Quality appraisal of clinical guidelines for the management of constipation according to AGREE II instrument

Constipation is one of the most frequent complaints in health establishments, affecting approximately 20% of the world population. The key to the successful management of this disorder is its early identification and immediate treatment with nondrug measures as the main therapeutic alternative. While clinical guidelines are informative documents that include recommendations aimed at optimizing patient care, this study aims to evaluate the quality of these documents available in the management of this condition. This study evaluated clinical guidelines available in databases, scientific societies and class councils, which had as their theme the management of constipation. The methodology used to evaluate the quality of these documents was the application of the AGREE II instrument. Kappa test of agreement between evaluators was applied to corroborate uniformity in application of the AGREE II instrument. Of the eleven documents analyzed, only three presented the criteria for recommendation proposed for use in clinical practice, with a general average of quality indexes ranging from 93 to 72%. One guideline was recommended with modifications in its structure, with a general average of quality indexes of 44%. The mean overall score among the analyzed documents was 49.01%. Methodological care and expression of information in a concise and clear way, both of the content and methodology employed, result in documents of great quality, however, no aspects were present in most of the guidelines analyzed. Furthermore, it is observed that there are still points that should be improved.


Introduction
Constipation is one of the most frequent health establishments, affecting approximately 20% worldwide. (Antunes et al., 2019;Garcia et al., 2016) In general, the prevalence ranges from 2 to 27% according to locality, population group and type of survey used for diagnosis, making its real prevalence difficult to determine. (Antunes et al., 2019;Silva & Sabino Pinho, 2016) It also affects up to 30% of children and represents approximately 3 to 5% of all visits to pediatricians, in addition to a quarter of referrals to pediatric gastroenterology. (DynaMed [Internet]. Ipswich (MA): EBSCO Information Services. 1995 -. Record No. T900171, 2018; National Collaborating Centre for Women's and Children 's Health, 2010;Paquette et al., 2016;Robin et al., 2018;Sood, n.d.;van den Berg et al., 2006) In the elderly population, it is estimated that 25 to 50% of all people over 65 years of age have this clinical condition. (American Gastroenterological Association, 2013;Choung et al., 2007;Donald et al., 1985;Emmanuel et al., 2017;Everhart et al., 1989;Harari et al., 1996;Junior et al., n.d.;Rao, n.d.;Sandler et al., 1990;N J Talley et al., 1996;Nicholas J Talley et al., 1992;Wald et al., 2008;Whitehead et al., 1989) The key to the successful management of this disorder is its early identification and immediate treatment, emphasizing holistic care and multidisciplinary support when necessary. (Greenwald, 2010) For the recovery of regular defecation habits, the literature addresses nonpharmacological care. In addition to being among the initial conducts, they are the main steps to prevent  Saúde, 2016) In addition to presenting quality information based on the best available scientific evidence, the guidelines should be easily accessible by healthcare professionals. Due to the large volume of information and variability in the quality of scientific information generated in the health area, there is a need to develop a synthesis that facilitates access to this knowledge and enables recommendations based on results from multiple sources, providing scientific support for decision making. (Ministério da Saúde, 2016) Considering all the aspects demonstrated above on the panorama of constipation in the world and the usefulness of clinical guidelines in the use of health care, this study aims to evaluate the quality of these available documents through the analysis of the format of construction of these tools and the level of scientific evidence present in them.

