Article Text

Visual identifiers for people with dementia in hospitals: a qualitative study to unravel mechanisms of action for improving quality of care
  1. Elizabeth Sutton1,
  2. Natalie Armstrong1,
  3. Louise Locock2,
  4. Simon Conroy3,
  5. Carolyn Tarrant1
  1. 1 Health Sciences, University of Leicester, Leicester, UK
  2. 2 HSRU, University of Aberdeen Institute of Applied Health Sciences, Aberdeen, UK
  3. 3 MRC Unit for Lifelong Health and Ageing, University College London, London, UK
  1. Correspondence to Dr Elizabeth Sutton, University of Leicester College of Life Sciences, Leicester, LE1 7RH, UK; es225{at}


Background Hospitalised people with dementia (PwD) experience worse care and more patient safety incidents than non-dementia patients. Visual identifiers are commonly used to identify patients who have a diagnosis of dementia, with the aim of promoting more personalised care. However, little is known about how they work in practice, nor about the potential unintended consequences that might arise from their use. We aim to identify the mechanisms through which visual identifiers could support good care for PwD, how and why their use may have negative consequences and the conditions for their effective use.

Methods We conducted interviews with 21 dementia leads and healthcare professionals, 19 carers and two PwD, and produced case studies of visual identification systems in four UK acute hospital trusts between 2019 and 2021. Analysis drew on the concept of classification to identify and explore mechanisms of action.

Results We identified four mechanisms through which visual identifiers could help towards providing good care for PwD: enabling coordination of care at organisational level; signalling eligibility for dementia-specific interventions; informing prioritisation of resources on wards; and acting as a quick reference cue for staff. But identifier effectiveness could be undermined by: lack of standardisation and consistency; a lack of closely coupled information about individual needs; and stigma associated with a dementia diagnosis. Identifier effectiveness was dependent on their implementation being supported through staff training, resources directed and efforts to develop a supportive culture for caring for this patient group.

Conclusion Our research highlights the potential mechanisms of action of visual identifiers and their possible negative consequences. Optimising the use of identifiers requires consensus on the rules of classification and the symbols used, and closely coupled patient information. Organisations need to provide support, offer the right resources and training and engage meaningfully with carers and patients about the use of identifiers.

  • Qualitative research
  • Healthcare quality improvement
  • Health services research

Data availability statement

Data are available upon reasonable request.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Hospitalised patients with dementia often experience poor care. Visual identifiers for this group are designed to increase awareness of a person’s dementia diagnosis and are already in routine use across hospitals in the UK and beyond. Although they can help staff identify patients with additional needs, their use has also been criticised for failing to bring about improvements and even contributing to poor care.


  • Using the concept of classification, and based on interviews with staff and people with dementia (PwD) and their relatives, we describe the mechanisms through which visual identifiers could help improve care for PwD, and how negative consequences can arise. We need to understand the mechanisms of effect of visual identifiers for PwD: how they can work to improve care, and how, why and under what circumstances their use can lead to negative unintended consequences. Without this understanding there is a risk that this widely used intervention will have a limited or even negative impact on the care of PwD in practice.


  • Pairing visual identifiers with well-designed personalised information documents, ensuring the right resources and training for staff and seeking genuine engagement with relatives and carers are critical for their effective use.


Approximately a quarter of hospital beds in the UK are occupied by someone living with dementia.1 Evidence suggests that people with dementia (PwD) experience longer stays, higher readmissions, greater numbers of patient safety incidents and a higher risk of mortality than those without dementia.2–6 PwD have specific additional needs when in hospital, which are not always met, including needs related to communication, nutrition, continence and managing confusion and distress.7–9 Importantly, PwD have a need for genuine connections and empathetic communications with staff to feel safe.10 Previous studies also highlight many missed opportunities in providing person-centred care for PwD in acute care settings.11

To address these concerns, many hospitals within the UK, and in other countries including Australia and the USA,12–15 have introduced visual identifiers for PwD. These typically take the form of symbols—such as a butterfly—added to wristbands or displayed on bedside signs with the aim of raising staff awareness that a patient may have additional needs. Identifiers are often used in combination with other interventions, such as staff training on providing care to PwD in an effort to promote an approach that better meets the needs of PwD and supports the delivery of person-centred care. While identifiers are hypothesised to have value for improving the care of PwD in hospitals, potential problems with their use have also been raised. These include risks of obscuring the person behind the diagnosis, resulting in less personalised care, problems with misclassification of patients16 and concerns about the consistency and reliability of their use.17

Over 90% of hospitals in England have identification systems in place for PwD.18 19 Despite their widespread use in practice, none of these schemes have been fully evaluated.20 Given that visual identifiers for PwD are in routine use across hospitals in the UK and beyond, there is a pressing need to understand their potential mechanisms of action: how they function to improve care. It is also important to pay attention to their potential ‘dark logic’; to consider how, why and under what circumstances their use might lead to negative unintended consequences, either on the outcomes of interest (‘paradoxical effects’) or other outcomes (‘harmful externalities’).21 This understanding will be valuable in informing decisions about whether and how to use identifiers and in what form, and in developing approaches to counter possible harms from their use.

