Method

This study systematically identified practice guidelines regarding occupational risk management of employed women during their pregnancies and evaluated their methodological quality.

The types of documents considered for inclusion were practice guidelines, position statements or technical reports including recommendations, targeting women who were employed and became pregnant.

Results

Search results and selection

The search identified 1024 references of which 603 were obtained from bibliographic databases and 421 from guideline-focused resources and 73 documents were selected for detailed assessment. After analyzing their full text, 42 documents were excluded for various reasons (supplementary material S4). After exclusions, we included 31 documents in the review (figure 1).

Figure 1

Eligibility flowchart.

Quality of documents

The methodological quality of guidelines included varied considerably across the AGREE II domains (table 4).

Table 4

Methodological quality of included documents, according to AGREE II scores, and ranked from highest to lowest quality according to the score received at the ‘Rigor’ domain 3.

Guideline		DOMAIN I Scope		DOMAIN II Stakeholders			DOMAIN III Rigor		DOMAIN IV Clarity	DOMAIN V Applicability	DOMAIN VI Independence	Overall quality* (0–7)
RCP 2009 (51)		94.44		65.27			85.41		94.44	52.08	75.00	6
SVS 2023 (35)		86.11		87.50			71.87		88.88	35.42	56.25	6
RCP 2013 (52)		77.77		75.00			57.81		75.00	44.79	31.25	6
EHRA 2017 (33)		86.11		62.50			53.64		73.61	31.25	41.66	6
NIOSH 1977 (48)		84.72		62.50			36.97		52.77	42.70	14.58	3
NIOSH 2013 (49)		86.11		37.50			35.41		72.22	23.95	50.00	4
AEEMT 2011 (27)		70.83		50.00			32.81		63.88	35.41	4.16	4
SOGC 2017 (53)		65.27		44.44			32.29		52.77	7.29	43.75	3
INSS 20112 (44)		54.16		51.38			31.77		54.16	13.54	14.58	3
CDC 2010 (31)		80.55		54.16			29.16		55.55	33.33	16.66	3
CSN 2016 (32)		68.05		37.50			27.60		63.88	50.00	4.16	4
IBV 2004 (39)		75.00		54.16			27.08		72.22	55.20	37.50	3
ASPB 2019 (30)		44.44		23.61			26.52		66.66	33.33	4.16	3
ANMTAS 2008 (29)		75.00		56.94			21.87		62.50	35.41	4.16	3
ACOG 733 (25)		41.66		40.27			20.83		65.27	33.33	0.00	3
ACOEM 2016 (23)		56.94		38.88			20.83		44.44	31.25	64.58	2
INSS 2008 (43)		52.77		43.05			19.79		56.94	8.33	12.50	2
AMAT 2008 (28)		45.83		18.05			19.27		56.94	28.12	4.16	2
ACOG 533 (24)		54.16		13.88			17.18		66.66	16.66	12.50	2
ACOG 832 (26)		41.66		48.61			16.66		52.77	17.70	50.00	2
OSALAN 2005 (50)		51.38		18.05			16.14		26.38	16.66	12.50	1
IMEX (41)		66.66		25.00			15.62		56.94	32.29	6.25	1
EHRA 2023 (34)		77.77		69.44			15.10		34.72	22.91	35.41	1
ISTAS 2013 (47)		65.27		56.94			12.50		37.50	17.70	16.66	1
ISTAS 2008 (45)		80.55		41.66			11.97		65.27	29.16	4.16	2
INSHT/INSST 2011 (42)		79.16		23.611			11.97		43.05	14.58	29.16	1
ISTAS 2009 (46)		62.50		38.88			8.33		44.44	17.70	2.08	1
HSE INDG334 (36)		59.72		12.50			7.29		59.72	22.91	0.00	1
ILO 2004 (40)		65.27		29.16			4.68		62.50	33.33	8.33	2
HSE INDG373 (37)		51.38		22.22			4.68		45.83	29.16	0.00	1
HSE web (38)		100		33.33			0.00		55.55	2.08	0.00	1

*Score assigned by consensus among reviewers.

Domain III (rigor in development) was the domain with the lowest scores among domains related to methodological quality (domains III and IV). Only four guidelines achieved scores >50% (33, 35, 51, 52), indicating at least a moderate level of methodological quality in their development. Of those, only two guidelines achieved scores >70% (35, 51).

