In Germany, guidelines are classified into four levels (S1, S2e, S2k, S3) based on the underlying methodology and are registered and published in a quality-assured guideline register of the Association of the Scientific Medical Societies (AWMF). In accordance with international methodological requirements, evidence- and formally consensus-based S3 guidelines have the highest methodological quality. They contain predominantly evidence-based recommendations for which a systematic review of the available evidence has been conducted. In addition, they include some recommendations derived from expert opinion on aspects for which there is little published evidence as well as some adaptations of other guidelines where these have been judged to be reliable and of high quality. All recommendations subsequently undergo a formal consensus process (Wissenschaftlichen et al. 2020).

For the development of this guideline on perioperative management of gastrointestinal tumors, a guideline panel was founded in 2019 by the leading professional societies (German Society of General and Visceral Surgery (DGAV) and German Society of Coloproctology (DGK)). This panel is composed of the coordinating team (MAW, SP, TOV), representatives of various professional societies, and selected experts in the field of perioperative care as well as patient representatives. The guideline panel is supported and supervised by representatives of the AWMF (MN) and the German Guideline Program in Oncology (GGPO) (MF and TL).

At a kick-off meeting in September 2020, an agreement was first reached on how to deal with conflicts of interest. Each member of the guideline group disclosed their conflicts of interest, which were then evaluated by the coordination team as either conflict-free or as low, medium, or high. Examples of how the conflicts were rated are given in Table 1, and the consequences are explained:

Table 1 Dealing with conflicts of interest

Subsequently, the research questions and patient-related outcome parameters developed by the coordination team were discussed and finally agreed upon by consensus. Furthermore, it was determined which guideline questions should be answered based on expert opinions and which should be evidence-based.

Subsequently, 6 working groups were founded to work on individual aspects of the project, whereby special attention was paid to the composition of the working groups: all groups were assembled interdisciplinary, containing patient representatives as well as known experts in the field. A list of the working groups with the respective PICO questions can be found in Supplement 1.

The evidence-based questions will be addressed by the individual working groups under the supervision of the coordinating team according to an 8-step process developed in line with the GIN-McMaster Guideline Development Checklist (Schünemann et al. 2014).

The guideline is registered in the guideline register of the AWMF (register number 088—010OL, available at https://www.awmf.org/leitlinien/detail/anmeldung/1/ll/088-010OL.html).

8-step process of evidence synthesis

Processing of the evidence-based questions follows a multi-stage scheme that defines the methodological procedure. Figure 1 summarizes the 8 steps of the evidence review and recommendation development. Below, the individual steps are explained in detail.

Fig. 1

Summary of the 8 steps of the evidence review and recommendation development

Step 1: Starting point

The predefined guideline questions are prepared for evidence-based answering. In some cases, guideline questions are further subdivided as only some aspects are answered in an evidence-based manner or the answer might differ for subgroups (e.g., intraoperative drains depending on the type of surgery performed).

The PICO schema (population, intervention, comparator, outcome) is used to frame the research question, and the predefined outcomes from the kick-off meeting are extended or adapted as necessary.

Step 2: Literature search

The strategies for the systematic literature search are being developed by the coordination team. Details on the search strategies, the timing of the literature search, and the databases searched will be presented in the evidence report of the corresponding chapters of the guideline.

If the guideline is not completed within 6 months of the search, updates will be conducted to ensure that no new literature is missing from the guideline.

The literature search is also conducted by the coordination team. First, a search for systematic reviews (SR) containing meta-analyses of randomized controlled trials (RCTs) is done. Afterwards, a search for RCTs is conducted, either as an update search for RCTs published after the baseline review or, if no high-quality review could be found, to conduct an own meta-analysis.

Step 3: Literature selection

The title and abstract screening is conducted independently by two members of the coordination team, with a third reviewer resolving any disagreements using the online screening tool Rayyan (Ouzzani et al. 2016). After obtaining the relevant full texts, these are made available to the working groups for full text screening. A pre-selection of the systematic reviews on which the answer to a guideline question could be based is made by two independent working group members on the basis of the pre-established inclusion and exclusion criteria. Disagreements are resolved by discussion and, if necessary, consultation with the guideline coordinator responsible for the working group.

The process of study selection will be presented in the evidence report of the guideline via flow diagrams (Page et al. 2021).

Step 4: Qualitative assessment of the included publications

Systematic reviews that meet the inclusion criteria are assessed by the working group according to AMSTAR 2 (Shea et al. 2017). The AMSTAR 2 assessment is used to select one or more reviews to serve as basis for answering the research question. For this purpose, an assessment sheet focusing on the aspects relevant to the guideline was prepared. Among others, an adequate assessment of the risk of bias of the included studies of the meta-analysis is essential for selection as baseline review. Additional RCTs are assessed with the Cochrane Risk of Bias 2 Tool (RoB2) (Sterne et al. 2019), whereby the assessment is performed individually for each of the predefined outcomes.

The AMSTAR 2 and RoB2 assessments of the baseline reviews and included RCTs will be reported in the evidence report of the guideline in the corresponding guideline chapter.

Step 5: Summery of findings

For all outcomes predefined in the kick-off meeting, summary of findings tables are created using GRADE proGDT (McMaster University and Evidence Prime 2022). All summary of findings tables will be published in the evidence report of the guideline. If no evidence can be found for a predefined outcome, it is nevertheless listed in the summary of findings table and the missing evidence is marked accordingly.

