Over the past 20 years, the incidence of certain morbidities reported in women during hospitalization for childbirth has increased, particularly for postpartum hemorrhage, hypertension, and gestational diabetes.1 There also have been documented changes in rates of rarer events, such as uterine rupture2 and blood transfusion.3 Because it is likely that these trends in North America are not attributable solely to changes in diagnostic reporting practice, they raise important questions concerning the patterns of maternal morbidities on an international scale.
The phrase maternal morbidity may imply moderate to severe conditions that potentially result in obstetric complications and is recognized as a useful indicator of maternal well-being and quality of obstetric care.4–8 Yet maternal morbidities encountered differ geographically as well as over time because of advances in treatment, health care delivery factors, changes in the childbearing population, and the nature of the morbidity itself. Understanding the incidence and change of maternal morbidities on a national scale is essential for health policy design, clinical recommendations, and targeted interventions.
Each year, approximately 70,000 women give birth in Ireland. Although Irish hospital-based studies of severe maternal morbidity have been conducted,9,10 these studies were neither nationally representative nor did they capture the breadth of clinically important obstetric and nonobstetric complications that may have an effect on pregnancy, delivery, and the puerperium. Advances in methods using hospital discharge data to identify population-based maternal morbidity rates have been made in other European countries, such as Norway,11 Finland,12 Denmark,13 and Scotland.14 We complement this research using hospital discharge data from 2005 through 2008 to conduct a nationally representative analysis of maternal morbidity in Ireland. Specifically, we investigated the incidence of maternal morbidities during childbirth while hospitalized and morbidity trends during the study period.
MATERIALS AND METHODS
We conducted a population-based retrospective cohort study of women delivering in hospitals in Ireland using the Hospital In-Patient Enquiry data set, which is the only national source of hospital discharge data from all public hospitals. Anonymized Hospital In-patient Enquiry data are maintained by the Economic and Social Research Institute; further details on data collection have been described elsewhere.15 In brief, for each hospital admission, Hospital In-Patient Enquiry contains information on patient demographics, hospitalization characteristics, and up to 20 diagnoses or procedures or both, coded according to the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Australian Modification (ICD-10-AM).16 All hospitals submitting information to the Hospital In-Patient Enquiry database use identical software, standardized for the entry of Hospital In-Patient Enquiry data, and clinical coders nationally undergo a centralized training program in the entry of morbidity data under supervision of the Economic and Social Research Institute. Furthermore, more than 140 validation checks are routinely performed on Hospital In-Patient Enquiry data, and 18 chart-based audits have been conducted since 2001. Independent reviews of the clinical coder training program and audit procedures have identified Hospital In-Patient Enquiry as an accurate and reliable data collection system.17,18
We included all singleton childbirth hospitalizations from 19 public maternity hospitals between January 1, 2005 and December 31, 2008 (N=253,792) because this time period reflected the most recently available data comprising a large number of births and permitted analysis of uncommon morbidities. Singleton childbirth hospitalizations were identified if the hospital record contained a diagnostic code for a single live birth (Z37.0) or a single stillbirth (Z37.1). Because pregnancy outcomes of interest were live and stillborn fetuses, we excluded records associated with spontaneous abortions and abdominal, tubal, and hydatidiform mole pregnancies (n=27). Given these criteria, our analysis included 252,276 live birth and 1,489 stillbirth deliveries. The study was exempt from Institutional Research Board review because data analyzed were de-identified, unlinked, and available for public access.
We modeled our definition of maternal morbidity after two studies from the U.S. Centers for Disease Control and Prevention (CDC), which categorized diagnostic codes into clinically relevant maternal morbidity groups.1,5 In parallel with the CDC studies, we defined maternal morbidity as moderate to severe conditions recorded during delivery hospitalization that may have adversely affected maternal physical or mental health. We included conditions that were specific to pregnancy (eg, gestational diabetes) and conditions that may have exacerbated pregnancy (eg, cystic fibrosis).
Because the two previous CDC studies1,5 were based on the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) and because the ICD-10-AM is more detailed than the ICD-9-CM, morbidity categorizations could not be directly translated. Therefore, we identified equivalent diagnoses using mapping tables available from the Australian National Centre for Classification in Health.19 For diagnoses without a direct match, we considered all potentially relevant ICD-10-AM codes within the pregnancy, childbirth, and puerperium (O00–O99) chapter as well as all other remaining chapters. Three clinicians, including two obstetricians (R.A.G., J.J.M.) and one nurse-midwife (M.T.L.R.), independently reviewed ICD-10-AM diagnoses, and conditions were included on consensus of the three reviewers. We grouped ICD-10-AM codes that met our definition of maternal morbidity into 38 major clinical categories. Similar to Berg et al,1 we added cesarean delivery, considered a major surgical procedure, as a maternal morbidity category for our analysis.
