This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Falls of individuals with dementia are frequent, dangerous, and costly. Early detection and access to the history of a fall is crucial for efficient care and secondary prevention in cognitively impaired individuals. However, most falls remain unwitnessed events. Furthermore, understanding why and how a fall occurred is a challenge. Video capture and secure transmission of real-world falls thus stands as a promising assistive tool.
The objective of this study was to analyze how continuous video monitoring and review of falls of individuals with dementia can support better quality of care.
A pilot observational study (July-September 2016) was carried out in a Californian memory care facility. Falls were video-captured (24×7), thanks to 43 wall-mounted cameras (deployed in all common areas and in 10 out of 40 private bedrooms of consenting residents and families). Video review was provided to facility staff, thanks to a customized mobile device app. The outcome measures were the count of residents’ falls happening in the video-covered areas, the acceptability of video recording, the analysis of video review, and video replay possibilities for care practice.
Over 3 months, 16 falls were video-captured. A drop in fall rate was observed in the last month of the study. Acceptability was good. Video review enabled screening for the severity of falls and fall-related injuries. Video replay enabled identifying cognitive-behavioral deficiencies and environmental circumstances contributing to the fall. This allowed for secondary prevention in high-risk multi-faller individuals and for updated facility care policies regarding a safer living environment for all residents.
Video monitoring offers high potential to support conventional care in memory care facilities.
A fall is defined as an “unexpected event in which the participant comes to rest on the ground, floor, or lower level” [
Detecting a fall early and in an ongoing manner provides significant potential for reduced morbidity and mortality in patients and system-wide savings [
A significant portion of recent health technology innovation regarding fall management has been driven by industry and has taken place in the commercial space. To date, the most well-known commercial solutions include wearable alert systems [
In this study, the video technology was used to review real-world falls in a single memory care facility, thus avoiding artificiality of simulated or acted falls carried out in a contained laboratory environment, as well as biased information about falls gathered from individuals’ recalling the fall or from administrative hospital record [
A holistic approach of the fall management was used in this paper. The objective of the study was to analyze how continuous video monitoring and video review of falls occurring in common spaces and private rooms of residents living in a memory care facility can support best quality of care.
This study reports on an ancillary study that is part of a larger project called SafelyYou. SafelyYou aims at developing deep learning (a subfield of machine learning) algorithms for automated real-life real-time fall detection in nursing and memory care facilities (http://www.safely-you.com). This pilot observational study was carried out between July and September 2016. Falls were video-captured in residents 24 hours a day, 7 days a week, and the video recordings were provided to the facility staff for video review. The study took place in a memory care facility that is part of the Memory Care Community in California and of the Integral Senior Living network, in which residents reside in a supportive ecosystem. The facility offers 40 individual bedrooms with individual bathrooms and common indoor areas (2 living rooms, 2 eating areas, and kitchens and hallways) where residents are allowed to walk and spend time freely. Residents of this memory care facility have all been diagnosed with dementia (Alzheimer disease and related dementias), had a mean age of 79.4 years (standard deviation [SD] 3.2), and were predominantly female (71.4%) at the time of the study inclusion.
The primary outcome measure is the count of the total number of residents’ falls occurring in the video-covered areas of the facility over the 3-month period of video recording (allowing us to compute a fall rate per month). This count is further compared with the cases of falls that the facility health board independently reported in its daily routine care for each known occurrence of fall (ie, administrative report regardless of the video recording) 2 months before video deployment (baseline occurrence, May-June 2016) and during the 3 months of study (July-September 2016).
The secondary outcome measures qualitatively assess the use of video recording and replay possibilities for care practice. This entails (1) acceptability of video monitoring by residents and facility staff and use of fall review by facility staff to support care practice and quality of care; and (2) the analysis of falls and of fall-related injuries, leveraging video replay to depict intrinsic and extrinsic factors, and environmental circumstances contributing to the falls Acceptability and impact of video review on care practice were assessed through semidirected interviews carried out during bimonthly meetings with the care facility staff over the 3 months of study. An adapted version of the 4-point Hopkins Falls Grading Scale [
A total of 43 wall-mounted cameras were deployed in all common areas and private rooms of consenting residents and families in accordance with the following privacy and ethical guidelines.