Based on a qualitative study with health professionals, PwD and their carers, we analyse mechanisms through which visual identifiers work in practice, what problems they can address and when and how they might generate negative consequences. This understanding will inform future efforts to use identifiers more effectively to support good care for PwD.


As part of a broader programme of research into the use of visual identification systems for PwD in hospital,20 22 we conducted a qualitative study involving in-depth case studies and interviews with health professionals, PwD and their carers between 2019 and 2021. We selected four acute care sites as case studies: two large and two smaller hospitals (based on the number of beds) in different regions of England. In consultation with experts in dementia care, sites were selected to include hospitals using a range of identifiers (including the national butterfly scheme17 and locally developed identifiers) and with varying levels of performance in implementing their dementia strategies. In each of the four sites, we interviewed the dementia lead or senior dementia nurse about their organisation’s approach to quality of care for PwD. We gathered documentation including dementia strategies and any dementia-related information displayed in each site. We also recruited front-line healthcare professionals with experience of working with PwD across diverse roles within the case study sites (three to seven per site). An additional five health professionals who worked with PwD were recruited through their willingness to take part in an interview after completing a national survey that formed part of the wider research programme20 to gain insight into practices beyond the four case study hospitals. Interviews with health professionals explored experiences of providing hospital care for PwD and how visual identifiers and related interventions worked in practice.

We also conducted interviews with PwD and their relatives. Participants were recruited through the ‘Join Dementia Research’ network23 and via social media. Interviews explored experiences of hospital care, and views on visual identifiers and related interventions. Patient and carer topic guides were piloted with a person with dementia and a carer (see online supplemental materials 1 and 2 for topic guides).

Supplemental material

Supplemental material

Due to the COVID-19 pandemic, all interviews were conducted by telephone or online, with written informed consent. All interviews were conducted by an experienced non-clinical qualitative researcher (ES), and lasted between 30 and 90 min. Interviewing continued until no new themes were identified. Interviews were audio recorded, transcribed verbatim and analysed using a thematic analysis approach.24 ES conducted initial open coding of transcripts by hand. This coding was inductive and included the way that identifiers could impact both positively and negatively on care quality. This informed the development of a coding framework, where codes were grouped into higher order themes (see coding frame in online supplemental material 3). This was refined and revised in discussion with CT, and used to code subsequent transcripts using NVivo V.12 software. We used compiled narrative case studies for each site. In each case, we mapped out the type of tools and approaches used to identify and support PwD. We interrogated the themes and case studies to explore the practices and meanings associated with the use of visual identifiers, and perceptions of their impact on quality of care for PwD. This was undertaken by comparing and contrasting the views of healthcare professionals, relatives of those with dementia and those living with dementia, and the documentation from each site. In particular, we contrasted the ‘official’ reports of how identifiers should be used in each site with what individuals told us about how they worked in practice and the consequences of their use.

Supplemental material


Twenty-one healthcare professionals, 19 relatives of PwD and two people living with dementia took part in the interviews. Quotation identifiers explain the type of person interviewed. For health professionals, identifiers also show either the case study they were from or that they were recruited from the survey. In this section, we first present findings on the case studies of how visual identifiers were being used in practice. We then articulate several ‘mechanisms of action’ through which the use of such identifiers could lead to better care and outcomes for PwD. Finally, we consider some of the potential failures and unintended consequences of such tools.

Visual identifiers: classifying patients and signalling category membership

The application of a visual identifier acted as a visible classification system25—identifying people as belonging to a specific group or category, that of ‘people with dementia’ (and/or with suspected dementia or delirium). Visual identifiers used across sites were multiple and varied, and included alerts on hospital electronic systems, stickers and magnets on notes and whiteboards, and bedside signs and wristbands displaying a symbol (eg, a forget-me-not flower or butterfly). Table 1, based on case study analysis of the four sites drawing on key documents and interviews with dementia leads, summarises the variation in use of identifiers across the sites.

