In total, 57 schools were eligible for the study of which 10 declined to participate (all in municipality one - M1), a decision taken by the school principal mainly due to lack of time. A flowchart of participating schools and school personnel over time is shown in Fig. 1 and characteristics of school personnel are shown in Table 3.
Table 3
Characteristics of school personnel pre-intervention by municipality.
Participant characteristics
|
Total
n = 150
|
Municipality 1 n = 45
|
Municipality 2 n = 41
|
Municipality 3 n = 64
|
P-value
|
Age mean years (SD)
|
46.4 (9.7)
|
48.5 (6.3)
|
46.4 (9.9)
|
44.9 (11.2)
|
0.16*
|
Female n (% in municipality)
|
140 (93)
|
41 (91)
|
39 (95)
|
60 (94)
|
0.39**
|
Professional title
|
|
|
|
|
|
Principal/Vice-Principal n (%)
|
39 (26)
|
11 (24)
|
11 (27)
|
17 (27)
|
0.86**
|
Teachers n (%)
|
72 (48)
|
23 (51)
|
17 (41)
|
32 (50)
|
|
School Nurses n (%)
|
39 (26)
|
11 (24)
|
13 (32)
|
15 (23)
|
|
Work experience in the field median years (IQR)
|
7 (3, 15)
|
11 (4, 20)
|
5 (2, 15)
|
6 (2,10.5)
|
0.02†
|
Work experience in the current school median years (IQR)
|
4 (2, 6)
|
4 (2, 6)
|
4 (2, 8)
|
3 (1, 6)
|
0.50†
|
SD: standard deviation; IQR: Interquartile range. Organizational instability per school was defined as any change to school personnel implementing the HSS program. *ANOVA; **Pearson’s Chi-Squared test; †Kruskal-Wallis |
Table 3. Characteristics of school personnel pre-intervention by municipality.
Figure 1: Flow chart describing the process of enrolling schools and personnel in the three municipalities
Out of the 229 individuals meeting the inclusion criteria, 150 (66%) filled the pre-intervention survey. Table 2 shows relevant characteristics of the school personnel. The only significant difference between the municipalities was the number of years of work experience among the staff which was significantly higher in M1 compared to M2 and M3 (p = 0.02). More than half of the schools (n = 7/13, 54%) in M2 experienced organizational instability compared to 40% (n = 6/15) in M1 and 16% (n = 3/19) in M3.
Readiness
When comparing differences between municipalities pre-intervention, Leadership readiness (LRIT) was high in M1 (39.5, IQR:30,43) and M3 (45.5, IQR: 40.5,50.5) (Fig. 2), however there was a significant decrease of 14 and 9 points respectively (p = 0.02 and 0.045) post-intervention for both municipalities. While M2’s leadership score was low compared to M1 and M3 at the pre-intervention time point (32, IQR:20,44), it significantly (p = 0.04) increased post-intervention (42.5, IQR:33,45).
Figure 2: Boxplot graph of Leadership Readiness to Implement (LRIT) score from pre-intervention to post-intervention in the three municipalities.
Likewise, staff readiness (SRIT) was also high in M1 (42, IQR:28,47) and M3 (42, IQR:38,47.5) pre-intervention, however it slightly decreased (p = 0.09) in M1 post-intervention (Fig. 3). M2 reported the lowest readiness (27.0, IQR:27,31) among staff post-intervention (p < 0.001).
Figure 3: Boxplot graph of Staff Readiness to Implement (SRIT) score from pre-intervention to post-intervention in the three municipalities.
The median score for leadership and staff readiness was 41 (IQR: 28,46). The readiness scores of all personnel were positively correlated with the ACC (0.66), APP (0.70), and FEAS (0.71) pre-intervention scores in the municipalities.
Acceptability within and between municipalities
The median score for ACC was 15 (IQR:12,18). When comparing differences between municipalities pre-intervention (Fig. 4), M2 reported significantly (p = 0.003) lower median ACC scores (12, IQR:10,16) compared to M1 (16, IQR:14,19) and M3 (16, IQR:14,18). Significant differences (p = 0.02) between municipalities post-intervention were also observed. When comparing changes between pre- and post-measurements, only M1 reported a significant 2 (IQR:-5,1) point decrease in the score (p = 0.003) from median 16 (IQR:14,19) to 12.5 (IQR:9,17).
Figure 4: Boxplot graph of Acceptability (ACC) score from pre-intervention to post-intervention in the three municipalities
Appropriateness within and between municipalities
The median score for APP was 16 (IQR: 12.5,18). When comparing differences between municipalities pre-intervention for APP (Fig. 5), M2 reported significantly lower scores (12.5, IQR:10,16.5, p = 0.004) compared to M1 (16, IQR:15,18) and M3 (16, IQR: 14,18). When comparing differences between municipalities pre- and post-intervention, M1 and M3’s score (16) remained the same while M2 increased non-significantly by one point (13.5, IQR:9.5,15). There were no statistically significant differences between the municipalities at post-intervention or over time.
