Risk assessment for people living with dementia: a systematic review

Abstract

Objective:

This systematic review identified key components of risk assessment for people with dementia, examined attitudes toward risk identification and risk assessment, and appraised existing risk assessment tools.

Methods:

Systematic searches of five databases on two platforms (EBSCO, OVID) and gray literature databases (Open Grey, Base) were conducted. Studies were screened for inclusion based on predetermined eligibility criteria and quality assessed using the Mixed Methods Appraisal Tool. Findings were tabulated and synthesized using thematic synthesis.

Results:

Our review found people with dementia, their family carers, and healthcare professionals differed in how risk is conceptualized, with views being shaped by media perceptions, personal experiences, socio-cultural influences, dementia knowledge, and dementia severity. We found that mobilization (causing falls inside and getting lost outside) is the most frequently identified risk factor. Our findings show people with dementia are generally risk-tolerant, while healthcare professionals may adopt risk-averse approaches because of organizational requirements. We found factors that disrupt daily routines, living and caring arrangements, medication management, and unclear care pathways contribute toward adverse risk events. We discovered that most studies about risk and risk assessment scales did not consider insight of the person with dementia into risks although this is important for the impact of a risk. No risk instrument identified had sufficient evidence that it was useful.

Conclusion:

Accurate risk assessment and effective communication strategies that include the perspectives of people with dementia are needed to enable risk-tolerant practice. No risk instrument to date was shown to be widely acceptable and useful in practice.

Introduction

Over 55 million people live with dementia globally and numbers are increasing as the population ages (WHO, 2021; Prince et al., 2015). Around two-thirds (61%) of people living with dementia (PLwD) are in the community, of which a third live alone (Prince et al., 2014). 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., 2019).

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, 2022).

As cognition deteriorates, PLwD often become more vulnerable through self-neglect, accidents, and a greater risk of abuse and exploitation (Manthorpe and Moriarty, 2010). 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, 2010; Morgan et al., 2004; Morgan and Williamson, 2014; 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., 2022).

It is essential to balance independence and risks when planning care (Mitchell and Glendinning, 2008; Taylor, 2006), and PLwD want to and should be involved in that decision-making process (Pel-Littel et al., 2021; Wied et al., 2019). 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., 2018; Pel-Littel et al., 2021). 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, 2013). 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, 2011; Lea et al., 2014). As to our knowledge, there is no systematic review on this subject, we aimed to systematically:

1. 1. Explore key components of risk assessment for PLwD in the community and within care homes.

2. 2. Report on the attitudes to risk identification and risk assessment of PLwD, their family carers, and health and social care professionals.

3. 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., 2015), adheres to the PRISMA 2020 statement (Page et al., 2021), 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 10^th June 2020 and again on 23^rd 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., 2016) 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. 1. Peer-reviewed primary research

2. 2. Involved PLwD regardless of sex, age, type, or stage of dementia, living in the community or in nursing homes.

3. 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. 4. Reported any outcome or component related to risk assessment in dementia (e.g. safety, risk of harm, neglect).

5. 5. Any study design – quantitative or qualitative.

Exclusion criteria

We excluded studies if they reported:

1. 1. Pharmaceutical interventions and risk

2. 2. Risk factors for dementia development

3. 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., 2018), 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., 2023). We used thematic synthesis to analyze the qualitative data extracted (Thomas and Harden, 2008). 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, 2008). 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., 2006).

Results

The PRISMA flow diagram (Page et al., 2021) 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., 2023).

Figure 1. 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, 2001; Evans and Means, 2006) 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, 2001 ^P1; Bantry White and Montgomery, 2016 ^P2; Behrman et al., 2017 ^P3; Benbow and Kingston, 2017 ^P4; Bourgeois and Couturier, 2009 ^P5; Choi et al., 2014 ^P6; Clarke et al., 2010 ^P7; Dickins et al., 2018 ^P8; Evans et al., 2016 ^P9; Evans and Means, 2006 ^P10; Lach et al., 1995 ^P11; Lee et al., 2019 ^P12; Pickering et al., 2019 ^P13; Poulin de Courval et al., 2006 ^P14; Sandberg et al., 2015 ^P15; Stevenson and Taylor, 2018 ^P16; Stevenson and Taylor, 2017 ^P17; Stevenson et al., 2019 ^P18; Struckmeyer et al., 2020 ^P19; Taylor et al., 2017 ^P20). Of the twenty studies included in the review, eleven were qualitative^{P1,P3,P4,P7-P10,P15-P18}, four used quantitative methods^{P6,P13,P14,P19,} and five used mixed methodology^{P2,P5,P11,P12,P20}.

Table 1. Study characteristics of included papers

Most studies were based in the UK (n = 11). Others were from America^P11,P13,P19, Canada^P12,P14, Australia^P8, France^P5, South Korea,^P6 and Sweden^P15. One study^P5 was translated from French to English by a bilingual researcher. The number of participants in Evans and Means (2006)^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 studies^{P2,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 PLwD^P8, five explored how HCPs construct and communicate risk assessment^{P1,P7,P17,P18}, and how risk is assessed and communicated^P20. Seven studies explored risk-related experiences and perceptions of risk in dementia with either PLwD^{P4,P5,P15,P18}, carers ^{P3-P6,P11,P13,P16,P17}, or HCPs^P3,P4,P9,P10. Most studies (n = 15) were conducted in community settings, and one study^P9 was in a care home. Four studies reported the development of scales by HCPs to assess risk and safety^{P2,P12,P14,P19}. Table 2 shows an overview of study outcome measures and findings of qualitative studies and Table 3 of quantitative studies.

