National trends of episiotomy in non‐instrumental vaginal deliveries (2013–2022): A multi‐data source approach

Abstract

Objective

Episiotomy and obstetric anal sphincter injuries (OASIS) are recognized indicators of intrapartum care quality. The aim of this study was to assess episiotomy trends in non‐instrumental vaginal deliveries across all Portuguese public maternities over the last decade. Rates of third‐ and fourth‐degree perineal lacerations were also evaluated.

Methods

A retrospective secondary data analysis was conducted using data from an official national health entity, including all non‐instrumental vaginal deliveries in the Portuguese National Health Service between January 1, 2013 and December 31, 2022. Information on episiotomy and third‐ and fourth‐degree perineal lacerations was extracted. Descriptive statistics were presented using absolute and relative frequencies. Linear regression was used to assess trends. Additional national databases were consulted to validate findings, and data quality was assessed using the intraclass correlation coefficient (ICC). A significance level of 5% was considered.

Results

Of 689 908 registered deliveries, 490 322 (71%) were vaginal, and 361 887 (52%) were non‐instrumental vaginal deliveries. Among the latter, the episiotomy rate was 42.5%. A significant decrease was observed in episiotomy rates, from 63% in 2013 to 21% in 2022, (P < 0.001), accompanied by a significant increase in third‐degree perineal lacerations (0.15% to 0.31%, P < 0.001), while fourth‐degree perineal lacerations remained stable (0.03% to 0.02%, P = 0.001).

Conclusion

A restrictive episiotomy policy was increasingly adopted in Portugal over the last decade. Although rare, third‐degree perineal lacerations increased during this period. Further research is needed to determine optimal intrapartum care practices regarding episiotomy use. Standardized data coding and improved access to anonymized national datasets are essential for accurate monitoring and cross‐country comparability.

Abbreviations

ACSS

Administração Central do Sistema de Saúde

CPDO

Portuguese Consortium of Obstetric Data

DGS

Directorate‐General of Health

ICD

International Classification of Diseases

INE

National Institute of Statistics

NICU

Neonatal Intensive Care Unit

OASIS

Obstetric Anal Sphincter Injuries

OECD

Organization for Economic Co‐operation and Development

1. INTRODUCTION

Prenatal and intrapartum care are continuously evolving, driven by evidence‐based practice and the ongoing revision of management guidelines and protocols. Disparities in obstetric practice raise concerns about healthcare equity across different regions of the world. ¹ , ²

Episiotomy remains one of the most common obstetric procedures, historically indicated to facilitate vaginal delivery and prevent severe perineal trauma, such as Obstetric Anal Sphincter Injuries (OASIS). ² , ³ , ⁴ , ⁵ , ⁶ Third‐ and fourth‐degree perineal lacerations are categorized as OASIS; in fourth‐degree perineal lacerations, the anal epithelium or rectal mucosa is also affected. ⁷ These injuries are considered severe obstetric complications and represent the leading cause of postpartum fecal incontinence. ⁸

However, since the late 1980s, routine use of episiotomy has been increasingly questioned, due to limited evidence of benefit and concerns over complications, including pain, infection, dyspareunia, and incontinence. ³ , ⁵ , ⁶ , ⁹ , ¹⁰

In response, several international guidelines—including those from the WHO ¹¹ —now recommend a restrictive approach. Episiotomy should be reserved for specific clinical situations, with decisions based on individualized judgment. ⁴ , ⁷ , ¹² Following these recommendations, episiotomy rates have significantly declined, particularly across Western European countries. ¹³ Although some studies have reported increases in OASIS with declining episiotomy use, this association remains controversial. ¹⁴ , ¹⁵

In Portugal, concerns over high episiotomy rates gained public attention after the 2010 European Perinatal Health Report, which identified Portugal as having the second‐highest rate (73%) among 20 countries. ¹⁴ , ¹⁶ This finding, amplified by social media, ignited a national debate about intrapartum care, particularly regarding non‐consensual or routine episiotomy and broader obstetric practices, with activists framing them as “obstetric violence”. ¹⁷

This growing public concern led to resolution no. 181/2021 by the Portuguese parliament, which called for the reduction of routine and non‐consented episiotomy practices. ¹⁸ In this context, the need for accessible, reliable data on obstetric care in Portugal became evident. However, as some physicians have pointed out, official data on patient experiences remain difficult to access. ¹⁹ , ²⁰

Although some maternity protocols were previously adopted, formal clinical national guidance in Portugal advocating the restrictive use of episiotomy was first published in 2023. ²¹

Monitoring episiotomy and OASIS rates is essential for assessing the quality and equity of obstetric care. ¹⁶ , ²² The aim of the present study was to analyze trends in episiotomy during non‐instrumental vaginal deliveries in Portugal between 2013 and 2022, and to evaluate the incidence of OASIS over the same period. We hypothesize a decrease in episiotomy rates without a corresponding rise in OASIS incidence. To our knowledge, no prior studies have examined this data in Portugal.