Loop equipment, including Internet Protocol (IP) cameras, network attached storage, Wi-Fi, secured storage on the university server, and phone apps.
The videos of fall events that had been depicted by the research team were made available to be viewed by the executive director of the facility who would decide to discuss them with her staff. The meetings between the facility staff and research team were carried out twice a month during the 3 months of the study in a rather flexible way and using semidirected interviews. The main purposes of these meetings were as follows: (1) to be sure that no unanticipated issues or concerns with residents, surrogates, and/or staff arose and (2) to observe the use (or no use) of the videos and what were the changes in care practice that were reported. During these meetings, the research team asked about the use of the videos in a neutral way (ie, observing the potential uptake of the recording without pushing attitude). The main focus of the first meeting concerned the confirmation of the resident-surrogate dyads who had agreed to participate, as well as the questions from the executive director. The final meeting focused on the removal of all the cameras of the facility and discussed the practice changes that the video recording had potentially triggered.
Privacy and consent procedures were developed with support from the institutional review board (IRB) of the University of California, Berkeley (http://cphs.berkeley.edu/), and following guidelines from California Department of Social Services Community Care Licensing Division (CDSS-CCLD). Approval of the study protocol was obtained from the Committee for Protection of Human Subjects of University of California, Berkeley, before starting the study (CPHS protocol number 2015-11-8119). Residents living within the care facility showed severe cognitive impairment related to Alzheimer disease and related dementias. Their capacity to consent to research according to the legal standards of informed consent was altered. As a consequence, surrogate consent was required for this pilot study. The legally authorized representatives of the facility residents were informed at a town hall meeting that a study on fall prevention would occur at the facility and were invited to participate in its presentation with their relative. The legally authorized representatives of the facility residents were given oral and written information about the purpose of the study, procedures, risks, and benefits as listed in the consent form. Those who would like to participate signed the self-certification document to confirm they were the legally authorized representatives and were provided the informed consent document provided by the research team. The study was explained to the affected individuals living in the facility. If affected individuals provided assent, they would be included in the study. If they provided any verbal or nonverbal indication that they do not wish to have the camera in their room or object to any other part of the study, they would not be included. The legally authorized representative was the one who could say yes to the study, thus providing informed consent, but the resident retained the right to say no to the study at any time, thus providing assent. If at any time, individuals expressed verbal or nonverbal indication that they would like the camera removed, personnel would remove the cameras. Participants or legally authorized representatives who originally assented or consented to the study and would later revoke consent would also have cameras removed and video data destroyed.
In private bedrooms, cameras were located high-up in a corner in the bedroom but not in the bathroom and remained visible to the participants. When cameras were not unplugged, they would show a small red light when motion is detected in a room. A sticker was positioned on the participants’ doors as a reminder to the residents, families, and facility staff that participants were being filmed in their private rooms. This physical sign on the door stating that video recording was in progress ensured that everyone entering the room was aware of the camera. Flyers that explained the goals of the research study, the length of the study, the use of wall-mounted cameras, and the generic email address and centralized phone number were positioned in several locations of the facility. The generic study email address and the centralized phone number were provided to respond to any withdrawal wish, expression of interest, or questions. Cameras were also equipped with an explanatory tag that described the goals of the research, the use of wall-mounted cameras, and the possibility to unplug the camera at any time and the way to do so, as well as the name of the principal investigator, the generic study email address, and the centralized phone number to be used in case of concerns. The guidelines from CDSS-CCLD were followed for the study protocol. Whereas the federal law requires that all residents have the right to privacy, the CDSS guidelines for use of the video surveillance state that recording in a common area does not require a waiver because there is no expectation of privacy in common areas (such as eating areas) [
