Introduction
Worldwide, over 55-million people live with dementia (Livingston et al., Reference Livingston, Huntley, Sommerlad, Ames, Ballard, Banerjee, Brayne, Burns, Cohen-Mansfield, Cooper, Costafreda, Dias, Fox, Gitlin, Howard, Kales, Kivimäki, Larson, Ogunniyi and Mukadam2020), and the number is projected to triple by 2050 (Nichols et al., Reference Nichols, Steinmetz, Vollset, Fukutaki, Chalek, Abd-Allah, Abdoli, Abualhasan, Abu-Gharbieh, Akram, al Hamad, Alahdab, Alanezi, Alipour, Almustanyir, Amu, Ansari, Arabloo, Ashraf and Vos2022). Dementia is a progressive medical condition associated with a decline in cognitive functions, including judgment, behaviour, language, and thinking (Duong et al., Reference Duong, Patel and Chang2017; Emmady et al., Reference Emmady, Tadi and Del Pozo2022). Persons living with dementia are exposed to multiple risks secondary to these cognitive challenges, including falls, injuries, becoming disoriented, and getting lost (Alzheimer Society of Ontario, 2022; Gilmour et al., Reference Gilmour, Gibson and Campbell2003; Lach, Reference Lach2017; Petersen et al., Reference Petersen, Siersma, dePont Christensen, Storsveen, Nielsen and Waldorff2018).
Persons living with dementia are at higher risk of getting lost and going missing than other populations (Neubauer et al., Reference Neubauer, Miguel-Cruz and Liu2021c), as their wayfinding abilities can be diminished due to cognitive decline (Liu et al., Reference Liu, Gauthier and Gauthier1991; Puthusseryppady et al., Reference Puthusseryppady, Coughlan, Patel and Hornberger2019). Indeed, missing incidents can occur during everyday activities (Rowe et al., Reference Rowe, Vandeveer, Greenblum, List, Fernandez, Mixson and Ahn2011), including walking, and driving, even while they are with a care partner (Kowalski, Reference Kowalski2020; MacAndrew et al., Reference MacAndrew, Schnitker, Shepherd and Beattie2018; Rowe et al., Reference Rowe, Greenblum, Boltz and Galvin2012a). According to the Alzheimer’s Association (2022), 6 in 10 persons living with dementia will wander at least once during their disease. Indeed, because of critical wandering, or wandering that results in walking with no orientation to time and place, older adults could go missing and become lost (Neubauer et al., Reference Neubauer, Azad‐Khaneghah, Miguel‐Cruz and Liu2018; Petonito et al., Reference Petonito, Muschert, Carr, Kinney, Robbins and Brown2013). However, the scholarly evidence (Kikuchi et al., Reference Kikuchi, Ijuin, Awata and Suzuki2019; Kowalski, Reference Kowalski2020; Kwok et al., Reference Kwok, Yuen, Ho and Chan2010) suggests that the estimated prevalence of missing incidents due to critical wandering in persons living with dementia is still unclear.
Persons living with dementia can go missing even in familiar places and be exposed to severe outcomes (Kikuchi et al., Reference Kikuchi, Ijuin, Awata and Suzuki2019; Kowalski, Reference Kowalski2020). The most common consequences include hypothermia and drowning (Byard & Langlois, Reference Byard and Langlois2019; Kikuchi et al., Reference Kikuchi, Ijuin, Awata and Suzuki2019). In severe cases, missing incidents can lead to death (Murata et al., Reference Murata, Takegami, Onozuka, Nakaoku, Hagihara and Nishimura2021). Mortality rates related to missing incidents are notable (Murata et al., Reference Murata, Takegami, Onozuka, Nakaoku, Hagihara and Nishimura2021); when disoriented, a person living with dementia could be easily hit by a car or a train (Rowe et al., Reference Rowe, Vandeveer, Greenblum, List, Fernandez, Mixson and Ahn2011; Young et al., Reference Young, Papenkov and Nakashima2018) and depending on the season and terrain (e.g., weather conditions, or getting lost in highways or crowded areas) could experience multiple sources of danger. Persons with dementia could have a reduced ability to process pain and thermoregulate (Fletcher et al., Reference Fletcher, Downey, Golden, Clark, Slattery, Paterson, Rohrer, Schott, Rossor and Warren2015), further increasing their risk of harm. Risks of going missing increase stress on care partners (Lim et al., Reference Lim, Son, Song and Beattie2008; Rolland et al., Reference Rolland, Gillette-Guyonnet, Nourhashémi, Andrieu, Cantet, Payoux, Ousset and Vellas2003; White et al., Reference White, Montgomeiy and McShane2010) and impose a high demand on police, public safety services, and resources (e.g., equipment, helicopters, tracking dogs, paramedics and search, rescue personnel, and search and rescue programs) (Neubauer et al., Reference Neubauer, Philip, Marshall, Daum, Perez, Miguel-Cruz and Liu2021a; Shalev-Greene & Pakes, Reference Shalev Greene and Pakes2014).