2. MATERIALS AND METHODS

A retrospective secondary data analysis was conducted using the National Inpatient Database of the Portuguese Central Administration of the Health System (Administração Central do Sistema de Saúde [ACSS]), ²² an agency under the Ministry of Health. The procedures and diagnosis of every hospital discharged patient are coded by medical staff specially prepared for this task, according to the International Classification of Diseases, 9th revision, Clinical Modification (ICD‐9‐CM), ²³ and more recently, the International Classification of Diseases, 10th revision, Clinical Modification and Procedure Coding System (ICD‐10‐CM/PCS). ²⁴ Data is then grouped according to the diagnosis‐related groups (DRG), ²⁵ currently by the All‐Patient Refined (APR‐DRG) version 31. Monthly data from all national public hospitals, including those in Portugal's autonomous regions (Madeira and Azores Islands), are collected. Data collection uses a unique patient identifier and a unique discharge number, both linked to a diagnostic coding based on the International Classification of Diseases (ICD).

This study is reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline.

Annual data of interest were requested and obtained via email from the National Inpatient Database of the ACSS ²² using an ICD classification list (available in Table S1), covering the period from January 1, 2013, to December 31, 2022. The data focused on non‐instrumental vaginal deliveries, episiotomy practices, and intrapartum perineal lacerations. The primary outcomes of interest were episiotomy rates and third‐ and fourth‐degree perineal lacerations (OASIS) rates, used to assess national trends over the 10‐year period under analysis. The national definition of birth, as adopted by the National Institute of Statistics (INE), ²⁶ is “The complete expulsion or extraction from the maternal body of one or more fetuses of 22 or more weeks of gestation, or weighing 500 g or more, regardless of whether they are born alive and whether the delivery is spontaneous or induced”. The WHO‐RCOG four‐degree classification of obstetric lacerations ²⁷ was applied.

In addition to the data provided by the ACSS, ²² we consulted other official online databases, including the National Institute of Statistics (Instituto Nacional de Estatística [INE]), ²⁶ the Portuguese Consortium of Obstetric Data (Consórcio Português de Dados Obstétricos [CDPO]), ²⁸ the Organization for Economic Co‐operation and Development (OECD), ²⁹ and the European Perinatal Health Report of the EURO‐PERISTAT's 2010. ¹⁶ These sources were consulted to cross‐check the data provided by the ACSS. No data from these external sources were merged with the ACSS data.

INE ²⁶ information is based on annual hospital data surveys. The number and mode of birth (eutocic/dystocic/cesarean/other) for each maternity unit across the study period (January 1, 2013, to December 31, 2022) were available. However, INE does not routinely publish data on episiotomy practices or intrapartum perineal lacerations. Therefore, in addition to ACSS data, ²⁹ these variables were obtained from the CPDO ²⁸ and OECD ²⁹ databases. Since 2019, the CPDO collects obstetric data from 13 public maternity hospitals that implemented the digital clinical records system “ObsCare”. This system is updated monthly, ensuring timely access to relevant information. The CPDO data used included the total number of births, mode of delivery, episiotomy rate in all vaginal deliveries and specifically in non‐instrumental vaginal deliveries, as well as the number of third‐ and fourth‐degree perineal lacerations. Data on OASIS rates were also compared with those of the OECD health statistics database, ²⁹ which is based on ICD‐10‐CM/PCS ²⁴ coding provided by each member country, including Portugal.

The anonymized data available from these sources enabled cross‐validation of ACSS statistics, particularly regarding the “total number of births,” “non‐instrumental vaginal deliveries,” “total number of episiotomies,” “episiotomies in non‐instrumental vaginal deliveries,” and “total number of third‐ and fourth‐degree intrapartum perineal lacerations” occurring in Portuguese public hospitals. Instrumental vaginal deliveries, non‐vaginal deliveries, and private or out‐of‐hospital births were excluded.

Independent variables included the year of birth (2013–2022) and mode of delivery (non‐instrumental vaginal delivery, with or without episiotomy). In summary, the primary outcome was the episiotomy rate in non‐instrumental vaginal deliveries in public hospitals. The secondary outcome was the rate of third‐ and fourth‐degree perineal lacerations (OASIS) in the same population. Trends in both outcomes over the 10‐year study period (2013–2022) were analyzed.

