Carer burden and its predictors have been widely studied in dementia.Reference Campbell, Wright, Oyebode, Job, Crome and Bentham1–Reference Etters, Goodall and Harrison3 Studies on the multidimensional measurement of carer burden often fail to consider the contribution of stigma to burden, or the positive aspects of caring that may act to protect family carers.Reference Saffari, Lin, Koenig, O'Garo, Broström and Pakpour4 Stigma directed at family carers of a stigmatised individual is construed as ‘stigma by association’ or ‘courtesy stigma’.Reference Östman and Kjellin5 When stigma by association becomes internalised, termed ‘affiliate stigma’, it can have negative affective, behavioural and cognitive consequences, such as unhappiness, withdrawal and sense of inferiority.Reference Su and Chang6 Important processes that contribute to or mitigate carer burden, such as stigma by association and positive aspects of caregiving, are absent from instruments used to examine affiliate stigma in carers of people living with dementia.Reference Werner, Mittelman, Goldstein and Heinik2,Reference Chang, Su and Lin7 One instrument that does incorporate these processes is the Family Stigma Instrument (FAMSI), devised for use with family carers of people with intellectual and developmental disabilities (IDDs).Reference Mitter, Ali and Scior8 The FAMSI is a conceptually driven instrument grounded in both stigma theory and positive psychology approaches to caregiving. As such, it potentially has added utility for understanding the fine balance between positive and negative consequences of dementia for caregiver burden. Our aim was to investigate the validity of the FAMSI in dementia care, and the extent to which experiences of stigma are endorsed in a UK population of family carers of people living with dementia.
Method
Setting and participants
Participants were a convenience sample in South-East England, UK. They were recruited via the Join Dementia Research database (https://www.joindementiaresearch.nihr.ac.uk/) and through social media, community advertising or researcher outreach activities where the study was presented to carers’ groups (e.g. Alzheimer's Society groups). Participants were included if they were aged ≥18 years; were a family carer for someone with a primary progressive dementia; and were able to understand, read and write the English language. Participants were excluded if they were a former carer; had a late-stage chronic, terminal medical condition or a significant sensory impairment that would preclude completion of the study measures; or lacked the capacity to consent. Sample size was calculated based on guidanceReference Terwee, de Vet, Knol, van der Windt, Knol and Dekker9 where seven multiplied by the number of FAMSI items gave the target sample size (e.g. 7 × 26 = 182 participants).
Data collection
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. All procedures involving human patients were approved by University College London Research Ethics Committee (approval number 11501_002).
Data were collected through face-to-face interviews and online questionnaires, according to participant preference and geographical locality. Potential participants were given a study information sheet and at least 24 h to consider participating, before taking part either face to face with a researcher or independently online. The corresponding author or one of two Master's students carried out face-to-face data collection. Qualtrics version 19 (Qualtrics, Provo, UT, USA, https://www.qualtrics.com) was used for online data collection, with participants accessing the participant information sheet, screening questions, consent form and study measures through a survey link. All participants were asked to complete a retest, for which they completed the measures again after a 2-week period (time point 2) in the same format in which they had completed them initially (time point 1). This retest was to assess the stability of the measure over a 2-week period, which is sufficient time to balance between recollection bias and unwanted change.Reference Marx, Menezes, Horovitz, Jones and Warren10
Measures
FAMSI
The FAMSIReference Mitter, Ali and Scior8 is a 26-item instrument with two stigma-focused subscales of stigma by association (eight items) and affiliate stigma (affective, behavioural and ‘perceived’ domains with four items each), and a third subscale concerned with positive aspects of caring (six items). Items relating to stigma by association began with the phrase, ‘some people might’, e.g. ‘ … feel embarrassed about associating with them’. Items on the affiliate stigma subscale began with the phrases, ‘I feel … ’, ‘I avoid …. ’ or ‘I am … ’ for the affective, behavioural and perceived domains, respectively. For example, ‘I feel embarrassed about them’ (affective), ‘I avoid introducing my friends to them’ (behavioural), or ‘I am excluded from activities when other people find out about their dementia’ (perceived). The affiliate stigma questions were framed such that ‘them’ or ‘their’ referred to the person living with dementia. The subscale pertaining to perceived affiliate stigma, which Mitter et alReference Mitter, Ali and Scior8 originally labelled ‘cognitive’, was changed to ‘perceived’ as items for this domain reflect family carers’ perceptions of how others treat them. Positive aspects of caregiving items began with the phrase ‘Caring for my family member .… ’, e.g. ‘ … has enabled me to develop a more positive attitude toward life’. Response options for all items were on a five-point scale from ‘strongly disagree’ (1) to ‘strongly agree’ (5), with a midpoint of ‘neither agree nor disagree’ (3). FAMSI instructions and items were adapted to be relevant for carers of people living with dementia by replacing the term ‘intellectual and developmental disability’ with dementia. The original validation of the FAMSI with family carers of people with IDDs demonstrated adequate to good internal consistency for all subscales (Cronbach's alpha range 0.77–0.91). However, only the positive aspects of caregiving and affective affiliate subscales were stable over time (intraclass correlation coefficient (ICC) >0.7).