Strategies exist to limit the risk of getting lost including locks, alarm systems, and location tracking devices (Adekoya & Guse, Reference Adekoya and Guse2019; Neubauer & Liu, Reference Neubauer and Liu2021b). While these systems can mitigate the risks of getting lost, they are usually adopted after a missing incident occurs (Bantry White & Montgomery, Reference Bantry White and Montgomery2015a; Neubauer & Liu, Reference Neubauer and Liu2021b). Despite an increasing body of knowledge about locator devices and tracking systems for use with persons who have dementia (Neubauer et al., Reference Neubauer, Miguel-Cruz and Liu2021c; Neubauer & Liu, Reference Neubauer and Liu2021a; Rasquin et al., Reference Rasquin, Willems, de Vlieger, Geers and Soede2007; Wojtusiak & Mogharab Nia, Reference Wojtusiak and Mogharab Nia2019; Emrich-Mills et al. Reference Emrich-Mills, Puthusseryppady and Hornberger2021), the usefulness of these devices in preventing missing incidents remains mixed (Neubauer et al., Reference Neubauer, Azad‐Khaneghah, Miguel‐Cruz and Liu2018). The effectiveness of these strategies could be restricted by the limited understanding of the risk factors for getting lost and going missing because of critical wandering among persons living with dementia. Thus, proactive approaches should be informed by a comprehensive understanding of risk factors associated with missing incidents (Neubauer & Liu, Reference Neubauer and Liu2021a) to contribute to predictive models (Barrett et al., Reference Barrett, Bulat, Schultz and Luther2018; Homdee et al., Reference Homdee, Alam, Hayes, Hamid, Park, Wolfe, Goins, Fyffe, Newbold, Smith-Jackson, Bankole, Anderson and Lach2019; Khaertdinov et al., Reference Khaertdinov, Semerci and Asteriadis2021) that can mitigate the risks for missing incidents. A combination of proactive strategies and predictive models would be a preventative approach that enhances the safety of persons living with dementia. This has the potential to reduce stress for care partners and possibly reduce the demand for public services involved in a dementia-related missing incident search and rescue operation.
An understanding of the associated risk factors contributes to the mitigation of adverse outcomes and the development of ways to manage the risks of missing incidents (Bantry White & Montgomery, Reference Bantry White and Montgomery2016; MacAndrew et al., Reference MacAndrew, Schnitker, Shepherd and Beattie2018; Murata et al., Reference Murata, Takegami, Onozuka, Nakaoku, Hagihara and Nishimura2021). A risk factor can be understood as the measurable characteristic of each person living with dementia in a specified population that precedes the outcome of interest (Kraemer et al., Reference Kraemer, Kazdin, Offord, Kessler, Jensen and Kupfer1997). Despite the known harms resulting from missing incidents, few studies have identified the risk factors associated with getting lost and going missing among persons living with dementia (Chung & Lai, Reference Chung and Lai2011; McShane et al., Reference McShane, Gedling, Keene, Fairburn, Jacoby and Hope1998), and to date, no comprehensive review has systematically described the risks factors associated with these incidents. To address this gap, we conducted a scoping review to investigate the risk factors associated with getting lost and going missing in persons living with dementia.
Methods
Design
Our scoping review drew on the approach described by Daudt et al. (Reference Daudt, van Mossel and Scott2013) and Arksey and O’Malley (Reference Arksey and O’Malley2005) to investigate the existing risk factors associated with getting lost and going missing due to critical wandering in persons living with dementia in the literature. This approach was to: (a) determine the research questions and search strategy using the Population, Concept, Context framework (Peters et al., Reference Peters, Godfrey, Khalil, McInerney, Parker and Soares2015); (b) identify relevant studies; (c) screen and select studies; (d) chart the data; (e) summarize and aggregate the data; and (f) report the results. This approach included an interprofessional team in step (b) and used a three-tiered approach to cross-check studies in step (c). The research question that guided this scoping review was: What are the existing risk factors (concept) associated with getting lost and going missing due to critical wandering (context) in persons living with dementia (population). Finally, this study aligns with what is suggested by the PRISMA extension for scoping reviews (PRISMA-ScR) (Tricco et al., Reference Tricco, Lillie, Zarin, O’Brien, Colquhoun, Levac, Moher, Peters, Horsley, Weeks, Hempel, Akl, Chang, McGowan, Stewart, Hartling, Aldcroft, Wilson, Garritty and Straus2018), and followed the suggested checklist accordingly.
Data sources and search strategy
We searched four electronic databases: Medline, CINAHL, EMBASE, and Scopus and examined peer-reviewed literature published from January 1980 to October 2020 to obtain information from the largest amount of available academic literature published in recent years. We utilized Medical Subject Headings, keywords, and combinations of both using operators (e.g., AND and OR) related to missing incidents, disappearing cases, getting lost, and wayfinding difficulties associated with critical wandering in persons living with dementia, such as boundary transgression, elopement behaviour, critical wandering, walkabouts, and actual missing incidents. We validated the search strategy through consultation with a health sciences librarian. The search occurred in October 2020. The search strategy is reported in Supplementary Table A1.
