Introduction
Over 55 million people live with dementia globally and numbers are increasing as the population ages (WHO, 2021; Prince et al., Reference Prince, Wimo, Guerchet, Ali, Wu and Prina2015). Around two-thirds (61%) of people living with dementia (PLwD) are in the community, of which a third live alone (Prince et al., Reference Prince2014). UK policy priority is to find the best ways to help PLwD live well at home for as long as possible (DH, 2016). PLwD have complex needs, including comorbid physical illness. They are supported mainly by their families but also use health and social care services (Wittenberg et al., Reference Wittenberg2019).
Practice and policy guidance require health and social care professionals to identify, manage, and reduce risk in dementia to enable PLwD to remain at home for longer (DH, 2009, 2016). Part of the initial clinical assessment is to identify, evaluate, and manage risk for this purpose. However, anxiety about the consequences of risk-related decisions is a major concern for practitioners (Ruston and Backhouse, Reference Ruston and Backhouse2022).
As cognition deteriorates, PLwD often become more vulnerable through self-neglect, accidents, and a greater risk of abuse and exploitation (Manthorpe and Moriarty, Reference Manthorpe and Moriarty2010). Positive risk management, also known as risk enablement, balances the advantages from taking risks against the harms done by trying to avoid risk that can lead to excessive restrictions. It can result in beneficial outcomes such as increased confidence, self-satisfaction, and preserved independence (Manthorpe and Moriarty, Reference Manthorpe and Moriarty2010; Morgan et al., Reference Morgan, Stewart, D’arcy and Werezak2004; Morgan and Williamson, Reference Morgan and Williamson2014; DH, 2016). A recent systematic review found interventions used to mitigate risk in dementia homecare had both favorable and unfavorable outcome in that they reduced risk, improved safety, and increased well-being but also created loss of autonomy and additional risks for clients and homecare workers (Backhouse et al., Reference Backhouse, Ruston, Killett, Ward, Rose-Hunt and Mioshi2022).
It is essential to balance independence and risks when planning care (Mitchell and Glendinning, Reference Mitchell and Glendinning2008; Taylor, Reference Taylor2006), and PLwD want to and should be involved in that decision-making process (Pel-Littel et al., Reference Pel-Littel2021; Wied et al., Reference Wied, Knebel, Tesky and Haberstroh2019). Effective risk identification and assessment can inform the most appropriate way to manage risk and promote risk enablement that involves balancing the rights of PLwD to retain as much control as possible, with those of society and families’ wishes to reduce risks and promote safety at home. Furthermore, differences often exist in the perceptions and attitudes toward risk held by PLwD to those held by their carers, and those of health and social care professionals (Daly et al., Reference Daly, Bunn and Goodman2018; Pel-Littel et al., Reference Pel-Littel2021). So, to ensure effective person-centered risk assessment it is important to understand how risk is conceptualized and viewed by PLwD, their family carers, and the healthcare professionals (HCP) involved (Cott and Tierney, Reference Cott and Tierney2013). Comprehending how risk is assessed, recognized, negotiated, managed, and enabled is a key part of changing policy and practice to implement personalization within care assessment, and enable health and social care professionals to develop appropriate interventions to promote safety, delay care home admissions, and decrease carer burden (Carr, Reference Carr2011; Lea et al., Reference Lea2014). As to our knowledge, there is no systematic review on this subject, we aimed to systematically:
-
1. Explore key components of risk assessment for PLwD in the community and within care homes.
-
2. Report on the attitudes to risk identification and risk assessment of PLwD, their family carers, and health and social care professionals.
-
3. Appraise existing risk assessment tools.
Methods
This systematic review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standardized reporting guidelines (Moher et al., Reference Moher2015), adheres to the PRISMA 2020 statement (Page et al., Reference Page2021), and is registered with PROSPERO (CRD42020187519).
URL: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID = 187519
Search strategy
Search terms were informed by discussions with clinical and academic dementia experts and librarians with expertise in database searching and literature review. The search terms were adapted for use in bibliographic databases with database-specific filters when available (e.g. Boolean operators). Subsequently, we combined MeSH terms and free-text words (keywords). We used search terms including: “risk assessment,” “risk assessment tool,” “risk identification” and terms were connected using AND with search terms for “dementia.” Figure S1, supplementary information, details the search strategy used in EBSCOhost and OVID interfaces.
Information sources
We searched five electronic databases from inception on 10th June 2020 and again on 23rd April 2022: MEDLINE, CINAHL, PsycInfo, EMBASE, EMCARE via the EBSCOHost and OVID interfaces. We also searched Cochrane Library and PROSPERO for ongoing or completed systematic reviews. We conducted forward and backward citation searching of included studies and examined the “cited by” function. Gray literature was searched for additional studies using BASE and OpenGrey, which included reviewing conference proceedings and dissertation thesis abstracts.
