Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Trends in postpartum haemorrhage

Trends in postpartum haemorrhage Carolyn A. Cameron Centre for Perinatal Health Services Research, University of Sydney, New South Wales Abstract Objective: To assess trends and outcomes of postpartum haemorrhage (PPH) in New South Wales (NSW). Methods: A population-based descriptive study of all 52,151 women who had a PPH either during the hospital stay for the birth of their baby or requiring a re-admission to hospital between 1994 and 2002. Data were obtained from the de-identified computerised census of NSW hospital in-patients and analysed to examine trends over time. The outcome measures included maternal death, hysterectomy, admission to intensive care unit (ICU), transfusion and major maternal morbidity, including procedures to reduce blood supply to the Christine L. Roberts Centre for Perinatal Health Services Research, University of Sydney, New South Wales, and Royal Prince Alfred Hospital, New South Wales Emily C. Olive, Jane B. Ford Centre for Perinatal Health Services Research, University of Sydney, New South Wales Wendy E. Fischer New South Wales Pregnancy and Newborn Services Network ostpartum haemorrhage (PPH) is the leading cause of direct maternal death in Australia.1 Eight deaths attributable to PPH were reported between 1997 and 1999, a rise from five in the previous triennium. Maternal morbidity resulting from PPH is harder to quantify, but it is estimated that between 5-15% of women giving birth in developed countries have excessive blood loss,2,3 with approximately 2% of postnatal women subsequently admitted to hospital with a PPH.4 Morbidity associated with PPH includes hysterectomy, other surgical procedures to reduce blood flow to the uterus, transfusion, renal failure and coagulation deficiencies. In 2000, a coronial inquest into a maternal death in New South Wales (NSW) sparked a review of hospital PPH policies. Deficiencies in existing clinical policies were identified and in November 2002, the NSW Department of Health issued a new evidence-based policy framework for the prevention, early recognition, management and monitoring of PPH. The policy recommends active management of the third stage of labour for all women, that is, the administration of a prophylactic oxytocic drug, early cord clamping and controlled cord traction to deliver the placenta. These measures have been shown to reduce the risk of PPH by more than 50%.5 The aim of this paper is to assess trends and outcomes of PPH prior to the introduction of the new policy framework. To date, there is little published data on the rate of PPH in Australia. uterus, acute renal failure and postpartum coagulation defects. Results: From 1994 to 2002 both the number and adjusted (for under-reporting) rate of PPH during the birth admission increased from 8.3% of deliveries to 10.7%. The rate of PPH adjusted for maternal age and mode of delivery was similar to the unadjusted rate. There was a sixfold increase in the rate of transfusions from 1.9% of women who haemorrhaged Methods Data source De-identified data were obtained from the NSW Department of Health’s computerised census of all in-patients treated in NSW hospitals – the Inpatient Statistics Collection (ISC). The ISC contains details of all inpatient separations (discharges, transfers and deaths), including information relating to the patient’s diagnosis and the procedures undertaken, from every public, private, and repatriation hospital, private day procedure centre and public nursing home in NSW.6 Each period of in-patient care generates a separate record regardless of the time between the date of separation and a to 11.7%. Hospital readmissions for PPH declined from 1.2% of deliveries to 0.9%. These were statistically significant changes. There were no significant changes in the rate of hysterectomies, procedures to reduce blood supply to the uterus, admissions to ICU, acute renal failure or coagulation defects. Conclusion: The increased rate of PPH during the birth admission is concerning. The increase in PPH could not be explained by increasing maternal age or caesarean sections. Linked birth and hospital discharge data could determine whether the increase in PPH is caused by other changes in obstetric practices or population. (Aust N Z J Public Health 2006; 30: 151-6) Correspondence to: Ms Carolyn Cameron, Centre for Perinatal Health Services Research, Building DO2 University of Sydney, New South Wales 2006. Fax: (02) 9351 7742; e-mail: ccameron@perinatal.usyd.edu.au Submitted: May 2005 Revision requested: July 2005 Accepted: September 2005 2006 VOL. 30 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Cameron et al. Article subsequent re-admission. The ISC is a financial year collection from 1 July through to 30 June of the following year. Herein, each financial year will be referred to by the second half of the financial year, so for example, the financial year 1993/94 will be known as 1994. The diagnostic and procedure information within the ISC is coded using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9)7 for the years to June 1998, and the International Classification of Diseases, Tenth Revision, Australian Modification (ICD-10)8 from July 1998 onwards. Each separation is classified using up to 40 diagnostic and 31 procedure