Participant’s characteristics
A total of 1795 hospital staff completed the survey. However, the sample size is reported in parenthesis if there was missing data. It was not possible to determine the response rate as the survey was distributed widely and the total number of survey recipients is unknown. The demographic characteristics of the participants are displayed in Table 1. Participants were mostly female (83.2%, n = 1791) and aged between 35 and 57 years (56.7%, n = 1246). Approximately a quarter of the participants did not provide their age range. The most common discipline was nursing (41.4%, n = 1783). We merged the occupations into meaningful groups for analysis, these being ‘doctors, nurses and healthcare attendants’ (50.8%, n = 1783), ‘allied professionals’ (11.8%, n=1783) and ‘general support staff’ (37.4%, n=1783). Most staff reported not having any previous dementia training (76.5%, n=1791) and more than half (62.1%, n=1474) indicated that they do not have a relative with dementia.
Table 1. Characteristics of hospital staff
Variable
|
n (%)
|
Age
-18-35 years
-35-55 years
-55-75 years
|
N = 1246
421 (33.79%)
706 (56.66%)
119 (9.55%)
|
Gender
|
N = 1791
300 (16.8%)
1491 (83.2%)
|
Job Role
- Doctors, nurses, healthcare attendants
- Allied professionals
- General support staff
|
N = 1783
906 (50.8%)
210 (11.8%)
667 (37.4%)
|
Previous dementia training
|
N = 1791
420 (23.5%)
1371 (76.5%)
|
Family member with dementia
|
N = 1474
558 (37.9%)
916 (62.1%)
|
Attitudes toward dementia
1. Descriptives
The average total ADQ score was 70.64 (Sd = 8.60) (maximum total ADQ score = 95), with an average item score of 3.72 (i.e. the total score divided by the number of items) (maximum average item score = 5). The average total hope subscale score was 24.88 (Sd = 5.41) (maximum total hope subscale score = 40) with an average item score of 3.11 (maximum average item score = 5) and the average total person-centred subscale score was 45.77 (Sd = 5.16) (maximum total person-centred subscale score = 55) with an average item score of 4.16 (maximum average item score = 5). A paired sample t-test on the average subscale scores revealed a significantly higher average score for the ADQ subscale person-centred attitude compared to hope attitude (t(1776 ) = -142.66, p < .001).
2. Multiple regression
The multiple regression analysis indicated that both ‘previous dementia training’ (β = 4.61, SE = .51, t = 9.10, p < .001) and ‘having a relative with dementia’ (β = 2.50, SE = .46, t = 5.40, p < .001) both significantly contributed to a positive attitude towards dementia as measured with the ADQ (R2 = .07, F(2,1458) = 56.46, p < .001).
3. Moderator effects
A series of two-way ANOVAs were conducted to investigate whether ‘previous dementia training’ or ‘having a relative with dementia’ had an impact on the participants’ attitude towards dementia (hope subscale score, person-centred subscale score and total ADQ score) within different demographical groups, namely gender (male versus female), age (18-54.9 years old versus 55-75 years old) and job roles (doctors, nurses and healthcare attendants versus allied professionals versus general support staff).
The two-way ANOVAs examining whether the interaction between ‘previous dementia training’ and gender had an impact on attitude towards dementia did not result in a significant interaction effect for the hope subscale score (F(1,1770) = 0.58, p = .45), but did show a significant interaction effect for the person-centred subscale score (F(1, 1772) = 6.56, p = .01) (see Figure 1). For the interaction between ‘dementia training’ and gender, contrast analysis revealed that males who did not have previous dementia training scored significantly lower on the person-centred subscale (M = 43.45, Sd = 5.67) compared to females who did not have previous dementia training (M = 45.52, Sd = 4.8) (F(1, 1772) = 32.48, p < .001). However, for participants who had previous dementia training, there was no difference in person-centred subscale scores found between males and females based on contrast analysis (F(1, 1772) = .02, p = .88). For the interaction ‘previous dementia training’ and gender on the total ADQ score the effect was close to significance (F(1, 1766) = 3.97, p = .05).