Screening and eligibility
HP exported all studies into reference management software (i.e., EndNote version 20) to remove duplicates. The remaining studies were uploaded to Covidence, a screening and data extraction tool, where further duplicates were removed automatically. Following this, reviewers were calibrated to ensure consistency in applying the inclusion and exclusion criteria (HP, AMC, and CD). Two additional reviewers (SM and EL) were trained to apply the inclusion and exclusion criteria independently before the full-text screening phase. Each study title, and abstract and full study, was reviewed independently by two team members. Then each reviewer voted independently to include or exclude each study (HP, EL, SM, CD and AMC). Reviewers engaged in discussions to resolve conflicts, and a third reviewer (CD and AMC) outside the discordant pair made the final decision to include or exclude the study. We obtained 74 per cent agreement during the full-text screening phase, which is considered a high level according to Garritty et al. (Reference Garritty, Gartlehner, Nussbaumer-Streit, King, Hamel, Kamel, Affengruber and Stevens2021).
Inclusion and exclusion criteria
Inclusion criteria
We included studies that:
-
1. reported risk factors (e.g., behavioural, cultural, environmental) associated with getting lost, going missing due to critical wandering, or resulting in a hazardous situation for persons living with dementia;
-
2. included persons living with dementia, Alzheimer’s disease, or cognitive impairment;
-
3. published in any language;
-
4. published from 1980 and onwards using any research design or method (e.g., quantitative, qualitative, mixed methods, literature reviews, and meta-analyses) regardless of the results.
Exclusion criteria
We excluded studies that:
-
1. did not involve primary data collection, secondary analysis, or a systematic way of collecting or synthesizing primary data (e.g., opinion papers, magazine articles, short papers, abstracts, and collections of opinions);
-
2. did not include persons living with dementia, Alzheimer’s disease, or cognitive impairment as the study population;
-
3. did not provide enough information for categorization or data extraction;
-
4. were not available in full text;
-
5. were out of the scope of this review (e.g., pharmacological intervention and genetics).
Data extraction
Three members of the research team (HP and EL) completed data extraction under the supervision of two senior authors (AMC and CD). We extracted data from studies in a spreadsheet where we operationalized the variables. We reviewed each study and extracted data according to the scope and objective of the scoping review (HP, EL, SM, AC, CD, and AMC). The entire team met regularly to discuss and resolve disagreements or uncertainties about the extracted information. Each study was reviewed independently to identify what factors were described, explained, associated, or related to critical wandering in persons living with dementia. During several rounds of discussion, the authors agreed on the risk factors that were identified. For each study, we extracted information on the following areas:
-
1. bibliometric information (e.g., year, country, type of document, and journal information);
-
2. participants’ information (e.g., type of participants involved, sample size, sex and age, and study population);
-
3. study population (e.g., objective, study design, design type, outcome variable(s), and data collection instruments);
-
4. definitions for critical wandering or getting lost in persons living with dementia;
-
5. description and operationalization of risk factors associated with critical wandering in persons living with dementia.
Data analysis and synthesis
Three research team members (AC, EL, and HPH) completed the data analysis. Before this began, we coded categorical variables numerically and categorized studies based on the primary impairment (e.g., dementia, cognitive impairment, and Alzheimer’s disease). To address the objective of this review, we searched and extracted risk factors associated with missing incidents of persons living with dementia, either reported lost or at risk of getting lost due to critical wandering (Algase, Reference Algase2006; Petonito et al., Reference Petonito, Muschert, Carr, Kinney, Robbins and Brown2013). In order to account for the inconsistent terminologies across the literature, which all address the same concept, we also included terms relevant to boundary transgression, elopement behaviour, critical wandering, walkabouts, and actual missing incidents. We used descriptive statistics to summarize and analyse bibliometric and sample details, study design, risk factors, and variables that were associated with a lost or missing incident. Due to the complexity of the included risk factors, we engaged in discussions with experts (i.e., nurses, occupational therapists, and engineers) to facilitate a synthesis of the identified risk factors and variables. Specifically, health-related professionals discussed how these factors are observed and related to persons living with dementia. In addition, they provided feedback during synthesizing risk factors and constructing definitions. For this paper, we defined risk factors are ‘clusters’ of variables, acting as a group of measurable variables or characteristics that increase or reduce chances of getting lost or going missing for a person living with dementia.
Operational definitions for the risk factors
According to the literature, it is a challenge to find consistent definitions of risk factors about missing incidents involving the general population, including persons living with dementia (Algase et al., Reference Algase, Moore, Vandeweerd and Gavin-Dreschnack2007; Ferguson, Reference Ferguson2022; Rowe et al., Reference Rowe, Houston, Molinari, Bulat, Bowen, Spring, Mutolo and McKenzie2015). Thus, using operational definitions is an important step to classify and describe risk factors. To define the risk factors and variables, we followed the best practice suggested by Beecher et al. (Reference Beecher, Devane, White, Greene and Dowling2019), including concept analysis, clarification, and exploration. First, we examined available definitions for concept analysis. Second, we clarified these definitions by integrating evidence extracted from the selected studies. Finally, we explored the definitions after several iterations with an interdisciplinary team and generated operational definitions for each risk factor. This approach was consistent with similar studies (Hummer et al., Reference Hummer, Hatch and Davison2020; Jogerst et al., Reference Jogerst, Daly, Brinig, Dawson, Schmuch and Ingram2011) that attempted to develop concepts where a topic was underexplored. We endeavoured to provide background and to inform the results of this scoping review using examples in the definitions of the risk factors. These definitions are intended to be illustrative and not exhaustive of each risk factor extracted in this review. The review included a range of types of studies. In Table 1, we provide characteristics of selected studies.