If a research question is answered based on only one meta-analysis, the pooled data from the meta-analysis is used to create the summary of findings table. However, if additional RCTs that meet the inclusion criteria but are not included in the meta-analysis are identified (for example, because they were published after the search period of the meta-analysis), the following procedure was defined by the coordination team in consultation with the AWMF and the GGPO:

Option 1: If additional RCTs substantially change the statement of the underlying review, a new meta-analysis is calculated

Option 2: If additional RCTs are congruent in statement with the underlying review, the following procedures are possible:

If the additional RCTs are rather small in size (number of patients) compared to the review or present a high risk of bias, a narrative mention of the new RCTs in the text of the guideline chapter is sufficient and only the review is presented in the summary of findings table

If the additional RCTs are large compared to the review and could improve the certainty of evidence according to GRADE, they shall be included in summary of findings table

Step 6: Assessment of the certainty of the evidence

The assessment of each outcome in the summary of findings table is made separately by two members of the working group. The following aspects are assessed, leading to an increase or decrease in the confidence of the evidence (Schünemann 2022; Schwenk 2009).

Risk of bias: A high risk of bias or even some concerns about the risk of bias in one or more of the included studies for the outcome may downgrade confidence in the evidence.

Inconsistency: Moderate or considerable heterogeneity between studies that cannot be explained by subgroup analysis may downgrade confidence in the evidence.

Indirectness: Differences between the original PICO question and the included studies regarding population, intervention, comparator, or outcomes may downgrade confidence in the evidence. In particular, when surrogate outcomes are used, transferability must be critically assessed and confidence in the evidence may need to be adjusted.

Imprecision: A small sample size or limited number of events as well as wide confidence intervals are indications of uncertainty about the magnitude of the effect and may lead to a downgrading of confidence in the estimated effect.

Large effect: If the effect is large (RR either > 2.0 or < 0.5 based on consistent evidence from at least 2 studies), this can lead to an increase in confidence in the evidence. In the case of a very large effect (RR either > 5.0 or < 0.2 based on direct evidence with no major threats to validity), even a twofold increase in confidence in the evidence is possible.

Plausible confounding: Circumstances of the studies not taken into account in the meta-analysis, which may lead to an overestimation or underestimation of the effect, may influence the confidence in the evidence.

Dose response gradient: The demonstration of a dose–response relationship can potentially increase the confidence in the evidence.

Our confidence in the evidence is then expressed as one of four GRADE levels of certainty (see Table 2) (Balshem et al. 2011):

Table 2 GRADE levels of certainty of the evidence (Balshem et al. 2011)Step 7: Development of a treatment recommendations (EtD)

To derive treatment recommendations from evidence, the GRADE Evidence to Decision (EtD) framework provides a systematic and transparent approach (Alonso-Coello et al. 2016). We developed a template of an EtD framework adapted to our guideline in which the following aspects of decision-making are taken into account:

Benefits of the intervention

Possible harm from the intervention

Reliability and quality of evidence

Preferences and acceptability

Resources and feasibility of the recommendation

According to the EtD principle, the pre-formulated questions are to be answered by the working groups and it is to be stated where the findings come from (evidence or expert opinion). Based on the answers given in the EtD framework, the working group then makes an assessment and derives a conclusion. This conclusion forms the basis for a recommendation text, which is drafted by the working group.

Step 8: Finalizing the guideline

Analogous to the consensus-based questions, the individual recommendations developed by the working groups are transferred by the coordinating group to the content management system (CMS) of the GGPO. The CMS offers the possibility for all voting members to review, comment on, or approve the recommendations. The comments received are then discussed again in the working group, and, if necessary, amendments are made to the recommendation or the background text. The basis for any discussion is the EtD framework completed in step 7.

Finally, all recommendations are discussed in a consensus conference with neutral moderation and formal anonymous voting according to the NIH methodology for consensus conferences (National Institutes of Heatlh n.d.). Here, the recommendations are presented by the respective working groups to the whole guideline group, and it is explained how the recommendation was developed using the EtD framework. Possible amendments are discussed and voted on with the aim to reach a consensus. The degree of agreement for a recommendation will be published in the written version of the guideline below the recommendation (> 95% = strong consensus/ > 75–95% = consensus). If no consensus is reached (> 50–75% = majority agreement/ ≤ 50% = no majority agreement), this is also explained.

Ethics and dissemination

As soon as the guideline is completed, the draft version will be made available to the (professional) public for external review over a period of 6 weeks. In addition, all organizations involved in the development of the guideline including patient representatives will be asked to circulate it to their members for review, including a structured comment form. Any changes resulting from the consultation phase are agreed upon within the guideline group and documented in the guideline report.

Subsequently, the guideline will be approved by the executive boards of all participating professional societies/organizations, and the fulfillment of the S3 requirements are verified before the guideline is published in the AWMF guideline register.

In addition to the long version with background information, an abridged version and a patient version as well as methods report with evidence summaries will also be published. This report—analogous to this protocol—is intended to ensure the transparency of the guideline development process and thus the trustworthiness of the guideline.

All guideline versions as well as the evidence report will be freely accessible via the AWMF and GGPO websites and via an app provided by the GGPO.