2.1. Statistical analysis

Variables are described using absolute and relative frequencies. Linear regression was applied to evaluate time trends in episiotomy and OASIS rates.

The reliability of the ACSS database was assessed by comparing it with data from the other data sources using the intraclass correlation coefficient (ICC). Consistent trends observed across the other databases support the reliability of the data source used. A 5% significance level was adopted.

3. RESULTS

Throughout the 10‐year study period, a total of 689 908 deliveries, encompassing all modes of delivery, occurred in public hospitals. Among these, 490 322 (71%) were vaginal deliveries, and 361 887 (52%) were non‐instrumental vaginal deliveries (Table 1). Overall, episiotomy was performed in 42.5% of non‐instrumental vaginal deliveries. This proportion declined significantly from 63% in 2013 to 21% in 2022, at an average annual decrease of approximately 4.5 percentage points (β = −4.549, P < 0.001). The rate of third‐degree perineal lacerations in non‐instrumental vaginal deliveries increased significantly, from 0.15% in 2013 to 0.31% in 2022, corresponding to an average yearly rise of 0.021 percentage points (β = 0.021, P < 0.001). In contrast, the rate of fourth‐degree perineal lacerations remained stable over the same period (β = −0.001, P = 0.580), with 0.03% recorded in 2013 and 0.02% in 2022. Annual rates of episiotomy and OASIS in non‐instrumental vaginal deliveries in Portuguese public hospitals are presented in Table 2 and illustrated in Figure 1 (trend analysis across the study period).

TABLE 1.

Annual number (n) of deliveries (all modes) in Portuguese public hospitals ^a and number (n) and percentage (%) of the non‐instrumental vaginal deliveries, from 2013 to 2022.

Year	Deliveries (all modes)	Non‐instrumental vaginal deliveries
	n	n	(%)
2022	67 576	34 272	(51)
2021	64 000	32 625	(51)
2020	68 564	34 947	(51)
2019	71 654	37 024	(52)
2018	72 350	38 136	(53)
2017	71 828	38 109	(53)
2016	72 105	39 119	(54)
2015	69 373	37 518	(54)
2014	65 212	35 212	(54)
2013	67 246	34 925	(52)
Total	689 908	361 887	(52)

Note: Based on the ACSS data.

Abbreviation: ACSS, Administração Central do Sistema de Saúde.

^a Including Madeira and Azores.

TABLE 2.

Annual number (n) of non‐instrumental vaginal deliveries in Portuguese public hospitals, total number (n) and percentage (%) of episiotomies, third‐ and fourth‐degree perineal lacerations, and OASIS ^a , from 2013 to 2022.

Year	Non‐instrumental	Episiotomies n (%)	Perineal lacerations
	Vaginal deliveries n		Third‐degree n (%)	Fourth‐degree n (%)	OASIS n (%)
2022	34 272	7137 (21)	105 (0.31)	7 (0.02)	112 (0.33)
2021	32 625	8597 (26)	101 (0.31)	9 (0.03)	110 (0.34)
2020	34 947	11 235 (32)	105 (0.30)	5 (0.01)	110 (0.31)
2019	37 024	12 847 (35)	115 (0.31)	8 (0.02)	123 (0.33)
2018	38 136	15 030 (39)	100 (0.26)	8 (0.02)	108 (0.28)
2017	38 109	17 341 (46)	68 (0.18)	9 (0.02)	77 (0.20)
2016	39 119	19 805 (51)	79 (0.20)	6 (0.02)	85 (0.22)
2015	37 518	19 659 (52)	73 (0.19)	14 (0.04)	87 (0.23)
2014	35 212	20 249 (58)	57 (0.16)	4 (0.01)	61 (0.17)
2013	34 925	21 917 (63)	53 (0.15)	9 (0.03)	62 (0.18)

Note: Based on the ACSS data.

Abbreviations: ACSS, Administração Central do Sistema de Saúde; OASIS, obstetric anal sphincter injuries.

^a OASIS = third‐ plus fourth‐degree perineal lacerations.

FIGURE 1.

Trends of episiotomy, and third‐ and fourth‐degree perineal laceration rates of non‐instrumental vaginal deliveries in Portuguese public hospitals, from 2013 to 2022.