A registered nurse was hired specifically for the study and was available to answer concerns from the participants, the families, and the facility staff, which could emerge before and during the study, including potential withdrawal from the study. If the participant or his/her legally authorized representative expressed willingness to withdraw from the study, they were to inform either the facility staff who would transmit this information to the nurse or the research team by directly using the generic study email address and/or the centralized phone number generated for the study. The possibility of participants’ withdrawal from the study at any point was mentioned at both oral and written levels during information and inclusion sessions. As mentioned on the camera laminated tag, the equipment could also be turned off at any time by simply unplugging it from the wall outlet. If the camera had been unplugged for over 24 hours, the team would figure out whether the participant or surrogate forgot to plug the camera back in or whether he/she would like to have the camera removed for the rest of the study. If a participant or his/her surrogate wished to withdraw the study at any time, all his/her video data would be destroyed. Video segments found improper by the review board were referred to the dementia care nurse of the team in case of content of potential physical or sexual abuse, neglect, sexual activity, or other actions that could imply abuse if taken out of context and other incriminating behaviors. Before deleting data, the dementia care nurse was responsible for determining whether the matter should be taken to facility management or to adult protective services. In accordance with Californian legislation [
A total of 15 out of 38 resident-family dyads (40%) were able to attend the information meeting about the research study, out of which 10 gave oral and written consent and volunteered for the research, and 5 did not wish to participate. Accordingly, the video recording in private rooms included 10 residents, and video recording in common spaces included the total of 38 residents in July and August, followed by 36 residents in September (because of a slight dip in facility occupancy rate).
No impact of the video deployment, recording, and review on the daily routine of the residents and professional caregivers was reported over the 3-month period. At the end of the study period and based on the preliminary results and care experience, the project partner of memory care facilities of Integral Senior Living network agreed to expand the protocol to 14 facilities.
Bimonthly follow-up interviews showed that, in the first 7 weeks of the study, no formal video review was carried out by facility staff despite the fact that video recordings from the previous 72 hours were easily available through secured mobile devices to facility management. Facility management reported hardly ever using the video feeds during this time because of the numerous other challenges faced with operating a memory care facility and the little obvious value granted to the video so far. After 7 weeks, a particularly severe fall incident was recorded during daytime in which the resident was lying on the ground for almost 3 hours without receiving assistance. In accordance with procedures approved by the IRB of the university, this incident was reported to facility management. After reviewing this fall, facility management showed increased interest in reviewing other falls, and the mobile device app provided to review videos proved to be accessible and easy to use to facility staff, who subsequently gained familiarity with it. Further interviews revealed that facility management found video replay useful to grade the severity of the injury and eventually screen patients in the future for external referral to the emergency unit in case of severe injury. In addition, interviews revealed that facility management carried out preventative care interventions, which they believed would address some of the causes of future falls. These preventive actions first included moving furniture and changing room layout based on potential tripping hazards and falls (noticed from videos). Second, changes to care policy that included additional checking on high-risk residents every hour instead of every 2 hours at night were instated following the review of the data.
During the 3-month intervention period, a total of 26 falls were reported in routine conventional care by facility staff for the whole facility (in both video-covered and video-uncovered areas;
In the 2 months before the video deployment, a total of 18 falls were administratively reported (11 in May and 7 in June), providing a prevideo intervention facility baseline fall rate of a mean of 9 falls per month. An expected facility fall rate adjusted for the number of residents of 12.7 and 12 falls per month was reported for comparison purpose in
As summarized in
Fall count display over video-covered and video-uncovered areas.
Fall rate per month displayed over the 3-month study period.
Characteristics of falls in a sample of 16 falls collected in 5 individuals over a 3-month period.
