Table 3.
Sub-national differences in the effect of diagnostic testing on paediatric fever treatment in studied countries
| Country | Outcome | Strata | n febrile under fives taken to any care | % with outcome (95% CI) | % with a diagnostic test (95% CI) | AOR (95% CI) | p value |
|---|---|---|---|---|---|---|---|
| Malaria risk | |||||||
| Benin | Any anti-malarial use | High risk | 403 | 50.8 (45.3-56.3) | 20.0 (16.1-24.4) | 2.30 (1.10-4.82) | 0.028 |
| Moderate risk | 212 | 41.4 (32.9-50.5) | 18.7 (11.9-28.1) | 0.32 (0.08-1.29) | 0.108 | ||
| Burundi* | Any antibiotic use | Moderate risk | 1,239 | 54.0 (49.8-58.2) | 37.7 (33.8-41.6) | 0.44 (0.32-0.60) | <0.001 |
| Low risk | 125 | 57.3 (42.7-71.8) | 37.1 (27.2-47.0) | 6.75 (1.30-35.00) | 0.023 | ||
| Malawi | Any anti-malarial use | High risk | 1,870 | 56.9 (53.8-60.1) | 19.5 (16.8-22.2) | 1.63 (1.26-2.11) | <0.001 |
| Moderate risk | 2,361 | 53.2 (49.5-56.9) | 22.8 (19.3-26.3) | 1.12 (0.85-1.47) | 0.410 | ||
| Uganda | ACT use | High risk | 1,730 | 46.1 (41.5-50.9) | 26.9 (23.3-30.8) | 0.95 (0.70-1.28) | 0.736 |
| Moderate risk | 625 | 51.6 (46.2-57.1) | 34.7 (28.9-41.1) | 0.67 (0.46-0.98) | 0.040 | ||
| Any antibiotic use | High risk | 1,730 | 36.6 (32.9-40.4) | 26.9 (23.3-30.8) | 1.24 (0.93-1.65) | 0.146 | |
| Moderate risk | 625 | 33.0 (28.3-38.1) | 34.7 (28.9-41.1) | 1.68 (1.12-2.52) | 0.012 | ||
| Source of care | |||||||
| Malawi | Any antibiotic use | Public | 2,978 | 27.6 (25.2-30.0) | 19.3 (17.2-21.4) | 0.85 (0.67-1.09) | 0.199 |
| Private | 1,358 | 34.0 (30.2-37.8) | 26.0 (21.7-30.4) | 1.36 (0.95-1.96) | 0.097 | ||
| Rwanda* | Any anti-malarial use | Public | 522 | 24.4 (20.8-28.4) | 40.3 (35.8-44.8) | 0.59 (0.31-1.14) | 0.119 |
| Private | 134 | 5.0 (2.3-10.8) | 22.4 (15.6-31.1) | 29.38 (2.25-383.63) | 0.010 | ||
| Symptoms | |||||||
| Benin* | Any anti-malarial use | Fever alone | 380 | 46.6 (40.9-52.5) | 15.5 (11.8-20.1) | 0.94 (0.39-2.29) | 0.893 |
| Fever, cough, rapid breaths | 239 | 50.5 (43.1-57.8) | 27.2 (21.0-34.5) | 5.48 (1.27-23.63) | 0.022 | ||
| Burkina Faso | Any anti-malarial use | Fever alone | 1,253 | 48.5 (44.9-52.2) | 6.9 (5.5-8.6) | 2.03 (1.13-3.66) | 0.018 |
| Fever, cough, rapid breaths | 570 | 46.7 (42.0-51.5) | 9.5 (6.9-12.8) | 0.69 (0.34-1.41) | 0.304 | ||
| Malawi | ACT use | Fever alone | 1,980 | 51.7 (48.4-54.9) | 21.9 (19.3-24.5) | 0.87 (0.66-1.13) | 0.298 |
| Fever, cough, rapid breaths | 2,355 | 42.2 (39.1-45.4) | 21.0 (18.1-23.9) | 1.40 (1.09-1.79) | 0.008 | ||
| Any antibiotic use | Fever alone | 1,980 | 18.2 (16.0-20.4) | 21.9 (19.3-24.5) | 1.46 (1.07-1.99) | 0.017 | |
| Fever, cough, rapid breaths | 2,355 | 39.2 (36.2-42.2) | 21.0 (18.1-23.9) | 0.77 (0.60-1.00) | 0.052 | ||
| Senegal* | ACT use | Fever alone | 527 | 5.6 (3.3-9.1) | 15.1 (11.5-19.6) | 75.42 (1.09-5212.78) | 0.045 |
| Fever, cough, rapid breaths | 747 | 5.5 (3.2-9.3) | 14.8 (11.1-19.4) | 1.69 (0.50-5.73) | 0.400 |
If evidence of an interaction was found (0.05 level) between diagnostic testing and investigated variables (malaria risk, transmission season, age, symptoms, source of care) within a country, the final model was stratified accordingly and results presented above. CI = confidence interval; AOR = adjusted odds ratio. AORs are based on mixed-effects logistic regression models for specified strata in each country dataset adjusted for data clustering and confounding covariates (malaria endemicity; transmission season; public/private source; level of care; child’s age and sex; maternal age and education; residence; household wealth and size; health care access (money); health care access (distance); symptoms; health card).
*Some results should be interpreted with caution due to few observations and few positive outcomes.