The aim of the study was to identify which groups of women contribute to interinstitutional variation of caesarean delivery (CD) rates and which are the reasons for this variation. In this regard, 15,726 deliveries from 11 regional centers were evaluated using the 10-group classification system. Standardized indications for CD in each group were used. Spearman’s correlation coefficient was used to calculate (1) relationship between institutional CD rates and relative sizes/CD rates in each of the ten groups/centers; (2) correlation between institutional CD rates and indications for CD in each of the ten groups/centers. Overall CD rates correlated with both CD rates in spontaneous and induced labouring nulliparous women with a single cephalic pregnancy at term (
Caesarean delivery (CD) rate is increasing worldwide [
At present, the CD rate in Italy is one of the highest in the world and represents a challenge for the National Public Health system. In 1980, the rate was 11% and since then it has increased, reaching 39% in 2008, with significant interregional variations, ranging from 24% to 62% [
A recent systematic review evaluated the advantages and deficiencies of these methodologies [
From its introduction in 2001 [
The aim of our study, carried out on more than 15 thousand deliveries from all of the 11 obstetric departments of our region, was to assess whether inter-institutional variation of overall CD rates correlated with the size or CD rates in selected obstetric populations of each institution, defined by using the TGCS. Moreover, we hypothesized that specific indications for CD in defined obstetric groups might account for overall variation of institutional CD rates.
An 18-month prospective study collected data on mode of delivery from all births of the 11 single-institutional obstetric cohorts of Friuli Venezia Giulia (range 369–1,810 deliveries/year/unit). Friuli Venezia Giulia is a region of northeastern Italy accounting roughly for 10,000 deliveries per year with one of the lowest overall regional CD rates in Italy (23.4% in 2010). The source institutions, referred as institutions A to M, are first-level departments serving low-risk pregnancies, except for centers I and M working for a mixed population with the availability of a neonatal intensive care unit (NICU, second referral units).
The units differed for number of deliveries/year as follows: units A, B, C, E, F, G, H, and L had less than 1,000 deliveries/year; center D accounted for 1,000–1,500 deliveries/year; 1,500–2,000 deliveries/year were assisted in institutions I and M.
To avoid potential information bias due to different definitions on collected data, we created a regional standardized computerized database with the collaboration of all centers. All centers approved and validated the data collection form. All obstetricians and midwives of all the centers were instructed to manage the database and to collect data. Information on institutional deliveries was prospectively collected at the time of delivery by obstetrician or midwife attending the delivery in each center. Collected data were systematically reviewed every month by the referent obstetrician of each center. Special attention was devoted to overall data completeness and accuracy. During the study period, two of the authors (GM and SA) organized periodical multicenter meetings to discuss the results and provide assistance. All women provided informed consent to include their records in the presentation of summary data for births.
The study was approved by the institutional review board of the coordinating center (Technical Scientific Committee (CTS), Institute for Maternal and Child Health (IRCCS Burlo Garofolo, Trieste, project 86/05) February 28, 2007) and access to the data was approved by all hospital trust administrations. According to the Italian law on privacy (Art. 20-21, DL 196/2003), data were anonymized at every institution where each patient was assigned a unique identifier. This identifier did not allow to trace the patient’s identity and other sensitive data.
The study population of each institution was evaluated using the TGCS (Table
10-group classification. Groups 2 and 4 were further divided respectively in to 2a, 2b and 4a, 4b.
Group | Classification |
---|---|
1 | Nulliparous, single cephalic, ≥37 weeks, in spontaneous labour |
2a | Nulliparous, single cephalic, ≥37 weeks, induced labour |
2b | Nulliparous, single cephalic, ≥37 weeks, CD before labour |
3 | Multiparous (excluding previous CD), single cephalic, ≥37 weeks, in spontaneous labour |
4a | Multiparous (excluding previous CD), single cephalic, ≥37 weeks, induced labour |
4b | Multiparous (excluding previous CD), single cephalic, ≥37 weeks, CD before labour |
5 | Previous CD, single cephalic, ≥37 weeks |
6 | All nulliparous breeches |
7 | All multiparous breeches (including previous CD) |
8 | All multiple pregnancies (including previous CD) |
9 | All transverse/oblique lies (including previous CD) |
10 | All preterm single cephalic, <37 weeks, including previous CD |
CD: caesarean delivery.