Distribution | Severity | Fall circumstances | |||||||
Subject (S#)aand fall | Location | Timeb | Body impact | Head injury | Severity grading | Activity performed |
Interaction with the living environment as a contributor to the fall: |
Got up alone | |
#1 | CSd | D | 0 | 0 | 2 | Transfer sit-to-stand while talking (W07e) | No extrinsic factor identified |
0 | |
#2 | CS | D | 0 | 0 | 2 | Slipping from chair (W07e) | No extrinsic factor identified | 0 | |
#3 | CS | D | 1 | 0 | 2 | Walking with caregiver (W03f, W04g) | Extrinsic obstacle (other resident in wheelchair in the pathway) |
0 | |
#4 | CS | D | 0 | 0 | 2 | Transfer sit-to-stand (W07e) | No extrinsic factor identified | 0 | |
#5 | CS | N | 0 | 0 | 2 | Transfer sit-to-stand (W07e) | No extrinsic factor identified |
1 | |
#6 | CS | D | 0 | 0 | 2 | Slipping from chair (W07e) | No extrinsic factor identified | 0 | |
#7 | CS | D | 0 | 0 | 3 | Walking (W01h) | No extrinsic factor identified |
0 | |
#8 | CS | D | 0 | 0 | 2 | Transfer sit-to-stand (W07e) | No extrinsic factor identified |
0 | |
#9 | CS | D | 0 | 0 | 2 | Transfer sit-to-stand (W07e) | No extrinsic factor identified |
1 | |
#10 | CS | D | 1 | 0 | 3 | Moving with wheelchair (W05i) | No extrinsic factor identified |
0 | |
#1 | BRj | D | 1 | 0 | 3 | Transfer stand-to-sit while dressing (W06k) | Environmental hazard (messy bed) |
0 | |
#1 | BR | N | 0 | 0 | 3 | Walking/loss of support (W03f, W06k) | Environmental stressor (subject pushed from other resident’s bed) |
0 | |
#2 | BR | D | 0 | 0 | 3 | Transfer stand-to-sit (W08l, W06k) | Environmental hazard (grabbing clothes on the floor) |
0 | |
#3 | BR | D | 1 | 0 | 4 | Transfer sit-to-stand (W06k) | Environmental hazard (slippery bed blanket/messy bed) |
0 | |
#1 | BR | N | 1 | 1 | 3 | Transfer stand-to-sit (W06k) | Environmental hazard (slippery bed sheet/messy bed and poor lighting) |
1 | |
#1 | BR | N | 0 | 0 | 2 | Transfer lay-to-sit (W06k) | Environmental hazard (slippery bed sheet/messy bed) |
0 |
aF indicates female and M indicates male.
bD indicates day and N indicates night.
cInternational Classification of Diseases, Tenth Edition (ICD-10).
dCS: common space.
eW07: fall involving chair.
fW03: other fall on same level due to collision with, or pushing by, another person.
gW04: fall while being carried or supported by other persons.
hW01: fall on same level from slipping, tripping, and stumbling.
iW05: fall involving wheelchair.
jBR: bedroom.
kW06: fall involving bed.
lW08: fall involving other furniture.
A video-witnessed pre-fall activity (subject 4, in his private bedroom). Reproduced with permission of the individual and his family.
A video-witnessed backward fall event (subject 4, in his private bedroom). Reproduced with permission of the individual and his family.
This observational study brings evidence that continuous video monitoring and video review of falls of residents in a memory care facility can support best quality of care. It was found in this pilot study that continuous video monitoring in common spaces and private bedrooms of such care facility and fall review were both feasible and acceptable by facility staff after a certain adoption period. Although these preliminary results need to be confirmed with a larger number of facilities and a larger sample of participants and fall cases in future studies, fall review appears as a valuable health care procedure that might contribute to improved safety in residents and yield better quality of care in facility practice. Fall review provides a unique access to the unpredictable unwitnessed history of a fall, thus supporting screening for the severity of the fall and fall-related injury at the acute phase. Video replay might also allow for secondary prevention in high-risk multi-faller residents with cognitive disorders and, more broadly, for updated facility care policies and preventative actions regarding the living environment of all residents.
