INTRODUCTION
Various technological innovations have emerged to facilitate real‐time data collection, offer remote/distance delivery of dementia care interventions that are traditionally offered in‐person, or serve as direct interventions themselves for people living with dementia and their caregivers.1, 2 Remote activity monitoring (RAM) is utilized to track and alert users (e.g., professional staff, caregivers) to behaviors and challenging events that may presage more adverse health outcomes such as hospitalizations or nursing home admissions.1, 3, 4 These systems often feature unobtrusive activity monitoring of people with dementia via the installation of various types of sensors throughout the living setting; algorithms are then applied to the sensor data to alert professionals or family caregivers when certain types of expected or unexpected activities occur. The aim of the present study was to assess whether RAM technology was associated with reductions in negative health transitions and service utilization for persons with Alzheimer's disease or a related dementia over an 18‐month period.
METHODS
The design, procedure, data collection, and sample characteristics of this 18‐month, mixed methods randomized controlled trial is described in detail elsewhere.3, 4, 5, 6 Briefly, the clinical trial enrolled 88 care recipients and their caregivers in the RAM intervention arm and 91 care recipients and their caregivers in the control arm. To analyze the effects of the RAM system on care recipient health transitions and service utilization, we focused on quantitative data collected on baseline, 6‐, 12‐, and 18‐month caregiver surveys. The treatment group had the RAM system installed in their home. The attention control group did not receive RAM technology, but did receive bi‐annual “check‐in” calls from project coordinators to maintain rapport and ensure follow‐up survey completion.
Baseline and follow‐up surveys assessed whether the care recipient had fallen or wandered in the past 6 months (yes/no). Falling was defined as an unintentional change in position to the floor or ground. Wandering was defined as an aimless or purposeful activity that causes a social problem, such as getting lost, leaving a safe environment or intruding in inappropriate places. Frequency of falling or wandering was captured categorically (1–2, 3–6, 7–9, or 10 or more times). 7
Caregivers were also asked whether the care recipient had used any of the following healthcare services in the past 6 months: nursing home stays, assisted living stays (including memory care), other residential care stays (e.g., family care home, adult foster care), hospital stays, or emergency room visits. 8
Pearson chi‐squared tests were used to conduct a descriptive analysis of outcomes. Multilevel mixed effects models were utilized to estimate odds ratios for binary and categorical outcomes. A random effect was included to account for variability between care recipients. In adjusted models, we included linear time, care recipient age and sex, and baseline level of the dependent variable if significant differences were observed at baseline. All analyses were conducted in Stata 15.1.
RESULTS
In unadjusted models (Table 1), reported emergency room visits and falls were significantly lower for care recipients in the intervention arm compared with the control arm in the months preceding the 18 month survey interval. In adjusted models (Table 2), emergency department visits were almost 50% lower in the intervention group compared with the control group. In addition, the odds of experiencing a higher frequency of falls (i.e., being in a higher response category) versus a lower frequency of falls was 0.36 (95% CI 0.15–0.85) for those in the intervention group compared with controls. The RAM technology did not have a statistically significant effect on any other outcome.
TABLE 1.