Groups 2 (induced labour or prelabour CD in singleton, cephalic presentation, at term, nulliparous women) and 4 (induced labour or prelabour CD in singleton, cephalic presentation, at term, multiparous women) were further divided into groups 2a and 2b, and 4a and 4b, according to whether they were induced or delivered by pre-labour CD.
In each group, the indications for induction of labour and CD were reported (Tables
Indications of induction of labour.
(1) Prelabour rupture of membranes | |
(2) Postterm (gestational age ≥ 41 weeks) | |
(3) Hypertensive disorders | |
(4) Other maternal reasons, for example, procedure done for the benefit of the mother* | |
(5) Fetal reasons, for example, procedure done for the benefit of the fetus* | |
(6) No absolute indications or no indication reported |
*Preexisting or gestational diabetes, preexisting maternal disease suggesting the termination of pregnancy, obstetric cholestasis, alloimmunisation, severe oligohydramnios, and intrauterine growth restriction.
Indications of caesarean delivery.
(1) Suspicious or pathological cardiotocography (CTG anomalies) [ |
|
(2) Other fetal reasons, for example, procedure done for the benefit of the fetus* | |
(3) Other maternal reasons, for example, procedure done for the benefit of the mother* | |
(4) Antepartal hemorrhage or placenta previa | |
(5) Preeclampsia or HELLP syndrome | |
(6) Breech presentation | |
(7) One previous caesarean delivery | |
(8) More than one caesarean delivery | |
(9) Dystocia-failed induction [ |
|
(10) Dystocia-failure to progress [ |
|
(11) No indication reported including maternal request |
*HIV, preexisting or gestational diabetes, preexisting maternal disease suggesting the termination of pregnancy, obstetric cholestasis, alloimmunisation, severe oligohydramnios, and intrauterine growth restriction.
CD was also classified as elective or emergency and before or during labour.
Information on maternal age, gestational age at delivery, neonatal birthweight, and perinatal mortality was also collected.
Difference in means between centers was analyzed using the ANOVA if data were normally distributed, or else with the nonparametric Kruskal-Wallis test. Post hoc analysis was carried out using Bonferroni’s correction. Difference in proportion between centers were analyzed using Pearson’s chi-squared test or Fisher exact test as appropriate, and the Bonferroni correction was applied in case of multiple testing. Spearman’s correlation coefficient was used to study the correlations between institutional CD rates and relative size/CD rates in each of the ten groups and to verify whether specific indications for CD in selected groups correlated with variation of overall CD rates.
Categorical variables were presented as frequencies and percentages, or as percentages and 95% confidence intervals; continuous variables were presented as mean and standard deviations.
Statistical analysis was carried out with the STATA statistical package (version 9.0) [
A total of 3,791 caesarean deliveries were registered among 15,726 deliveries, giving an overall CD rate of 24.1%. CD rates differed significantly among institutions (range, 14.3–34.1%). Table
Descriptive analysis of study population by center.