Table 1. Characteristics of selected studies (n = 73)
Potential for bias assessment
For this scoping review, we made a substantial effort to reduce the potential for bias in the study selection process and during the data extraction and analysis phases (Liberati et al., Reference Liberati, Altman, Tetzlaff, Mulrow, Gotzsche, Ioannidis, Clarke, Devereaux, Kleijnen and Moher2009). First, we searched different academic databases. Second, we included studies regardless of positive or negative results. Third, we included studies in various languages, not only English. Fourth, we included various publications, such as master’s or Ph.D. theses and journal articles. Fifth, we addressed the raters’ biases by having multiple rounds of screening and selection. Each study was screened twice independently, and we used a third rater to resolve disagreements. Our approach was consistent with methods for reducing the potential for bias reported in the literature (Miguel Cruz et al., Reference Miguel Cruz, Daum, Comeau, Salamanca, McLennan, Neubauer and Liu2023; Page et al., Reference Page, McKenzie and Higgins2018; Rios Rincon et al., Reference Rios Rincon, Miguel Cruz, Daum, Neubauer, Comeau and Liu2021).
Results
Bibliometric analysis
Figure 1 shows the scoping review process. The initial search identified 3,376 studies, and after deduplication, 1,638 (48.51%, 1,638/3,376) studies were screened by title and abstract and 1,365 (40.43%, 1,365/3,376) were excluded. We could not retrieve 21 studies (0.62%, 21/3,376) because we did not have full-text access, even after attempting to obtain interlibrary loans through two universities. Then 252 (7.76%, 252/3,376) studies were assessed in full, and 179 (71%, 179/252) studies were excluded based on the inclusion and exclusion criteria. Finally, 73 (2.16%, 73/3,376) studies were included in the data analysis and synthesis phase and reported. The list of selected studies, including study titles, is reported in Supplementary Table A2.
Figure 1. Scoping review process
Studies were published by 93 authors from 17 different countries, most after 1997 (91.78%, 67/73). The majority were from the United States (46.24%, 43/93), Australia (10.75%, 10/93), Great Britain (9.68%, 9/93), and Korea (6.45%, 6/93). Most of the selected studies were published as journal articles (93.15%, 68/73), and were published as research papers (71.23%, 52/73), literature reviews (21.92%, 16/73), and theoretical models and seminal papers (6.85%, 5/73).
For this scoping review, we extracted 39 studies published in quartile 1 (54.93%, 39/73), 17 studies published in quartile 2 (23.94%, 17/73), 7 studies published in quartile 3 (9.86%, 7/73), and 4 studies published in quartile 4 (5.63%, 4/73) of the Scimago Journal and Country Rank (SCImago, n.d.), which is a ranking that allows researchers to measure the scientific influence of the scholarly published literature. For six studies (8.45%), the quartile score was not reported or impossible to estimate. Additionally, we used Sackett’s (Reference Sackett2000) approach to classifying the level of evidence. Of our studies, only seven were systematic reviews (9.5%, 7/73), two were case studies (2.73%, 2/73), one was single case study (1.36%, 1/73), and one was case–control study (1.36%, 1/73).
Study features and settings
Table 2 shows a detailed description of the specific research design and methods of the selected studies. The majority were quantitative (63.01%, 46/73), followed by literature reviews (15.07%, 11/73), mixed methods studies (9.59%, 7/73), and qualitative studies (4.11%, 3/73). Six studies (8.21%, 6/73) did not report methods, or the method used could not be determined.
Table 2. Summary of risk factors and variables reported in the literature
In terms of population of the selected studies, 27 studies (36.99%, 27/73) reported persons living with dementia as the primary participant type, 19 studies (26.03%, 19/73) reported persons living with dementia and care partners (e.g., carers, care partners, and nursing staff) as participants, a small group of the studies reported care partners and nursing staff as participants (5.48%, 4/73), and other types of participants (i.e., dementia researchers, health or social care workers, and residents of long-term care facilities in general) (4.11%, 3/73). Twenty studies (27.40%, 20/73) did not report the types of participants. In addition, most of the selected studies reported dementia (46.58%, 34/73) as a primary medical condition of the participants, followed by a combination of dementia and Alzheimer’s disease (24.66%, 18/73) and Alzheimer’s disease exclusively (13.70%, 10/73). The sample details of the selected studies classified by the primary medical condition of participants are provided in Supplementary Table A3.
Risk factors for getting lost and going missing due to critical wandering in persons living with dementia
Naturally, different studies reported the same risk factors using multiple names and concepts. HP and EL worked with CD and AMC to review the definitions or context provided in each study to cluster the risk factor into a common name. For instance, age was reported as age, older age, or older than 65 years old. Further, many studies reported more than one variable, which we extracted and counted separately. We identified 27 variables that were grouped into three risk factor domains. The domains were: (a) demographics and personal characteristics, which included the core characteristics of persons living with dementia, such as age, sex, race, and education, also called sociodemographic data (Jelastopulu et al., Reference Jelastopulu, Giourou, Argyropoulos, Kariori, Moratis, Mestousi and Kyriopoulos2014) and personal characteristics (i.e., individual attributes of persons living with dementia, including changes in moods and traits) (National Institute on Aging, 2022), individual history of locomotion and wandering, and one’s ability to perform activities of daily living; (b) health conditions and symptoms, which include physical or mental features of a disease condition (National Institute of Aging, 2022), such as cognitive impairment, neurologic and circadian changes, memory impairment and executive functioning and attention issues, and health conditions, that is, the health status of a persons living with dementia, including the physical, mental, and psychosocial health, type of dementia, and other mental and physical medical conditions, precipitating factors and medication side effects; and (c) environmental and contextual antecedents that relate to the immediate surroundings or situation of persons living with dementia, such as the physical environment, unmet needs, situational environment, concentration of services and resources, living situation, population density, and accessibility to support.