variables. Study population PPH may occur during the birth admission or women may be admitted with a PPH subsequent to the birth. Thus the study population included all women discharged from a NSW hospital between 1 July 1993 and 30 June 2002 following a birth, where a diagnosis of PPH was recorded in the hospital separation data, or with a primary or stay diagnosis of delayed or secondary postpartum haemorrhage (a postpartum admission for PPH). Homebirths and births to NSW women in hospitals outside NSW were not available. major maternal morbidity and maternal death per 100 PPHs. We separately report the incidence of PPH during the birth admission and PPH requiring a subsequent admission to hospital. PPH is known to be under-reported by 41%.10 We determined PPH rates adjusted for under-reporting. To assess any contribution of increases in both maternal age and the caesarean section rate, we generated rates of PPH adjusted for maternal age and mode of delivery. We used the direct method of standardisation with 2002 as the standard population. Changes in other outcomes over time were investigated using χ² for trend statistic. The level of significance for trend was set at p<0.01 because of the large number of births and statistical comparisons. Analyses were conducted using the SAS System for Windows via the NSW Department of Health’s HOIST (Health Outcomes Information and Statistical Toolkit) data warehouse system. The study was approved by the University of Sydney Human Research Ethics Committee. Results A total of 775,073 women gave birth in a NSW hospital over the nine-year study period and 52,151 women had a postpartum haemorrhage recorded in their NSW hospital separation data (ISC record). Of these, 43,505 women (5.6% of those giving birth) had a PPH during the birth admission and 8,646 women (1.1% of those giving birth) were readmitted to hospital with a PPH. Definitions and outcomes Birth admissions were identified by a diagnosis of delivery or if the Australian National Diagnosis Related Group code (a classification system used to assist with resource allocation) indicated a delivery (see Table 1 for ANDRG codes). PPH is defined in the NSW Department of Health’s policy Framework For Prevention, Early Recognition And Management Of Postpartum Haemorrhage as any blood loss where a clinical diagnosis of PPH is recorded, or if blood loss is reported as greater than 500 ml following a vaginal delivery or a post-caesarean blood loss of more than 750 ml.9 This is consistent with both ICD-9 and ICD-10 definitions for PPH7,8 (see Table 1 for ICD-10 codes). Maternal outcomes among women with PPH included transfusion (including whole blood, red cells, platelets, coagulation factors, plasma and blood expanders), dilatation and curettage (D&C), hysterectomy, admission to an intensive care unit (ICU), procedures to reduce blood supply to the uterus (uterine artery or internal iliac artery ligation or embolisation), acute renal failure, disseminated intravascular coagulopathy (DIC), postpartum coagulation disorders and maternal death. We defined major maternal morbidity as any procedure to reduce blood supply to the uterus, acute renal failure following labour and delivery, DIC or a postpartum coagulation disorder. Other maternal factors available for analysis were age (categorised as <20, 20-34 and ≥35) and delivery (categorised as vaginal and caesarean section). Table 1: Framework for identifying and reporting PPH and associated morbidities using hospital discharge data (ICD-10 codes).8 PPH during birth admission 1. Select Delivery Z37.0 – Z37.9 (in any of the diagnosis fields) or ANDRG V4.2: O01A – D, O02Z, O60A – D and 2. PPH O72 (in any of the diagnosis fields) PPH during the birth admission and associated morbidity 1. As above and 2. Hysterectomy 35653 (in any of the procedure fields) or 3. Major maternal morbidity (in any of the procedure/diagnosis fields) a. Procedures to reduce blood supply to the uterus 34103-12, 35321-00 or b. DIC/Postpartum coagulation defects D65, O72.3 or c. Acute renal failure following delivery O90.4 or 4. Transfusion 13706, 92060-92064 (in any of the procedure fields) Readmission for PPH 1. PPH O72.2 (as the principal [icd10d1] or stay diagnosis [icd10d2]) Readmission for PPH and associated morbidity 1. As above and 2. Postpartum evacuation of uterus (dilatation and curettage/ suction curettage) 16564 (in any of the procedure fields) or 3. Transfusion 13706, 92060-92064 (in any of the procedure fields) Analysis We estimated the annual incidence of PPH per 100 deliveries and the incidence of hysterectomy, transfusion, admission to ICU, AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2006 VOL. 30 NO. 2 Vulnerable People Trends in postpartum haemorrhage The number and rate of PPHs occurring during the birth admission increased significantly over time from a crude rate of 4.9% in 1994 to 6.3% in 2002 (see Table 2, Figure 1). After adjusting for under-reporting, PPHs during the birth admission increased from 8.3% in 1994 to 10.7% in 2002 (see Table 2). Postpartum haemorrhage rates adjusted for maternal age and mode of delivery were similar to the unadjusted PPH rates (see Figure 1). There was a significant increase in transfusion associated with PPH in both the birth admission and subsequent hospitalisation. In 1994, less than 1.9% of