Other variables of interest included marital status, which was classified as single, married, or other. Medical card status identified whether a patient was in possession of a general medical card, which entitles low-income individuals to select free health care services.
The rates of maternal morbidities over the course of the 4-year period were defined as the number of women who met at least one criterion of the morbidity category per 100 deliveries (shown as a percentage). Overall morbidity rates were calculated both exclusive and inclusive of cesarean delivery. Morbidities that affected at least 1.0% of total deliveries also were examined by live birth and stillbirth, and significant differences by birth outcome were assessed using χ2 tests. Temporal trends for overall morbidities as well as for the most common morbidities (1.0% or more of total deliveries) were explored. To determine significant trends, crude incidence rates were derived per 1,000 singleton deliveries and a Cochran-Armitage test was performed. Analyses were conducted using SAS 9.2.
There were several differences in the distribution of patient characteristics relative to maternal morbidity diagnosis (Table 1). Compared with women without a morbidity diagnosed, women with a morbidity diagnosed were more likely to remain hospitalized for at least 5 days, to be admitted to the intensive care unit, to have a cesarean delivery, and to have a stillborn fetus.
Rates for the 38 maternal morbidity categories undergoing review and the ICD-10-AM codes used to derive each category are shown in Table 2. Exclusive of cesarean delivery, approximately one in six women (17.2%) had a maternal morbidity diagnosed during delivery in the hospital. When cesarean delivery was included as an additional indicator, more than one third (35.6%) had a maternal morbidity diagnosed. The percentage of women with either hemorrhage and genital tract trauma (6.5%) or pregnancy-induced conditions (6.4%) diagnosed were similar. Overall, 4.5% of women had nonacute or chronic conditions diagnosed, 1.6% had infections diagnosed, and 0.6% had acute medical conditions diagnosed. Between 2005 and 2008, the five most common maternal morbidities were cesarean delivery (24.5%), gestational hypertensive disorders (5.0%), postpartum hemorrhage (2.9%), pelvic and perineal trauma (2.3%), and anemia (2.0%).
The singleton stillbirth rate among women delivering in hospitals was 5.9 per 1,000 singleton deliveries (data not shown). Patterns in several of the most common conditions (those affecting 1% or more of total deliveries) varied according to birth outcome (Table 3). Notable differences of live birth compared with stillbirth, respectively, were found among cesarean delivery (24.6% compared with 6.9%), pelvic and perineal trauma (2.3% compared with 0.6%), and antepartum hemorrhage (1.2% compared with 7.4%). Women with live births experienced a lower total morbidity, excluding cesarean delivery, of 17.0% compared with 24.2% among women with stillbirths. However, the pattern reversed when total morbidity including cesarean delivery was considered. The incidence of total morbidity including cesarean delivery was highest among women with live births relative to stillbirths (35.5% compared with 28.1%, respectively).
Between 2005 and 2008, the rate of cesarean delivery among all singleton deliveries remained at approximately 25% (data not shown). Although no significant trends were found over the 4-year study period for antepartum hemorrhage, gestational hypertensive disorders, or anemia, rates did significantly increase for postpartum hemorrhage, pelvic and perineal trauma, and gestational diabetes (P<.001; Fig. 1). Moreover, the overall rate of maternal morbidity excluding cesarean delivery per 1,000 singleton deliveries increased significantly from 171.5 cases in 2005 to 178.9 cases in 2008 (P<.001; data not shown). Similarly, the rate of maternal morbidity including cesarean delivery per 1,000 singleton deliveries increased from 353.3 cases in 2005 to 362.9 cases in 2008 (P<.001; data not shown).
In our analysis of Ireland's nationally representative cohort of women giving birth in hospitals, we identified the incidence of maternal morbidity and change over time during a 4-year period. We offer three key findings. First, maternal morbidity affected a large proportion of women, with nearly one in six women experiencing a potentially detrimental complication during delivery. This rate increased to more than one in three deliveries when cesarean delivery was included as an additional measure of morbidity. Second, the presence and pattern of maternal morbidity distinctly varied for women with live births compared with stillbirths. Third, among all deliveries, the incidence of postpartum hemorrhage, pelvic and perineal trauma, and gestational diabetes increased from 2005 to 2008.