Data management and selection process
We used EndNote X7.0.2 (https://www.myendnoteweb.com/) and Rayyan (Ouzzani et al., Reference Ouzzani, Hammady, Fedorowicz and Elmagarmid2016) to store, organize, and manage references and abstracts. After removing duplicates, three reviewers (EP, LW, MM) worked as pairs to independently screen all titles and abstracts and select articles potentially meeting the eligibility criteria for full-text review. They then independently screened the full text of the remaining articles to determine if the articles met the inclusion criteria. Any disagreements were discussed with the third reviewer and study lead (JH), and a decision was reached.
Inclusion criteria
No restrictions were applied in terms of language, country, or publication year. Studies were included if they were:
-
1. Peer-reviewed primary research
-
2. Involved PLwD regardless of sex, age, type, or stage of dementia, living in the community or in nursing homes.
-
3. Provided evidence of assessing risk for people living at home with dementia or reported on perspectives of risk assessment and attitudes toward risk assessment for HCPs, PLwD, and their carers.
-
4. Reported any outcome or component related to risk assessment in dementia (e.g. safety, risk of harm, neglect).
-
5. Any study design – quantitative or qualitative.
Exclusion criteria
We excluded studies if they reported:
-
1. Pharmaceutical interventions and risk
-
2. Risk factors for dementia development
-
3. Only risk enablement, management, and strategies implemented in dementia.
Data extraction
We developed a data extraction template in Excel. EP and CK piloted it independently for two included studies, compared the data extracted, and refined the template. Data were then extracted and entered for all included papers. This comprised information on study characteristics: authors, year of publication, setting, study design, sample size, and participants’ age, sex, dementia type and severity, the objectives of the review, risk assessments, key elements of risk, and attitudes to assessing risk in dementia. Where a study had incomplete data, reviewers attempted to contact the authors to obtain more study data but did not receive any responses.
Quality assessment
Two reviewers (EP and CK) independently assessed the quality of the included studies using the Mixed Method Appraisal Tool (MMAT: Hong et al., Reference Hong2018), and any discrepancies were resolved by consulting a third reviewer (JH). Using one tool for the range of methods in included papers maximized appraisal consistency and across studies comparison.
Data synthesis and analysis
We calculated Cohen’s kappa to measure the level of agreement for the screening and selection of the papers (Higgins et al., Reference Higgins2023). We used thematic synthesis to analyze the qualitative data extracted (Thomas and Harden, Reference Thomas and Harden2008). Once familiar with the data, two reviewers (EP & CK) created a coding frame and regularly met to discuss their coding and amend the framework if needed. Once coded the team met to explore relationships and hierarchies between codes. Extracted data were organized into descriptive themes and then finally into analytical themes (Thomas and Harden, Reference Thomas and Harden2008). Five primary themes emerged from the data extraction, corresponding to the systematic review’s objectives.
Quantitative data were extracted and tabulated using a separate form. Where appropriate, data were synthesized narratively as the data heterogeneity made it unsuitable for meta-analysis (Popay et al., Reference Popay2006).
Results
The PRISMA flow diagram (Page et al., Reference Page2021) displays the systematic search and screening results (Figure 1). The search yielded 9,328 citations after duplicate removal, out of which 20 papers met the eligibility criteria for the review. The inter-rater reliability score for study selection was κ = 0.973 suggesting “near-perfect agreement” (Higgins et al., Reference Higgins2023).
PRISMA flow diagram.
Quality assessment of the included studies
Fourteen studies met all MMAT quality criteria and four studies met over half of the criteria. Two studies (Adams, Reference Adams2001; Evans and Means, Reference Evans and Means2006) were rated as lower quality due to inadequate detail about the methods used and data synthesis. Table S1: supplementary information, shows the quality assessment of the included studies.
Characteristics of included studies
The characteristics of the 20 studies are reported in Table 1 and referenced P1–P20 (Adams, Reference Adams2001 P1; Bantry White and Montgomery, Reference Bantry White and Montgomery2016 P2; Behrman et al., Reference Behrman, Wilkinson, Lloyd and Vincent2017 P3; Benbow and Kingston, Reference Benbow and Kingston2017 P4; Bourgeois and Couturier, Reference Bourgeois and Couturier2009 P5; Choi et al., Reference Choi2014 P6; Clarke et al., Reference Clarke2010 P7; Dickins et al., Reference Dickins, Goeman, O’Keefe, Iliffe and Pond2018 P8; Evans et al., Reference Evans, Perkins, Clarke, Haines, Baldwin and Whittington2016 P9; Evans and Means, Reference Evans and Means2006 P10; Lach et al., Reference Lach, Thomas Reed, Smith and Carr1995 P11; Lee et al., Reference Lee2019 P12; Pickering et al., Reference Pickering, Yefimova, Maxwell, Puga and Sullivan2019 P13; Poulin de Courval et al., Reference Poulin de Courval2006 P14; Sandberg et al., Reference Sandberg, Rosenberg, Sandman and Borell2015 P15; Stevenson and Taylor, Reference Stevenson and Taylor2018 P16; Stevenson and Taylor, Reference Stevenson and Taylor2017 P17; Stevenson et al., Reference Stevenson, Savage and Taylor2019 P18; Struckmeyer et al., Reference Struckmeyer, Pickens, Brown and Mitchell2020 P19; Taylor et al., Reference Taylor, Stevenson and McDowell2017 P20). Of the twenty studies included in the review, eleven were qualitativeP1,P3,P4,P7-P10,P15-P18, four used quantitative methodsP6,P13,P14,P19, and five used mixed methodologyP2,P5,P11,P12,P20.
