Study population
At the time of the survey, 924 ALPHIV (12–19 years) were receiving ART in the 19 participating hospitals. Among them, 709 (77%) were interviewed with their caregiver/institutional staff, 573 living in family settings and 136 in institutions. Two adolescents living with family were excluded because they had no appropriate control in the general population. Therefore, we analysed data for 707 ALPHIV, including 571 living in family settings and 136 in institutions. The control group was composed of 689 adolescents, including 571 also living in family settings and 118 in institutions (Figure 1).
Insert Figure 1 here
Table 1 presents the adolescents’ sociodemographic, educational, school-life, and health characteristics. Cases and controls are compared, distinguishing those living in family settings and in institutions.
Table 1. Sociodemographic characteristics, education, school life, and health situation of adolescents: Comparison of ALPHIV (cases) and controls, distinguishing adolescents living in family and in institutions.
|
|
Adolescents living in family
|
|
Adolescents living in institutions
|
Variables (n in case of missing data)
|
|
ALPHIV N=571 (%)
Median (sd)
|
Controls
N=571 (%)
Median (sd)
|
p-value
|
|
ALPHIV
N=136 (%)
Median (sd)
|
Controls
N=118 (%)
Median (sd)
|
p-value
|
SOCIO-DEMOGRAPHIC CHARACTERISTICS
|
|
|
|
|
|
|
|
|
Sex (boy)
|
|
238 (41.7)
|
238 (41.7)
|
|
|
68 (50.0)
|
59 (50.0)
|
|
Age
|
|
14.4 (2.9)
|
14.4 (2.9)
|
|
|
14.8 (3.1)
|
14.6 (3.4)
|
|
Orphan (one or both parents)
|
|
493 (86.3)
|
65 (11.4)
|
|
|
131 (96.3)
|
88 (74.6)
|
|
Ethnic minority (475,571,129,116)
|
|
15 (2.6)
|
29 (5.1)
|
|
|
11 (8.1)
|
37 (31.4)
|
|
Type of caregiver (560, 570)
|
|
|
|
|
|
|
|
|
Parent
|
|
179 (31.3)
|
415 (72.7)
|
|
|
-
|
-
|
|
Grandparent
|
|
213 (37.3)
|
89 (15.6)
|
|
|
-
|
-
|
|
More distant relative or guardian
|
|
168 (29.4)
|
66 (11.6)
|
|
|
-
|
-
|
|
Caregiver’s level of education
|
|
|
|
|
|
|
|
|
Secondary school and above
|
|
108 (18.9)
|
211 (37.0)
|
|
|
-
|
-
|
|
Primary school
|
|
379 (66.4)
|
325 (56.9)
|
|
|
-
|
-
|
|
Never attended school
|
|
84 (14.7)
|
35 (6.1)
|
|
|
-
|
-
|
|
Household’s financial situation difficult
|
|
207 (36.3)
|
109 (19.1)
|
|
|
-
|
-
|
|
Living area rural (569,569)
|
|
442 (77.4)
|
458 (80.2)
|
|
|
-
|
-
|
|
|
|
|
|
|
|
|
|
|
EDUCATION
|
|
|
|
|
|
|
|
|
Current school attendance
|
|
472 (82.7)
|
538 (94.2)
|
|
|
130 (95.6)
|
118 (100)
|
|
Disrupted school trajectory (564,571,135,118)
|
157 (27.5)
|
32 (5.6)
|
|
|
101 (74.3)
|
52 (44.1)
|
|
Academic delay (≥ 1 year) (564,571,135,118)
|
|
96 (16.8)
|
21 (3.7)
|
|
|
96 (70.6)
|
52 (44.1)
|
|
Early school dropout
|
|
61 (10.7)
|
11 (1.9)
|
|
|
5 (3.7)
|
0 (0)
|
|
|
|
|
|
|
|
|
|
|
SCHOOL LIFE
|
|
|
|
|
|
|
|
|
In public school (570,571,132,115)
|
|
532 (93.2)
|
523 (91.6)
|
|
|
82 (60.3)
|
90 (76.3)
|
|
Attendance in extra-curricular school (569,571,135,118)
|
306 (53.4)
|
378 (66.2)
|
|
|
85 (62.5)
|
84 (71.2)
|
|
No or few friends at school (567,571,136,118)
|
|
64 (11.2)
|
23 (4.0)
|
|
|
17 (12.5)
|
31 (26.3)
|
|