Center | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
A | B | C | D | E | F | G | H | I | L | M | Overall | |
Deliveries, |
1291 | 950 | 957 | 1791 | 1328 | 1642 | 1054 | 868 | 2745 | 579 | 2521 | 15726 |
MA, yrs* | 31.8 ± 5.0 | 31.7 ± 5.4 | 31.2 ± 5.3 | 31.7 ± 5.4 | 31.3 ± 5.1 | 31.8 ± 5.0 | 31.1 ± 5.2 | 31.6 ± 5.2 | 32.3 ± 5.1 | 31.8 ± 5.1 | 31.8 ± 5.4 | 31.7 ± 5.2 |
GA, wks ± days | 39 ± 1 | 39 ± 1 | 39 ± 1 | 38 ± 6 | 39 ± 2 | 39 ± 1 | 39 ± 0 | 39 ± 2 | 39 ± 0 | 39 ± 0 | 38 ± 4 | 39 ± 0 |
GA, wks-range** | 26–42 | 32–42 | 33–42 | 26–42 | 34–42 | 29–42 | 24–42 | 28–42 | 24–42 | 33–42 | 21–42 | 21–42 |
BW, grams* | 3324 ± 469 | 3276 ± 452 | 3337 ± 451 | 3277 ± 531 | 3362 ± 447 | 3325 ± 481 | 3321 ± 460 | 3315 ± 493 | 3301 ± 560 | 3310 ± 443 | 3256 ± 631 | 3304 ± 520 |
PMR, |
1 (0.1) | 0 (0) | 1 (0.1) | 8 (0.5) | 3 (0.2) | 4 (0.3) | 2 (0.2) | 1 (0.1) | 15 (0.6) | 0 (0) | 16 (0.6) | 51 (0.3) |
CD rate, |
184 (14.3) | 281 (29.6) | 266 (27.8) | 519 (29.0) | 255 (19.2) | 285 (17.4) | 230 (21.8) | 158 (18.2) | 598 (21.8) | 156 (26.9) | 859 (34.1) | 3791 (24.1) |
CD, 95% |
12.4–16.3 | 26.7–32.6 | 25.0–30.8 | 26.9–31.1 | 17.1–21.4 | 15.6–19.3 | 19.4–24.4 | 15.7–20.9 | 20.3–23.4 | 23.4–30.8 | 32.2–36.0 | 23.4–24.8 |
*ANOVA’s
MA: maternal age; GA: gestational age at delivery; Wks: weeks; BW: birthweight; PMR: perinatal mortality rate; CD: caesarean delivery; CI: confidence intervals. Maternal age, gestational age, and birthweight are expressed as mean ± standard deviation.
Distribution of the relative sizes and CD rates in each of the ten groups among centers are described in Supplemental Tables 1 and 2 (supplementary material available online at
Pairwise comparisons among centers showed that differences in the distribution of ten groups relative sizes existed. Nulliparous women at term, cephalic presentation, spontaneous or induced labour (groups 1 and 2a, resp.) and multiparous at term, cephalic presentation, spontaneous labour or with past CD (groups 3 and 5, resp.) were the most represented groups in all the centers (Supplemental Table 1). CD rates in groups 1, 2a, 4a, 5, 8, and 10 differed among institutions (Supplemental Table 2).
As for the correlation between the relative size of the ten groups and the overall CD rates, only groups 2b (nulliparous women, single cephalic, ≥37 weeks, CD before labour; Spearman’s rho 0.90,
Overall CD rates correlated with CD rates in groups 1 (nulliparous women, single cephalic, ≥37 weeks, spontaneous labour; Spearman’s rho 0.77;
Correlation between overall inter-institutional CD rates and (a) CD rates in group 1 (nulliparous women at term, cephalic presentation, spontaneous labour), (b) CD rates in group 2a (nulliparous women at term, cephalic presentation, induced labour), and (c) CD rates and (d) relative size of group 5 (multiparous women at term, cephalic presentation, past CD). CD: caesarean delivery.
Looking at the indications for CD and to their association with overall institutional CD rates, we decided to focus our attention on nulliparous women, single cephalic, ≥37 weeks, spontaneous or induced labour (groups 1-2a) and multiparous women with past scar, single cephalic, ≥37 weeks (group 5). Nulliparous and multiparous women at term, cephalic presentation, CD before labour (groups 2b and 4b, resp.) were excluded from further analysis, firstly because the sizes of these groups were not clinically relevant (supplemental Table 1, fourth column) and secondly because, in most cases, the indications for CD were unlikely to be susceptible to modification of management (i.e., elective or emergency prelabour CD for absolute fetal or maternal indications).
Looking at the indications in group 1 (spontaneous labouring nulliparous women with a single cephalic pregnancy, at term), “failure to progress” and “CTG anomalies” were both significantly correlated with overall variation of CD rates (Spearman’s rho 0.77;
Correlation between indications in selected groups and overall inter-institutional CD rates. (a) Failure to progress and (b) CTG anomalies in group 1 (nulliparous women at term, cephalic presentation, spontaneous labour); (c) failed induction in group 2a (nulliparous women at term, cephalic presentation, induced labour); (d) one previous CD in group 5 (multiparous women at term, cephalic presentation, past caesarean delivery). CD: caesarean delivery; CTG: cardiotocographic.