Table 1

Variation in identifiers used across the case study sites

Mechanisms of action: how visual identifiers could support good care

We identified several mechanisms through which visual identifiers could positively support the delivery of good care for PwD: through enabling care coordination at organisational level, signalling eligibility for dementia-specific initiatives, allowing allocation of resources at ward level and acting as a quick reference cue to prompt assessment of needs. For detailed quotes see table 2.

Table 2

Mechanisms of action: how visual identifiers could support good care

The ability to mark people as belonging to the category ‘person with dementia’ opened up possibilities to improve coordination and delivery of care at organisational level. For example, flags on electronic systems could help identify and locate PwD within the hospital, which could then help to support the coordination of their care pathways, including avoiding unnecessary moves around the hospital. Identifiers also helped with efforts to audit, monitor and improve the quality of care for this patient group as a whole.

Identifiers on wristbands and patient documents also played a role in the delivery of hospital-wide initiatives for enhanced care for this patient group—such as meaningful activity coordinators (site 2), dementia volunteer support, fast tracking through X-ray pathways or additional support for nutritional or other needs. Identifiers acted as a marker that signalled their eligibility for these dementia-specific initiatives. Visual identifiers also acted to alert individual ward staff interacting with a patient, at the point of care, that a patient had dementia. This was perceived to have value, by both staff and relatives, in prompting ward staff to consider the needs and limitations associated with the membership of the category ‘person with dementia’. While this often meant drawing on generalised knowledge about dementia, rather than information about the individual patient, being able to identify someone as belonging to this category had the potential to improve interactions and quality of care. It could help staff to make sense of a person’s behaviour in the context of a dementia diagnosis, tailor their communication approach to avoid distress and take into account likely limitations in a person’s memory or communication ability.

At ward level, visual identification systems were seen as particularly helpful in enabling managers to organise and plan their resources for day-to-day care within the ward. A visual overview of the number of patients on the ward with dementia—from the e-system, whiteboard magnets or bedside identifiers—could provide accessible information to inform assessments of availability of staff with relevant expertise and guide allocation of staff or requests for additional staffing.

By providing a quick reference cue, identifiers on wristbands also supported efforts to protect the safety of patients in this group, reflecting the particular vulnerabilities of many PwD, for example, prompting staff to consider fall risks, or enabling identification of patients who may have left their ward and become lost.

Using visual identifiers: potential failures and unintended consequences

Although visual identifiers were recognised as having the potential to improve the coordination and delivery of care for PwD, we also identified challenges with the use of tools that could undermine their effectiveness in practice, and potentially generate negative consequences. For detailed quotes see table 3.

Table 3

Potential failures and unintended consequences of visual identifiers

The value of the identifier was dependent on collective understanding of its meaning and what it signified. But we found a lack of standardisation of symbols used across different sites which undermined their value as a quick reference tool (table 1 and table 3 for quotes), and no universal understanding of what the symbols meant across the workforce within and across sites. This lack of standardisation was perceived as particularly problematic when staff were locums or bank staff who might work at several different hospitals. The confusion that different identifiers, and ways of using them, might cause to patients and their families was also a concern. Additional resources were in place alongside visual identifiers in all hospitals we studied to provide staff with this personal information for PwD (in the form of a booklet or poster). These resources were designed to help staff better understand the interests and individual needs of a person. Staff felt, however, that personal information documents were poorly completed and used inconsistently. In addition, some accompanying personal information booklets, originally designed to be used in long-term care, were relatively detailed and dense. They focused on providing rich details about the person behind the diagnosis of dementia. Although specialist dementia staff and volunteers reported this was valuable in helping them find out more about the person, this format did not meet the needs of clinical staff for quick and easy access to key information. Booklets tended to be put away at the back of patients’ notes or at the bedside, and clinical staff tended not to prioritise taking time to read detailed personal information documents in the context of a time-pressured environment. In contrast, posters that provided personal information ‘at a glance’ provided limited information for dementia specialists and volunteers, but were felt to work well as a means of communicating relevant information about a patient to busy clinical, ward and support staff.

There were practical challenges with labelling patients, particularly in relation to the use of wristbands, bedside signs and whiteboard magnets. Across all the sites, many healthcare professionals were uncertain about whose responsibility it was to apply an identifier, and at what point in the patient’s journey this should happen. Staff reported a lack of consistency about the point in the patient’s journey at which they were applied and variable engagement across different types of ward. In practice, visual identifiers were not used consistently to flag eligible patients for the additional support that they needed. As a result of this lack of consistency, patients who should have benefited from additional dementia support could be overlooked. By default, the lack of an identifier implicitly classified them into the ‘non-dementia’ category, leading to inequalities in care provision. There were also ambiguities and differences of opinion about who was, and who should be, eligible to have a visual identifier applied (only patients with confirmed dementia, or patients with suspected dementia and delirium). This resulted in inconsistencies across hospitals in terms of which types of patients would have access to the pathways and initiatives designed for patients with additional needs.