Seventy-one percent of the documents (N=22) presented important methodological limitations related to the lack (or limited description) of the methods used to formulate the recommendations, obtaining scores <30% (23–26, 28–32, 34, 36–43, 45–47, 50). The limitations most frequently identified were the lack of an explicit description of the methods used to conduct the literature search (24–26, 36–38, 40, 41, 45, 46) and a description of the methodology used to formulate the recommendations or limitations of the scientific literature used in all 22 documents. In 17 documents, the relationship between the scientific evidence and the proposed recommendations was not clearly explained (27, 28, 30, 32, 36–41, 44–46, 50, 53). Likewise, most guidelines provided little or no information on the procedures planned for future updating.

The scores for domain IV (clarity of presentation) were generally higher than those described for domain III. Twenty-four guidelines (77.4%) achieved scores ≥50%, indicating that, in general, the recommendations were presented in a clear and structured manner (24–33, 35, 36, 38–41, 43–45, 48, 49, 51–53). However, only 14 guidelines (45.2%) achieved scores >60%, demonstrating a higher level of clarity and greater ease in identifying the key recommendations (24, 25, 27, 29, 30, 32, 33, 35, 39, 40, 45, 49, 51, 52).

Despite these relatively favorable scores, some shortcomings were observed. Eight guidelines did not formulate their recommendations using specific and clear statements (23, 26, 31, 33, 37, 45, 47, 53), whereas in 11 documents the recommendations were not easily identifiable within the text (23, 31, 34, 36–38, 42, 45–48).

The scores obtained by the documents in domain I (scope and purpose) were generally higher. Only four guidelines scored <50%, mainly because they did not clearly define their objectives (25, 26, 28) or did not specify the clinical questions they intended to address (30).

Applicability (domain V) was one of the lowest-scoring domains. None of the guidelines provided comprehensive tools to facilitate the implementation of their recommendations in practice. Only three guidelines achieved scores ≥50% (32, 39, 51), mainly because they addressed certain aspects related to implementation, such as resource implications or auditing criteria, rather than providing practical implementation tools. In addition, 26 guidelines did not include criteria for monitoring or auditing the implementation of their recommendations.

Regarding the overall quality assessment, four guidelines received the highest score of six out of seven points (33, 35, 51, 52). In contrast, more than half of the guidelines included (51.6%) received overall quality scores of 1–2 (on a 7-point scale).

Discussion

Our systematic evaluation using the AGREE II instrument of the 31 documents included in this study revealed substantial variability in their quality. Although most documents clearly stated their objectives and presented their recommendations in a clear manner, only four achieved adequate levels of methodological rigor during their development and overall quality. Except for three documents, the rest raised serious concerns about their applicability. Most guidelines achieved better results in the other domains.

The most relevant finding of this study was the generally low performance in domain III (rigor of development). This reflects the limitations in reporting description of the methods used to search for and synthesize scientific evidence, as well as the processes used to formulate recommendations and the procedures planned for updating them. These shortcomings are particularly important given the central role of domain III in assessing the methodology used to develop guidelines and formulate recommendations, as well as their transparency and reproducibility. The lack of information on these processes could undermine the scientific credibility of the recommendations, reducing the practical value of these documents as tools to support their use in clinical practice, particularly in the field of occupational health. However, the AGREE II instrument assesses the methodological quality and transparency of guidelines’ development process rather than the scientific validity or clinical effectiveness of the recommendations themselves. Consequently, guidelines with limited methodological rigor may nevertheless contain recommendations supported by robust scientific evidence, whereas methodologically well-developed guidelines may rely on outdated or limited evidence. Accordingly, low scores in domain III should be interpreted primarily as reflecting insufficient reporting or methodological transparency rather than necessarily indicating that the recommendations themselves are inappropriate. Another important limitation was the absence of regular and explicit updating procedures in many of the included guidelines. As scientific evidence and working conditions continue to evolve, guidelines that are not updated have the risk of becoming obsolete and potentially unsafe. Similar findings have been highlighted in broader methodological literature, where updating procedures were infrequently performed as poorly reported across clinical fields (54, 55).

The scores obtained for domain IV (clarity) were more encouraging. In most guidelines, the recommendations were presented in a clear and structured manner. This suggests that recommendations can appear clear despite being supported by a scarce description of their development methods. However, only about half of the guidelines allowed readers to easily identify the recommendations within the document. This suggests that greater attention was given to the wording of recommendations than to facilitating their understanding by end users, indicating that the usability of these documents in routine occupational health practice could be further improved.