Rosenberg Self-Esteem Scale
The Rosenberg Self-Esteem Scale (RSES)Reference Rosenberg11 was used to measure self-esteem of carers. Previous literature has documented an inverse relationship between self-esteem and stigma in dementia, HIV/AIDs and cancer.Reference Burgener and Berger12–Reference Bhatt, Stoner, Scior and Charlesworth14 Therefore, it was hypothesised that stigma by association and affiliate stigma would be negatively correlated with self-esteem, whereas the positive aspects of caregiving subscale would be positively correlated.
Demographics questionnaire
The following demographic information was collected through carer report: gender, age, kin relationship to person living with dementia, ethnicity, cohabitation status, employment status and whether English is their first language. In addition, participants were asked the age and dementia subtype of the care recipient, and time since diagnosis.
Data analysis
Data distributions of the FAMSI (floor/ceiling effects, normality), internal consistency (Cronbach's alpha cut-off value ≥0.70, indicating adequate internal consistency), test–retest reliability (ICC cut-off value ≥0.70, indicating adequate test–retest reliability) and convergent validity with the RSES were analysed in accordance with psychometric guidelines,Reference Terwee, de Vet, Knol, van der Windt, Knol and Dekker9 using SPSS version 26 for Windows. It was hypothesised that stigma by association and affiliate stigma would be negatively correlated with RSES score, and positive aspects of caregiving would be positively correlated with RSES score. Percentages for each response option were compared for all FAMSI items, to explore which FAMSI domains were endorsed over others. Relationships between domains were explored through Pearson's correlations, as per previous analysis.Reference Mitter, Ali and Scior8
Results
Sample characteristics
Seventy carers of people living with dementia met the eligibility criteria and provided informed consent. Four participants took part face to face and 66 participants took part online (see Table 1). An additional three online participants dropped out during screening and one online participant was excluded because of a high level of incomplete data (>30% of items unanswered). Across all measures, levels of missing data were low (<10%), indicative of high completion rates and acceptability and suitability of the measures in this population. A Little's missing completely at random test was non-significant for each measure (p = 0.623), indicating data were missing completely at random. Therefore mean imputation at an item level was performed to deal with the low levels of missing data.Reference Graham15,Reference Eekhout16
Participants were on average 60 years of age and predominantly White (92%), female (76%) and caring for their own or a spouse's parent (54%). The person they cared for was most commonly living with Alzheimer's disease (37%) or dementia of mixed aetiology (24%).
Assessment of psychometric properties of FAMSI domains
FAMSI subscales were normally distributed, with the exception of the stigma by association subscale where data were moderately negatively skewed (−1.255), although this was below the absolute skew limit of −2 (see Table 2). Analysis indicated some potential floor effects of the affective affiliate stigma and behavioural affiliate stigma subscales, with 32.9% and 30% of the sample achieving the lowest score for these measures, respectively. All FAMSI scales had adequate test–retest reliability and internal consistency.Reference Terwee, de Vet, Knol, van der Windt, Knol and Dekker9 No significant correlations between the FAMSI scales and RSES were observed, indicating a lack of convergent validity.
FAMSI, Family Stigma Instrument; ICC, intraclass correlation coefficient.
a. Use of non-parametric test.
Endorsement of FAMSI domains
Response options endorsed by carers for each item of the FAMSI are presented in Table 3. Overall, carers agreed with statements in the stigma by association scale. Overall, the participants rejected the affective and behavioural domains of affiliate stigma, suggesting that they were able to resist the potential negative emotional impact of stigma and did not change their own behaviour when in social situations. In contrast, they tended to endorse items in the perceived domain, indicating that they perceived a change in others’ behaviour toward them as a result of their relative's diagnosis of dementia. Although there was some agreement that caring for someone with dementia allowed them to form friendships with others in a similar situation, no other positive aspects of caring were either strongly endorsed or rejected by the sample as a whole.
FAMSI, Family Stigma Instrument.
Relationships between FAMSI domains
A moderate significant correlation between stigma by association and both affiliate stigma (r = 0.489, P < 0.001) and perceived affiliate stigma (r = 0.624, P < 0.001) was observed, indicating that carers who endorsed stigma by association items were also more likely to experience affiliate stigma. Perceived affiliate stigma was positively correlated with affective (r = 0.272, P < 0.05) and behavioural family stigma (r = 0.300, P < 0.05), and behavioural and affective family stigma were positively correlated with each other (r = 0.670, P < 0.001).