Table 2. Qualitative results: findings from qualitative and mixed method studies

Table 3. Quantitative results: findings from quantitative and mixed method studies

Data synthesis

Five themes emerged from the thematic synthesis of the included studies:

1. 1. Conceptualization of risk – individual perceptions of risk, including how different individuals define, construct, and identify risk situations.

2. 2. Components of risk – key elements included in risk and safety assessments.

3. 3. Contributors to risk – factors that impact the risk level and how risk assessments are conducted.

4. 4. Perspectives on risk assessment – how individuals assess risk and approaches to risk management.

5. 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 anxiety^P15,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 positively^P18.

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 happening^P16,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 organizations^P2,P17. Furthermore, risk is perceived to be ongoing and continuously changing as dementia progresses^P17,P18, while level of risk is often determined by the PLwD’s living situation and availability of family support^P15 and their willingness to engage with the support offered^P17. 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 dementia^P8,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 emergencies^P8,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 anxiety^P18. 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 cultures^P9,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 others^P1. 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 missing^P1,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 HCPs^{P2,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.

Table 4. Identified risk factors from included studies

Eight studies identify fire-related risks either in general or from electrical and gas appliances^{P5,P6,P8,P10,P11,P14,P16,P17}. Five studies identify risk related to financial exploitation^{P4,P12,P16,P17,P20}. Some studies identify risk factors related to emotional well-being, inadvertent and intentional self-harm^P6, neglect of personal hygiene^P17,P20 or accidents while showering or bathing^P10,P16, food and alcohol consumption or malnourishment^{P6,P11,P12,P14,P17,P18,P20}. Less commonly identified were accidents related to self-regulation of temperature (clothing, heating)^P11, leaving hot water running^P6, use of sharps or firearms^P11; incidents involving the police^P4; mood disorders^P20; and loss of important documents^P17.

Theme 3 – Contributors to risk

Eight papers examine factors that impact on risk levels and risk assessment^{P3,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 harm^{P3,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 neglect^P13. Also, unclear pathways through support services increase risk for PLwD and their carers^P3. 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 dementia^P7,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 risks^P8. Risk enablement can support people to live at home longer, while poor service utilization or provision increases carer stress and hastens care home placement^P8. 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 practice^P8. Organizational cultures impact risk assessment through nurturing either risk-averse or risk-tolerant approaches^P7. 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 strategies^P7. Professionals’ own risk-averse attitudes also contribute to problem-oriented thinking, increasing the likelihood of risk-averse organizational and team cultures^P7. Decisions regarding risk are also influenced by staff workload and the availability of services^P7.

Effective risk management requires communication between the assessing professional, PLwD, and carers. Taylor et al. (2017)^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 PLwD^P8,P10,P18, carers^{P1,P3,P8-P10,P16}, and HCPs involved in dementia care^{P3,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 purpose^P18. 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 lost^P8.

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 care^P16.

Carers raise risk concerns they want help with, while many concerns are not reported to HCPs^P3 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 well^P16. 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 carers^P1. 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 risks^P3.

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 active^P16. However, HCPs recognize the need to accept some level of risk to enable PLwD to maintain autonomy^P7,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 studies^{P1,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 environment^P15.

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 tasks^P11,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 response^P3.

Community psychiatric nurses report that following the risk assessment, action is usually taken to mitigate the risk following discussion with other HCPs or family members^P1. 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 PLwD^P3. Overall, HCPs believe that people who raise concerns regarding risk receive a better level of care and support^P16.

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 (2016)^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 plan^P12. It is a comprehensive framework designed to encourage awareness of existing risks among clinical teams, promote independence, and involve PLwD in care planning^P12. 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, 2002) 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, 2013; McDermott, 2010). 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., 2021) 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., 2021). 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. (2019) 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, 2017). 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, 2016). 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., 2021; Tagai et al., 2020). 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., 2021).

Taking risks is traditionally perceived as a negative action that should be minimized or eliminated to avoid harm (Clarke and Mantle, 2016). 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, 2017; Morgan and Williamson, 2014). 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., 2020). 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., 2015).

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, 2009). 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, 2012; Manthorpe and Moriarty, 2010; Tuijt et al., 2021). 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, 2006). This may be exacerbated by a lack of service availability or service utilization for carers (Clarke et al., 2010). Effective communication ensures that information, preferences, and concerns are shared and enables informed decision-making (Fetherstonhaugh et al., 2013, 2016). 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., 2016). 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, 2009).

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., 2020). 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, 2017; 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., 2020). Yet, despite the range of technologies available for PLwD and carers there is very little evidence of widespread practical application (Lorenz et al., 2019). 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., 2021; Kruse et al., 2020).

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, 2016). 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, 2016). 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, 2016; Stevenson et al., 2019). Shared approaches to risk should balance rights and protection by developing outcomes-focused, person-centered safeguarding practices (Manthorpe and Moriarty, 2010; Morgan and Andrews, 2016; Stevenson et al., 2018). 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, 2010; Morgan and Andrews, 2016). 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., 2005). 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., 2019; Struckmeyer et al., 2020; Bantry White and Montgomery, 2016; Poulin de Courval et al., 2006). 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., 2020; Bantry White and Montgomery, 2016). 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.