Lack of school life enjoyment (566,571,136,118)
|
75 (13.1)
|
56 (9.8)
|
|
|
19 (14.0)
|
25 (21.2)
|
|
Self-reported academic performance (564,571,136,118)
|
Mediocre / poor
|
|
26 (4.6)
|
13 (2.3)
|
|
|
12 (8.8)
|
11 (9.3)
|
|
Fair
|
|
282 (49.4)
|
279 (48.9)
|
|
|
59 (43.4)
|
56 (47.5)
|
|
Good / excellent
|
|
258 (45.2)
|
279 (48.9)
|
|
|
65 (47.8)
|
51 (43.2)
|
|
Absenteeism for medical reasons (560,570,135,118)
|
Rarely
|
|
195 (34.2)
|
332 (58.1)
|
|
|
60 (44.1)
|
67 (56.8)
|
|
Sometimes
|
|
234 (41.0)
|
175 (30.6)
|
|
|
52 (38.2)
|
33 (28.0)
|
|
Regularly / for a long time
|
131 (22.9)
|
63 (11.0)
|
|
|
23 (16.9)
|
18 (15.3)
|
|
Grade repetition (567,571,136,118)
|
|
87 (15.2)
|
19 (3.3)
|
|
|
-
|
-
|
|
Higher education aspirations
|
Yes
|
|
285 (49.9)
|
411 (72.0)
|
|
|
74 (54.4)
|
75 (63.6)
|
|
No
|
|
104 (18.2)
|
36 (6.3)
|
|
|
21 (15.4)
|
16 (13.6)
|
|
Don’t know
|
|
182 (31.9)
|
124 (21.7)
|
|
|
41 (30.1)
|
27 (22.9)
|
|
Stigmatization experience(s) at school (564)
|
134 (23.5)
|
-
|
|
|
-
|
-
|
|
|
|
|
|
|
|
|
|
|
HEALTH
|
|
|
|
|
|
|
|
|
History of hospitalizations (570,571,136,118)
|
347 (60.8)
|
225 (39.4)
|
|
|
76 (55.9)
|
36 (30.5)
|
|
Neurocognitive difficulties (481,571,136,118)
|
|
87 (15.2)
|
5 (0.9)
|
|
|
25 (18.4)
|
8 (6.4)
|
|
HAZ < -2 (569,128)
|
|
240 (42.0)
|
-
|
|
|
46 (33.8)
|
-
|
|
Age at ART initiation (549,121)
|
|
9.0 (3.1)
|
-
|
|
|
8.6 (3.5)
|
-
|
|
Age at diagnosis (563,106)
|
|
7.5 (3.8)
|
-
|
|
|
8.3 (4.0)
|
-
|
|
NNRTI based ART (571,131)
|
|
421 (73.7)
|
-
|
|
|
87 (64.0)
|
-
|
|
Viral load ≥1000 cp/mL (569,130)
|
|
103 (18.0)
|
-
|
|
|
3 (2.2)
|
-
|
|
CD4 count <20% (568,130)
|
|
120 (21.0)
|
-
|
|
|
13 (9.6)
|
-
|
|
†: Mc Nemar test; ‡: Fisher’s exact test (when conditions for the application of Chi-squared test are not met); §:Chi-squared test; ¶:Wilcoxon signed-rank test; £: Mann-Whitney-Wilcoxon test; ¥: Stuart-Marxwell test (for matched analyses with variables with more than 2 categories)
Sociodemographic profile of ALPHIV and controls
Adolescents living in family settings
Both groups (ALPHIV and controls) were composed of boys (42%) and girls (58%) with a median age of 14.4 years.
Most ALPHIV (86%) had lost one or both parents, while most controls had not (11%; p < 0.001). Among ALPHIV, only one-third (31%) had a parent as the primary caregiver compared to 73% of controls. Educational level of caregivers was significantly lower among ALPHIV than among controls, with 19% of ALPHIV caregivers having had attended secondary school or above versus 37% of the controls’ caregivers. The household financial situation was significantly more often perceived as “difficult” among ALPHIV than controls (36% vs. 19%, p < 0.001).
Adolescents living in institutions
Both groups were composed in equal proportions of boys and girls with a median age of 14.8 years. Almost all ALPHIV had lost one or both parents, whereas the proportion was approximately three-quarters for the control group (p < 0.001). Controls were significantly more often members of ethnic minorities than were ALPHIV (31% vs. 8%, p < 0.001).