Discussion
Stigma in dementia caregiving is an emerging area of research.Reference Su and Chang6,Reference Chang, Su and Lin7 This study is, to our knowledge, the first quantitative exploration of stigma by association, affiliate stigma and positive aspects of caregiving with family carers of people living with dementia in the UK. This study adds to the notion of the ‘double effect’ of dementia on experiences of stigma, where people living with dementia experience stigmaReference Bhatt, Stoner, Scior and Charlesworth14 and the majority of their carers also experience stigma. Our results show that carers endorsed stigma by association, suggesting they were aware that others viewed or treated them in a stigmatising fashion; however, the lack of endorsement of affiliate stigma constructs (e.g. behavioural and affective) suggests that carers may not have experienced the internalised consequences of stigma as strongly, and reasons for this are discussed below. Alongside stigma, many carers reported positive experiences within caregiving. This suggests a complex relationship between stigma experiences and caregiving, whereby carers experience and are aware of both positive and negative aspects of their role, and therefore positive aspects of caregiving may mediate or buffer against stigma. This confirms other work noting the coexistence of positive and negative aspects in caregiving.Reference Crellin, Orrell, McDermott and Charlesworth17
Polarising versus balancing the experiences of carers
Three instruments are available to quantify affiliate stigma, namely the Family Stigma in Alzheimer's Disease Scale,Reference Werner, Mittelman, Goldstein and Heinik2 the Affiliate Stigma ScaleReference Chang, Su and Lin7 (developed for caregivers of people with mental health challenges) and the FAMSI (developed for people with IDDs).Reference Mitter, Ali and Scior8 Limitations of the first two instruments include a focus on negative experiencesReference Werner, Mittelman, Goldstein and Heinik2 or neglect of stigma by association.Reference Chang, Su and Lin7 To further our understanding of carer burden, stigma and the balance of negative and positive experiences in families, we chose to examine the psychometric properties of the FAMSI as it is the only instrument that encompasses stigma by association and affiliate stigma, as well as positive aspects of caregiving. The present psychometric evaluation of the FAMSI in dementia care suggests good validity for use with families. Analysis indicates that it has acceptable internal consistency and content validity, with the latter evidenced by significant correlations between total FAMSI scores and its subscales.
Our validation of the FAMSI measure in dementia caregiving offers new avenues for further exploration of the wide-ranging experiences of stigma, burden and emerging positive psychology approaches.Reference Quinn, Clarke and Wolverson18 For example, future research with the FAMSI could investigate how positive frames in caregiving may counteract stigma by association, stigmatised beliefs or feelings of shame that contribute to increased burden and reports of behavioural and psychological symptoms in dementia.Reference Feast, Orrell, Charlesworth, Melunsky, Poland and Moniz-Cook19 Although stigma resistance has been explored across mental health conditions, there are no such explorations in the field of dementia caregiving. The FAMSI is an appropriate tool for further examination of this and the relationship between positive aspects of caregiving and stigma by association in dementia.
The theoretical model of sense of competence, which incorporates burden and positive aspects of caregiving, has been prevalent in the dementia caregiving literature for some time. However, despite significant experience of stigma by association, as documented here, this concept has yet to be included in formal models of caregiving. As such, this research represents the first step in quantitatively documenting stigma by association in caregivers and its relationship with positive aspects of caregiving.
Relevance of stigma to carers of people living with dementia
This study provides scope for further use of the FAMSI in conceptual work, as well as being mindful in practice that stigma is often felt by those living with dementia and those caring for them.20 Experiences of perceived affiliate stigma were more heavily endorsed compared with affective or behavioural affiliate stigma. Carers were almost eight times more likely to report perceived affiliate stigma compared with other types of affiliate stigma. Of note is our adaption of the label, but not the item descriptions, of the affiliate stigma scale, i.e. ‘cognitive’ altered to the ‘perceived’ domain. This is in line with the dementia literature where perception is seen as a particular aspect of cognition relating to a person's view or understanding of their experience.Reference Feast, Orrell, Charlesworth, Melunsky, Poland and Moniz-Cook19 In this case, items of this domain reflected the perceptions of carers and how they experienced the behaviours of others as stigmatising (e.g. I am treated different by some people when I am with them), rather than reflecting the cognitions of carers that may be internalised as a result of being stigmatised.