women who haemorrhaged following delivery received a transfusion. This proportion rose more than sixfold to 11.7% in 2002 (see Table 2). A total of 226 women had a hysterectomy, ranging between 19 and 31 per annum (2-4 per 1,000 deliveries). Among women with a PPH the rate of admission to intensive care increased slightly over the study period but did not reach statistical significance (p=0.04). There was no change in the proportion of women who suffered major maternal morbidity associated with PPH. Twenty-four women underwent procedures to reduce blood flow to the uterus between 1998 (when these codes were introduced) and 2002, 51 women (range 2-10 per annum) were reported to have acute renal failure following delivery and 334 women (range 21-47 per annum) were reported to have DIC or postpartum coagulation defects. There was a significant decrease in the number and proportion of women readmitted to hospital with a PPH over the study period (see Table 2). Among the women readmitted, the proportion undergoing a D&C also decreased from 74% in 1994 to 54% in 2002. Among women who had a PPH either during the birth admission or a subsequent admission for PPH, there were five maternal deaths recorded during the nine-year period, four during the birth admission and one during a readmission for PPH. Table 1 gives ICD-10 codes used to identify maternal morbidity associated with PPH that could be used at hospital, local, State and national levels to ensure uniform reporting. Discussion We report trends in PPH and our findings show that the rate of PPH during the birth admission increased between 1994 and 2002 to more than 6% of all deliveries, prior to the introduction of the NSW Department of Health’s policy framework on PPH. Possible explanations for the increase may be attributed to changes in three main areas – the ascertainment and reporting of PPH, variations in obstetric practices, and changes in maternal characteristics. Determining the quantity of blood loss at delivery is almost always a subjective measure relying primarily on visual estimation by caregivers and is underestimated11 and consequently underreported. We know of no reason to expect a change in PPH reporting during the study period, although it has probably been systematically under-reported. A recent validation study of the ISC for birth admission diagnostic codes compared with patient notes found that although most reported PPH cases were true cases (positive predictive value 94%), PPH was under-numerated (sensitivity 58.6%, [95% CI 38.9-76.5]), indicating that these rates are an underestimate.10 (This is consistent with a 1990 validation of population birth data (sensitivity 66%)).12 Although the administrative coding system used by hospitals was updated to ICD-10 in July 1998, the definition of PPH did not alter and a trend of increasing PPH during the birth admission was evident Table 2: Maternal morbidity and mortality associated with PPH between 1 July 1993 and 30 June 2002. Obstetric event or procedure PPH at delivery (n) (% of deliveries) PPH rate adjusted for age and mode of delivery (% of deliveries) PPH rate adjusted for under-reporting (% of deliveries) Hysterectomy (n) (% of PPH) Transfusion (n) (% of PPH) Admission to ICU (n) (% of PPH) Major maternal morbidity (n) (% of PPH) Readmission for PPH (n) (% of deliveries) Transfusion (n) (% of PPH) D&Cb (n) (% of PPH) 4,281 (4.9) 4.8 8.3 20 (0.5) 81 (1.9) 45 (1.1) 47 (1.10) 1,067 (1.2) 7 (0.7) 793 (74.3) 4,106 (4.6) 4.6 8.1 22 (0.5) 121 (3.0) 45 (1.1) 40 (0.97) 975 (1.1) 7 (0.7) 700 (71.8) 4,377 (5.1) 5.0 8.8 24 (0.6) 211 (4.8) 67 (1.5) 45 (1.03) 1,062 (1.3) 20 (1.9) 728 (68.6) 4,786 (5.5) 5.2 9.4 26 (0.5) 314 (6.6) 46 (1.0) 36 (0.75) 1,009 (1.2) 24 (2.4) 717 (71.1) 4,800 (5.6) 5.5 9.5 31 (0.7) 429 (8.9) 55 (1.2) 51 (1.06) 998 (1.2) 25 (2.5) 698 (69.9) 5,104 (5.9) 5.8 10.1 31 (0.6) 491 (9.6) 53 (1.0) 56 (1.10) 991 (1.2) 26 (2.6) 581 (58.6) 5,374 (6.2) 6.1 10.6 27 (0.5) 548 (10.2) 57 (1.1) 29 (0.54) 893 (1.0) 29 (3.3) 570 (63.8) 5,378 (6.4) 6.2 10.9 19 (0.4) 560 (10.4) 59 (1.1) 42 (0.78) 929 (1.1) 38 (4.1) 530 (57.1) 5,299 (6.3) 6.3 10.7 26 (0.5) 619 (11.7) 65 (1.2) 56 (1.06) 722 (0.9) 44 (6.9) 388 (53.7) p valuea <0.001 0.43 <0.001 0.04 0.21 <0.001 <0.001 <0.001 Notes: (a) p value for χ²trend (b) Postpartum evacuation of uterus (dilatation and curettage/suction curettage). 2006 VOL. 30 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Cameron et al. Article prior to the introduction of ICD-10. The introduction of the new policy framework and focus on PPH may increase PPH reporting even if practices to reduce PPH are more rigorously instituted. Changes in obstetric practice that have an impact on known risk factors for PPH, such as alterations to the therapeutic regimen and techniques used in induction, delivery and management of the third stage of labour, have occurred during the study period. Although we do not have information on practices among individual clinicians or hospitals (and do not know if practice reflects the available evidence), a Cochrane review first published in 1997 and updated in 2000 found active management of the third stage of labour more effective than expectant management in terms of blood loss, postpartum haemorrhage and other serious complications.5 A systematic review comparing prophylactic syntometrine and oxytocin