Although the fall rate is quite high in long-term care facilities [
The video footage gave access to unrivaled data that were explored from a multidisciplinary perspective, thanks to the combination of the information gathered during the meetings with the facility staff and the analyses of the videos carried out by the researchers. A first finding is that rapid postfall review provides a unique access to the ever-unpredictable “unwitnessed” hidden and silent event of the fall. Access to the natural history of the fall is all the more challenging because individuals suffering from cognitive impairment including memory loss are usually unable to recall the fall [
This study makes it also challenging to analyze the complex multifactorial falling patterns through video in the particular perspective of cognitively impaired older adults. Factors that contribute to the risk of falls in patients have traditionally been classified as intrinsic (individual predisposition), extrinsic (environmental hazard), and situational (related to the activity being done) [
As previously stated, a fall is usually multifactorial and happens as a result of a complex interaction between the individual and his or her living environment [
A video-witnessed post-fall recuperation (subject 4, in his private bedroom). Reproduced with permission of the individual and his family.
A video-witnessed post-fall activity (subject 4, in his private bedroom). Reproduced with permission of the individual and his family.
This study needs to be replicated and results confirmed over a larger sample size of individuals and memory care facilities and over a longer period of time to control for size effect, to measure long-lasting effects, and to allow for meaningful examination of the relation between decrease in fall rate and the proposed intervention. Recommendation for future research include (1) upgrading computational deep-learning algorithms to provide an automated diagnosis (or assumption) of real-time fall, as well as an at-risk screening scale estimating the fall risk in every resident, thanks to an automatized set of video-based biomarkers; (2) measuring time spent lying on the floor (time-to-event between the fall and caregiver intervention); (3) conduction of further studies (if possible randomized) comparing conventional care with real-time utilization of an interactive assistive video diagnostic of falls; (4) proposing a cost-effectiveness analysis of using such technology in memory care facilities; (5) conducting interviews within focus groups using medical anthropology approaches to get a deeper understanding about professional caregivers’ perspective on the video monitoring; (6) increasing knowledge about fall epidemiology and falling patterns regarding cognitive functioning of the individuals in particular (including distinct pathologies such as Alzheimer disease, Parkinson disease, Lewy body disease, and frontotemporal dementia); and (7) deploying and testing the device in other settings such as individual homes.
Falls and fall-related injuries are frequent and potentially preventable causes of morbidity, functional decline, and increased health care use and mortality among individuals suffering from Alzheimer disease and related disorders. The findings of this study highlight the potential of video-monitoring deployment to support fall diagnostic and fall-related injuries and suggest that video review can have a positive impact on quality of care in memory care facilities. Given the growing demand for assisted living in elderly and persons with dementia, video monitoring appears as a promising assistive tool to support health care organizations and possibly complement existing conventional care for both detection and prevention of falls. But more data are needed to validate that the fall rate in managed care facilities can be reduced and safer care provided through interactive video review of falls.
California Department of Social Services Community Care Licensing Division
International Classification of Disease, Tenth Edition
network attached storage
standard deviation
The authors would like to sincerely thank the care community for their support. It was truly inspirational to work with a team so dedicated to improving the quality of care and that takes such pride in the quality of care provided; it is really commendable that the facility management was open to letting a research group record video on their premises. The authors thank Oriana Peltzer, Oumaima Makhlouk, Pierre-Louis Ehret, and Casey Maas for supporting installation. This research would not be possible without them. The authors are very grateful to Dr Bruce Miller and collaborators at the Memory and Aging Center at the University of California, San Francisco, for the rich interaction they provided in the field of functional monitoring and health technology in patients suffering from dementia. This research has been supported by a grant from the Center for Information Technology Research in the Interest of Society (CITRIS Seed Funding Opportunities 2016), a grant from the National Science Foundation (Industrial Innovation and Partnerships grant 2016, STTR PHASE I #1648753), by a research grant from Nokia, and by a grant from the France-Berkeley Fund. The first author thanks the Global Brain Health Institute, the Fulbright-Foundation Monahan program, the Fondation des “Gueules Cassées,” and the Société Française de Médecine Physique et de Réadaptation for supporting her work.
None declared.