Johns Hopkins Nursing Evidence-Based Practice Thrid Edition: Model and Guidelines

(^132) Johns Hopkins Nursing Evidence-Based Practice: Model and Guidelines, Third Edition
As the group gains experience in reading and appraising research, the members’
comfort level in making this determination will likely improve.
Meta-Analyses (Level I Evidence)
The strength of the evidence on which recommendations are made within a
meta-analytic study depends on the design and quality of studies included in the
meta-analysis as well as the design of the meta-analysis itself. Factors to consider
include sampling criteria of the primary studies included in the analysis, quality
of the primary studies, and variation in study outcomes between studies.
First, the EBP team looks at the types of research designs (level of evidence)
included in the meta-analysis. Meta-analyses containing only randomized con-
trolled trials are Level I evidence. Some meta-analyses include data from quasi-
experimental or nonexperimental studies; hence, evidence would be at a level
commensurate with the lowest level of research design included (e.g., if the meta-
analysis included experimental and quasi-experimental studies, the meta-analysis
would be Level II).
Second, the team should look at the quality of the individual studies included in
the meta-analysis. For an EBP team to evaluate evidence obtained from a meta-
analysis, the report of the meta-analysis must be detailed enough for the reader
to understand the studies included.
Third, the team should assess the quality of the meta-analysis. The discussion sec-
tion should include an overall summary of the findings, the magnitude of the ef-
fect, the number of studies, and the combined sample size. The discussion should
present the overall quality of the evidence and consistency of findings (Polit &
Beck, 2017). The discussion should include a recommendation for future research
to improve the evidence base.
Experimental Studies (Level I Evidence)
Level I evidence, the highest in the rating hierarchy, is derived from a well-
designed RCT on the question of interest (refer to Table 6.7). Level I evidence is
also derived from a meta-analysis of RCTs that support the same findings in dif-