Certainty of evidence and level of recommendation for complex versus simple interventions in Chilean clinical guidelines.

Tags: Poster
Seron P1, Oliveros M1, Fuentes R1
1Universidad de La Frontera

Background: Complex interventions are characterized by the inclusion of a set of components that interact with each other in a complex and dynamic health system, with flexible designs that adapt to the implementation context and with outcomes of interest that are also diverse, multiple and often far from a direct causal pathway. In addition to being frequently implemented, they challenge the researchers who seek to evaluate them, both because of the difficulties in minimizing the drawbacks that impact the internal validity of the studies and their generalizability or external validity. This could influence the strength of the associated recommendations, among other things, by assessing the certainty of the evidence supporting them. The Chilean Ministry of Health is updating the clinical guidelines using the GRADE approach and the recommendations include different interventions.

Objectives: To compare the certainty of evidence and the strength of recommendation of complex versus simple interventions contained in Chilean clinical guidelines designed with the GRADE approach.

Methods: All Chilean guidelines using GRADE approach in their design were included. Then, the interventions derived from the questions formulated by the panel, were classified as complex if they including at least four complexity characteristics: Population-level intervention, ‘‘Proactive’’ treatment, Flexibility in the implementation, Multiple health and social outcomes, Longer causal pathway, and Multiple implementation setting as the mandatory characteristic. Additionally, data about the evidence certainty and strength of the recommendation was extracted from guidelines. All process was performed by two independent and blinded researcher, and disagreement were solved by consensus. A descriptive analysis of the data was performed comparing both types of interventions.

Results: 49 guideline were included with 299 recommendations intervention associated. From these, 32 (10,7%) were complex compared with 267 (89,3%) simple. The certainty of evidence was HIGH in the 4% vs 4,3% (p>0,05), MODERATE in the 8% vs 19,8% (p>0,05), LOW in the 32% vs 26,1% (p>0,05); or VERY LOW in the 56% vs 49,8% (p>0,05) of the complex compared with simple intervention respectively. The strength of recommendation was STRONG in the 6,3% vs 13,5% (p>0,05), CONDITIONAL in the 65,6% vs 82,3% (p