Study characteristics of included papers
Most studies were based in the UK (n = 11). Others were from AmericaP11,P13,P19, CanadaP12,P14, AustraliaP8, FranceP5, South Korea,P6 and SwedenP15. One studyP5 was translated from French to English by a bilingual researcher. The number of participants in Evans and Means (Reference Evans and Means2006)P10 study was unclear, while the total number of participants in the remaining 19 studies was 739. These comprised PLwD (n = 291), their carers (n = 268), social or HCPs supporting PLwD (n = 221), and healthy older people (n = 20). Sex of participants was not reported in six studiesP2,P7,P10-P12,P19. In the remaining sample of 648 participants, there were more females (n = 458) than males (n = 190).
All included papers focused on risk assessment in dementia. One explored how risk is conceptualized by PLwDP8, five explored how HCPs construct and communicate risk assessmentP1,P7,P17,P18, and how risk is assessed and communicatedP20. Seven studies explored risk-related experiences and perceptions of risk in dementia with either PLwDP4,P5,P15,P18, carers P3-P6,P11,P13,P16,P17, or HCPsP3,P4,P9,P10. Most studies (n = 15) were conducted in community settings, and one studyP9 was in a care home. Four studies reported the development of scales by HCPs to assess risk and safetyP2,P12,P14,P19. Table 2 shows an overview of study outcome measures and findings of qualitative studies and Table 3 of quantitative studies.
Qualitative results: findings from qualitative and mixed method studies
Quantitative results: findings from quantitative and mixed method studies
Data synthesis
Five themes emerged from the thematic synthesis of the included studies:
-
1. Conceptualization of risk – individual perceptions of risk, including how different individuals define, construct, and identify risk situations.
-
2. Components of risk – key elements included in risk and safety assessments.
-
3. Contributors to risk – factors that impact the risk level and how risk assessments are conducted.
-
4. Perspectives on risk assessment – how individuals assess risk and approaches to risk management.
-
5. Risk reduction – strategies to mitigate risk following an assessment.
In addition to the five themes identified, we appraised the four risk assessment scales.
Theme 1 – Conceptualization of risk
This theme explores differences in how PLwD, carers, and HCPs conceptualize and define risk; there are three subthemes: definition, construction, and identification of risk.
Definition of risk
PLwD perceive risk to be associated with a lack of psychological or physical safety. Examples include unfamiliar settings or situations that create disorientation and feelings of anxietyP15,P18. Consequently, risk is associated with feeling loss of control, fear of the unknown, and uncertainty about what might happen. Feeling safe, particularly while maintaining independence and autonomy is valued positivelyP18.
Carers associate risk with exposure to harm and accidents. While both carers and HCPs conceptualize risk as the consequence of a potential action and are concerned by severity rather than likelihood of it happeningP16,P17. Risk susceptibility is determined by whether the risk is judged as being acceptable, primarily by the HCP but influenced by the risk management culture of their organizationsP2,P17. Furthermore, risk is perceived to be ongoing and continuously changing as dementia progressesP17,P18, while level of risk is often determined by the PLwD’s living situation and availability of family supportP15 and their willingness to engage with the support offeredP17. Table S2, supplementary information, identifies factors influencing definitions of risk.
Construct of risk
Concepts of risk are constructed by socio-cultural factors interlinked with health, psychosocial, and environmental influences, knowledge, and understanding of dementiaP8,P10,P18. Socio-cultural factors include life history (e.g. previous occupation), media representations of dementia, personal experiences, the individual’s personality, and emotional responses to prior events, such as falls, going missing, or health emergenciesP8,P18. Consequently, risk is viewed as being idiosyncratic and constructed in the context of multiple influences on the individual and associated with emotions such as anxietyP18. In addition, HCP’s concepts of risk are influenced by concerns about harm to the individual and others, resource, and environmental limitations, while attempting to balance the rights of individuals and others alongside organizational risk assessment culturesP9,P17. Table S2, supplementary information, identifies factors influencing the conceptualization of risk.
Identification of risk
Risk identification explores how carers identify concerns about PLwD causing harm to self or others or harm resulting from the actions of othersP1. In practice, carer’s accounts of daily life experiences and disease progression enable HCPs to detect and acknowledge if a situation constitutes a risk, for example, someone with dementia going missingP1,P16.