Even if appropriately applied, identifiers in and of themselves did not guarantee better care. The identifier directed attention to the dementia diagnosis, but did not provide any information about the individual person behind the diagnosis, their abilities, needs and preferences. Staff described how classifying and identifying people as belonging to the category ‘person with dementia’ could lead to over-reliance on assumptions based on category membership. While awareness of the dementia diagnosis could alert staff to use a more dementia-friendly approach in their communication with a patient (as described above), it could also result in staff acting on assumptions about the limitations of patients with dementia, which could disempower patients and impact negatively on their independence and functional ability. Relatives recognised the limitations to what a visual identifier on its own could achieve, unless it was coupled with efforts to enable a more personalised response based on information about individual needs.

Although the staff, relatives and PwD we interviewed mostly felt that identifiers were potentially of benefit (even if this benefit was not fully realised in practice), some family members felt that their loved one would not want their diagnosis to be made visible. The use of an identifier was seen by some as having the potential to lead to negative consequences due to negative cultural representations of dementia. Some staff, relatives and PwD had concerns that marking patients with an identifier could result in prejudice and discrimination within the healthcare setting, if staff held negative beliefs about patients with dementia. Some relatives described their loved one experiencing discrimination in practice.

Both staff and relatives recognised that identifiers alone had limited potential to improve the care of PwD. Lack of in-depth training (particularly for bank or locum staff) on the skills and knowledge required to communicate and support PwD in hospitals, lack of local expertise in dementia and limited resources to provide the additional support required all made it more likely that an identifier would be ineffective in improving care. The extent to which identifiers could enable better care was dependent on their use being supported through staff training, resources directed at the needs of PwD and efforts to develop a more informed and supportive culture for caring for PwD.


Our research shows how classifying people as belonging to the ‘dementia’ category, and visibly signalling this category membership, could support improvements in the quality of care delivery for PwD in hospital. In particular, our participants perceived such classificatory tools as having the potential to play an important role in: enabling coordination of care at organisational level; signalling eligibility for dementia-specific interventions; informing prioritisation of resources on wards; and acting as a quick reference cue for staff interactions with patients. Provided staff have the skills and time to engage effectively with PwD, and are working within a supportive and well-resourced environment, they can use this information to tailor their interactions with patients in line with the broad needs of PwD. But such tools can fail to deliver on their promise, and potentially have negative consequences, when there is a lack of standardisation and consistency, a lack of closely coupled information about individual needs and a risk of discrimination associated with making the condition visible.

Featherstone and colleagues have been particularly critical of the use of signs and symbols for PwD in hospitals. In line with some of our findings, they argue that these ‘technologies of attention’ obscure the person, risk misclassification and can lead to depersonalised care.16 A recent national survey on the use of visual identifiers20 found drawbacks to the use of visual identifiers, such as concerns about disclosure and discrimination as well as practical challenges to their use. There is a danger, however, of arguing against the use of visual identifiers as an approach to optimising care for PwD on the basis of their potential to generate negative consequences. It is not that the principles of classification and labelling themselves are problematic—our research shows that being able to classify and mark patients as belonging to a vulnerable group with additional needs can (at least potentially) bring significant benefits in terms of optimising the organisation and delivery of care. Rather, we argue that we need instead to understand when and why such classification systems can generate negative consequences (their potential ‘dark logic’), and consider how to mitigate against these. This is particularly important as our and others’ work would seem to suggest that unintended consequences in this context are primarily on the outcomes of interest (so-called ‘paradoxical effects’), meaning that there is a very real risk that care for this vulnerable group of patients is made worse rather than better. Inconsistencies in the application of, and attribution of meaning to, the marker could create confusion and inequalities in care. We recommend that, at a national level, healthcare organisations gain consensus on the rules of classification, the process of applying the marker and the symbol to be used. To overcome the risks of assumptions and depersonalised care associated with a classification system, any markers applied to patients should act as a flag to staff to seek out personal information, rather than merely operate as a label to show a patient is a member of a category.