Our study examines population-based incidence rates of maternal morbidity over time Ireland and contributes to a growing body of literature that has utilized hospital discharge data to better-understand patterns in maternal morbidities.1,3,12,20–22 Specifically, our work is congruent with and builds on the research of Lynch et al,9 who analyzed audit data to investigate severe maternal morbidity from 1999 to 2003, and Murphy et al,10 who analyzed severe maternal morbidity cases from three large Dublin tertiary hospitals. Both studies found that obstetric hemorrhage was a major maternal morbidity.
We found that the differences in maternal morbidity patterns between women with live and those with stillbirths reinforced the credibility of the Hospital In-Patient Enquiry data set because, as expected, antepartum hemorrhage and placental abruption (data not shown) were notably higher among stillbirth compared with live birth deliveries. Furthermore, documenting maternal morbidities by birth outcome complemented recent work published by the CDC.5 We found, for example, that the rate of gestational hypertensive disorders was almost identical between stillbirth and live birth deliveries in Ireland (5.3% compared with 5.0%, respectively), although rates of hypertensive disorders were slightly lower for stillbirth deliveries compared with live birth deliveries in the United States (9.9% compared with 12.0%, respectively).
Increases in the rates of postpartum hemorrhage and gestational diabetes observed in this study were consistent with international systematic reviews from other high-income countries,23,24 and specifically with findings presented by Berg et al.1 Many factors may have contributed to these increases; given the association between both conditions and obesity,25 these trends may reflect the increasing prevalence of obesity among Irish women, which was recently estimated at 23%.26 Among all deliveries, the rate of pelvic and perineal trauma increased, which is counterintuitive given the high rate of cesarean delivery. This finding is in contrast to the findings of Berg et al,1 who reported a decrease in perineal trauma in the United States. The increase in pelvic and perineal trauma may be attributable to increasing rates of operative vaginal vacuum-assisted delivery in Ireland from 12.5% in 2005 to 13.7% in 2008. In addition, the lower rates for stillbirth deliveries may reflect a higher incidence of small-for-gestational-age newborns among this cohort and, hence, less perineal sequelae.
However, trends must be interpreted with caution because of the relatively short time span of the study period. Trends potentially may be an artifact of improved diagnostic reporting. In 2007, the Royal College of Obstetricians and Gynaecologists issued clear guidelines for identifying pelvic and perineal trauma27; in 2006, Bose et al28 highlighted significant underestimation of postpartum blood loss by clinicians. Although both publications consequently affected clinical audit of these conditions Ireland, the extent to which they have influenced reporting on hospital discharge records has not been determined. Thus, to substantiate or refute trends suggested in this study, additional longitudinal analysis is warranted, specifically focusing on postpartum hemorrhage, gestational diabetes, and pelvic and perineal trauma.
Our study focused on the mapping of maternal morbidity outcomes between the ICD-9-CM and the ICD-10-AM and on the modification of previously established morbidity categories based on ICD-9-CM codes. Numerous diagnoses in the ICD-10-AM are more specific than its predecessor; conditions, such as morbidly adherent placenta, are uniquely identified in the ICD-10-AM. Therefore, we were able to discern with a greater degree of precision which morbidities should be included or excluded from our definition of maternal morbidity. Given that diagnostic categories between the ICD-9-CM and ICD-10-AM were extensively reviewed, this preliminary work can facilitate future studies transitioning between ICD systems. Our work has relevance for at least 17 countries currently licensed to use the ICD-10-AM version29 and also may benefit countries, such as the United States, that will be adopting the ICD-10-CM in October 2013.30
Using hospital discharge data to identify maternal morbidity has potential limitations. Sensitivity studies of pregnancy-related hospital discharge records have shown that certain morbidities are prone to underestimation, particularly nonobstetric or chronic conditions, and these underestimations may be more frequent among deliveries with less complications.31,32 While independent reviews of the Hospital In-Patient Enquiry data set have found that its coding entries are highly accurate17,18 and our overall rates of postpartum and antepartum hemorrhage and uterine rupture were similar to those reported in a U.S. study using hospital discharge data,1 there is an inherent need to conduct validation studies to identify potential reporting bias. Although underreporting of maternal morbidity in hospital-discharge data can be decreased by linking hospital-discharge and birth certificate data,32–34 Ireland's data-protection legislation currently prohibits these types of linkages. Because we were unable to link data sets, we lacked data on parity and neonatal birth weight, which precluded examination of these factors. However, presently there is no alternative national data source to Hospital In-Patient Enquiry routinely collecting information on maternal morbidities in Ireland. Given that 99% of births in Ireland are hospital-based,10 and that Hospital In-Patient Enquiry data are recorded for 98% of all hospitalized births, our findings represent nearly all women giving birth in Ireland.