Unadjusted proportion of outcomes in prior 6 months
| RAM intervention | Control | p‐value | |
|---|---|---|---|
| Outcomes | |||
| Falls | |||
| Baseline (T1), n (%) | 47 (53.4%) | 52 (57.1%) | 0.615 |
| 6 months (T2), n (%) | 38 (45.8%) | 45 (52.3%) | 0.217 |
| 12 months (T3), n (%) | 31 (39.7%) | 39 (46.4%) | 0.498 |
| 18 months (T4), n (%) | 36 (46.8%) | 50 (61.7%) | 0.037 |
| Wandering | |||
| Baseline (T1), n (%) | 26 (29.6%) | 15 (16.5%) | 0.037 |
| 6 months (T2), n (%) | 21 (25.3%) | 12 (14.0%) | 0.051 |
| 12 months (T3), n (%) | 20 (25.6%) | 19 (22.6%) | 0.650 |
| 18 months (T4), n (%) | 23 (29.9%) | 15 (18.5%) | 0.086 |
| Nursing home admission | |||
| Baseline (T1), n (%) | 3 (3.4%) | 2 (2.2%) | 0.623 |
| 6 months (T2), n (%) | 4 (4.8%) | 3 (3.5%) | 0.664 |
| 12 months (T3), n (%) | 10 (12.8%) | 3 (3.6%) | 0.030 |
| 18 months (T4), n (%) | 4 (5.2%) | 6 (7.4%) | 0.395 |
| Other residential admission | |||
| Baseline (T1), n (%) | 3 (3.4%) | 6 (6.6%) | 0.330 |
| 6 months (T2), n (%) | 9 (10.8%) | 13 (15.1%) | 0.409 |
| 12 months (T3), n (%) | 8 (10.3%) | 14 (16.7%) | 0.234 |
| 18 months (T4), n (%) | 9 (11.7%) | 8 (9.9%) | 0.713 |
| Hospitalization | |||
| Baseline (T1), n (%) | 18 (20.5%) | 12 (13.2%) | 0.193 |
| 6 months (T2), n (%) | 11 (13.3%) | 13 (15.1%) | 0.729 |
| 12 months (T3), n (%) | 7 (9.0%) | 6 (7.1%) | 0.668 |
| 18 months (T4), n (%) | 4 (5.2%) | 11 (13.6%) | 0.072 |
| Emergency room visit | |||
| Baseline (T1), n (%) | 35 (39.8%) | 24 (26.4%) | 0.057 |
| 6 months (T2), n (%) | 20 (24.1%) | 23 (26.7%) | 0.693 |
| 12 months (T3), n (%) | 9 (11.5%) | 19 (22.6%) | 0.062 |
| 18 months (T4), n (%) | 11 (14.3%) | 22 (27.2%) | 0.047 |
Note: Sample sizes for remote activity monitoring (RAM) intervention group: T1 N = 88, T2 N = 83, T3 N = 78, T4 N = 77. Sample sizes for control group: T1 N = 91, T2 N = 86, T3 N = 84, T4 N = 81. Pearson chi‐square p‐values assess differences in outcomes by treatment group at each time point. Comparisons were deemed statistically significant at p < .05, with differences also reported at the p = .05 level, represented in bold.
TABLE 2.
Adjusted odds of outcomes in prior 6 months (RAM intervention vs. control)
| Odds ratio (SE) | 95% CI | p‐value | |
|---|---|---|---|
| Falls | 0.48 (0.18) | (0.23–1.00) | 0.051 |
| Wandering | 1.65 (0.79) | (0.65–4.22) | 0.290 |
| Nursing home admission | 1.58 (0.71) | (0.65–3.82) | 0.314 |
| Other residential care admission | 0.73 (0.24) | (0.38–1.40) | 0.346 |
| Hospitalization | 0.77 (0.23) | (0.43–1.39) | 0.384 |
| Emergency room visit | 0.51 (0.17) | (0.27–0.97) | 0.041 |
Note: Sample sizes for remote activity monitoring (RAM) intervention group: T1 N = 88, T2 N = 83, T3 N = 78, T4 N = 77. Sample sizes for control group: T1 N = 91, T2 N = 86, T3 N = 84, T4 N = 81. All models adjusted for time, care recipient age, and sex. Wandering also adjusted for “wandering in the last 6 months” at baseline because significant differences were observed between treatment and control groups. Comparisons were deemed statistically significant at p < .05, with differences also reported at the p = .05 level, represented in bold.
Abbreviations: CI, confidence interval; SE, standard error.
DISCUSSION
Although RAM did not provide direct support for the management of behaviors for persons with AD/ADRD, the findings imply that this technology may prevent some adverse health events for people living with dementia in the community. The ongoing, unobtrusive monitoring and system alerts of RAM may have resulted in caregivers identifying activity or the lack thereof that may have prevented falls and wandering events. In turn, emergency room use among persons with dementia may have been avoided. Although other trials have reported null or nonsignificant findings, the 18‐month follow‐up period may have allowed us to identify the influence of RAM on health service use or other events that are more likely to occur over time.9, 10 There are several important limitations, however. The sample was not representative and researcher blinding was not possible. Although the technology was used in at‐home settings, in several instances, dementia caregivers simply stopped using systems when they were perceived as not helpful, thus threatening internal validity.