Theme 2 – Components of risk assessment
This theme concerns the identification of risk factors and actual risks that form the components of risk assessment. Twelve studies explore perspectives of PLwD, carers, and HCPsP2,P4,P5,P8,P10-P12, P14,P16,P18,P20.
Risk components relate to external (i.e. the physical environment, such as living arrangements or external support) and internal (i.e. the individual, such as progression of dementia) factors. These often have a harmful impact on PLwD, their carers and/or family members, and other members of the wider community.
Risk components are categorized into eight overarching domains: constant/pervasive risks; safety inside the home; safety in the community; food and nutrition; substances/dangerous objects; emotional and behavioral issues including self-care; interpersonal/social; financial. Table 4 reports the risk components identified in each study. Mobilization both inside and outside of the home is the most frequently identified risk and is discussed in all papers. Mobilization is associated with falling or slipping. Other common risk factors are associated with driving and road safety, navigation outside of the home and orientation. The most frequently identified risks within the home are related to food preparation and safety in the kitchen and taking medication.
Identified risk factors from included studies
Eight studies identify fire-related risks either in general or from electrical and gas appliancesP5,P6,P8,P10,P11,P14,P16,P17. Five studies identify risk related to financial exploitationP4,P12,P16,P17,P20. Some studies identify risk factors related to emotional well-being, inadvertent and intentional self-harmP6, neglect of personal hygieneP17,P20 or accidents while showering or bathingP10,P16, food and alcohol consumption or malnourishmentP6,P11,P12,P14,P17,P18,P20. Less commonly identified were accidents related to self-regulation of temperature (clothing, heating)P11, leaving hot water runningP6, use of sharps or firearmsP11; incidents involving the policeP4; mood disordersP20; and loss of important documentsP17.
Theme 3 – Contributors to risk
Eight papers examine factors that impact on risk levels and risk assessmentP3,P7-P9,P13,P16,P17,P20. Two subthemes are identified, factors that increase risk likelihood and severity, and factors that influence risk assessment.
Factors that affect the likelihood or severity of risk
Five studies identify factors contributing to risk by either influencing the likelihood that an adverse event will occur or the resultant harmP3,P9,P13,P16,P17. These factors broadly cover daily routines, dementia progression, living arrangements, caring arrangements, and medication management. For example, increased carer stress and disruption to daily routines reportedly increase the likelihood of abusive or neglectful behavior, whereas carers and PLwD participating in social activities together acts as a protective factor against neglectP13. Also, unclear pathways through support services increase risk for PLwD and their carersP3. Table S3, supplementary information, identifies factors contributing to risk and their outcomes.
Factors that influence risk assessment
Three papers outline factors that contribute, either positively or negatively, to the assessment of risk in dementiaP7,P8,P20. Negative ramifications of risk assessment are that personal responsibilities and freedoms are often reduced for PLwD and restrictions put in place. PLwD and carers perceive risk as an accepted part of daily life but frequently following a diagnosis, capacity may be questioned by HCPs without assessment and assumed to be lacking regarding decisions that may result in risksP8. Risk enablement can support people to live at home longer, while poor service utilization or provision increases carer stress and hastens care home placementP8. Staff acknowledge risk-averse practices may cause harm to PLwD through reducing independence, but also may reduce adverse consequences and professional responsibilities limit their ability to change practiceP8. Organizational cultures impact risk assessment through nurturing either risk-averse or risk-tolerant approachesP7. For example, work cultures supportive of person-centered care are viewed as enabling risk tolerance, while those which emphasize the deficits of PLwD limit creative solutions to manage risk, and staff ability to advocate for risk-tolerant strategiesP7. Professionals’ own risk-averse attitudes also contribute to problem-oriented thinking, increasing the likelihood of risk-averse organizational and team culturesP7. Decisions regarding risk are also influenced by staff workload and the availability of servicesP7.
Effective risk management requires communication between the assessing professional, PLwD, and carers. Taylor et al. (Reference Taylor, Stevenson and McDowell2017)P20 outline challenges to communicating risk following assessment, which from most challenging to least challenging are: conflicting ideas on client’s best interests; PLwD lacking insight into the risk; unrealistic expectations of services; risk of damage to professional or client relationship; lack of data on likelihoods and fear of complaint.
Theme 4 – Perspectives on risk assessment
Nine papers discuss perspectives on risk assessment, which include the perceptions of PLwDP8,P10,P18, carersP1,P3,P8-P10,P16, and HCPs involved in dementia careP3,P7,P8,P10,P17. Table S4, supplementary information, outlines individual attitudes toward risk assessment and the approaches they use to make judgments and decisions regarding risk.
People living with dementia
PLwD regard themselves as being largely risk-tolerant, considering some risk as acceptable. Maintaining independence in continuing activities is perceived to support their mental and physical well-being. PLwD judge that maintaining their quality of life, through engaging in meaningful activities enables a sense of purposeP18. Regarding risky situations, PLwD emphasize that they often have strategies in place to manage or resolve any perceived risk, for example, seeking familiar objects or asking for help if they became lostP8.