Under Irish governance (Health Act, No. 1/1970, Irish Statute Book), all pregnant women in Ireland are entitled to free peripartum care, 6 months of maternity leave, and additional leave to attend obstetric appointments and antenatal classes. However, despite these progressive measures, the 2006 publication of the Lourdes Hospital Inquiry35 identified shortcomings in Irish obstetric care and subsequently recommended the establishment of the National Perinatal Epidemiology Centre (http://www.ucc.ie/en/npec/) to conduct national-level research on the health of mothers and newborns. Thus, in an effort to meet the recommendation set forth by the Lourdes Hospital Inquiry, we have found that the Hospital In-Patient Enquiry database has the potential to serve as a national monitoring system of maternal health during childbirth hospitalization because of its representativeness of the childbearing population and its extensive quality-control protocols. Nonetheless, to improve the utility of the Hospital In-Patient Enquiry database for epidemiologic research and to identify potential reporting bias, hospital discharge records should be linked with birth certificate information when data linkage restrictions are overturned, which is currently anticipated to occur in 2011. However, continued retrospective analysis of hospital discharge data would complement the Severe Maternal Morbidity Surveillance System, which is a prospective monitoring system expected to commence Ireland in June 2011 under the auspices of National Perinatal Epidemiology Centre. Future cross-referencing of these two data sources would facilitate quality assessment of both data collection systems.
In conclusion, our study, undertaken in Ireland, provides national-level information on maternal morbidity, thus providing clinicians, public health officials, and policy makers with data to describe and understand the spectrum of clinically important maternal morbidities encountered by the Irish childbearing population during labor, delivery, and the puerperium. We have found that maternal morbidities in Ireland are common and changing, underscoring the benefits of continuous comprehensive examination of maternity care services for all women during childbirth to address treatment of morbidities and potentially prevent new morbidities.
1. Berg CJ, Mackay AP, Qin C, Callaghan WM. Overview of maternal morbidity during hospitalization for labor and delivery in the United States: 1993–1997 and 2001–2005. Obstet Gynecol 2009;113:1075–81.
2. Wen SW, Huang L, Liston R, Heaman M, Baskett T, Rusen ID, et al. Severe maternal morbidity in Canada, 1991–2001. CMAJ 2005;173:759–64.
3. Callaghan WM, Mackay AP, Berg CJ. Identification of severe maternal morbidity during delivery hospitalizations, United States, 1991–2003. Am J Obstet Gynecol 2008;199:133 e1–8.
4. Geller SE, Cox SM, Callaghan WM, Berg CJ. Morbidity and mortality in pregnancy: laying the groundwork for Safe Motherhood. Womens Health Issues 2006;16:176–88.
5. Bruce FC, Berg CJ, Hornbrook MC, Whitlock EP, Callaghan WM, Bachman DJ, et al. Maternal morbidity rates in a managed care population. Obstet Gynecol 2008;111:1089–95.
6. Roberts CL, Cameron CA, Bell JC, Algert CS, Morris JM. Measuring maternal morbidity in routinely collected health data: development and validation of a maternal morbidity outcome indicator. Med Care 2008;46:786–94.
7. Zhang W-H, Alexander S, Bouvier-Colle M-H, Macfarlane A, MOMS-B Group. Incidence of severe pre-eclampsia, postpartum haemorrhage and sepsis as a surrogate marker for severe maternal morbidity in a European population-based study: the MOMS-B survery. BJOG 2005;112:89–96.
8. Say L, Pattinson RC, Gulmezoglu AM. WHO systematic review of maternal morbidity and mortality: the prevalence of severe acute maternal morbidity (near miss). Reprod Health 2004;1:3.
9. Lynch CM, Sheridan C, Breathnach FM, Said S, Daly S, Byrne B. Near miss maternal morbidity. Ir Med J 2008;101:134–6.
10. Murphy CM, Murad K, Deane R, Byrne B, Geary MP, McAuliffe FM. Severe maternal morbidity for 2004–2005 in the three Dublin maternity hospitals. Eur J Obstet Gynecol Reprod Biol 2009;143:34–7.
11. Bjørstad AR, Irgens-Hansen K, Daltveit AK, Irgens LM. Macrosomia: mode of delivery and pregnancy outcome. Acta Obstet Gynecol Scand 2010;89:664–9.
12. Pallasmaa N, Ekblad U, Gissler M. Severe maternal morbidity and the mode of delivery. Acta Obstet Gynecol Scand 2008;87:662–8.