The most common risk factors reported in the selected studies were (a) cognitive impairment (11.85%, 41/73) (e.g., cognitive decline, deficits and functioning, decreased cognitive ability, and lower scores on the Mini-Mental State Exam); (b) physical environment (8.38%, 29/73) (e.g., ambiance, monotonous architecture, environment modifications and stimuli, stressful lights and noises, security systems, crowdedness, architectural design, and person–environment interactions); (c) responsive behaviours (7.23%, 25/73) (e.g., abusive behaviour, aggression and agitation, and behavioural disturbances); (d) personal characteristics (6.94%, 24/73), including premorbid personality and extraverted personality, agreeableness, conscientiousness, emotions, need for security, and negative emotional states; (e) unmet needs (5.20%, 24/73) (e.g., expressed interest to go home, hunger, lack of exercise or meaningful activities, looking for someone or something familiar, need for security, need to use the toilet, pain, and physical or emotional needs); and (f) advancing age (5.20%, 24/73). A summary of the risk factors and variables reported in the literature is presented in Table 3.
Table 3. Definitions of risk factors and variables
Discussion
Summary of evidence
This scoping review identified the risk factors associated with getting lost or going missing due to critical wandering in persons living with dementia. We included a total of 73 studies that reported risk factors. Overall, we identified 27 variables that were grouped into three risk factors domains: (a) demographics and personal characteristics; (b) health conditions and symptoms; and (c) environmental and contextual antecedents. In this scoping review, cognitive impairment was the most frequent variable reported in the selected studies related to missing incidents due to critical wandering in persons with dementia. Cognitive impairment was observed and linked to other variables in previous studies (Hong & Song, Reference Hong and Song2009; Kwok et al., Reference Kwok, Yuen, Ho and Chan2010; Marquardt, Reference Marquardt2011; Rowe et al., Reference Rowe, Houston, Molinari, Bulat, Bowen, Spring, Mutolo and McKenzie2015; Song & Algase, Reference Song and Algase2008).
The second risk factor most mentioned in the selected studies was the physical environment (Ferguson, Reference Ferguson2022; Taylor et al., Reference Taylor, Bradbury-Jones, Hunter, Sanford, Rahilly and Ibrahim2014). Existing evidence suggests that modifications to the surrounding physical environment such as light intensity, variations in temperature, increase or decrease of noise, humidity levels, and the appearance of visual stimuli can have an impact on the behavioural and psychological symptoms of dementia and the spatial behaviour of persons living with dementia (Algase et al., Reference Algase, Beattie, Antonakos, Beel-Bates and Yao2010; Bautrant et al., Reference Bautrant, Grino, Peloso, Schiettecatte, Planelles, Oliver and Franqui2019; Carlson et al., Reference Carlson, Fleming, Smith and Evans1995; Caspi, Reference Caspi2014; Hodgkinson et al., Reference Hodgkinson, Koch, Nay and Lewis2007; Mazzei et al., Reference Mazzei, Gillan and Cloutier2014; Yao, Reference Yao2004). For example, when a physical environment is modified, even with the intention of making it safer, the space can become unfamiliar to persons living with dementia and pose a risk for disorientation (Chaudhury et al., Reference Chaudhury, Hung, Rust and Wu2017; Marquardt, Reference Marquardt2011). This influence, in turn, can trigger a missing incident (Ferguson, Reference Ferguson2022; Lai & Arthur, Reference Lai and Arthur2003; Puthusseryppady et al., Reference Puthusseryppady, Manley, Lowry, Patel and Hornberger2020; Rowe, Reference Rowe2003). In this case, the physical environment is something external to an individual living with dementia, and the alteration of this environment might be out of an individual’s control.
