Antibiotic prescribing and dispensing for acute respiratory infections in children: effectiveness of a multi-faceted intervention for health-care providers in Vietnam

Figures & data

Figure 1. Flow chart of the study.

Figure 2. The posters for the health care facilities in Bavi district.

Table 1. Socio-demographic characteristics of the participating health-care providers (HCPs) in the intervention and control arms.

	Absolute number and percentage of HCPs in each group
	Intervention (n = 130)	Control (n = 154)
Characteristics	n (%)	n (%)
Age (years)
≤ 29	25 (19)	34 (23)
30–39	32 (25)	42 (27)
40–49	39 (30)	39 (25)
50+	34 (26)	39 (25)
Sex
Male	43 (33)	55 (36)
Female	87 (67)	99 (64)
Educational level*
Basic^†	29 (22)	14 (9)
Intermediate^††	89 (69)	122 (79)
University^†††	12 (9)	18 (12)
Practice type
Public health facilities	74* (57)	108* (70)
Private clinics	33* (25)	17* (11)
Drug stores	23* (18)	29* (19)

Notes: *Significant difference using chi-square test (p < 0.05), comparing intervention and control arms.

^†Basic drug seller, basic nurse, village health worker or no medical training.

^††Assistant doctor, assistant pharmacist or middle nurse.

^†††Medical doctor or pharmacist.

Table 2. Percentages of participants with correct answers and improvements for HCPs’ knowledge and practical competences regarding appropriate antibiotic use for ARI treatment.

Questions regarding HCP knowledge and practical competence	Percentage of participants with correct answers
	ARI intervention group				Control group
	Before	After	Difference	p-value^†	Before	After	Difference	p-value^†
Q1. Aetiology of ARI	21	49	28	0.000	31	21	–10	0.192
Q2. Distinguishing between pneumonia and a common cold	53	59	6	0.453	59	68	9	0.255
Q3. Symptoms don’t require antibiotics: cough and runny nose	75	85	10	0.172	70	74	4	0.505
Q4. Symptoms require antibiotics: cough, fever and fast breathing	75	89	14	0.015	83	87	4	0.509
Q5. Symptoms don’t require antibiotics: cough, no fever or chest contraction	75	89	14	0.043	47	76	29	0.000
Q6. Symptoms don’t require antibiotics: cough and fever	55	70	15	0.068	58	45	–13	0.123
Q7. Case scenario: probable diagnosis?	80	88	8	0.266	70	83	13	0.070
Q8 Case scenario: treatment for a child with cough and fever, not fast breathing?	37	57	20	0.023	32	43	11	0.059

Note: ^†p-value in the McNemar test comparing the percentages of correct answers of HCPs before and after the interventions in each group, adjusted with respect to intra-cluster correlation (ICC).

Table 3. Mean of total knowledge scores for HCPs regarding appropriate antibiotic use for ARI treatment in children under five, divided according to characteristics.

Characteristics	Mean of total knowledge score (range: 0–8)								Improvement difference between intervention and control groups	p-value ^††
	ARI intervention				Control
	Before	After	Improvement	p-value^†	Before	After	Improvement	p-value^†
Age (years)
≤ 29	4.68	6.16	1.48	0.001	4.24	4.56	0.32	0.369	1.16	0.078
30–39	4.94	6.19	1.25	0.019	4.71	5.33	0.62	0.188	0.63	0.309
40–49	4.64	5.38	0.74	0.113	4.85	5.28	0.43	0.283	0.31	0.540
50+	4.56	5.91	1.35	0.003	4.10	4.62	0.52	0.276	0.84	0.117
Sex
Male	4.81	5.72	0.91	0.045	4.69	4.95	0.26	0.592	0.65	0.212
Female	4.64	5.94	1.30	0.000	4.37	4.98	0.62	0.046	0.68	0.116
Education
Basic	4.00	5.17	1.17	0.019	4.71	4.07	–0.64	0.292	1.82	0.026
Intermediate	4.80	6.07	1.27	0.000	4.41	5.04	0.63	0.028	0.68	0.093
University	5.67	6.08	0.41	0.620	4.83	5.17	0.34	0.363	0.08	0.847
Health facilities
Public	5.14	6.18	1.04	0.003	4.70	5.10	0.40	0.157	0.64	0.158
Private	4.15	5.52	1.37	0.005	4.88	4.41	–0.47	0.373	1.84	0.004
Drug stores	4.09	5.39	1.30	0.013	3.45	4.79	1.34	0.031	–0.04	0.827
Total	4.70	5.87	1.17	0.000	4.49	4.97	0.48	0.068	0.69	0.054

Notes: ^†p-value of Wilcoxon’s test (adjusted with respect to intra-cluster correlation – ICC) comparing total knowledge scores between the characteristic groups before and after the interventions within the intervention/control groups.

^††p-value of Mann–Whitney test (adjusted with respect to intra-cluster correlation – ICC) comparing total knowledge scores between the characteristic groups before and after the interventions between the intervention group and control group.

Table 4. HCPs’ prescribing/dispensing patterns before and after the interventions.

Prescriptions/dispensing (N)		Antibiotics  used (%)	Multilevel logistic regression
			I/C AOR (p-value) ^1	A/B AOR (p-value)^2
Mild ARIs
Control (339)	Before (193)	140 (73)	1	1
	After (146)	102 (70)		0.310 (0.000)
	After – Before	(–3)		NA
Intervention (711)	Before (326)	275 (84)	1.003 (0.994)	1
	After (385)	215 (56)		0.310 (0.000)
	After – Before	(–28)		NA
Severe ARIs
Control (115)	Before (49)	39 (80)	1	1
	After (66)	62 (94)		3.313 (0.101)
	After – Before	(14)		NA
Intervention (101)	Before (36)	36 (100)	30.113 (0.048)	1
	After (65)	64 (99)		3.313 (0.101)
	After – Before	(–1)		NA
Other diseases
Control (288)	Before (143)	100 (70)	1	1
	After (145)	101 (70)		0.806 (0.235)
	After – Before	(0)		NA
Intervention (467)	Before (273)	205 (75)	1.204 (0.630)	1
	After (194)	133 (69)		0.806 (0.235)
	After – Before	(–6)		NA
All diseases
Control (742)	Before	279 (73)	1	1
	After	265 (74)		0.556 (0.000)
	After – Before	(1)		NA
Intervention (1279)	Before	516 (81)	1.045 (0.876)	1
	After	412 (64)		0.556 (0.000)
	After – Before	(–17)		NA

Notes: ¹I/C AOR: Adjusted odds ratio comparing intervention arm vs. control arm.

²A/B AOR: Adjusted odds ratio comparing after vs. before intervention acting. NA: non applicable.