Figure 5: Boxplot graph of Appropriateness (APP) score from pre-intervention to post-intervention in the three municipalities
Feasibility within and between municipalities
The median score for FEAS was 15 (IQR: 12,17). When comparing differences between municipalities pre-intervention for FEAS (Fig. 6), M2 scored significantly lower (12, IQR:8,16 p < 0.001) compared to M1 (16, IQR:14,18) and M3 (16, IQR:13.5,17.5). Significant (p = 0.02) differences between municipalities post-intervention were also observed. M1 reported 15 (IQR:9,17), M2 reported 13 (IQR:9,16.5), while M3 stayed at the same high level (16, IQR:14,19.5). When comparing changes between pre-to post measurements for each municipality, M1 significantly decreased (p = 0.01) by one point (IQR:-4,0) compared to the other two municipalities (p = 0.03).
Figure 6: Boxplot graph of Feasibility (FEAS) score from pre-intervention to post-intervention in the three municipalities
Differences between and within professional groups
When comparing changes in ACC, APP and FEAS pre- and post-intervention between the three professional groups irrespective of municipality, only minor changes were found (Table 3). There were no statistically significant differences in the ACC or APP scores between the different professional groups either at pre- or post-intervention. However, there was a statistically significant decrease of two points in principals’ ACC scores (p = 0.04) and decrease of one point in teachers’ ACC scores (p = 0.01). Regarding the FEAS score, school nurses reported slightly lower FEAS scores (14, IQR:10,16) than principals (16, IQR:14,18) and teachers (16, IQR:12,18) pre-intervention approaching statistical significance (p = 0.059), which was similar post-intervention as well (Table 4).
Table 4
Median score of ACC, APP, FEAS for pre- and post-intervention and the change among professional groups
|
n
|
Principal/VP
|
n
|
Teacher
|
n
|
School Nurse
|
p-value*
|
ACC Pre-intervention Score, median (IQR)
|
39
|
15 (12, 18)
|
72
|
16 (12.5,18)
|
39
|
15 (10, 16)
|
0.38
|
ACC Post-intervention Score, median (IQR)
|
26
|
15.5 (11, 18)
|
44
|
14 (10, 17)
|
30
|
14.5 (11, 16)
|
0.86
|
Change in ACC over time, median (IQR)
|
26
|
-2 (-3,1)**
|
44
|
-1 (-5,1)**
|
30
|
0 (-4,2)
|
0.32
|
APP Pre-intervention Score, median (IQR)
|
38
|
16 (14, 17)
|
71
|
16 (12, 18)
|
39
|
16 (12, 17)
|
0.77
|
APP Post-intervention Score, median (IQR)
|
25
|
16 (12, 19)
|
44
|
15 (13, 17)
|
29
|
14 (10, 16)
|
0.14
|
Change in APP over time, median (IQR)
|
25
|
-1 (-4,0)
|
44
|
0 (-4,3)
|
29
|
0 (-4,1)
|
0.71
|
FEAS Pre-intervention Score, median (IQR)
|
38
|
16 (14, 18)
|
71
|
16 (12, 18)
|
39
|
14 (10, 16)
|
0.059
|
FEAS Post-intervention Score, median (IQR)
|
25
|
16 (12, 20)
|
44
|
16 (12.5,19)
|
29
|
13 (9, 16)
|
0.059
|
Change in FEAS over time, median (IQR)
|
25
|
-1 (-4,1)
|
44
|
1 (-2,4)
|
29
|
-1 (-3,2)
|
0.24
|
ACC = Acceptability, APP = Appropriateness, FEAS = Feasibility, IQR = Interquartile range, VP = Vice Principal; *P-value is comparing between professional groups using the Kruskal-Wallis test, ** Indicates a statistically significant decrease (P < 0.05) within the group using the Wilcoxon signed-rank test. |
No statistically significant differences for ACC, APP and FEAS were found among personnel from organizationally stable or instable schools.
Table 4. Median score of ACC, APP, FEAS for pre- and post-intervention and the change among professional groups
Figure 7 shows a graphical representation of pre-intervention Readiness in relation to the ACC score. Fifty percent of participants from M2 scored in quadrant I “Low Acceptability and Readiness” compared to 25% and 23% in M1 and M3, respectively. In quadrant IV “High Acceptability and Readiness”, 45% and 53% of M1 and M3 scored there, while only 26% of M2 scored in quadrant IV. Corresponding plots between APP and Readiness and FEAS and Readiness show similar results and are given in the additional files 3 and 4. All M3 schools decided to continue with the program at the end of year one, which was not the case in M1 and M2.
Figure 7: Correlation (r = 0.66) between pre-intervention Acceptability (ACC) scores and Readiness scores for all school personnel. The dotted lines indicate the median value of each variable. LRIT; Leader Readiness to Implement Tool, SRIT: Staff Readiness to Implement