The third most common variable was responsive behaviours such as abusive behaviour, aggression and agitation, angering situations, and hyperactivity (Algase et al., Reference Algase, Son, Beattie, Song, Leitsch and Yao2004; Bowen et al., Reference Bowen, McKenzie, Steis and Rowe2011; Chung & Lai, Reference Chung and Lai2011; Dawson & Reid, Reference Dawson and Reid1987; Detweiler et al., Reference Detweiler, Murphy, Myers and Kim2008; Dewing, Reference Dewing2005; Goldsmith et al., Reference Goldsmith, Hoeffer and Rader1995; Kiely et al., Reference Kiely, Morris and Algase2000; Lester et al., Reference Lester, Garite and Kohen2012; Rowe et al., Reference Rowe, Greenblum, Boltz and Galvin2012; Rowe & Glover, Reference Rowe and Glover2001; Volicer et al., Reference Volicer, van der Steen and Frijters2013). Responsive behaviours are especially relevant to missing incidents as persons living with dementia in advanced stages have been reported to try to leave the place they occupy in response to agitation or anxiety-provoking events (Dawson & Reid, Reference Dawson and Reid1987; Goldsmith et al., Reference Goldsmith, Hoeffer and Rader1995; Rowe et al., Reference Rowe, Greenblum, Boltz and Galvin2012), or due to their place of residence no longer being familiar to them (Carlson et al., Reference Carlson, Fleming, Smith and Evans1995). Authors have described the importance of addressing responsive behaviours in persons living with dementia. For example, Wilkinson et al. (Reference Wilkinson, Kanik, O’Neill, Charoenkitkarn and Chignell2017) described using nonpharmacological strategies, such as music therapy, to manage responsive behaviours. Responsive behaviours may act as a moderator and proxy for other variables, such as medication side effects, locomotion, and physical environment (Holt et al., Reference Holt, Hoben, Weeks and Estabrooks2021; Wilkinson et al., Reference Wilkinson, Kanik, O’Neill, Charoenkitkarn and Chignell2017). The literature suggests that responsive behaviours are a response to something else, for example, unmet needs, physical environment, or emotional, intellectual, or social stimuli (Algase et al., Reference Algase, Son, Beattie, Song, Leitsch and Yao2004; Detweiler et al., Reference Detweiler, Murphy, Myers and Kim2008). Responsive behaviours should be addressed in a supportive and positive manner by speaking calmly or ensuring that the needs of persons living with dementia are understood and managed (Hartung et al., Reference Hartung, Freeman, Grosbein, Santiago, Gardner and Akuamoah-Boateng2020). It stands to reason that the efficient management of this variable could reduce the risk of getting lost in persons living with dementia (Neubauer, Reference Neubauer and Liu2020).
The fourth most common variable was personal characteristics and history. This variable includes personality types, emotions, and moods such as openness, agreeableness, extraversion, negativity, neuroticism, outgoing personality, passivity (Kiely et al., Reference Kiely, Morris and Algase2000; Lee, Reference Lee2011; MacAndrew et al., Reference MacAndrew, Schnitker, Shepherd and Beattie2018; Sutin et al., Reference Sutin, Stephan, Luchetti and Terracciano2018; Thomas et al., Reference Thomas, Glogoski and Johnson2006), and methods of coping with stress (Algase, Reference Algase2008; Gu, Reference Gu2015; Hodgkinson et al., Reference Hodgkinson, Koch, Nay and Lewis2007; Klein et al., Reference Klein, Steinberg, Galik, Steele, Sheppard, Warren, Rosenblatt and Lyketsos1999; Sheehan et al., Reference Sheehan, Burton and Mitchell2006; Thomas et al., Reference Thomas, Glogoski and Johnson2006). For example, life events include changes in routines or relocation to another residence (Beattie et al., Reference Beattie, Song and LaGore2005; Dewing, Reference Dewing2005; Hong & Song, Reference Hong and Song2009; Jeong et al., Reference Jeong, Song and Park2016; Lester et al., Reference Lester, Garite and Kohen2012). In fact, in a study conducted by Sutin et al. (Reference Sutin, Stephan, Luchetti and Terracciano2018), participants who scored higher in neuroticism were at greater risk of experiencing behavioural and psychological symptoms of dementia, including getting lost in familiar places and wandering. Also, agreeableness was associated with less risk of getting lost in familiar, thus explaining how personalities may influence the risk of getting lost for persons living with dementia.
The fifth most common variable was unmet needs. These include expressed interest in going home, engaging in past activities, hunger, lack of exercise and meaningful activities, looking for someone familiar, pain and physical discomfort, physical and biological needs, and need for security (Algase et al., Reference Algase, Son, Beattie, Song, Leitsch and Yao2004; Carlson et al., Reference Carlson, Fleming, Smith and Evans1995; Chung & Lai, Reference Chung and Lai2011; Cipriani et al., Reference Cipriani, Lucetti, Nuti and Danti2014; Dewing, Reference Dewing2005; Goldsmith et al., Reference Goldsmith, Hoeffer and Rader1995; Gu, Reference Gu2015; Klein et al., Reference Klein, Steinberg, Galik, Steele, Sheppard, Warren, Rosenblatt and Lyketsos1999; Lai & Arthur, Reference Lai and Arthur2003; Lester et al., Reference Lester, Garite and Kohen2012; Lucero, Reference Lucero2002; MacAndrew et al., Reference MacAndrew, Fielding, Kolanowski, O’Reilly and Beattie2017b; Rowe et al., Reference Rowe, Greenblum, Boltz and Galvin2012; Thomas et al., Reference Thomas, Glogoski and Johnson2006; Volicer et al., Reference Volicer, van der Steen and Frijters2013). Unmet needs could trigger missing incidents, as persons living with dementia may be inclined to take action to meet their needs without weighing the risk, for example, going out without notifying care partners. In circumstances where certain needs are unmet, such as urination or bowel movement and hunger, they find ways to meet these needs, which can lead them to leave their home (Algase et al., Reference Algase, Son, Beattie, Song, Leitsch and Yao2004; Carlson et al., Reference Carlson, Fleming, Smith and Evans1995; MacAndrew et al., Reference MacAndrew, Fielding, Kolanowski, O’Reilly and Beattie2017b). Based on the identified risk factor and available literature, there is evidence to suggest that many of these factors, such as responsive behaviours, personal characteristics and history, and unmet needs paired with problems related to cognitive impairment, could lead to a missing incident among persons living with dementia. However, the specific relationships between these risk factor categories are poorly understood.