Educational and school-life characteristics of ALPHIV and controls
Adolescents living in family settings
At the time of the survey, ALPHIV were less likely to be attending school than their peers in the general population (83% vs. 94%, p < 0.001). Disrupted school trajectories were more frequent among cases than controls (27% vs. 6%, p < 0.001). Most (> 90%) cases and controls living in family settings were enrolled in public schools, but ALPHIV were less likely to attend extracurricular programmes (53% vs. 66%, p < 0.001). Cases reported having few or no friends more often than controls (11% vs. 4%, p < 0.001). However, school-life enjoyment and self-reported academic performance did not differ between cases and controls. Significantly more cases than controls mentioned frequent absenteeism for medical reasons (23% vs. 11%, p < 0.001), and they were more likely to have repeated a grade (15% vs. 3%, p < 0.001). ALPHIV were more often reluctant or unsure about pursuing higher education than were controls (50% vs. 28%, p < 0.001). Finally, according to their caregiver, 24% of ALPHIV had experienced stigmatization at school in the form of bullying, violent or humiliating behaviours or exclusion.
Adolescents living in institutions
At the time of the survey, almost all adolescents (> 95%) living in institutions were currently attending school. Unlike the group of adolescents living in family settings, most school-life characteristics did not differ between cases and controls. Disrupted school trajectories were, however, more frequent among cases than controls, although the percentages were high in both groups (74% vs. 44%, p < 0.001). Interestingly, cases were more likely to report having many friends at school (88% vs. 74%, p = 0.01) and to be attending a private school (36.8% vs. 21.2%, p = 0.01). The control group comprised a small number (n = 16) of HIV-exposed uninfected (HEU) adolescents (born from HIV-infected mothers). The percentage of school trajectory disruption among HEU was slightly lower than among the HIV non-exposed uninfected controls (31% vs. 45%, data not shown).
Health situation of ALPHIV and controls
Adolescents living in family settings
Having a hospitalization history and neurocognitive difficulties were more frequent among ALPHIV than controls (61% vs. 39%, p < 0.001 and 15% vs. 1%, p < 0.001, respectively). Moreover, 42% of ALPHIV suffered from growth delay (HAZ < 2). Median ages at HIV diagnosis and ART initiation were 7.5 and 9.0 years, respectively.
At the time of the survey, 74% of ALPHIV were receiving NNRTI-based ART and 18% of ALPHIV had a viral load ≥ 1000cp/mL and 21% a CD4 count < 20%.
Adolescents living in institutions
Here also, a history of hospitalizations and neurocognitive difficulties were more frequent among ALPHIV than controls according to institutional staff (56% vs. 31%, p < 0.001 and 18% vs. 6%, p < 0.001, respectively). Thirty-four percent of ALPHIV had an HAZ of < -2. The median age at HIV diagnosis was 8.3 and 8.6 years at ART initiation.
At the time of the survey, 69% of ALPHIV were receiving NNRTI-based ART. Only 2% of ALPHIV had a viral load ≥ 1000cp/mL and 9% had a CD4 count < 20%.
Factors associated with school trajectory disruption
Among all adolescents surveyed: cases or controls (Table 2)
The multivariable analysis indicated that the risk of having a disrupted school trajectory was 5 times higher among ALPHIV ( = 5.15 [3.72-7.23], p < 0.001) compared to controls. Other factors associated with school trajectory disruption were: male sex ( = 1.76 [1.32–2.35], p < 0.001), living in an institution ( = 10.98 [7.72–15.80], p < 0.001), or being cared for by a distant relative or guardian ( = 1.49 [1.03–2.15], p = 0.03), compared to being cared by parents and grand-parents, and history of hospitalizations ( = 1.46 [1.09–1.97], p = 0.01). Age was also significantly associated with a disrupted school trajectory, with a 22% increase per year ( = 1.22 [1.13–1.31], p < 0.001).
We estimated that about half of school trajectory disruptions could be attributed to HIV-infection (AF= 0.47 [0.38–0.55], p < 0.001).