• Identification and selection of guidelines
This methodological-type study evaluated clinical guidelines available in databases, scientific societies and class councils, which had as their theme the management of constipation. From September to November 2019, the literature was searched in the following databases: PubMed, Cochrane, Biblioteca Virtual de Saúde do Brasil, Google Scholar and Scientific Societies (in addition to those recommended by Up To Date), National Institute of Care and Excellence (NICE), and on sites of professional categories, using the descriptors "constipation AND guidelines". These sources of information were chosen based on the possible quality of evidence and access to health professionals.
Titles, abstracts and full texts were analyzed to select clinical guidelines according to the inclusion and exclusion criteria. The inclusion criteria for the bibliography researched were i) to be a document with general aspects about constipation (evaluation, differential diagnosis, prevention, treatment, management); ii) to cover the general public, since these are potential targets of health care in relation to constipation; iii) present diversity of localities of origin; iv) have been published in the last 11 years considering the research period of the literature of this study. Therefore, clinical guidelines were included that addressed general aspects of the clinical picture and whose focus on the management of constipation was at the level of primary care/selflimited disorder.
Regarding the exclusion criterion, we first adopted the fact that it was in duplicate and later was a literature that focused on a) chronic constipation, b) themes related to hospital application (e.g., constipation induced by opioids and other drugs, chemotherapy and palliative care, fecal disimpaction), c) irritable bowel syndrome, d) sacral neuromodulation disorders, e) documents that did not have intestinal constipation as the main subject.
• Instrument for guidelines evaluation -AGREE II The group of four evaluators composed of a graduate student (BMCSA), a pharmacist (LBPB) and two pharmacy undergraduates (APC/BFRL) evaluated the clinical guidelines selected using the AGREE II instrument (Appraisal of Guidelines for Research & Evaluation). This evaluation occurred between May and June 2020. AGREE II is a tool developed to assess the variability of the quality of clinical guidelines. Its validation was based on analyses of more than 100 guidelines selected and independently evaluated by more than 200 evaluators from different parts of the world. [28,29] It is therefore capable of evaluating, if described in the document, the methodological rigor and transparency with which a clinical guideline is developed. The domains analyzed are a) scope and purpose of the guideline, b) involvement of stakeholders, c) rigor of methodological development, d) clarity of presentation, e) applicability, and f) editorial independence. (Consórcio Agree, 2009) Research, Society andDevelopment, v. 11, n. 3, e57911327086, 2022 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v11i3.27086 4 In the end, this instrument allows two types of analysis. The first is related to the classification of the overall quality of the guideline, observed by the average of all domains evaluated of each guideline. Second, the author's conclusion about which documents are recommended for use. [30] For the interaction of the results, it is worth noting that the instrument does not define minimum scores for the domains or score patterns to characterize literature quality. Consequently, it is free to interpret the evaluator as long as it takes into account the context to which the tool is being applied.(Consórcio Agree, 2009) • Data analysis The data obtained from the guidelines after the evaluation by AGREE II were analyzed through six domains that are divided by items that, in all, add up to 23 questions. In each, a score is elected ranging from 1 (totally disagree) to 7 (totally agree). Finally, a quality index is calculated for each of the six domains of AGREE II,(Consórcio Agree, 2009) ranging from 0 to 100% according to the relationship between the score obtained and the maximum score that can be achieved.
For the definition of the documents recommended, considering the great emphasis of this research to evaluate the quality of clinical guidelines based on scientific evidence, Santana et al.'s (2018) methodology with alterations was used. (Santana et al., 2018) to consider a recommended or recommended guideline with modifications: 1) For recommended guidelinesquality index above 50% in development rigor and publication average; b) For recommended guidelines with modificationsdevelopment rigor and presentation clarity above 50%. The other guidelines that did not meet these criteria were not recommended. The interpretation of the results was performed in October 2020.
In the same period in which the evaluation of clinical guidelines was realized, the Kappa test of agreement between evaluators was applied using the free online platform VassarStats. Among the forms of kappa coefficient calculation, quadratic weighting was used, and for this study, an acceptable coefficient of agreement above 0.4 was considered. The choice was because this form is the most appropriate way to evaluate the agreement between examiners because it allows the interpretation equivalent to the intraclass correlation coefficient. (De & Ronsoni, 2013;Griep et al., 2003;Jakobsson & Westergren, 2005;Sim & Wright, 2005;Tooth & Ottenbacher, 2004)