Carers
Similarly, carers often think that positive risk-taking is beneficial and it is better to manage risks rather than trying to eliminate them, therefore balancing the need for autonomy with safety. They also believe restricting freedoms may impact negatively on the PLwD, being potentially counterproductive by creating more risks through their relative becoming inactive or uncooperative and unwilling to accept careP16.
Carers raise risk concerns they want help with, while many concerns are not reported to HCPsP3 due to differences between HCP and carers’ perceptions of safety or anxiety about care packages being withheld. Risks that are discussed with HCPs include medication, support, and safety inside the home. Conversations about risk with PLwD are dependent on the ability of the care recipient to comprehend and have insight into the issues discussed and carers feel reluctant to discuss risk that may cause anxiety, when their relative does not understand the danger wellP16. Carers were often unsure whether PLwD would report risk concerns to them.
Healthcare professionals
HCPs identify a need for collaborative decision-making in risk assessment. Community psychiatric nurses regularly discuss risk and risk management strategies with carersP1. Challenges arise where HCPs and carers have differing views about responsibility for decision-making in risk management. However, HCPs highlight the potential for collaborative working with carers to promote autonomy and minimize risk, by developing systems for carers to identify and report safety risksP3.
HCPs identify challenges in implementing risk assessments, including adopting risk-averse approaches to ensure PLwD will be safe. Judgment on risks balance two needs: firstly, the need to reduce harm, and secondly, the fear of being judged and being held accountable for potential harm to PLwD. This was sometimes problematic, as it created conflict with PLwD who wanted to remain independent and activeP16. However, HCPs recognize the need to accept some level of risk to enable PLwD to maintain autonomyP7,P9. Context and circumstances dictate this process, and it is important to consider life history, preferences, and current circumstances.
Theme 5 – Risk reduction
The final theme refers to the strategies individuals use to reduce or avoid risk following the identification of risk, reported in five studiesP1,P3,P11,P15,P16. PLwD report utilizing strategies to minimize risk, for example, when they are in an unfamiliar setting or a confusing and risky situation. PLwD use the following four strategies to mitigate risk: a) seek something recognizable, b) convince themselves that the situation is as it should be, c) avoid the risky situations, d) accept assistance from one’s environmentP15.
Carers report discussions with PLwD primarily involve increasing awareness of the risk, such as reminding their relative that they may fall. Other strategies include accompanying the PLwD to activities, increased vigilance (monitoring environment), home adaptations, hiding objects, general supervision, and taking over activities or tasksP11,P16. Carers may be reluctant to report concerns to HCPs for fear that the PLwD would receive substandard care, to avoid bothering the professional, or to avoid receiving an overly bureaucratic responseP3.
Community psychiatric nurses report that following the risk assessment, action is usually taken to mitigate the risk following discussion with other HCPs or family membersP1. HCPs follow formal procedures, aligned with local safeguarding guidelines to report, and manage identified risks. Risk-averse and over-bureaucratic processes can result in difficulties with implementing person-centered approaches for minimizing risk and optimizing quality of life for PLwDP3. Overall, HCPs believe that people who raise concerns regarding risk receive a better level of care and supportP16.
Risk assessment scales
Four studies report instruments to evaluate safety and measure risk; see Table S5: supplementary information, for included domains and key components.
Bantry White and Montgomery’s (Reference Bantry White and Montgomery2016)P2 scale measures safety when walking outdoors for PLwD and the overarching construct is prevention and protection from harm while wandering or being lost. The scale consists of 38 items divided into five domains. Perceptions of the structured assessment schedule were mixed, as clinical reviewers considered some components useful and others too vague. Some clinicians wanted a more flexible approach to assessing wandering and safety and had difficulty in distinguishing the objective versus the subjective risks associated with danger or harm to the person’s safety while wandering or going missing.
The person-centered risk assessment framework (PCRAF) consists of two components: the PCRAF framework and My Plan, to engage PLwD in risk assessment and create a positive person-centered care planP12. It is a comprehensive framework designed to encourage awareness of existing risks among clinical teams, promote independence, and involve PLwD in care planningP12. The seven domains include risky situations (e.g. falls, driving) that may result in destabilization of co-existing chronic illnesses that lead to emergency department visits, hospitalizations, and premature care home placement. Despite HCPs reporting high satisfaction it has several limitations and challenges to implementation being time-consuming and not yet fully integrated into the assessment process. Furthermore, it requires an established relationship between PLwD and HCPs. Therefore, an initial assessment or diagnostic visits may not be the best time to administer the PCRAF as PLwD and their carers are still grappling with the diagnosis.
The Safety Assessment Scale (SAS)P14 was developed for use by HCPs working in community settings to assess home safety in PLwD, to minimize the risk of accidents, provide recommendations to carers, and enhance case management. It consists of 32 items with nine domains related to risks of accidents and includes open-ended questions related to the person’s behavioral and environmental risks. SAS shows good validity and reliability and can be used by a range of different HCPs. A shorter version has been developed with 19 items maintaining seven out of the initial nine sections.