Although we identified three risk factor domains, we acknowledge that the risk factors can be presented in alternate ways, such as internal or personal, external, and fixed and variable risk factors (Kraemer et al., Reference Kraemer, Kazdin, Offord, Kessler, Jensen and Kupfer1997). The literature suggests that some risk factors and variables overlap, moderate the effect of other risk factors, mediate or intervene in the outcome (i.e., getting lost), and act as a proxy for other risk factors (Ferguson, Reference Ferguson2022). This is because a factor comprises a ‘cluster of variables that are correlated among themselves’ (Portney & Watkins, Reference Portney and Watkins2008). Indeed, we observed how some variables included in the risk factors might overlap with other factors, such as cognitive impairment, which can moderate or mediate spatial navigation (Vlček & Laczó, Reference Vlček and Laczó2014). Thus, risk factors domains and variables should not be analysed in isolation (Ferguson, Reference Ferguson2022; Kraemer et al., Reference Kraemer, Stice, Kazdin, Offord and Kupfer2001), and should focus on examining how risk factors may overlap proxy risk factors or how risk factors and variables act as moderators on the effect of other risk factors and mediate another risk factors as related to getting lost in persons living with dementia. This scoping review confirms that risk factors might be interrelated and may require additional exploration in practice (Ferguson, Reference Ferguson2022). This has implications for how we should collect data to prevent future missing incidents because the information collected on missing incidents involving persons living with dementia is limited and, in some cases, non-existent (Ferguson & Huey, Reference Ferguson and Huey2020; Neubauer et al., Reference Neubauer, Philip, Marshall, Daum, Perez, Miguel-Cruz and Liu2021a).
Neubauer and Liu (Reference Neubauer and Liu2021b), Rowe et al. (Reference Rowe, Houston, Molinari, Bulat, Bowen, Spring, Mutolo and McKenzie2015) and Yevchak et al. (Reference Yevchak, Steis and Evans2012) have proposed models that describe risk factors associated with getting lost in persons living with dementia. However, these models have yet to be validated. Grant et al. (Reference Grant, Collins and Nashef2018) noted that validation studies that determine the discriminative and face validity, calibration, and clinical effectiveness are essential to determine the risk prediction model’s usefulness. In addition, all three models capture a portion but not all existing risk factors. Neubauer and Liu (Reference Neubauer and Liu2021b) included the culture and geography of the individual as risk factors, whereas Rowe et al. (Reference Rowe, Houston, Molinari, Bulat, Bowen, Spring, Mutolo and McKenzie2015) captured contextual, situational, and neurocognitive antecedents, and Yevchak et al., (Reference Yevchak, Steis and Evans2012) identified antecedent and precipitating factors. Thus, the description and analysis of risk factors associated with missing incidents in persons living with dementia due to critical wandering have not been comprehensive.
In addition to a need for a comprehensive and validated risk model, there is a need to investigate and understand how specific risk factors evolve over time with the progression of dementia. For example, specific demographics can be considered an initial risk factor and medical conditions later in time (Ferguson, Reference Ferguson2022; Kraemer et al., Reference Kraemer, Kazdin, Offord, Kessler, Jensen and Kupfer1997). This would enable appropriate and customized interventions to mitigate the risks of getting lost and going missing based on the individual circumstances of a person living with dementia. Consequently, the inclusion of more specific risk factors or the exploration of combinations of risk factors that determine the risk would greatly enhance the practical relevance and applicability of our findings.
In this scoping review, we attempted to provide operational definitions for the risk factors, acknowledging that the lack of definitions is a common challenge related to risk factors associated with missing incidents of persons living with dementia (Algase et al., Reference Algase, Moore, Vandeweerd and Gavin-Dreschnack2007; Rowe et al., Reference Rowe, Houston, Molinari, Bulat, Bowen, Spring, Mutolo and McKenzie2015), which in turn is also an issue for the analysing of risk factors (Ferguson, Reference Ferguson2022). Thus, defining variables and risk factors by proposing operational definitions is an essential step in analysing and describing risk factors. These definitions provide a solid foundation for additional research into each risk factor and for developing predictive models.
Very few of the 73 included papers systematically validated the proposed risk factors. Additionally, the level of scientific evidence in the included studies was low (Straus et al., Reference Straus, Glasziou, Richardson and Haynes2018). However, while the evidence may be low, given the nature of a scoping review to summarize all existing literature, this is not prohibitive to our results. Most selected studies included data from police reports and newspapers; their credibility is as good as the quality of the data (Güss et al., Reference Güss, Tuason and Devine2020; Miguel-Cruz et al., Reference Miguel-Cruz, Marshall, Daum, Perez, Hirdes and Liu2022; O’Connor et al., Reference O’Connor, Ng, Hill and Frederick2021). Interestingly, few studies included the perspectives of persons living with dementia and care partners. We consider this as an opportunity for future studies.
