Does the Use of Dihydroartemisinin-Piperaquine in Treating Patients with Uncomplicated falciparum Malaria Reduce the Risk for Recurrent New falciparum Infection More Than Artemether-Lumefantrine?

Table 9

GRADE profile of evidence quality on outcome measures.

Study characteristics			Quality assessment				Finding summary	Grading
Study design	No. of participants	Outcome measure	Limitation	Consistency	Directedness	Precision	Outcome: pooled RR (95% CI) and 𝑃 value	Quality remark
RCT	3,172; (number of studies = 6)	Total treatment failure (PRC-corrected) by day 28	No serious limitations involved^1	No serious inconsistency^2	Direct^3	Serious imprecision (−1)^4	RR: 0.45 (0.20 to 1.01); 𝑃 value = 0.05. This summary result indicates that DP treatment was associated with 55% reduction in risk for actual treatment failure compared with AL; highest plausible reduction is 80%. Upper CI limit includes number of no effect, 1	Moderate quality: (+3)
RCT	2,340; (number of studies = 4)	Total treatment failure (PRC-corrected) by day 42	No serious limitations^5	Serious inconsistency^6 (−1) [𝐼^2=70]	Direct^3	Serious imprecision (−1)^7	RR: 0.56 (0.27 to 1.14); 𝑃 value = 0.11. Pooled RR indicates that DP treatment was associated with a 44% reduction in risk of treatment failure with highest plausible reduction being 73% compared with AL. Upper CI limit crosses line of no effect and includes 14% risk for treatment failure.	Low quality: (+2)
RCT	2,340; (number of studies = 4)	PCR-new falciparum infections by day 28	No serious limitations^8	No serious inconsistency^9 [𝐼^2=0.00]	Direct^10	Precise^11	RR: 0.21 (0.14 to 0.32); 𝑃 value < 0.001. This pooled result means that when DP is used for treatment of falciparum malaria it is able to reduce risk of new infections by 79% relative to AL within 28 days. Plausible risk reduction ranges from 68% to 86% as defined by the 95% CI. Result is statistically significant.	High quality: (+4)
RCT	2,662; (number of studies = 4)	PCR-new falciparum infections by day 42	No serious limitations^8	Serious Inconsistency (−1)^12 [𝐼^2=83]	Direct^7	No Serious Imprecision^13	RR: 0.56 (0.34 to 0.91); 𝑃-value = 0.02. The pooled RR means that treatment of falciparum malaria using DP is associated with 44% reduction in risk for new infection within 42 days post treatment compared with AL. Range of reduction is between 9% and 66%. Result is statistically significant.	Moderate quality: (+3)^14

Note: RR = risk ratio, RCT = randomized control trial, CI = confidence interval.
1No serious limitations: randomization and concealment were judged to low risk of bias; laboratory personnel and investigators who assessed study outcome were all blinded to avoid measurement bias.
²No serious inconsistency: heterogeneity as indicated by 𝐼² was 47 and this is classified as moderate heterogeneity.
³Directedness: all trials used in the analysis were conducted in countries in sub-Sahara Africa where malaria falciparum transmission is mostly high.
⁴Precision: there is an imprecision because the 95% CI crosses line of no difference and includes increased risk of 1%.
⁵No serious limitations: randomization and concealment were judged to pose low risk of bias; laboratory personnel and investigators who assessed study outcome were all blinded to avoid measurement bias.
⁶Serious inconsistency: there is a substantial statistical heterogeneity as indicated by a very high value of the 𝐼² of 70.
⁷Precision: there is a serious imprecision because the interval of the 95% CI crosses line of no effect and also involves an increased risk of 14% for actual treatment failure within 42 days after treatment.
⁸No serious limitations: randomization and concealment were judged to pose low risk of bias and laboratory personnel and investigators were blinded.
⁹No important inconsistency: there was no statistical heterogeneity as shown in the 𝐼² being 0.00.
¹⁰Directedness: there is no important indirectedness because all trials used in the analysis were conducted in countries in sub-Sahara Africa where malaria falciparum transmission is mostly high.
¹¹Precision: there is no imprecision in 95% CI of the pooled RR because the CI did not cross the line of no difference hence all benefits were in favour of DP than AL.
¹²Serious inconsistency: 𝐼²=83 implying that there is a substantial statistical heterogeneity.
¹³Precision: there is no imprecision in 95% CI of the pooled RR because the CI did not cross the line of no difference; hence all benefits were in favour of DP than AL.