WHO/HIV/2013.79
© World Health Organization 2013
GRADE Table: PICO a – Should nurses or clinical officers initiate and maintain ART among people living with HIV compared to doctors? Author(s): Tamara Kredo, Moses Bateganya, Elizabeth D. Pienaar, Folasade B. Adeniyi Date: 2012-10-22 Question: Should nurses or clinical officers initiate and maintain ART among people living with HIV compared to doctors? Settings: Ethiopia, South Africa Bibliography: : Assefa 2012, Bedelu 2007, Fairall 2012
Quality assessment
No. of patients
Effect
No. of studies
Design
Doctor versus nurse or clinical Risk of Other Inconsistency Indirectness Imprecision officer for initiating bias considerations and maintaining ART
Quality Control Relative (95% CI) Absolute
Importance
Death (RCT) (follow-up 12 months) 1 randomized trials no serious risk of bias no serious inconsistency no serious indirectness no serious none imprecision1 299/1614 (18.5%) 224/1156 (19.4%) RR 0.96 8 fewer per (0.82 to 1000 (from 35 1.12) fewer to 23 more) ⊕⊕⊕⊕ HIGH CRITICAL
Death (cohort) (follow-up 12 months) 2 observational no studies serious risk of bias2 no serious inconsistency no serious indirectness no serious imprecision none 798/6801 (11.7%) 2973/32359 RR 1.23 21 more per (9.2%) (1.14 to 1000 (from 13 1.33) more to 30 more) ⊕⊕ΟΟ LOW CRITICAL
Lost to care (RCT) (follow-up 12 months) 1 randomized trials no serious risk of bias no serious inconsistency no serious indirectness serious3 none 91/1614 (5.6%) 89/1156 (7.7%) RR 0.73 (0.55 to 0.97) 21 fewer per CRITICAL ⊕⊕⊕Ο 1000 (from 2 MODERATE fewer to 35 fewer)
Lost to care (cohorts) (follow-up 12 months)
This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS 1
WHO/HIV/2013.79
© World Health Organization 2013
2
observational no studies serious risk of bias
no serious inconsistency
no serious indirectness
serious4
none
1420/6797 (20.9%)
9618/32359 (29.7%)
RR 0.3 208 fewer per CRITICAL ⊕ΟΟΟ (0.05 to 1000 (from VERY LOW 1.94) 282 fewer to 279 more)
Death or lost to care (RCT) 1 randomized trials no serious risk of bias no serious inconsistency no serious indirectness no serious imprecision none 390/1614 (24.2%) 313/1156 (27.1%) RR 0.89 30 fewer per (0.79 to 1000 (from 57 1.01) fewer to 3 more) ⊕⊕⊕⊕ HIGH CRITICAL
Death or lost to care (cohorts) (follow-up 12 months) 2 observational no studies serious risk of bias no serious inconsistency5 no serious indirectness serious6 none 2218/6801 (32.6%) 12591/32359 RR 0.72 109 fewer per CRITICAL ⊕ΟΟΟ (38.9%) (0.48 to 1000 (from VERY LOW 1.07) 202 fewer to 27 more)
The confidence interval is narrow and does not include appreciable harm or benefit. Not downgraded for risk of bias. Two retrospective cohorts provided data. Bedelu (2007) included patients with higher CD4 counts at the health centre. Since this is likely to favour the intervention, we did not downgrade for risk of bias. 3 Downgraded by one for imprecision. There was a low number of events after adjusting for clustering (<300 events) 4 Downgraded by one for imprecision. The confidence interval includes both appreciable harm and appreciable benefit. 5 Not downgraded for inconsistency. Despite quantitative heterogeneity, both studies showed that attrition was decreased with task shifting of ART initiation and maintenance to nurses or clinical officers. 6 Downgraded by one for imprecision. 95% CI includes appreciable benefit and null value. 2
1
This work was commissioned by the World Health Organization and carried out by The University of California, San Francisco (UCSF), Cochrane Review Group on HIV/AIDS 2