Results
As observed in the flowchart (Figure 1), we initially identified 82 documents on constipation in the databases. Five documents were excluded by duplicity. Of the remaining documents, only 11 documents were elected for quality assessment after selection according to the criteria of inclusion and exclusion (Table 1).
It was also evaluated the presence of the use and mention of systems of classification of degree of recommendation and evidence present. In this sense, 27.3% (n=3) of the analyzed studies used the GRADE system (Grades of Recommendation Assessment, Development and Evaluation). A single guideline associated the Oxford system with the GRADE system. Research, Society and Development, v. 11, n. 3, e57911327086, 2022 (CC BY 4. The results of AGREE II application in the selected guidelines are grouped in Table 2. The application of the Kappa test of agreement between evaluators in the sum of the final results obtains a coefficient of 0.5799, meaning a moderate value, although within the stipulated range, to accept the results obtained (Kappa >0.4). (De & Ronsoni, 2013;Landis et al., n.d.;Molino et al., 2016) It should also be considered that the AGREE II instrument does not necessarily seek agreement among the evaluators.
Its proposal to address the variability of the analyzed guidelines is obtained precisely by the flexibility and subjectivity it provides to each evaluator. (Ronsoni et al., 2015) 4. Discussion As seen in Table 2, 100% of the evaluated documents obtained results in a score above 50%. The average of the results was close to 80%, and the guideline that had the lowest score presented a value of 53% (CG 02). This can demonstrate that the guidelines analyzed clearly bring the objectives they want and to whom the recommendations are applied.
Another aspect that should be observed is that the highest scores (>90%) were attributed to publications aimed at the treatment of children. Understanding that domain 1 also aims to clarify the possible impact on the health of the target audience, observing the clear definition of who is intended for the material, it is evident a greater prudence for this purpose in the case of children. It demonstrates, perhaps, that pediatric professionals require this attention and seek more assertive answers. From the obtained results, only three documents had evaluations above 50% (CG 03, 05 and 10), especially CG 05, which presented all requirements with high quality, obtaining 100% evaluation. Once again, they are items that are more completely present in pediatric guidelines, reinforcing the idea of materials elaborated more rigorously for this public. For the other guidelines that showed little score, the value obtained by AGREE II ranged between 22-43%, suggesting a possible lack of quality or presentation of this information throughout their literature, making it impossible to completely evaluate.

• Domain 3 -Rigour of development
The third domain concerns the process used to collect and synthesize the evidence and the methods used to formulate the recommendations and update them.(Consórcio Agree, 2009)The whole methodological description must appear objectively and completely. This stage of analysis stands out not only for the number of items to be judged in comparison to the other domains but also because it is one of the fundamental evaluations for determining works that are based on scientific and quality evidence.
Of the literature evaluated, only four reached more than half of the total points for this domain (CG 01, 03, 05, 10), consequently evidencing some level of methodological rigor. The remaining works scored between 10-17%, except for CG 02, Research, Society andDevelopment, v. 11, n. 3, e57911327086, 2022 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v11i3.27086 9 which obtained results of 41% at this stage of the evaluation. These results, from more than half (54.5%) of the guidelines analyzed on constipation, derive from some scenarios that were frequent in these guidelines. Aspects such as the lack of the approach on the methodology used for its construction, the inaccessibility of this methodological information, especially in full, or the delivery of this information described in general generating gaps of information pertinent to this analysis were constantly identified. In all cases, these limitations interfere with the evaluation of its reproducibility, as well as the quality of recommendations and evidence.

• Domain 4 -Clarity of presentation
The fourth domain of AGREE II concerns the language, structure and format of the guideline. (Consórcio Agree, 2009)This is how the text and ideas are organized and presented throughout the guideline. It also evaluates how the features of tables, images, charts, graphics and highlights snippets are enhanced to make key information easier to understand and identify.
Among the clinical guidelines evaluated, all obtained scores above 50% among the items evaluated in this domain, with emphasis on CG 05 for obtaining 100% of the score among the four evaluators. This result may suggest that, among the authors, there are great efforts in the presentation of the approached content. Consequently, this would result in negligence in detailing the methodologies employed.

• Domain 5 -Applicability
The fifth domain concerns likely facilitating factors and barriers to the implementation of recommendations, strategies to improve implementation, and involvement of resources related to the use of the guideline. (Consórcio Agree, 2009) Regarding this topic, it was noticed that the guidelines, in general, presented low-quality indexes. CG 05 was the only exception, with 71% grade in this domain. It was the only guideline that brought more evident the discussion on this topic and the factors implicating in its reproduction. As seen in domain 3 on development rigor, these are points that lack information that would add greater value to the guideline, making them a more robust instrument. Moreover, they direct their users to an appropriate application of the content, increasing the support to the suggested treatment and the probability of successful results.

• Domain 6 -Editorial independence
The domain related to editorial independence concerns the formulation of recommendations. It avoids bias arising from conflicting interests between authors and collaborators involved in the construction of the clinical guideline. (Consórcio Agree, 2009) In this last topic, the heterogeneity of the results generated stands out. Guidelines such as CG 01, 04, 09 and 11 had no score, caused by the evident unavailability of information on conflict of interest and opinions of the funding agency if it existed.
This shows the necessity, as in previous domains, for this information to be given expressly, ensuring transparency in the product preparation process and greater reliability in its content.

