Table 4.

GLS linear regression: clinician fever case management scores

	Unadjusted	Model 2 (adjusted)a
	Percentage point change (95% CI)	Percentage point change (95% CI)
Covariates
Prescriber adherence to negative malaria tests (score)	0.2 [0.1, 0.2]***	0.1 [0.0, 0.2]**
Has ≥ 1 otoscope
Yes	5.4 [1.3, 9.6]*	4.4 [0.0, 8.7]*
No	Reference	Reference
Has ≥ 1 ophthalmoscope
Yes	4.7 [0.6, 8.8]*
No	Reference
Has ≥ 1 medical pen light
Yes	6.7 [2.7, 10.6]**	4.0 [− 0.2, 8.2]
No	Reference	Reference
Has properly completed/maintained outpatient register
Yes	2.1 [− 2.4, 6.6]
No	Reference
Has properly completed/maintained inpatient register
Yes	4.2 [− 0.7, 9.0]†	− 1.5 [− 6.6, 3.6]
No	Reference	Reference
Has properly completed/maintained clinical reports
Yes	8.4 [2.2, 14.5]**	4.0 [− 2.6, 10.5]
No	Reference	Reference
Has properly completed/maintained drug stock cards
Yes	11.4 [1.9, 20.9]*	1.8 [− 7.3, 10.8]
No	Reference	Reference
Has malaria case management and treatment guidelines
Yes	10.5 [6.2, 14.8]***	8.5 [4.1, 13.0]***
No	Reference	Reference
Had an RDT stock-out of ≥ 7 days in the 3 months preceding OTSS visit
Yes	5.9 [− 0.8, 12.6]†	3.5 [− 3.4, 10.5]
No	Reference	Reference
Had a malaria microscopy consumables/reagents stock-out of ≥ 7 days in the 3 months preceding OTSS visit
Yes	− 5.4 [− 9.5, − 1.3]*	− 1.0 [− 5.4, 3.3]
No	Reference	Reference
Periodicity of OTSS assessments
All assessments occurred within a 1 year period	− 4.4 [− 10.7, 1.9]
All assessments occurred in > 1 year period	Reference
Number of OTSS assessments
1	Reference	Reference
2	− 5.9 [− 10.8, − 0.9]*	− 2.4 [− 8.5, 3.7]
3	− 0.8 [− 5.8, 4.1]	0.0 [− 5.8, 5.8]
4	6.0 [1.1, 11.0]*	5.1 [− 0.9, 11.0]

Statistically significant results are in italics

CI confidence interval

Model 1: Wald X2 = 68.7 (p < 0.0001); 196 observations

*** p < 0.001; ** p < 0.01; * p < 0.05; ^† p < 0.1

^aAn ex-ante pairwise correlation test was run to detect collinearity; covariates indicating problematic correlation coefficients (≥ 0.55) were not included in the fully adjusted model