Regarding the interaction ‘previous dementia training’ and ‘job role’, a significant interaction effect was demonstrated for the hope subscale (F(2,1761) = 5.56, p = .004) (see Figure 2). For the interaction of ‘previous dementia training’ and job role for the hope attitude, contrast analysis showed that there was a significant difference in hope subscale scores between having had previous training versus having no previous training for the doctors, nurses and healthcare attendants (F(1,1761) = 51.26, p< .001) and for the allied professionals (F(1,1761) = 5.37, p = .002), where the hope subscale scores were higher for the group that had previous dementia training. For staff who had previous dementia training, contrast analysis only showed differences between allied professional and general support staff (F(1,1761) = 7.22, p = .007), where allied professionals scored significantly higher on the hope subscale score (M = 27.45, Sd = 5.56) compared to the general support staff (M = 24.83, Sd = 6.19). No differences were found in the hope subscale score between the other job roles for staff who had previous dementia training using contrast analysis: doctors, nurses and healthcare attendants versus allied professionals (F(1,1761) = 4.42, p = .12) and comparing doctors, nurses and healthcare attendants versus general support staff (F(1,1761) = 3.62, p = .06). For staff who did not have previous dementia training, differences in hope subscale score were demonstrated using contrast analysis when comparing doctors, nurses and healthcare attendants with allied professionals (F(1,1761) = 15.95, p < .001) and when comparing doctors, nurses and healthcare attendants with general support staff (F(1,1761) = 18.11, p < .001), where doctors, nurses and healthcare attendants (M = 23.63, Sd = 5.86) scored significantly lower compared to the other job roles (allied professionals (M = 25.62, Sd = 5.52) and general support staff (M = 24.93, Sd = 4.36)), while no differences in hope subscale scores were found between allied professionals and general support staff (F(1,1761) = 1.91, p = .17).
Perceived dementia knowledge
1. Descriptive results
The average score on the perceived dementia knowledge questions was 4.63 (Sd=1.82) on a total scale of 10.
2. Multiple regression
Results of the multiple regression analysis indicated that both ‘previous dementia training’ (β = 1.49, SE = .10, t = 14.87, p < .001) and ‘having a relative with dementia’ (β = .36, SE = .09, t = 3.92, p < .001) both significantly contributed to their perceived dementia knowledge (R2 = .14, F(2, 1455) =118.71, p < 0.001).
3. Moderator effects
A series of two-way ANOVAs were conducted to investigate whether ‘previous dementia training’ or ‘having a relative with dementia’ have an impact on their perceived dementia knowledge within different demographic groups, namely gender (male versus female), age (18-54.9 years old versus 55-75 years old) and job roles (doctors, nurses and healthcare attendants versus allied professionals versus general support staff).
The two-way ANOVA examining whether the interaction between ‘previous dementia training’ and gender had an impact on dementia knowledge resulted in a significant effect (F(1,1455) = 9.54 , p = .002) (Figure 3). Contrast analysis demonstrated that for staff who did not have previous training males had a significantly lower perceived dementia knowledge score (M = 3.97, Sd = 1.63) compared to females (M = 4.29, Sd = 1.65) (F(1,1455) = 4.68, p = .03). For staff who had previous dementia training females had a significantly lower perceived dementia knowledge score (M = 5.66, Sd = 1.79) compared to men (M = 6.2, Sd = 1.80) (F(1,1455) = 5.26, p = .02).
Regarding the interaction between ‘previous dementia training’ and age (F(1,1322) = .002, p = .96) and the interaction between ‘previous dementia training’ and job role (F(2,1457) = .13, p = .88), no interaction effects were demonstrated on perceived dementia knowledge. Furthermore, investigating the interactions between ‘having a relative with dementia’ and gender (F(1,1454) = 1.60, p = .21), between ‘having a relative with dementia’ and age (F(1,1321) = .27, p = .60), and between ‘having a relative with dementia’ and job role (F(2,1456) = .34, p = .72), did not result in any significant interaction effects on perceived dementia knowledge.