Table 2. Factors associated with school trajectory disruption among all adolescents (cases and controls) surveyed: bivariate and multivariable analyses
|
|
Disrupted school trajectory
|
|
Bivariate analysis
|
|
Multivariable analysis
|
|
No
N=1046
N (%) / Median (sd)
|
Yes
N=342
N (%) / Median (sd)
|
|
|
|
|
|
|
|
ORB (IC95%)
|
|
ORA (IC95%)
|
p-value†
|
HIV status
|
|
|
|
|
|
|
|
|
Controls
|
|
605 (87.8)
|
84 (12.2)
|
|
1
|
|
1
|
|
ALPHIV
|
|
441 (63.1)
|
258 (36.9)
|
|
4.21 [3.20-5.55]
|
|
5.15 [3.72-7.23]
|
<0.001
|
|
|
|
|
|
|
|
|
|
Sex
|
|
|
|
|
|
|
|
|
Girl
|
|
628 (79.6)
|
161 (20.4)
|
|
1
|
|
1
|
|
Boy
|
|
418 (69.8)
|
181 (30.2)
|
|
1.69 [1.32-2.16]
|
|
1.76 [1.32-2.35]
|
<0.001
|
|
|
|
|
|
|
|
|
|
Type of caregiver
|
|
|
|
|
|
|
Parent or grandparent
|
|
765 (85.9)
|
126 (14.1)
|
|
1
|
|
1
|
|
More distant relative or guardian
|
172 (74.1)
|
60 (25.9)
|
|
2.12 [1.49-3.00]
|
|
1.49 [1.03-2.15]
|
0.03
|
Institution staff
|
|
100 (39.5)
|
153 (60.5)
|
|
9.29 [6.78-12.72]
|
|
10.98 [7.72-15.80]
|
<0.001
|
|
|
|
|
|
|
|
|
|
History of hospitalizations
|
No
|
|
567 (79.9)
|
143 (20.1)
|
|
1
|
|
1
|
|
Yes
|
|
479 (70.8)
|
198 (29.2)
|
|
1.64 [1.28-2.10]
|
|
1.46 [1.09-1.97]
|
0.01
|
|
|
|
|
|
|
|
|
|
Age (years)
|
|
14.6 (3.4)
|
15.2 (3.8)
|
|
1.19 [1.11-1.26]
|
|
1.22 [1.13-1.31]
|
<0.001
|
†:Wald test
Among all ALPHIV living in family settings or in institutions (Table 3)
Sociodemographic factors such as male sex ( = 1.57 [1.09–2.27], p = 0.02), age ( = 1.21 [1.08–1.35], p < 0.001), and living in an institution ( = 9.96 [6.20–16.37], p < 0.001) were significantly associated with disrupted school trajectories. Adolescents who mentioned a history of hospitalizations were more likely to experience school trajectory disruption ( = 1.52 [1.05–2.22], p = 0.03), as were those with neurocognitive difficulties ( = 3.29 [2.11–5.19], p < 0.001) and those with a HAZ <-2 ( = 1.81 [1.25–2.62], p = 0.001). Age at ART initiation, type of ART (first- or second-line), and type of caregiver were not associated with disrupted school trajectories, after adjustment for the other factors.
Table 3. Factors associated with school trajectory disruption among ALPHIV living in family settings and institutions: bivariate and multivariable analyses
|
|
Disrupted school trajectory
|
|
Bivariate analysis
|
|
Multivariable analysis
|
|
No
N=441
N (%) / Median (sd)
|
Yes
N=258
N (%) / Median (sd)
|
|
|
|
|
|
|
|
|
ORB (IC95%)
|
|
ORA (IC95%)
|
p-value†
|
Sex
|
|
|
|
|
|
|
|
|
Girl
|
|
273 (68.8)
|
124 (31.2)
|
|
1
|
|
1
|
|
Boy
|
|
168 (55.6)
|
134 (44.4)
|
|
1.76 [1.29-2.40]
|
|
1.57 [1.09-2.27]
|
0.02
|
|
|
|
|
|
|
|
|
|
Age (years)
|
|
14.4 (3.5)
|
15.1 (3.5)
|
|
1.20 [1.10-1.30]
|
|
1.21 [1.08-1.35]
|
<0.001
|
|
|
|
|
|
|
|
|
|
Type of caregiver
|
|
|
|
|
|
|
|
|
Parent or grandparent
|
|
287 (74.2)
|
100 (25.8)
|
|
1
|
|
1
|
|