3.1. Data quality

The analysis was conducted using data provided by the ACSS, ²² as it was the only official source offering complete information on all relevant variables across the 10‐year period evaluated. To assess the quality of ACSS data, the total number of deliveries was compared with delivery counts reported by INE, ²⁶ the episiotomy rate was cross‐checked against CPDO ²⁸ data, and the rate of perineal lacerations was compared with data from both CPDO ²⁸ and OECD. ²⁹

According to INE, between 2013 and 2022, there were 705 065 deliveries, encompassing all modes of delivery, in public hospitals, compared to 689 908 reported by ACSS—an estimated difference of around 2%. Despite this, agreement between the data sources was strong (ICC = 0.916). INE reported 52.13% non‐instrumental vaginal deliveries, while ACSS reported 52.45%. These small discrepancies may reflect differences in data collection methods: annual hospital surveys versus clinical coding. Nonetheless, despite differing approaches, the reliability between sources was excellent, and the proportion of non‐instrumental vaginal deliveries remained highly consistent.

Regarding the number of episiotomies, ACSS data were compared with those from the CPDO, which includes data from 13 public hospitals (about one‐third) using the ObsCare system, justifying some variation in episiotomy rates. Based on CPDO data from 2019 to 2022, episiotomy rates in non‐instrumental vaginal deliveries were 36.08%, 31.98%, 25.71%, and 20.99%, respectively. These rates were very similar to those recorded by the ACSS for the same years: 34.70%, 32.15%, 26.35%, and 20.82%. Moreover, both sources showed the same decreasing trend over these 4 years, supporting the consistency of the information. Reliability between these sources was excellent (ICC = 0.991).

For non‐instrumental vaginal deliveries, reliability of third‐ and fourth‐degree perineal lacerations data between the OECD and ACSS was very poor (ICC = 0.000): OECD reported 0.5%, 0.4%, 0.4%, 0.5%, 0.5%, and 0.5% for 2019, 2018, 2017, 2016, 2015, and 2013, respectively; ACSS reported 0.3%, 0.3%, 0.2%, 0.2%, 0.2% and 0.2%. Despite low agreement, both sources indicated stable rates over time.

In conclusion, reliability between CPDO and ACSS was excellent for episiotomy rates, whereas reliability between OECD and ACSS was poor for third‐ and fourth‐degree perineal lacerations. ACSS estimates were slightly lower, though both sources showed consistent trends, with discrepancies likely due to differing data collection methods.

4. DISCUSSION

A retrospective secondary data analysis was conducted to evaluate trends in episiotomy use in non‐instrumental vaginal deliveries in Portugal over the last decade, aiming to provide reliable data and to clarify national obstetric practices often perceived as less women‐centered.

This study revealed a significant national decline in episiotomy rates, from 63% in 2013 to 21% in 2022. These findings align with Teixeira et al., ³⁰ who reported a comparable downward trend between 2000 and 2015. The slight variation in results may reflect differences in study populations, as our analysis included Portugal's autonomous regions, unlike that of Teixeira et al. ³⁰ Nonetheless, both studies highlight efforts by Portuguese birth attendants to align with international guidelines recommending restrictive episiotomy use, ⁴ , ¹¹ suggesting that these recommendations have been gradually integrated into national practice. However, the most recent rate still remains above the goal of WHO of phasing out routine episiotomy. ¹¹

Our results also indicated a modest but significant increase in third‐degree perineal lacerations, from 0.15% in 2013 to 0.31% in 2022, while fourth‐degree lacerations remained unchanged. The 2018 report from the Observatório Português dos Sistemas de Saúde, ³¹ had already noted a rise in OASIS rates in non‐instrumental vaginal deliveries between 2000 and 2015, ranging from 0.11% to 0.38%.

The protective role of episiotomy against severe perineal trauma remains a subject of ongoing debate. While some studies suggest episiotomy may play a protective role against OASIS, ⁹ others report no benefit, ¹³ or even a potential increased risk. ⁵ , ¹⁴ Its role under specific circumstances—such as instrumental vaginal deliveries ¹² , ¹³ , ³² or in women with type 3 female genital mutilation ¹⁰ —also remains debated.

Despite the mixed evidence, the decline in episiotomy use in Portugal reflects an international trend, ¹ , ¹⁴ , ¹⁵ particularly among Western and European countries. For example, Nordic countries such as Sweden, Denmark, and Iceland have achieved low episiotomy rates (5%–7%) in non‐instrumental vaginal deliveries, ¹⁴ , ¹⁵ but with markedly higher OASIS rates (2%–4%) than those observed in our study. This inverse correlation has also been reported in studies based on EURO‐PERISTAT data ¹⁴ and in France. ³³ However, these differences may be influenced by episiotomy technique variations (e.g., lateral, mediolateral, or midline), often not distinguished in available data. ⁶ , ⁸ For example, Finland exclusively performs lateral episiotomy, with an OASIS rate of 1.1%. ¹⁵

The wide international variation in OASIS rates warrants cautious interpretation. Low rates may reflect underdiagnosis, ³⁴ while higher rates could stem from overdiagnosis. ³⁰ , ³³ To address this issue, the implementation of targeted training programs to improve diagnostic accuracy has been recommended.

