involved in the design, development, or testing of the GPS devices In 12 (75%) [22,24,25,28-36] of the included studies, social care professionals were involved in the testing of the GPS devices to a varied extent		12 (75)		—b
	Relevance to current pathways in health or social care system	Minimum and best practice evidence standards show that GPS devices have been successfully piloted or implemented in social care systems. This was described in 10 (63%) of the included studies. Of these 10, 3 were performed in Sweden [22,32,33], 6 in Norway [28-31,34,35], and 1 in Denmark [36]. All of them were part of larger projects supporting development of products, services, and decision-making processes to support OAsc and their families in their homes. Most of those projects were part of national government programs that aimed to stimulate the use of welfare technology		10 (63)		—
	Acceptability with users	Best practice evidence shows that representatives from the intended user groups (persons with dementia and OAs) were involved in the design, development, or testing of the DHTd and to show that users were satisfied with the DHT Representatives from the intended user groups (OAs in general or persons with dementia) were involved in testing of the GPS alarms in 15 (94%) of the studies (ie, all the included studies except [29]) Six (38%) of these studies showed that the users were satisfied with the alarms: 77% of the CGse of persons with dementia stated that they would recommend the use of GPS alarms in the Pot et al study [26]; 97% of the OAs who participated in the Røhne et al study [28] and 90% of the OAs in the Røhne et al study [35] stated that they were satisfied with the alarm All older users in the Ausen et al study [30] would recommend others in similar situations to use the GPS alarm User satisfaction was confirmed in the interviews in the Milne et al study [24] and in the values identified in the Boysen et al study [31]		6 (38)		—
	Equalities considerations	No information retrieved from included studies. Socioeconomic aspects were not addressed However, persons with dementia and OAs may be considered vulnerable groups		0 (0)		—
	Accurate and reliable measurements (if relevant)	No information retrieved from included studies. Technical validations were not included in the review. All included studies used commercial products		0 (0)		—
	Accurate and reliable transmission of data (if relevant)	No information retrieved from included studies. Technical validations were not included in the review. All included studies used commercial products		0 (0)		—
Tier 2
	Reliable information content	Minimum and best practice standards category is not relevant for GPS devices because they do not provide general information or advice to users concerning health, healthy living, lifestyle, diseases, illnesses, or conditions However, for reliability of information on user position and emergency situations, data on user testing were provided by 1 study (6%) [28], and data on CGs’ perceptions of the accuracy of the GPS information were provided by 2 (13%) studies. For example, relatives and staff in the Øderud et al study [34] had experienced slow or unreliable information on the user’s position. Moreover, [30] reported cases of poor mobile coverage that had resulted in failures in updating user position		0 (0)		—
	Ongoing data collection to show use	Cannot be assessed from the included studies. Evidence of ongoing data collection (required according to evidence standards for the category) was not reported in the included studies However, 10 (63%) of the included studies presented data on use on specific occasions related to the interventions In all, 3 studies (19%) reported quantitative data on usage period: [30] and [34] presented the number of participants who had used GPS trackers for up to 1 year and between 1 and 2 years, respectively. [33] reported the number of days that each participant had used the GPS trackers (mean 158 days, median 210 days, and range 37-260 days) A total of 3 (19%) studies [22,25,28] included system logs in the collection of data to investigate use. Interestingly, [22] saw that the extent to which persons with dementia used mobile phone–based GPS varied widely among the participants. Moreover, [28] described that the logs from the technical systems were thoroughly analyzed to understand the role and function of users, alarm units, response center, CGs, and relatives In all, 4 (25%) studies [23,24,26,27] based the data collection of use on the recall of the users or their CGs, and 2 of these [26,27] reported that the persons with dementia did not always take along the GPS devices (mobile phone or tracker worn on the belt) when going out and that the devices were not always switched on		0 (0)		—
