Skip to main content

Table 6 GLS linear regression: RDT use observation scores

From: Outreach training and supportive supervision for malaria case management in Zambia: the effects of focused capacity building on indicators of diagnostic and clinical performance

 

Unadjusted

Model 2 (adjusted)a

Percentage point change (95% CI)

Percentage point change (95% CI)

Covariates

 Malaria blood slide preparation observation score

0.5 [0.4, 0.6]***

0.5 [0.4, 0.6]***

 Malaria blood slide staining/reading score

0.4 [0.3, 0.5]***

 

 All SOPs for malaria microscopy are present

  Yes

Reference

 

  No

− 2.9 [− 8.9, 3.2]

 

 All bench aids for malaria microscopy are present

  Yes

Reference

 

  No

3.5 [− 2.1, 9.0]

 

 Conducts malaria species identification

  Yes

1.6 [− 3.2, 6.3]

 

  No

Reference

 

 Uses parasites/µL for parasite quantitation

  Yes

3.4 [− 0.7, 7.5]

 

  No

Reference

 

 Uses “plus system” for parasite quantitation

  Yes

− 6.7 [− 10.6, − 2.8]**

0.1 [− 4.1, 4.2]

  No

Reference

Reference

 Uses microscopy positive controls

  Yes

2.2 [− 1.7, 6.0]

 

  No

Reference

 

 Stores stained microscopy slides for cross-checking

  Yes

− 1.2 [− 5.0, 2.6]

 

  No

Reference

 

 Stock-outs of malaria microscopy consumables

  Yes

− 3.2 [− 6.7, 0.2]

1.1 [− 2.8, 5.1]

  No

Reference

Reference

 Stock-outs of RDTs

  Yes

0.5 [− 4.9, 5.9]

 

  No

Reference

 

 Periodicity of OTSS assessments

  All assessments occurred within a 1 year period

9.4 [2.9, 16.0]**

4.9 [− 0.9, 10.7]

  All assessments occurred in > 1 year period

Reference

Reference

 Number of OTSS visits

  1

Reference

Reference

  2

4.8 [0.4, 9.1]*

1.1 [− 4.7, 6.9]

  3

4.3 [0.4, 9.0]*

1.3 [− 4.5, 7.1]

  4

9.5 [5.1, 13.7]***

4.9 [− 1.1, 10.9]

  1. Statistically significant results are in italics
  2. CI confidence interval
  3. Model 3: Wald X2 = 90.7 (p < 0.0001); 191 observations
  4. *** p < 0.001; ** p < 0.01; * p < 0.05; p < 0.1
  5. aAn ex-ante pairwise correlation test was run to detect multicollinearity; covariates indicating problematic correlation (≥ 0.55) were not included in the fully adjusted model