Similar to previous findings in the IDD field,Reference Mitter, Ali and Scior8 more than half of the carers perceived the role of being a family carer as stigmatising. Many carers reported positive aspects of their caring roles. It is possible, for example, that if carers frame their role positively, they may be more likely to be able to resist shame and other negative feelings that have been associated within their reports of ‘behavioural and psychological symptoms in dementia’Reference Feast, Orrell, Charlesworth, Melunsky, Poland and Moniz-Cook19,Reference Quinn, Nelis, Martyr, Morris, Victor and Clare21 during interactions with their relative with a dementia or within their caring role. Thus, in their journey of dementia care, they may become less vulnerable to these aspects of affiliate stigma. This may also explain why, as compared with other domains, there was strong endorsement of perceived affiliate stigma, since positive framing of a caring role or stigma resistanceReference Firmin, Luther and Lysaker22 may be less likely to affect experiences of affiliate stigma. These findings suggest that affective and behavioural affiliate stigma reflect the carer's own feelings and responses, whereas perceived affiliate stigma is more reliant on external factors such as the carer's perception of the behaviour of others toward them within their social environment.
Limitations
Our study has several limitations that should be considered for future research. First, despite efforts to recruit a large and diverse sample, we did not meet our recruitment target, and the majority of participants were White and lacked ethnic diversity. Given knowledge of stigma experiences in family carers and ethnicity in the UK,Reference Mackenzie23 it will be important to examine the validity of the FAMSI with distinct groups of family carers in the UK. Also the validity of the FAMSI cannot be generalised to family stigma and dementia care in other countries and cultures.Reference Leung and Cohen24
Second, the FAMSI was originally developed to understand the experiences of family carers of people with IDDs and adapted for the dementia field, and not developed as a scale unique to the experiences of family carers of people living with dementia. It is conceivable that some of their specific concerns in relation to stigma may have been missed as a result.
Finally, self-esteem scores were not significantly correlated with stigma by association as measured by the FAMSI. This may have been because of the small sample size, but since the link between the concepts of self-esteem and stigma is based on research on the individual with a stigmatising condition, it is also possible that self-esteem is not theoretically related to stigma by association.Reference Burgener, Buckwalter, Perkhounkova and Liu25 For instance, carers are fundamental to ensuring that the needs of the person they care for are met on a daily basis. This has strong connotations with a sense of purpose and it is possible that, having been primed with the FAMSI, carers completed the RSES foregrounding a salient carer identity rather than other personal identities. Future studies could explore measurement of other concepts related to stigma by association, such as its role in the management of unmet need in people with dementia and behavioural and psychological symptoms in dementia. Methodology could also involve randomisation of the order in which measures are presented to participants.
Implications for clinical practice and future research
This research has laid the foundation for use of the FAMSI in dementia care, offering an instrument to further our understanding of stigma felt by family caregivers. Prevalent models of burden neglect stigma and its potential consequences for family caregiving. Theoretical models should be further developed to investigate stigma and its relationship with distress and burden, as well as positive psychological care, in family care settings.
Our study notes that the majority of carers experience stigma, but there are no widely implemented interventions targeting this in dementia care. This study has implications for health and social care practice, where the experience of carers should be acknowledged in formal care pathways and relevant support offered. For example, psychosocial factors have a known effect on behavioural and psychological symptoms in dementia (e.g. neuropsychiatric disturbances, distress in families),Reference Feast, Orrell, Russell, Charlesworth and Moniz-Cook26 where the roles of stigmatised beliefs and shame have also been implicated.Reference Feast, Orrell, Charlesworth, Melunsky, Poland and Moniz-Cook19 Using the FAMSI in formal assessments provides scope for understanding the hidden needs of families. Tailored stigma-neutralising interventions for family carers vulnerable to the effects of stigma need to be developed, and their impact on ongoing burden evaluated within longitudinal studies.
Future research
Further research in this field should aim to identify the impact of positive aspects of caregiving, stigma by association and affiliate stigma on carers, to determine vulnerabilities that may contribute to caregiver burden. Studies could examine hypotheses predicting that greater perceptions of positive aspects of caregiving have health and social benefits for carers. Recent literature has found that positive aspects of caregiving have been associated with benefits for carers, but not people living with dementia.Reference Quinn, Nelis, Martyr, Morris, Victor and Clare21 However, past literature has not fully incorporated stigma into both parts of the dyad or examined its relative contribution to quality of life for people living with dementia and their families. In addition, future research should aim to test the factor structure of the FAMSI.
Data availability
The data that support the findings of this study are available from the corresponding author, J.B., upon reasonable request.
Acknowledgements
We thank the carers of people living with dementia who participated in this research. We would also like to thank Rebecca Light and Ciara Rafter for their assistance in data collection.
Author contributions
J.B., K.S. and G.C. conceptualised/designed the study. J.B. acquired the data. J.B. and C.R.S. analysed the data. E.M.-C. interpreted the data. All authors drafted the manuscript.
Funding
This work was supported by the Alzheimer's Society (grant number AS-PhD-16–022 to G.C.), Economic and Social Research Council (ESRC; grant number ES/J500185/1) and Promoting Independence in Dementia Project (ESRC grant number ES/L001802/2 to E.M.-C. and G.C.).
Declaration of interest
None.
eLetters
No eLetters have been published for this article.