It remains unclear whether extremely low episiotomy rates may compromise safety in some clinical contexts. ¹⁴ , ¹⁵ Therefore, determining the optimal technique and clear indications is crucial when considering restrictive versus zero‐use episiotomy policies.

There is also growing resistance from pregnant women toward episiotomy, even when medically advised. ³⁵ The ongoing debate, lack of consensus, and pressure to make rapid decisions during labor place significant strain on birth attendants.

Although current guidelines stress the importance of informed consent and discussing risks and benefits, ¹¹ this can be difficult in urgent situations. ⁴ , ³⁵

A major strength of this study is the use of national datasets, including the official ACSS database. This enabled a comprehensive overview of episiotomy practices on non‐instrumental vaginal deliveries across the Portuguese public health system (free of charge for the patient), offering a nationwide perspective.

However, there are inherent limitations to a retrospective data analysis. The absence of key demographic and clinical variables, such as maternal or fetal risk factors, may act as confounders. Additionally, methodological differences between datasets (e.g., annual data hospital surveys [INE] vs. hospital clinical coding based on the ICD [ACSS]) and changes in coding systems—like the transition to ICD‐10‐CM/PCS in 2017—could affect the consistency of results. Although hospital discharge coding follows standardized guidelines, underreporting or misclassification of OASIS cannot be excluded, due to reliance on the accuracy of clinical records. Additionally, given the retrospective design, unmeasured factors such as provider experience, local protocols, and institutional culture may have influenced episiotomy practices and OASIS rates. The ACSS's data‐collection methodology and restricting the analysis to public maternities render duplication unlikely. While data from the different sources show consistently similar trends, minor discrepancies in absolute values persist.

In Portugal, the 2023 publication of a national guideline ²¹ recommending a restrictive use of episiotomy will promote more uniform practices and enable future research to assess maternal and perinatal outcomes.

Despite consensus favoring restrictive episiotomy use, no universal threshold or precise indications have been established. Evidence of benefits in specific clinical circumstances remains limited, ² , ¹¹ and potential harms—such as postpartum hemorrhage, particularly with lateral or mediolateral episiotomy—must also be considered.

Severe perineal trauma was the primary outcome. Future research should assess additional maternal and perinatal adverse outcomes (e.g., postpartum hemorrhage, Apgar scores, NICU admission) and satisfaction with the quality of care.

This national study provides a starting point for broader research and highlights the need for a more comprehensive approach to developing future recommendations.

5. CONCLUSIONS

The adherence of Portuguese birth attendants to guidelines recommending a restrictive use of episiotomy in non‐instrumental vaginal deliveries is corroborated in our analysis. Over the last decade, there has been a significant decrease in episiotomy practice in Portugal (within free‐of‐charge deliveries), accompanied by a significant increase in third‐degree perineal lacerations.

Key challenges in implementing either a restrictive or no‐episiotomy policy include evaluating the risk–benefit ratio of reduced episiotomy rates, defining clear clinical indications for the procedure and identifying the optimal surgical technique.

There is an urgent need for clear clinical guidelines and targeted educational programs to support best intrapartum practices, prioritizing both clinical outcomes and respectful care. This study underscores the importance of monitoring national intrapartum practice patterns, assessing their impact, and ensuring that future research is built on standardized coding and accessible, anonymized datasets to enable reliable, cross‐country comparisons.

Promoting safe, respectful, and evidence‐informed childbirth demands not only clinical vigilance but also a national investment in transparency, professional training, and patient‐centered care.

AUTHOR CONTRIBUTIONS

Mariana Bandeira did the literature research and contributed to data acquisition. Mariana Bandeira and Cristina Costa‐Santos coordinated the data management. Cristina Costa‐Santos performed the statistical analysis, conceived the graphs, and wrote the study results. Mariana Bandeira wrote the first draft of the manuscript. Fernando Lopes did the coding list to obtain the data requested to ACSS, available in Table S1. João Bernardes and Ana Reynolds conceived the idea for the study. Ana Reynolds coordinated the research activity planning and supervised its execution. Ana Reynolds and Mariana Bandeira conceived and designed the study, with input from all other authors. All authors interpreted the data and critically reviewed the manuscript. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.

CONFLICT OF INTEREST STATEMENT

The authors have declared that no competing interests exist.

Supporting information

Data S1.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.