Review Approach

The Brown Evidence-based Practice Center conducted this systematic review (SR) based on the Agency for Healthcare Research and Quality (AHRQ) Methods Guide for Effectiveness and Comparative Effectiveness Reviews (available at https://effectivehealthcare.ahrq.gov/topics/cer-methods-guide/overview). This SR also reports in accordance with the Preferred Items for Reporting in Systematic Reviews and Meta-Analyses (PRISMA),⁵⁵ A Measurement Tool to Assess Systematic Reviews (AMSTAR 2),⁵⁶ and any relevant extension statements.

A more detailed version of the SR methodology used can be found in Appendix A. Excluded studies are listed in Appendix B. Search results and descriptive data for all included studies are included in Appendix C. Extracted study results are in Appendix D.

The topic of this report and preliminary Key Questions (KQs) arose through a process involving the nominators (the American College of Physicians), a panel of Key Informants (KI), a Technical Expert Panel (TEP), the public, and AHRQ. Initially, the KI panel gave input on the KQs, including the outcomes, to be examined. AHRQ then posted these KQs and solicited public comment through its Effective Health Care (EHC) Program website and on the Federal Register. No comments were received. The TEP provided high-level content and methodological expertise throughout development of the review protocol. The final protocol was posted on the EHC website at https://effectivehealthcare.ahrq.gov/products/diverticulitis/protocol on September 12, 2019. We submitted the protocol for registration in PROSPERO in November 2019. On April 29, 2020, PROSPERO published the protocol with registration number CRD42020151246.

Study Selection

Literature searches were conducted in Medline^® (via PubMed^®), the Cochrane Register of Clinical Trials, the Cochrane Database of Systematic Reviews, Embase^®), CINAHL^®), and ClinicalTrials.gov, restricted to 1990 through June 1, 2020. The search was restricted to recent studies (since 1990) based on important changes in diagnosis and clinical management of diverticulitis based on increased use of CT imaging since the 1990s.

Table 1 presents the major eligibility criteria for each KQ. More detailed criteria are presented in Appendix A. We included randomized controlled trials (RCTs), nonrandomized comparative studies (NRCSs), single group studies (noncomparative between interventions), and existing SRs.

Table 1

Study eligibility criteria by Key Question.

Risk of Bias Assessment

We evaluated each study for risk of bias and methodological quality. Because we included a variety of study designs, we incorporated items from three different commonly-used tools and tailored the set of items for each study design.

For RCTs, we used all the items from the Cochrane Risk of Bias tool,⁵⁷ including randomization and allocation concealment methodology; blinding; completeness of data reporting; and selective reporting.

For NRCSs, we used specific elements from the Risk Of Bias In Non-randomised Studies - of Interventions (ROBINS-I) tool related to confounding and selection bias.⁵⁸ We also used items from the Cochrane Risk of Bias tool that were not specific to randomized trials.

For single-group studies, we used the items from the above-mentioned tools that related to participant loss to followup, incomplete outcome data, selective reporting, and adequacy of descriptions of study eligibility criteria, interventions, and outcomes.

Data Synthesis and Analysis

Within the main report, data are summarized either in succinct tables that focus on outcome, interventions, and comparative (when applicable) results or in forest plots or succinct summary tables (for most topics). Appendix D contains the succinct summary tables for the antibiotics Key Question because of their large number and length. The rest of Appendix D includes the more detailed, study-level results for each topic. Appendix C contains detailed tables that describe study and participant characteristics, intervention (and comparator) details, outcomes (and definitions), and arm- and comparison-level results. Appendix C also includes tables providing study-level risk of bias assessments.

When feasible and appropriate, we conducted random effects model pairwise meta-analyses. Details are in Appendix A. Of note, for harms data related to KQ 4 (elective surgery), we meta-analyzed adverse event rates (proportions) when two or more studies reported sufficiently similar adverse events. The goal of these meta-analyses was to allow concise presentation of the adverse event results data; thus, we did not restrict these meta-analyses based on the similarities of the investigated surgeries or on the statistical heterogeneity among included studies (the differences in adverse event rates). To indicate the heterogeneity across studies, we also report the range of adverse event rates across studies.

Grading the Strength of the Body of Evidence

We evaluated the strength of evidence (SoE) addressing each major analysis for each KQ. We graded the SoE as per the AHRQ Methods Guide.^{59, 60} For each SoE assessment, we considered the number of studies, the study limitations, the directness of the evidence to the KQs, the consistency of study results, the precision of any estimates of effect, and other limitations (particularly sparseness of evidence). Based on these assessments, we assigned a SoE rating as being either high, moderate, low, or insufficient to estimate an effect. Outcomes with highly imprecise estimates (95% confidence interval extends beyond both 0.50 and 2.0), highly inconsistent findings across studies, or with data from only one study were deemed to have insufficient evidence to allow a conclusion.