Our findings suggest that documents such as the ‘This is me’ booklet, designed to provide this personal information, were not immediately accessible for busy clinical staff and were rarely consulted by the staff providing care. Bedside posters provided a more easily accessible reference point for busy hospital staff. We suggest a need for more closely coupled personalised information to be integrated with the visual identifier. Such a resource should include the key information about individual needs and preferences that is most relevant to the person’s hospital care, and should be immediately accessible to all staff.

Simply implementing dementia-friendly initiatives such as visual identifiers and patient profiles without supporting staff with resources or training will fail to bring about good dementia care. The extent to which identifiers could enable better care was dependent on their use being supported through staff training, resources directed at the needs of PwD and efforts to develop a supportive culture for caring for PwD. Others have argued that dementia-friendly interventions such as visual identifiers need to be accompanied by efforts to build staff understanding of dementia, a framework for providing good dementia care and support from local change agents to bring about good care.26 Hospitals have to provide a supportive context for good dementia care to take place, and dementia care has to be embedded in an organisational commitment to optimise the delivery of care and ensure that systems are in place to support it. There is a real need for regular and in-depth training to enable effective responses to the classification of dementia. This training should seek to provide understanding of the common issues but also the spectrum of diversity in the group to avoid stereotyping, discrimination and stigma based on group membership. Impactful training should include input from patient and relative experiences and highly skilled facilitators.27 It should be available for all staff, including locum and agency staff. Finally, and very importantly, organisations need to ensure the engagement of patients and their carers in terms of their attitudes towards classification and their consent to be labelled.22 It is vital that discussions are held with relatives and patients as to the implications of being classified, having a marker applied and addressing concerns about stigma and discrimination.

Strengths and limitations

Our study included a small sample of four hospitals yet these all used a varying range of identifiers. In addition, our approach to site selection ensured we included sites with diversity in levels of engagement with their dementia strategies. Interviewees included dementia leads who were involved with the implementation of identifiers. While these staff are likely to have positive views of identifiers, we also interviewed front-line staff involved in the care of PwD to explore the realities of the way identifiers worked in day-to-day practice. We were only able to recruit two PwD to participate in interviews: we intended to recruit via in-person visits to hospitals, local groups and dementia cafes, but this was not possible due to the COVID-19 pandemic. We also were unable to undertake planned visits to hospitals to observe the use of identifiers in practice. Observation is a powerful method in uncovering the (often unconscious) gap between rhetoric and reality in dementia care,28 and future observational research would be of value. However, we were able to produce comprehensive case studies of local practices across four hospitals by interviewing a range of staff and drawing on a range of documentation used in each site.

Characterising visual identifiers as an example of a classification system provides us with new insights into the underlying principles of their use and the mechanisms through which they might work. Classification is argued to involve segmenting the world to conduct some form of work—in this case, the intention is to delineate a population of patients who are likely to require (and are eligible for) additional support while in hospital.25 However, as we have shown, the classificatory act in and of itself is not sufficient to deliver better care to those identified. In order to deliver meaningfully on their promise, classificatory tools such as visual identifiers need to be used in contexts which are capable of effectively delivering the work required to generate the desired outcomes of having classified.


Interviews with staff, PwD and their relatives reveal the mechanisms through which visual identifiers could play an important role in improving the quality of care for PwD. Characterising visual identifiers in terms of classification can help in understanding the mechanisms through which they can generate positive or negative consequences, as well as informing efforts to optimise their effectiveness in practice.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and ethical approval was obtained from the University of Leicester Medical and Biological Sciences Research Ethics Committee (28963), and governance approvals from the Health Research Authority (20/HRA/0653). Participants gave informed consent to participate in the study before taking part.


We thank all the participants for taking the time to talk to us, Janet Willars for her help in conducting the interviews and Karen Wildsmith and David Allen from our PPI group for their advice and guidance. We also thank Karolina Kuberska and Graham Martin for their valuable comments and insights.


Supplementary materials


  • Twitter @LizSule, @drnatarmstrong, @carolynctarrant

  • Contributors ES conducted the analysis and wrote the initial draft. NA, LL, SC and CT contributed to editing and revising the draft. CT led on the qualitative methodology and design and is guarantor.

  • Funding This study was funded by The Healthcare Improvement Studies Institute (THIS Institute), University of Cambridge. THIS Institute is supported by the Health Foundation, an independent charity committed to bringing about better health and healthcare for people in the UK. NA is supported by a Health Foundation Improvement Science Fellowship and also by the National Institute for Health and Care Research (NIHR) Applied Research Collaboration East Midlands (ARC EM).

  • Disclaimer The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

Linked Articles