These findings are consistent with several recent systematic appraisals of general pregnancy-related clinical practice guidelines using AGREE II. Systematic reviews of guidelines addressing COVID-19 management, nutrition and other areas of maternity care have consistently reported that, although many guidelines score highly in scope and purpose and clarity of presentation, the majority perform poorly in rigor of development, applicability and editorial independence (56–60). This body of evidence shows that methodological limitations in the development of guidelines are not only presented in the field of occupational health but reflect a condition in many areas of medicine and pregnancy-related clinical practice.

Scores in domain V (applicability) were generally low. The absence of practical tools to facilitate the implementation of recommendations was observed in most of the included documents. In the management of occupational risks during pregnancy, this issue is especially relevant, because implementation of the recommendations depends on collaboration among clinicians, occupational health professionals, employers, and workers. Domain II (stakeholder involvement) also received relatively low scores, indicating limited participation of experts from different disciplines in the development of the documents included. The involvement of multiple stakeholders is therefore a key element for improving methodological quality of guidelines and their applicability and acceptance by both healthcare professionals and employers.

Regarding overall quality, ratings were closely associated with domain III scores. Although AGREE II does not establish official cutoff values, the strong relationship observed between overall quality ratings and methodological rigor suggests that reviewers considered the rigor of guideline development to be a major determinant of their global assessment. This finding is consistent with the purpose of the AGREE II instrument, which emphasizes methodological quality as the foundation for trustworthy recommendations.

These findings suggest that guidelines may influence the work-life balance of pregnant workers. Decisions based on low-quality guidelines can lead to unnecessary or poorly managed absences from work, which in turn can negatively affect workers’ professional development, their income, the financial stability of their families (15) and for employers. The correct implementation of clinical practice guidelines could mitigate these negative effects (15). Besides, emerging evidence indicates that organizational support, appropriate workload, and workplace accommodations improve maternal well-being and facilitate return to work after maternity leave (1, 4, 61, 62). Consequently, future guidelines should better integrate these psychosocial aspects into occupational risk management.

Going forward, guideline development should prioritize transparent evidence synthesis, explicit grading of recommendations, multidisciplinary participation, practical implementation tools, and regular updating procedures. Together, these elements would contribute to improving the relevance, applicability, and long-term effectiveness of guidelines in occupational and public health practice (18, 54, 55).

The findings of this study should be interpreted considering several limitations. First, although the AGREE II instrument was originally developed to evaluate clinical practice guidelines, its domains and items are sufficiently generic to allow its application to guidance documents in other fields, including occupational health. The AGREE II instrument does not recommend universal cutoff values to classify guidelines according to methodological quality. Likewise, there is no validated evidence regarding the sensitivity or discriminatory performance of specific cutoff values for distinguishing between high- and low-quality guidelines. Consequently, the 50% threshold adopted in this study should be interpreted as a pragmatic criterion established a priori to facilitate descriptive comparisons among documents rather than as a validated standard of guideline quality. Four reviewers independently conducted this appraisal, three of whom had formal training in the use of the AGREE II instrument, following standardized procedures consistent with the AGREE II methodology. Nevertheless, some degree of subjectivity is inherent to this type of appraisal and cannot be completely eliminated, which may explain the variability observed in individual item scores.

Second, the project originated within the Spanish occupational health context, which may have introduced some bias in the identification of relevant documents from other countries. However, the search strategy was specifically designed to identify documents published in English, Italian, Portuguese, and Spanish, and targeted searches were conducted on the websites of relevant national and international organizations. Moreover, the methodological quality of the Spanish guidelines was comparable to that of most guidelines identified from other countries.

Third, the review was restricted to publicly available documents, and therefore internal or unpublished guidelines may not have been identified.

Finally, differences in legal frameworks, occupational health systems, and maternity protection policies across countries may limit the direct transferability of some recommendations to other settings.

Despite these limitations, this study has several important strengths. To our knowledge, it is the first systematic review to identify and critically appraise the methodological quality of guidelines specifically addressing occupational risk management during pregnancy using the AGREE II instrument. The inclusion of documents from multiple countries provides a comprehensive overview of current international practice and enhances the relevance of the findings beyond a single national context. By identifying systematic methodological weaknesses across guidelines, this review provides a foundation for improving guideline development and promoting more transparent, evidence-based recommendations to support occupational health practice for pregnant workers .

Supplementary material