The analysis of indications for CD in nulliparous women at term, in induced labour, showed that no specific indication correlated significantly with overall CD rates. Only “failed induction” demonstrated a correlation with overall CD rates, but it was not statistically significant probably due to the limited sample size of the group (Spearman’s rho 0.55;
In the group of multiparous women, cephalic presentation at term, with at least one previous scar (group 5), there was a highly significant correlation found with the indication “one past CD” (correlation with overall CD rates: Spearman’s rho 0.86;
Our prospective evaluation of more than 15 thousand deliveries in a region with overall low CD rates offers new insight into the application of the TGCS. By assessing the mutually exclusive obstetric populations and providing an accurate registration of the main indications for CD, we were able to identify which groups contributed significantly to variation of overall CD rates and why overall CD rates differed among institutions.
As suggested by the TGCS, we assessed firstly whether overall institutional CD rates were correlated with relative sizes or CD rates in specific groups. Overall CD rates correlated with CD rates in group 1 (nulliparous women, at term, single cephalic, spontaneous labour), in group 2a (nulliparous women, at term, single cephalic, induced labour), and with relative size and CD rates in group 5 (multiparous women, at term, cephalic presentation, with previous scar). These results strengthen the evidence that overall inter-institutional differences in CD rates depended on CD rates variations in these groups and correlated significantly with the size of multiparous women with past CD (group 5), a group at high risk for repeat CDs. Our observation confirmed that in order to reduce the overall CD rate, limiting CD rate in nulliparous women with a single cephalic pregnancy at term is the key to lowering the trend of overall increased abdominal deliveries. The decrease of CD rates in this group will consequently reduce the number of multiparous women with a previous CD (group 5) and hence the repeat CDs. These findings are in agreement with the results and conclusions of other studies. Brennan et al. and Delbaere et al. observed similar results, but they did not provide information about the causes of CD in these groups [
After having identified the groups that contributed to the overall variation of inter-institutional CD rates, we focused our attention on indications leading to CD in these cohorts. The aim of this evaluation was to understand whether specific indications leading to CD in selected groups correlated with inter-institutional variation of CD rates. We realized the requirement for clear, unambiguous, and precise definitions for common obstetrical diagnoses and procedures. Standardization of these definitions was an essential step to prospectively collect reliable data and to allow a consistent comparison among institutions. This was necessary to overcome the potential bias generated by differences in definitions or coding and to improve inter-institutional reproducibility of information about the obstetric conditions leading to CD [
In the group of nulliparous women, at term, single cephalic, spontaneous labour (group 1), the indications “CTG anomalies” and “dystocia-failure to progress” contributed significantly to the overall variation of institutional CD rates. This evidence might support the hypothesis that both the management of suspicious or pathological fetal heart rate tracings and of abnormal labour differed among centers. Barber et al. found similar results. They observed that among the primary caesarean deliveries, more subjective indications (“nonreassuring fetal status” and “arrest of dilation”) contributed significantly more than other more objective indications (malpresentation, maternal-fetal, and obstetric conditions) [
Another condition contributing to the inter-institutional variation of CD rates was the elective indication “one past CD” in the group of multiparous, at term with previous CD. A policy of elective CD in women with previous CD was significantly associated with the increase of overall CD rate. Clearly, the mode of delivery of women with past CD was planned differently among institutions. Some centers, with low elective CD rates for this indication, offered trial of labour, while others opted for elective CD, probably because they considered a failed trial of labour after previous CD was associated with more complications than elective repeat CD [
