Distinguishing between Knowledge Gaps and Misconceptions of Alzheimer’s disease among caregivers in the UK

A popular scale for assessing knowledge about Alzheimer‘s disease is the Alzheimer‘s Disease Knowledge Scale (ADKS). The aim of the study was to investigate the effect of adding  ̳don‘t know‘ to the original  ̳true‘ or  ̳false‘ response option. It was assumed that this modification would provide insight into the reasons underlying incorrect responses and could distinguish between misconceptions and knowledge gaps. To investigate this, carers (care home carers and informal carers) and members of the general population were recruited. The results showed that percentage correct responses was lower than previously reported, suggesting potential inflation of knowledge by guesses without the  ̳don‘t know‘ option. Moreover, care-home workers were more likely to select the incorrect response than  ̳don‘t know‘ compared to informal carers for several items related to the earlier stages of AD, suggesting a higher level of misconceptions around this topic and highlighting potential training needs for care home carers. DOI : COMING SOON Corresponding author: Petra M. J. Pollux (PhD), School of Psychology, University of Lincoln, Brayford Pool, Lincoln, LN6 7TS, E-mail: ppollux@lincoln.ac.uk, Tel: +44(0)1522886360


Introduction
With a growing global population, prevalence of dementia is increasing world-wide. It is estimated that approximately fifty percent of all residents in long-term care settings have dementia [1, 2] and the resulting heterogeneity in behaviours of residents can challenge the demands on knowledge and skills of carers in residential care homes [3, 4,5]. Knowledge of dementia and Alzheimer's disease (AD) has been shown to vary among caregivers in residential settings and among informal carers [6,7]. Poor understanding of the impact of dementia on patient behaviour can result in suboptimal care-choices and can increase agitation or anxiety unnecessarily [8,9,10]. Reliable assessment of knowledge among carers is therefore paramount for patient well-being and for guiding decisions about training strategies by residential care homes or for development of training programs for informal carers [8,11].
A popular scale for assessing knowledge and misconceptions of Alzheimer's disease is the Alzheimer's Disease Knowledge Scale (ADKS [11]). The ADKS consists of 30 ‗true/false' statements related to seven ‗knowledge domains': risk factors, symptoms, assessment and diagnosis, course of the disease, treatment and management, caregiving and life impact.
So far, the questionnaire has been used to assess knowledge in the general population or among students (UK [12]; Norway [13]; USA [14,15,16]; South Korea [17] among health professionals and care staff [18], among psychologists [19] and it has been used as an outcome measure for validation of an online training program for care-workers [20].
One potential limitation of the ADKS is that ‗true/false' responses in knowledge questionnaires provide only limited information about the reasons why an incorrect response is given [21,22,23,24]

Materials
The ADKS (Carpenter et al, 2009) assesses knowledge about seven knowledge domains: risk factors (e.g., ‗Genes can only partially account for the development of AD'), symptoms (e.g., ‗Most people with AD remember recent events better than things that happened in the past'), assessment and diagnosis (e.g., ‗AD is one type of dementia'), caregiving (e.g., ‗People with AD do best with simple instructions, given one step at a time'), treatment and management (e.g., ‗AD cannot be cured'), life impact (e.g., ‗It is safe for people with AD to drive, as long as they have a companion in the car at all times') and course of the disease (e.g., In rare cases, people have recovered from AD').

Group formation
For analysis of ADKS responses as a function of experience with AD, participants were organised into four groups: Group 1 consisted of people who work in care homes as caregivers (70 females, 1 male), group 2 were all caregivers recruited via the AD society who at the time of participation cared for a spouse with Alzheimer's disease at home (informal carers: 14 females, 6 males). Of the volunteers recruited via care homes, 63.6% were caregivers whereas the remaining participants were either involved in housekeeping or had other roles (e.g. hair-dressing). We categorized this latter group as participants who have contact with people with Alzheimer's disease at their workplace but who are not caregivers. These volunteers were combined with participants from the general population who also have contact with AD patients at work but who are not involved in direct care of people with AD on a daily basis (e.g. social workers, nurses, police officers).
All participants in this combined group answered ‗yes' to E-Q3 and ‗no' to E-Q2 (in total 51 females, 11 males).
The fourth group consisted of volunteers recruited from the general population who were not in the caregiving profession and who had no contact with people with AD at work or cared for people with AD at home (37 females, 33 males): All volunteers in this group answered ‗no' to E-Q2 and E-Q3. Details about the four groups are presented in Table 1.

Response profiles for each knowledge domain
Percentages correct, incorrect and ‗don't know' responses were compared between the four groups, separately for each knowledge domain, using one-way ANOVA and post-hoc pair-wise comparisons with Bonferroni corrections (see Figure 1).  with percentage correct responses lower than 55% (Q3, 7,9,12,14,18,20,22,26). Associations between these frequency distributions and the variable Group were analysed using chi-square tests of associations. Fisher's exact test was used when any expected cell value was lower than 5. These analyses showed significant

Additional analyses
Correlations: Table 3 shows that ADKS scores correlated significantly with age and self-reported knowledge but not with education. A few correlations may be specific to the sample used in the present study: Age of the participant correlated positively with ADKS score, negatively with education and positively with selfreported knowledge (see Discussion).

Effect of the 'don't know' option on percentage correct and incorrect responses: Percentages correct and incorrect responses of Groups 3 and 4 were
directly compared with responses collected previously from a sample from the general population using the ADKS version with only ‗true' and ‗false' response options [12].

Discussion
The present study used a modified version of the ADKS to assess knowledge about AD among carers and in a sample of the general population. Inclusion of the ‗don't know' option (selected ~23% of the time) reduced both correct responses (-10%) and incorrect responses (-11%) in the sample from the general population compared to percentages reported in our previous study [12] suggesting that percentage correct responses may be a better index of true knowledge about AD when the ‗don't know' option is included.   and non-carers alike. These findings seem to suggest that scientific advances on the relationship between cholesterol, hypertension and AD [26], including new insights into the potential positive effect of statins in treatment plans [27,28] are not effectively communicated to informal carers, care home workers and the general population. Given that both our studies were conducted in the UK implies little awareness about these factors in the British population, despite ongoing national dementia awareness campaigns [29,30].
A few results were consistent with previous findings [18,9]  Higher ADKS responses were also associated with those who rated their own knowledge higher, replicating our previous findings and those obtained from Australian health professionals [12,18]. Conversely, a few correlations may have been unique to the sample used in the current study and require further consideration. Table 3 shows that age of the participant correlated positively with ADKS score, negatively with education and positively with self-reported knowledge. This pattern of correlations is likely due to differences between groups in terms of age (reducing from Group 1 to 4) and level of education (increasing from Group 1 to 4) (see Table 1), which could be explained by the relatively high proportion of university students in this group (N=26) compared Group 3 (N=6) and Groups 1 or 2 (none).
Although multivariate regression analysis would have been informative for exploring the predictive value of group membership (Group 1-4), age, education and selfreported knowledge, multicollinearity can be problematic for interpretation of regression models [32]. Future studies investigating knowledge of AD in carers may consider matching the control group more carefully with the carers in terms of age and education.

Conclusion
The present study demonstrates that inclusion of the ‗don't know option in the ADKS provides valuable insight into the reasons why respondents select incorrect responses (misconceptions versus knowledge gaps). The study revealed that care-home carers were more likely to select the incorrect response with confidence for several items, suggesting a higher level of misconceptions than for informal carers. Possible itemspecific explanations were considered, highlighting training needs for improving knowledge of care home workers about the impact of AD on behaviour at earlier stages of the disease.