13. Nohr EA, Vaeth M, Baker JL, Sorensen T, Olsen J, Rasmussen KM. Combined associations of prepregnancy body mass index and gestational weight gain with the outcome of pregnancy. Am J Clin Nutr 2008;87:1750–9.
14. Pasupathy D, Wood AM, Pell JP, Fleming M, Smith GC. Rates of and factors associated with delivery-related perinatal death among term infants in Scotland. JAMA 2009;302:660–8.
16. National Centre for Classification in Health. The International Statistical Classfication of Diseases and Related Health Problems. Tenth Revision. Australian Modification. 4th edition. Sydney, Australia: National Centre for Classification in Health; 2004.
17. Bramley M, Reid BA. Morbidity data quality initiatives in Ireland. HIM J 2005;34:47–53.
18. Bramley M, Reid BA. Clinical coder training initiatives in Ireland. HIM J 2005;34:40–6.
20. Zwart JJ, Richters JM, Ory F, de Vries JL, Bloemenkamp KW, van Roosmalen J. Severe maternal morbidity during pregnancy, delivery and puerperium in the Netherlands: a nationwide population-based study of 371,000 pregnancies. BJOG 2008;115:842–50.
21. Brace V, Penney G, Hall M. Quantifying severe maternal morbidity: a Scottish population study. BJOG 2004;111:481–4.
22. Souza JP, Cecatti JG, Faundes A, Morais SS, Villar J, Carroli G, et al. Maternal near miss and maternal death in the World Health Organization's 2005 global survey on maternal and perinatal health. Bull World Health Organ 2010;88:113–9.
23. Knight M, Callaghan WM, Berg C, Alexander S, Bouvier-Colle M-H, Ford JB, et al. Trends in postpartum hemorrhage in high resource countries: a review and recommendations from the International Postpartum Hemorrhage Collaborative Group. BMC Pregnancy Childbirth 2009;9:55.
24. Hunt KJ, Schuller KL. The increasing prevalence of diabetes in pregnancy. Obstet Gynecol Clin North Am 2007;34:173–99, vii.
25. O'Sullivan JB, Mahan CM, Charles D, Dandrow RV. Screening criteria for high-risk gestational diabetic patients. Am J Obstet Gynecol 1973;116:895–900.
26. Harrington J, Perry I, Lutomski J, Morgan K, McGee H, Shelley E, et al. SLÁN 2007: survey of lifestyle, attitudes and nutrition in Ireland. Dietary habits of the Irish population, Department of Health and Children. Dublin, Ireland: The Stationary Office; 2008.
27. Royal College of Obstetricians and Gynaecologists. The management of third- and fourth-degree perineal tears. Green-top Guideline No 29. London (UK): Royal College of Obstetricians and Gynaecologists; 2007.
28. Bose P, Regan F, Paterson-Brown S. Improving the accuracy of estimated blood loss at obstetric haemorrhage using clinical reconstructions. BJOG 2006;113:919–24.
30. Centers for Disease Control and Prevention/National Center for Health Statistics. International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM). Available at: http://www.cdc.gov/nchs/icd/icd10cm.htm#10update
. Retrieved August 10, 2010.
31. Hadfield RM, Lain SJ, Cameron CA, Bell JC, Morris JM, Roberts CL. The prevalence of maternal medical conditions during pregnancy and a validation of their reporting in hospital discharge data. Aust N Z J ObstetGynaecol 2008;48:78–82.
32. Yasmeen S, Romano PS, Schembri ME, Keyzer JM, Gilbert WM. Accuracy of obstetric diagnoses and procedures in hospital discharge data. Am J Obstet Gynecol 2006;194:992–1001.
33. Lydon-Rochelle MT, Holt VL, Cardenas V, Nelson JC, Easterling TR, Gardella C, et al. The reporting of pre-existing maternal medical conditions and complications of pregnancy on birth certificates and in hospital discharge data. Am J Obstet Gynecol 2005;193:125–34.
34. Lydon-Rochelle MT, Holt VL, Nelson JC, Cardenas V, Gardella C, Easterling TR, et al. Accuracy of reporting maternal in-hospital diagnoses and intrapartum procedures in Washington State linked birth records. Paediatr Perinat Epidemiol 2005;19:460–71.
35. Harding Clark M. The Lourdes Hospital inquiry: An inquiry into peripartum hysterectomy at Our Lady of Lourdes Hospital, Drogheda. Department of Health and Children. Dublin, Ireland: The Stationary Office; 2006.