In summary, through this scoping review, we confirm that there is a need to increase the level of evidence that identifies clear outcomes pointing to risk factors in this population and recommend this to be addressed by future research. The lack of studies backed with evidence raises questions about the degree of credibility of each associated risk factor to be used in predictive models to estimate and mitigate the risk of someone living with dementia getting lost and going missing. An increased focus on high-quality validation studies would enable researchers to develop and combine predictive risk models with available, proactive approaches to enhance the autonomy and safety of persons living with dementia. High-quality validation studies could also reduce stress for care partners and minimize the high demand for public services, especially those related to search and rescue processes.
Future research
We support the literature that recommends the creation of models that incorporate associated risk factors for getting lost and going missing due to critical wandering in persons living with dementia. For example, Neubauer and Liu (Reference Neubauer and Liu2021b) recommend building a mathematical predictive model to quantify risk factors that would enable those living with dementia and family care partners to understand which risk factors they should focus on when choosing a strategy to manage missing incidents. In addition, Rowe et al. (Reference Rowe, Houston, Molinari, Bulat, Bowen, Spring, Mutolo and McKenzie2015) suggest that understanding how different types of personal antecedents interact in their model can promote more accurate preventative and response strategies. These authors recommended that future research confirm the antecedents and how they can lead to missing incidents. Finally, Yevchak et al. (Reference Yevchak, Steis and Evans2012) suggest that building a model that can weigh and quantify different sundown behaviours into the risk of going missing would be beneficial in preventing the risk of going missing among persons with cognitive impairments.
Future work may include developing assessment tools, validating risk factors, and, finally, developing predictive models to individually assess the risk of getting lost and going missing in persons living with dementia due to critical wandering, considering the risk factors identified in this review. As well, future research should delve deeper into identifying and elucidating these specific risk factor combinations that determine the risk of individuals with dementia going missing. By examining and highlighting these interdependencies, we can provide a more meaningful and practical understanding of the factors contributing to missing incidents. Additionally, more research is needed to describe related risk factors to inform preventive strategies for persons living with dementia, care partners, and health care professionals. Also important, we should consider the feasibility of data collection and propose alternative approaches, such as collaborations with police or first responders’ organizations and health care institutions, to gather more comprehensive and individualized data. Lastly, future research should aim to validate the observed risk factors with persons living with dementia, care partners, and health care professionals and develop prospective studies to describe further and understand the lost person’s behaviour, exploring multiple sources of information, including police records, or cross-referencing vulnerable person registries with police and search and rescue records. This analysis could inform best practices for prevention, risk assessment, and risk mitigation.
Limitations
After thoroughly searching the available literature, we extracted and described risk factors, categorized into three risk factors domains associated with missing incidents due to critical wandering in persons living with dementia. Despite our rigorous search strategy, we may have missed relevant studies because of the inconsistency in how authors utilized terms, such as ‘getting lost’ and ‘going missing’, and how the risk factors were understood and reported. We recognize that a duplication of studies included could happen when we include primary data articles and secondary sources, but this was not the case for this study. Our inclusion and exclusion criteria, as well as selection and extraction process, were designed to ensure a comprehensive and inclusive review of the literature, having in mind integrating as much relevant and informative sources as possible, without compromising the clarity and focus of the scoping review. A limitation of this study was that we did not assess the risk of bias due to the heterogeneous nature of the study designs included. Lastly, our scoping review could not evaluate the identified risk factors due to the lack of experimental evidence.
Conclusions
Despite the vast literature on missing persons (Greene & Alys, Reference Greene and Alys2016; C. Taylor et al., Reference Taylor, Woolnough and Dickens2019), few studies were identified in our review that explored the risk factors associated with missing incidents due to critical wandering among persons living with dementia (Ali et al., Reference Ali, Luther, Volicer, Algase, Beattie, Brown, Molinari, Moore and Joseph2016; Bantry White & Montgomery, Reference Bantry White and Montgomery2015b; Barnard-Brak et al., Reference Barnard-Brak, Richman and Owen2018). Of those that were included, three risk factor domains emerged: (a) demographics and personal characteristics, (b) health conditions and symptoms, and (c) environmental and contextual antecedents. This scoping review identified that the existing literature also contains weak empirical evidence about the risk factors for getting lost and going missing in persons living with dementia.
Persons living with dementia are at risk of getting lost and going missing in their communities. These incidents are a threat to their safety. Many adverse outcomes associated with persons living with dementia who go missing and become lost have been reported. As the number of persons living with dementia and experiencing cognitive impairment continues to grow (Hallam et al., Reference Hallam, Petersen, Cooper, Avgerinou and Walters2022; Nichols et al., Reference Nichols, Steinmetz, Vollset, Fukutaki, Chalek, Abd-Allah, Abdoli, Abualhasan, Abu-Gharbieh, Akram, al Hamad, Alahdab, Alanezi, Alipour, Almustanyir, Amu, Ansari, Arabloo, Ashraf and Vos2022), an understanding of the risk factors related to missing incidents is necessary to mitigate negative outcomes. This can also inform predictive models of risk, tools, and strategies to support the decision-making processes of persons living with dementia and care partners can support their quality of life and safety.
Acknowledgements
We thank Emily Rutledge and Adebusola Adekoya for their support in discussing the definitions presented in this manuscript. Librarian Jackie Stapleton assisted us with validating the search strategy.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0714980823000776.
Financial support
This research was supported by AGE-WELL NCE (Grant No. AWCRP-08) and MITACS (Grant No. IT18937).
Competing interests
The authors declare no competing interests.
Open access