The Home Environment Assessment Protocol-Revised (HEAP-R)P19 is a modified version of the HEAP scale (Gitlin et al, Reference Gitlin, Schinfeld, Winter, Corcoran, Boyce and Hauch2002) that assesses home safety for PLwD through observation and questions. HEAP-R consists of five domains: hazards, adaptions, physical cues, clutter, and comfort, reducing the original number of items from 192 to 60. The HEAP-R was pilot-tested in a study with three interrelated phases: assessment of content validity by experts, criterion validity by carer-PLwD dyads, and test-retest reliability by occupational therapists. The HEAP-R is used by occupational therapists and demonstrates good validity and reliability in measuring safety for clutter, hazards, and adaptions. However, its usability remains untested with other allied HCPs. Further research is indicated to determine clinical cut-points that accurately and meaningfully enable decision-makers to assess safety and risks within someone’s home.
Discussion
We identified only 20 relevant papers on risk assessment but were able in this first systematic review to identify how risk is conceptualized, the key components of risk assessment, perspectives about how risk is assessed, obstacles to conducting risk assessments, distinguish factors that contribute toward risk events and differing attitudes to risk identification and assessment in dementia.
This review found that PLwD, carers, and HCPs conceptualize risk differently. Risk is defined as being part of everyday life and the construction of risk is influenced by media perceptions, personal experiences, socio-cultural influences, and dementia severity. Conceptualization, therefore, transcends individual constructs and represents a complex and individualized view (Lupton, Reference Lupton2013; McDermott, Reference McDermott2010). PLwD associate risk with unusual situations that create feelings of fear and uncertainty and want to feel safe while retaining their independence. PLwD particularly value feeling safe at home (Bamford et al., Reference Bamford2021) and may perceive being out of their home as more unsafe. Carers perceive risk as exposure to danger and harm and prefer to manage rather than eliminate risk. Post-diagnostic support and education is needed to help families with decision-making around risk (Bamford et al., Reference Bamford2021). HCPs judge risk in dementia according to the severity of the consequences of an activity or situation rather than the likelihood of it occurring. They adopt risk-averse approaches because of the certainty of reduced harm for PLwD, and the fear of being held accountable. This contrasts with PLwD, who express concerns about diminished personal freedom and autonomy in decision-making following a diagnosis of dementia, when susceptibility to risk may be presumed and if present the PLWD may not have insight.
Our review showed HCPs were sensitive to PLwD’s need for autonomy and independence and supported risk enablement in care planning where possible. Similarly, Wied et al. (Reference Wied, Knebel, Tesky and Haberstroh2019) found that HCPs do not deny PLwD their right to make their own decisions and they are supported with decision-making in terms of care. A key challenge for HCPs is communication with PLwD and carers when perceptions of risk differ or there is lack of insight (Stevenson and Taylor, Reference Stevenson and Taylor2017). Within this review, scant attention is given to insight and exploring the PLwD’s level of awareness about their impairments and how these impact on their safety. Only one risk assessment scale included a question about insight, which was linked to wandering and whether the PLwD was “aware that they get lost?” (Bantry White and Montgomery, Reference Bantry White and Montgomery2016). Some PLwD do not acknowledge their condition or lack awareness of deficits in sensory, perceptual, motor, affective, or cognitive functioning and may underestimate their dementia-related impairments (Alexander et al., Reference Alexander, Martyr, Savage, Morris and Clare2021; Tagai et al., Reference Tagai, Nagata, Shinagawa and Shigeta2020). Self-reported awareness of impairments in mild-moderate dementia can change over time, so lack of insight is not always explained by cognitive or functional deficits and awareness should be measured at different timepoints in the illness trajectory (Alexander et al., Reference Alexander, Martyr, Savage, Morris and Clare2021).
Taking risks is traditionally perceived as a negative action that should be minimized or eliminated to avoid harm (Clarke and Mantle, Reference Clarke and Mantle2016). However, participant groups were largely unified in their views surrounding action versus inaction and both PLwD and carers believe that positive risk-taking is necessary to maintain quality of life and enable engagement in meaningful and enjoyable activities (Mapes, Reference Mapes2017; Morgan and Williamson, Reference Morgan and Williamson2014). Risk enablement helps maintain independence by focusing on the skills and assets of PLwD, as “building on preserved activities” can yield positive and beneficial outcomes (Rapaport et al., Reference Rapaport2020). However, difficulty acknowledging risks within familiar settings or activities where the person feels they have control may create obstacles for implementing strategies to reduce risk. Risk exposure is ongoing and changes as dementia progresses. Consequently, enabling positive risk-taking becomes challenging as dementia severity and vulnerability increase and the likelihood of risk-averse practices increases (Berry et al., Reference Berry, Apesoa-Varano and Gomez2015).