• Recommendation of clinical guidelines -Global analysis
Regarding the general quality of the clinical guidelines evaluated, it is perceived that the results presented are heterogeneous. When these eleven results were averaged, we obtained a result of 49.01%, which may suggest a moderate general quality of these available clinical guidelines. However, it is worth mentioning some documents, such as CG 05, that obtained the highest average result in the overall evaluation with 93%, followed by CG 10 and CG 03 with 74 and 72%, respectively.
According to the criteria established by this study, CG 03, 05 and 10 are considered "recommended". As mentioned, the three guidelines have as their target audience the children's and youth range.
Only CG 01 showed a compatible result with "recommended with modification". Although its development rigor reached satisfactory results, the publication average did not obtain performance that fit the first decision criterion. It is observed that the impact of the quality of the publication as a whole is due to the results obtained mainly in domains such as editorial independence and applicability, where this information was not clearly mentioned throughout the literature.
For the other clinical guidelines (CG 02, 04, 06-09 and 11), due to their performances after the application of the AGREE II instrument, they were classified as "not recommended". These documents did not present clear information about their elaborations from an evidence-based health perspective.
• Other aspects observed in the clinical guidelines There is a unique aspect related to CG 10 compared to the other studies evaluated that should be highlighted. Although descriptors were used to search for clinical guidelines, when analyzing this document, it was noted that it was a methodological guideline for the construction of clinical guidelines on the subject of constipation in children and young people. Nevertheless, taking into account its high performance in the application of AGREE II, we can admit that such document, although not a clinical guideline in its essence, did not fail to present sufficient quality in information and construction to be used as a tool for conduct guiding.
Additionally, it was observed that only three of the eleven guidelines had some recommendation classification system, which were the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) and Oxford (Oxford Centre for Evidence-Based Medicine: Levels of Evidence). These systems are strategies to categorize the degree of a recommendation (very or little recommended) and the level of evidence it has to support its strength. In this study, unlike what was stipulated, it was not possible to observe a significant repercussion of the presence of recommendation classification systems with the decision to recommend the guidelines analyzed according to the global evaluation (Table 3). However, it should be considered that the sample number involved in this analysis is relatively low for a more robust conclusion. Table 3. Comparison between recommended guidelines and the presence of a recommendation classification system. * Recommended with modifications. Source: Authors.
Two main events enabled this outcome. First, the impact of poor performance on the other items of the development rigor domain (observed mainly in CG 02), especially with regard to discussing the methodology of search and selection of evidence and limitations of their studies. The second factor involves the guidelines recommended in this study regardless of the presence of recommendation classification systems (CG 05 and 10), which is associated with its high quality of construction and methodological rigor. In them, their methodologies approach the formulation of questions as guidelines for the selection of evidence for the recommendations and analysis of the evidence selected by a committee and registered external stakeholders.(National Institute for Health and Care Excellence, n.d., 2014) Finally, the reasons that implied quality clinical guidelines were those that presented objectivity in the general information as well as in their proposals, richness of content and indication of tools for application, complete and descriptive methodology and variety of specialists present in the working group. There is also a very common problem, especially with regard to Brazilian documents on the definition of a single terminology for these guidelines, as mentioned in the study by Santana et al. (2018), limiting access to or meeting these literatures. (Santana et al., 2018) Despite the limitations presented, it is noteworthy that this study is the first to assess the methodological quality of guidelines on constipation available in databases used worldwide by health professionals. Thus, it supports the conduct of professionals in relation to the use of guidelines for decision-making in clinical practice.

Conclusion
The analysis of these obtained results allows us to visualize an overview of the literature that is available to health professionals as a tool to support conduct for the management of constipation. Certainly, these documents help greatly in providing information about the clinical condition and treatment alternatives. However, with this study, we can observe gaps and points that can be optimized when using AGREE II so that health professionals can use the best information and recommendations with well-defined levels of scientific evidence. In this way, they can guarantee the offer to the patient of excellent health care.
Based on this scenario, where the intervention of the health professional in the care of the self-limited health problem would generate benefits for the resolution of the condition and which instruments of synthesis of scientific evidence play an important role for the guidance of conduct, this study would also justify the proposal of a clinical guideline. Directed by the AGREE II evaluation performed previously, it is suggested that the guideline addresses the management of constipation as a self-limited disorder, especially in pharmaceutical care where this health professional has as a favorable point the ease of daily contact with the patient in pharmacies to solve minor disorders.
Ultimately, the AGREE II evaluation carried out by this study aims to promote that future researchers can develop guidelines with the best information and recommendations with well-defined levels of scientific evidence to support care activities by health professionals.