	Ongoing data collection to show value	Cannot be assessed from the included studies. Evidence of ongoing data collection to show value (required according to evidence standards for the category) was not reported However, 11 (69%) of the included studies presented data on use for values related to the health and welfare outcomes of users (OAs, persons with dementia, and CGs of persons with dementia) on specific occasions related to the interventions; one (6%) study [22] identified that CGs experienced that the persons with dementia had become more independent in outdoor activity; 1 (6%) study [24] identified that CGs and staff saw that GPS trackers could give persons with dementia in milder stages of dementia and their CGs increased freedom and decreased stress and anxiety; and 1 (6%) study [26] identified values perceived by some of the CRsf, including increased freedom and decreased worries and fewer conflicts with CGs when going outside alone. Moreover, the CGs experienced that they gave more freedom to the CR and some experienced fewer conflicts with the CR Another study (6%) [28] identified that more than 50% of the users thought that the GPS alarm helped to increase their freedom One study (6%) [29] noted that more than 50% of the participating staff perceived that GPS trackers for persons with dementia could, to some degree, free up time for service providers by reducing the number of inspections they carried out to see if the person is well, driving to and from the user and following the user on walks; [30] identified that all persons with dementia thought that GPS trackers enabled them to increase or maintain physical activity, to increase freedom in outdoor activities, and that all relatives experienced that the GPS trackers increased their feelings of safety when leaving the person with dementia by themselves; and [31] identified positive values of GPS trackers both in shared housing for persons with dementia, including freedom for persons with dementia, decreased stress and anxiety for employees, time savings for staff and cost reduction, and for home users, including increased security, with, in some cases, increased outdoor activity and CG relief Another study [32] identified that 5 of the 8 GPS tracker users experienced increased security and could continue to live at home for a longer periods. In addition, 5 of the 8 relatives experienced fewer concerns and worries; [34] noted that most of the users perceived that the GPS trackers provide security (for the user, CG, and staff), increase freedom for the user and sometimes also the CG, as well as help the user to be physically active and maintain their activity level; [35] identified that most of the GPS alarm users experienced that it increased their safety and freedom in daily life; and [36] identified that the GPS tracker increased the security and quality of life of persons with dementia and their CGs		0 (0)		—
	Quality and safeguarding	Cannot be assessed from the included studies However, 3 (19%) of the included studies had a study aim or presented data related to system-level quality and safeguarding: [30] and [34] presented the service model for implementation of GPS trackers in the homes of older adults, which included safeguarding measures taken by the municipality. Moreover, [34] presented data on the roles of different actors (users, relatives, and alarm centers) in charging and administration of the alarm as well as locating and retrieving the user, if necessary One study (6%) [28] described the establishment of an initial test routine to encourage users to regularly trigger the alarm when out walking		0 (0)		—
Tier 3a
	Demonstrating effectiveness in outcomes or improvements in outcomes	Effectiveness is not demonstrated in outcomes or improvements in outcomes according to best practice standards: no increase in the frequency of OAs going outside; no significant differences in changes in fear of falling, feelings of unsafety, or quality of life [27] Effectiveness is not demonstrated in outcomes or improvements in outcomes according to minimum evidence standards. Indications were identified for the following: Decrease in time searching for person with dementia (from a mean of 3-4 hours per event to 40 minutes) [24] Increase in the number of days that person with dementia was engaged in independent outdoor activity (three cases, no statistics available) [25] Decrease in role-overload of CGs of persons with dementia (P=.126; d=–0.25 for all CGs, and P=.119; d=–0.34 for CGs who could reach CR with the mobile alarm) and in feelings of worry (P=.08; d=–0.32 for all CGs, and P=.057; d=–0.46 for CGs who could reach CR with the alarm) [26] Reduction in costs for care of persons with dementia because of prolonged time that the person could live independently instead of in special housing (up to 3 months) [32,33,36] Difference in mean CG burden between relatives of persons with dementia using and not using GPS (P=.04) was indicated in small samples because a crossover design was used [37] Outcomes investigated with negative results: Activity of person with dementia: reduced because of disease progression [22] Burden and quality of life for CGs of persons with dementia: no significant changes [23]		1 (6) 7 (44) 2 (12)		High High High
	Use of appropriate behavior change techniques (if relevant)	Not relevant for GPS devices: no behavior change techniques used		0 (0)		—
Tier 3b
	Demonstrating effectiveness: improvements in outcomes	Effectiveness is not demonstrated in improvements in outcomes according to minimum evidence or best practice standards (see above)		0 (0)		—

^aRisk-of-bias assessment: a “high” risk of bias is equivalent to “high” for the Cochrane Risk-of-Bias Tool 2.0 (randomized studies) or “serious/critical” for the Risk of Bias in Non-randomized Studies of Interventions Tool (nonrandomized studies), and a “low” risk of bias is equivalent to “low” for the Cochrane Risk-of-Bias Tool 2.0 (randomized studies) or “low/moderate” for the Risk of Bias in Non-randomized Studies of Interventions Tool (nonrandomized studies).

^bNot available.

^cOA: older adult.

^dDHT: digital health technology.

^eCG: caregiver.

^fCR: care receiver.