Our review found that PLwD are generally risk-tolerant and report strategies to manage risk within their daily lives. Assessment and management often occur simultaneously, with risk reduction strategies being implemented at the time risks appear. However, susceptibility to risk is not always evident until a person’s behavior causes an accident and it becomes apparent that they can no longer safely carry out an activity (Bourgeois and Couturier, Reference Bourgeois and Couturier2009). Similarly, carers adopt largely informal approaches to risk identification and management. As potential risks arise, pragmatic and measured approaches are taken to reduce risk. Of note, carers can be reluctant to discuss some risk issues with their relative or with HCPs to avoid intrusive interventions. The carers’ role is pivotal to facilitating engagement between PLwD and HCPs and these differing perspectives about what constitutes an acceptable risk and reticence can lead to inadequate knowledge, assessment, and delays in accessing support (Bodemer and Gaissmaier, Reference Bodemer, Gaissmaier, Roeser, Hillerbrand, Sandin and Peterson2012; Manthorpe and Moriarty, Reference Manthorpe and Moriarty2010; Tuijt et al., Reference Tuijt2021). In contrast, HCPs are bound by organizational rules and regulations surrounding risk processes. However, these policies may underestimate the complexity of risk assessment judgments and mean decision-making centers around what is defensible as well as PLwD and carer preferences (Taylor, Reference Taylor2006). This may be exacerbated by a lack of service availability or service utilization for carers (Clarke et al., Reference Clarke2010). Effective communication ensures that information, preferences, and concerns are shared and enables informed decision-making (Fetherstonhaugh et al., Reference Fetherstonhaugh, Tarzia and Nay2013, Reference Fetherstonhaugh, Tarzia, Bauer, Nay and Beattie2016). However, lack of consensus and standardization in defining and measuring risk behaviors, such as wandering, distorts how risk factors are reported and risk reduction measures are dependent on accurate assessment (Ali et al., Reference Ali2016). While structured risk assessments are useful for systematically identifying safety hazards, high scores may be insufficient for predicting likelihood of an accident, but additional indicators can be assessed through dialog with carers (Bourgeois and Couturier, Reference Bourgeois and Couturier2009).
Risks were related to physical and emotional harm to PLwD and others, such as fires, falls, safety crossing roads, driving safety, exposure to cold, not taking medication correctly, vulnerability through exploitation and social isolation, and exposure was mostly gradual and persistent rather than sudden or unexpected. The pervasive nature of exposure to risks suggests that strategies can be introduced to target key areas and mitigate risk in dementia. Mobilization inside and outside the home was the most frequently identified risk factor. The risk of falls is significantly associated with impaired cognition, increased frailty, gait changes, reduced ADL ability, and depression in PLwD (Park et al., Reference Park, Lee and Kang2020). These have a significant impact on quality of life, health and healthcare costs, and expedite care home placement and mortality (NICE, 2013). Strategies such as fall prevention programs, exercise, vitamin D supplementation, and reducing hazards in the home environment (flooring, lighting, furniture, and fittings) can reduce the risk and be incorporated into care plans (Moncada and Mire, Reference Moncada and Mire2017; NICE, 2013). Other concerns about safety within the home include food preparation and medication administration. Assistive technology advancements which support independent living may help mitigate risk and allow PLwD to remain at home (Kruse et al., Reference Kruse, Fohn, Umunnakwe, Patel and Patel2020). Yet, despite the range of technologies available for PLwD and carers there is very little evidence of widespread practical application (Lorenz et al., Reference Lorenz, Freddolino, Comas-Herrera, Knapp and Damant2019). This may be due to conflicting evidence about their clinical and cost-effectiveness and barriers such as PLwD being reluctant to use assistive technology aids due to lack of familiarity with equipment or feeling stigmatized (Howard et al., Reference Howard2021; Kruse et al., Reference Kruse, Fohn, Umunnakwe, Patel and Patel2020).
Implications for practice
This review highlights the need to develop clear and transparent risk communication pathways to enable better-informed assessments and individualize care planning for risk management. Focusing on safety by avoiding risky situations, as opposed to enablement can result in approaches that may hasten an individual’s cognitive decline through inactivity (Clarke and Mantle, Reference Clarke and Mantle2016). Thus, risk assessment and management processes should focus as much as possible on the adaption and support that can sustain activities, as opposed to avoidance (Clarke and Mantle, Reference Clarke and Mantle2016). Modifying organizational risk cultures away from risk aversion and toward positive risk management would support professionals when making decisions with patients and families to strike the right balance between safety and autonomy. Effective risk communication is key and needs to be informed by a framework of policy and best practice that make PLwD central to the decision-making process (Morgan and Andrews, Reference Morgan and Andrews2016; Stevenson et al., Reference Stevenson, Savage and Taylor2019). Shared approaches to risk should balance rights and protection by developing outcomes-focused, person-centered safeguarding practices (Manthorpe and Moriarty, Reference Manthorpe and Moriarty2010; Morgan and Andrews, Reference Morgan and Andrews2016; Stevenson et al., Reference Stevenson, McDowell and Taylor2018). Where concerns about insight and capacity make this more complicated, positive risk-taking can be achieved through personalizing approaches that use the least restrictive practices and are in line with the principles of the Mental Capacity Act (DH, 2005). This assumes the PLwD has capacity unless it is established otherwise; and enables good, calculated decisions about risk (Manthorpe and Moriarty, Reference Manthorpe and Moriarty2010; Morgan and Andrews, Reference Morgan and Andrews2016). Additionally, carer knowledge and skills at recognizing and managing their relative’s exposure to risk are also as a primary factor in risk modification (Horvath et al., Reference Horvath2005). The value of maintaining autonomy through risk enablement with risk-taking is part of maintaining quality of life rather than something to be avoided. Risk assessment processes should therefore promote and maintain improvements in quality of life.