More distant relative or guardian
|
112 (67.5)
|
54 (32.5)
|
|
1.38 [0.93-2.06]
|
|
1.30 [0.85-1.98]
|
0.22
|
Institution staff
|
|
34 (25.2)
|
101 (74.8)
|
|
8.52 [5.43-13.4]
|
|
9.96 [6.20-16.37]
|
<0.001
|
|
|
|
|
|
|
|
|
|
History of hospitalizations
|
|
|
|
|
|
No
|
|
192 (68.1)
|
90 (31.9)
|
|
1
|
|
1
|
|
Yes
|
|
249 (59.9)
|
167 (40.1)
|
|
1.43 [1.04-1.96]
|
|
1.52 [1.05-2.22]
|
0.03
|
|
|
|
|
|
|
|
|
|
Neurocognitive difficulties
|
|
|
|
|
|
|
No
|
|
325 (65.0)
|
175 (35.0)
|
|
1
|
|
1
|
|
Yes
|
|
45 (40.9)
|
65 (59.1)
|
|
2.76 [1.76-4.09]
|
|
3.29 [2.11-5.19]
|
<0.001
|
|
|
|
|
|
|
|
|
|
HAZ
|
|
|
|
|
|
|
|
|
≥ -2
|
|
277 (68.4)
|
128 (31.6)
|
|
1
|
|
1
|
|
< -2
|
|
162 (57.0)
|
122 (43.0)
|
|
1.63 [1.19-2.23]
|
|
1.81 [1.25-2.62]
|
0.001
|
|
|
|
|
|
|
|
|
|
Age at ART initiation
|
|
|
|
|
|
|
|
|
< 9 years-old
|
|
226 (69.3)
|
100 (30.7)
|
|
1
|
|
1
|
|
≥ 9 years-old
|
|
201 (59.8)
|
135 (40.2)
|
|
1.52 [1.10-2.09]
|
|
1.43 [0.93-2.19]
|
0.10
|
|
|
|
|
|
|
|
|
|
ART type
|
|
|
|
|
|
|
|
|
NNRTI based
|
|
333 (66.1)
|
171 (33.9)
|
|
1
|
|
1
|
|
PI based (or other)
|
|
108 (57.4)
|
80 (42.6)
|
|
1.44 [1.02-2.03]
|
|
1.42 [0.95-2.12]
|
0.09
|
†: Wald test
Among ALPHIV living in family settings (Table 4)
In this subpopulation, age and male sex were also significantly associated with disrupted school trajectories, as well as having neurocognitive difficulties and growth delay. Additional contextual factors, such as being from an ethnic minority ( = 3.14 [1.00–9.75], p = 0.05), having a caregiver with less than secondary-level education ( = 2.05 [1.12–3.88], p = 0.02 for primary school dropout and = 2.37 [1.05–5.41], p = 0.04 for those who had never attended school), and experience of stigmatization at school = 1.94 [1.21–3.10], p = 0.01) were significantly associated with a disrupted school trajectory. Adolescents on second-line treatment were more likely to experience school trajectory disruption ( = 1.66 [1.04–2.64], p = 0.03), but there was no significant difference according to age at ART initiation. There was no significant association with the type of caregiver, household’s financial situation, type of school, type of living area, and history of hospitalization.
Table 4. Factors associated with school trajectory disruption among ALPHIV living in family settings: bivariate and multivariable analyses
|
|
Disrupted school trajectory
|
|
Bivariate analysis
|
|
Multivariable analysis
|
|
No
N=407
N (%) / Median (sd)
|
Yes
N=157
N (%) / Median (sd)
|
|
|
|
|
|
|
|
ORB (IC95%)
|
|
ORA (IC95%)
|
p-value†
|
Sex
|
|
|
|
|
|
|
|
|
Girl
|
|
256 (77.8)
|
73 (22.2)
|
|
1
|
|
1
|
|
Boy
|
|
151 (64.3)
|
84 (35.7)
|
|
1.79 [1.17-2.73]
|
|
1.81 [1.18-2.79]
|
0.01
|
|
|
|
|
|
|
|
|
|
Age (years)
|
|
14.4 (3.4)
|
15.2 (3.5)
|
|
1.23 [1.12-1.36]
|
|
1.27 [1.12-1.45]
|
<0.001
|
|
|
|
|
|
|
|
|
|
Neurocognitive difficulties