It is well known that despite the presence of specific guidelines for managing anomalies of fetal heart patterns, abnormal labour, and women with a previous CD, inter-institutional variations in clinical evaluation and management of these conditions will exist [
As for the group of nulliparous women at term, cephalic presentation, induced labour (group 2a), our data showed that induced labour in this cohort was at least three times as likely to result in CD than spontaneous labour (29.4% and 9.5%, resp.). However “failed induction” did not correlate significantly with overall CD rates probably because the contribution of this indication to the overall CD rate was only 5%. The evidence that the rate of CD for “failed induction” in this group differed among institutions might support the opinion that the decision to proceed with CD for this indication might be not based on uniform criteria and that clinical impatience may play a role in the decision making regarding the mode of delivery. Details on management of induction of labour were not available, but this finding might be useful to audit in each center marginal indications of induction, implementing an effective management to reduce unnecessary interventions [
The main limitation of our study was considering crude CD rates in each group without taking into account a number of variables that have been associated with high CD rates such as obesity, advanced maternal age, or clinical conditions defining the pregnancy as at risk [
Our results confirm that the TGCS represents a simple method that allows comparison of CD rates among institutions. The analysis of prospectively collected data, using standardized definitions of indications leading to CD and induction of labour in identified obstetric cohorts might be helpful to monitor and provide feedback to clinicians, identifying those procedures occurring without accepted medical indications [
Our conclusions are in agreement with the remarks recently made by Robson et al. in their editorial and with the recommendations of the Society for Maternal Fetal Medicine and the American College of Obstetricians and Gynecologists. In this regard, every effort should be made to prevent the first CD in the nulliparous population by focusing the attention on caesarean deliveries occurring after labour inductions, those labeled as for “nonreassuring fetal status,” and those occurring for “labour arrest” or “failed induction” without meeting accepted criteria [
All authors declare that there is no conflicts of interests.
Gianpaolo Maso and Salvatore Alberico conceived and designed the study. Gianpaolo Maso, Salvatore Alberico, Luca Ronfani, Monica Piccoli, Sara Parolin, Carmine Gigli, Daniele Domini, Claudio Fiscella, Sara Casarsa, Carlo Zompicchiatti, Michela De Agostini, Attilio D’Atri, Rafaela Mugittu, Santo La Valle, Cristina Di Leonardo, Valter Adamo, Mara Fracas, Giovanni Del Frate, Monica Olivuzzi, Silvio Giove, Maria Parente, Daniele Bassini, Simona Melazzini, Silvio Giove, Secondo Guaschino, Caterina Businelli, Franco G. Toffoletti, Sergio Demarini, Laura Travan, Diego Marchesoni, Alberto Rossi, Giorgio Simon, Sandro Zicari, and Giorgio Tamburlini contributed to conducting the research and collecting data. Marcella Montico and Lorenzo Monasta analysed the data. Gianpaolo Maso, Marcella Montico, and Luca Ronfani interpreted the data. Gianpaolo Maso drafted the paper in collaboration with Monica Piccoli, Marcella Montico, Lorenzo Monasta, Luca Ronfani, Sara Parolin, and Salvatore Alberico. Gianpaolo Maso, Monica Piccoli, Salvatore Alberico, Marcella Montico, Lorenzo Monasta, Luca Ronfani, Sara Parolin, Carmine Gigli, Daniele Domini, Claudio Fiscella, Sara Casarsa, Carlo Zompicchiatti, Michela De Agostini, Attilio D’Atri, Rafaela Mugittu, Santo La Valle, Cristina Di Leonardo, Valter Adamo, Mara Fracas, Giovanni Del Frate, Monica Olivuzzi, Silvio Giove, Maria Parente, Daniele Bassini, Simona Melazzini, Silvio Giove, Secondo Guaschino, Caterina Businelli, Franco G. Toffoletti, Sergio Demarini, Laura Travan, Diego Marchesoni, Alberto Rossi, Giorgio Simon, Sandro Zicari, and Giorgio Tamburlini met the ICMJE criteria for authorship.
This study was financed by the Grant no. 86/05 of the Institute for Maternal and Child Health (IRCCS) Burlo Garofolo, Trieste, Italy. The authors are very grateful to midwives and registrars of Friuli Venezia Giulia Institutions for their contribution to data collection and to Michael S. Robson FRCS, MRCOG, FRCPI (National Maternity Hospital, Dublin-Ireland) for his support to revise the draft.