Implications for research
This review included studies that evaluated risk assessment tools and all reported good reliability and validity (Lee et al., Reference Lee2019; Struckmeyer et al., Reference Struckmeyer, Pickens, Brown and Mitchell2020; Bantry White and Montgomery, Reference Bantry White and Montgomery2016; Poulin de Courval et al., Reference Poulin de Courval2006). Scales represent valuable tools to develop person-centered risk assessments. However, some scales were too specific to be used pragmatically across all clinical environments and multi-disciplinary professionals (Struckmeyer et al., Reference Struckmeyer, Pickens, Brown and Mitchell2020; Bantry White and Montgomery, Reference Bantry White and Montgomery2016). Our findings identify the key components of risk assessment, which supports the content validity for developing future risk assessment tools. Scales should be able to work across health and social care and acceptability and implementation of risk assessment should be the focus of future research. To date, three of the four scales we identified have very narrow risk assessments and none have been evaluated completely for usability and whether they can help produce strategies which are acceptable to PLwD and their carers and reduce risk.
Strengths and limitations
The strengths of this review include the rich range of evidence, including papers in all languages from databases inception and including the gray literature. Multiple reviewers were involved in all stages of the review process. It collated multiple perspectives from individuals with dementia and those involved in their care. By triangulating these perspectives, we were able to build a holistic view of risk assessment in dementia.
Studies within this review were conducted in western, high-income countries. This limits the generalizability of our findings to low and middle-income countries. Most studies were observational and there was little testing of observations, but most studies were of good quality. Overall, the evidence collated and summarized enables clear and coherent messages to be drawn out, contributing to the meaningfulness of this review.
Conclusion
In conclusion, our systematic review summarizes the major components of risk to be included in a risk assessment tool for PLwD and offers insight into the complexity of risk and the diversity of perspectives on risk assessment and management. Safety concerns remain paramount, and accurate risk assessment and excellent communication are needed to enable risk-tolerant practices and ensure these approaches can be evaluated effectively. PLwD and carers should play an active and prominent role in the risk assessment process. We did not find any instrument which to date had been shown to be widely acceptable and useful in practice. The review clearly demonstrates the need for further research and evidence-building about risk assessment in dementia care.
Conflict of interest
G.L. and SCG are supported by University College London Hospitals’ National Institute for Health Research (NIHR) Biomedical Research Centre, and GL is also supported by North Thames NIHR Applied Research Collaboration and as an NIHR Senior Investigator.
Source of funding
This paper presents independent research funded by the Dunhill Medical Trust (grant reference number: RPGF1711\10). The views expressed in this publication are those of the authors and do not necessarily reflect those of the Dunhill Medical Trust.
Description of authors’ roles – CRediT author statement:
Conception and design of the work – Juanita Hoe, Elena Profyri, Gill Livingston, Sergi Costafreda
Data collection – Elena Profyri, Charlotte Kemp, Monica Manela, Lucy Webster
Data analysis and interpretation – Juanita Hoe, Elena Profyri, Charlotte Kemp, Justine Anthony, Frank Arrojo, Helen Souris, Gill Livingston, Sergi Costafreda,
Drafting the article – Juanita Hoe, Elena Profyri, Charlotte Kemp, Justine Anthony
Critical revision of the article – Juanita Hoe, Gill Livingston, Sergi Costafreda, Lucy Webster, Frank Arrojo, Helen Souris.
Final approval of published version – Juanita Hoe, Gill Livingston, Sergi Costafreda, Frank Arrojo, Helen Souris.
Funding acquisition – Juanita Hoe, Gill Livingston, Sergi Costafreda, Frank Arrojo, Helen Souris.
Acknowledgements
The authors thank Stephen O’Driscoll and Chris Thorpe (Research Librarians, City, University of London) for their support and guidance in the search process and the Risk Assessment and Increasing Safety in Dementia (RAISe-Dementia) study steering committee for their helpful feedback and advice on earlier drafts of this review.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1041610223004398.