|
|
|
|
|
|
|
|
|
No
|
|
295 (75.6)
|
95 (24.4)
|
|
1
|
|
1
|
|
Yes
|
|
41 (48.2)
|
44 (51.8)
|
|
3.33 [2.05-5.41]
|
|
3.34 [2.00-5.61]
|
<0.001
|
|
|
|
|
|
|
|
|
|
HAZ
|
|
|
|
|
|
|
|
|
≥ -2
|
|
254 (78.4)
|
70 (21.6)
|
|
1
|
|
1
|
|
< -2
|
|
151 (63.4)
|
87 (36.6)
|
|
2.09 [1.44-3.04]
|
|
1.80 [1.18-2.76]
|
0.01
|
|
|
|
|
|
|
|
|
|
Ethnic origin
|
|
|
|
|
|
|
|
|
Thai
|
|
331 (72.7)
|
124 (27.3)
|
|
1
|
|
1
|
|
Ethnic minority
|
|
7 (46.7)
|
8 (53.3)
|
|
3.05 [1.08-8.59]
|
|
3.14 [1.00-9.75]
|
0.05
|
|
|
|
|
|
|
|
|
|
Caregiver’s level of education
|
|
|
|
|
|
|
Secondary school and above
|
|
89 (82.4)
|
19 (17.6)
|
|
1
|
|
1
|
|
Primary school
|
|
261 (70.0)
|
112 (30.0)
|
|
2.01 [1.17-3.46]
|
|
2.05 [1.12-3.88]
|
0.02
|
Never attended school
|
|
57 (68.7)
|
26 (31.3)
|
|
2.14 [1.08-4.21]
|
|
2.37 [1.05-5.41]
|
0.04
|
|
|
|
|
|
|
|
|
|
Stigmatization experience(s) at school
|
|
|
|
|
|
|
No or don’t know
|
|
326 (75.8)
|
104 (24.2)
|
|
1
|
|
1
|
|
Yes once or more
|
|
81 (60.4)
|
53 (39.6)
|
|
2.05 [1.36 ;3.09]
|
|
1.94 [1.21-3.10]
|
0.01
|
|
|
|
|
|
|
|
|
|
ART type
|
|
|
|
|
|
|
|
|
NNRTI based
|
|
310 (74.2)
|
108 (25.8)
|
|
1
|
|
1
|
|
PI based (or other)
|
|
97 (66.4)
|
49 (33.6)
|
|
1.45 [0.96-2.18]
|
|
1.66 [1.04-2.64]
|
0.03
|
|
|
|
|
|
|
|
|
|
Age at ART initiation
|
|
|
|
|
|
|
|
|
< 9 years
|
|
208 (78.5)
|
57 (21.5)
|
|
1
|
|
1
|
|
≥ 9 years
|
|
186 (67.1)
|
91 (32.9)
|
|
1.79 [1.21 ;2.63]
|
|
1.49 [0.90-2.46]
|
0.12
|
|
|
|
|
|
|
|
|
|
Type of caregiver
|
|
|
|
|
|
|
Parent
|
|
140 (78.2)
|
39 (21.8)
|
|
1
|
|
1
|
|
Grandparents
|
|
147 (70.7)
|
61 (29.3)
|
|
1.49 [0.94-2.37]
|
|
1.03 [0.60-1.77]
|
0.93
|
More distant relative or guardian
|
112 (67.5)
|
54 (32.5)
|
|
1.73 [1.07-2.80]
|
|
1.53 [0.89-2.63]
|
0.12
|
|
|
|
|
|
|
|
|
|
Household’s financial situation
|
|
|
|
|
|
|
|
|
Fair / good / very good
|
|
259 (71.9)
|
101 (28.1)
|
|
1
|
|
1
|
|
Difficult / very difficult
|
|
148 (72.5)
|
56 (27.5)
|
|
0.97 [0.66-1.42]
|
|
0.92 [0.59-1.42]
|
0.70
|
|
|
|
|
|
|
|
|
|
Public school
|
|
|
|
|
|
|
|
|
Yes
|
|
383 (72.5)
|
145 (27.5)
|
|
1
|
|
1
|
|
No
|
|
24 (66.7)
|
12 (33.3)
|
|
1.32 [0.64-2.71]
|
|
0.99 [0.42-2.20]
|
0.98
|
|
|
|
|
|
|
|
|
|
Type of living area
|
|
|
|
|
|
|
|
|
Rural
|
|
317 (72.4)
|
121 (27.6)
|
|
1
|
|
1
|
|
Urban
|
|
88 (71.0)
|
36 (29.0)
|
|
1.07 [0.69-1.67]
|
|
1.49 [0.90-2.45]
|
0.12
|
|
|
|
|
|
|
|
|
|
History of hospitalizations
|
|
|
|
No
|
|
173 (77.9)
|
49 (22.1)
|
|
1
|
|
1
|
|
Yes
|
|
234 (68.6)
|
107 (31.4)
|
|
1.61 [1.09-2.39]
|
|
1.40 [0.91-2.17]
|
0.13
|
†:Wald test
In the sensitivity analyses, using 2 or more years of academic delay as a threshold or the age–grade delay, results were similar to those obtained for our main outcome (Additional file 2, 3, 4). Also, after excluding the HEU from the control group, results remained unchanged (Additional file 5).