The clinical and public health needs to develop and implement strategies that effectively support people living with dementia at home and their unpaid caregivers are pressing. Technology solutions that: (a) supplement extensive, unpaid assistance from family members and others; and (b) prevent or delay the onset of negative health events could address dementia care challenges, as suggested by our study.
AUTHOR CONTRIBUTIONS
Joseph E. Gaugler conceptualized the study, drafted and edited the entire manuscript, and prepared the manuscript for final dissemination. Christina Rosebush and Rachel Zmora conducted the empirical analyses and finalized the results in text and tabular format. Elizabeth A. Albers assisted with data analysis and manuscript review.
FUNDING INFORMATION
This study was supported by grant R18 HS22836 from the Agency for Healthcare Research and Quality (Principal Investigator: Joseph E. Gaugler, PhD).
CONFLICT OF INTEREST
The authors declare no conflicts of interest.
SPONSOR'S ROLE
None.
ACKNOWLEDGMENTS
The authors would like to thank the families and people living with dementia who dedicated their time to participate in our study. The authors would also like to thank Sharon Blume and the Lutheran Home Association in Belle Plaine, Minnesota for their partnership on this project. The authors also thank HealthSense, Inc. (later GreatCall, Inc.), which was the company we initially engaged with when conducting this project.
Funding information Agency for Healthcare Research and Quality, Grant/Award Number: R18 HS22836
REFERENCES
- 1. Demiris G. Technology‐based solutions to address the family care gap challenge. In: Gaugler JE, ed. Bridging the Family Care Gap. Academic Press; 2021:367‐386. [Google Scholar]
- 2. Lindeman DA, Kim KK, Gladstone C, Apesoa‐Varano EC. Technology and caregiving: emerging interventions and directions for research. Gerontologist. 2020;60(Suppl 1):S41‐S49. doi: 10.1093/geront/gnz178 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mitchell LL, Peterson CM, Rud SR, et al. “It's like a cyber‐security blanket”: the utility of remote activity monitoring in family dementia care. J Appl Gerontol. 2020;39(1):86‐98. doi: 10.1177/0733464818760238 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Zmora R, Mitchell LL, Bustamante G, Finlay J, Nkimbeng M, Gaugler JE. Dementia caregivers' experiences and reactions to remote activity monitoring system alerts. J Gerontol Nurs. 2021;47(1):13‐20. doi: 10.3928/00989134-20201208-03 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Gaugler JE, Zmora R, Mitchell LL, et al. Six‐month effectiveness of remote activity monitoring for persons living with dementia and their family caregivers: an experimental mixed methods study. Gerontologist. 2019;59(1):78‐89. doi: 10.1093/geront/gny078 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Gaugler JE, Zmora R, Mitchell LL, et al. Remote activity monitoring for family caregivers of persons living with dementia: a mixed methods, randomized controlled evaluation. BMC Geriatr. 2021;21:715. doi: 10.1186/s12877-021-02634-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of admission to a nursing home. N Engl J Med. 1997;337(18):1279‐1284. doi: 10.1056/NEJM199710303371806 [DOI] [PubMed] [Google Scholar]
- 8. Newcomer R, Spitalny M, Fox P, Yordi C. Effects of the Medicare Alzheimer's Disease Demonstration on the use of community‐based services. Health Serv Res. 1999;34(3):645‐667. [PMC free article] [PubMed] [Google Scholar]
- 9. Finch M, Griffin K, Pacala JT. Reduced healthcare use and apparent savings with passive home monitoring technology: a pilot study. J Am Geriatr Soc. 2017;65(6):1301‐1305. doi: 10.1111/jgs.14892 [DOI] [PubMed] [Google Scholar]
- 10. Gathercole R, Bradley R, Harper E, et al. Assistive technology and telecare to maintain independent living at home for people with dementia: the ATTILA RCT. Health Technol Assess. 2021;25(19):1‐156. doi: 10.3310/hta25190 [DOI] [PMC free article] [PubMed] [Google Scholar]