for delivery of the placenta found syntometrine was associated with less blood loss (when blood loss was moderate) but more side-effects (nausea, vomiting and hypertension) than oxytocin.13 The Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG) issued a statement in 2003 indicating that third-stage management may be either active or expectant but recommending active management, although clinicians are free to practice at their own discretion.14 Publication of major clinical trials has been shown to influence treatment practices,15 however we do not know what impact the Cochrane reviews and RANZCOG statement had on third-stage management and the ISC does not include data on medications used in labour. Other changes in obstetric practice may have occurred that have an effect on PPH risk factors, such as an increase in inductions of labour.16,17 Figure 1: Trends in PPH rates in NSW, 1994-2002. Increasing maternal age has been one of the most significant changes in maternal characteristics over the past decade.16,17 In 1994, 13% of women giving birth were 35 years or older and by 2002 this proportion had risen to more than 18%.18,19 Older women are at increased risk of adverse maternal and perinatal outcomes including multiple gestation (associated with advances in assisted reproductive technologies), breech presentation and caesarean section.20-22 Caesarean sections have increased from 17.3% of all births in NSW in 1994 to 24.9% in 2002.18,19,23 We assessed the contribution of these factors to the PPH rate. However, the similarity between the adjusted and unadjusted rates suggests that the observed increase in PPHs could not be attributed to the changes in maternal age and caesarean section rates during the study period. Descriptive maternal data including information on parity, length and type of labour, and medical or obstetrical complications, which would enable the increase in PPH rates to be investigated, are not included in the ISC. The usual source of this information is the NSW Midwives Data Collection (MDC), however PPH has not been recorded on the MDC since 1998, thus necessitating the use of the ISC to determine PPH rates during the birth admission. In any case, postpartum haemorrhages requiring a readmission to hospital would have to be ascertained from the ISC as they are separate to the admission for delivery. The use of linked birth and hospital discharge data could determine whether the increase in PPH is caused by changes in obstetric practices or population. There was no significant increase in the rates of hysterectomy and major maternal morbidity over the study period. Surgical procedures such as internal iliac artery ligation and hysterectomy PPH as a % of all births 0 1994 1995 1996 1997 1998 Year PPH at delivery PPH rate adjusted for under-reporting Readmission for PPH PPH rate adjusted for age and mode of delivery 1999 2000 2001 2002 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2006 VOL. 30 NO. 2 Vulnerable People Trends in postpartum haemorrhage are used to manage severe PPH. Uterine artery embolisation is a relatively new procedure requiring the use of angiographic facilities, which are not available in all NSW hospitals. We found no significant change in the rate of admission to intensive care, although this may be an underestimation of the severity of PPH. A recent descriptive study from Scotland found that only about one-third of women with severe maternal morbidity are transferred to intensive care units – suggesting that as many maternity units provide high-dependency care there are fewer admissions to ICU24 and consequently the severity of the morbidity may be underestimated. The rates of acute renal failure and postpartum coagulation defects are among all women who had a PPH. One of the limitations of cross-sectional data is the inability to establish causation. Thus we can only determine that these outcomes occurred among women who had a PPH and not whether PPH, preeclampsia or multiple factors contributed to these outcomes. The significant increase in the rate of transfusion among both women who haemorrhaged during the birth admission and those subsequently admitted with a PPH may indicate that blood loss is increasing or there may be a lowering of the threshold to transfuse. Alternatively, it may reflect an increase in confidence both caregivers and patients have in the quality and safety of blood and blood products. Possible explanations for the reduction in readmissions for PPH may include changes in admission policy, with a higher threshold for readmission. Our findings suggest that women were more likely to receive treatment such as transfusion or D&C and less likely to be admitted for general care and monitoring. However, the number of women readmitted to hospital with a PPH may be overestimated as the reported figures for subsequent admissions are based on admissions, not on patients. Therefore a woman may have more than one admission for the same complication and would be counted every time she was admitted to hospital. There is considerable variation between jurisdictions in the definition, ascertainment and reporting of PPH.25 The NSW rate (6.3% during the birth admission in 2002) is lower than the recently reported rates in Victoria (9.2%)26 and New Zealand (8.7%),27 but comparable to rates in South Australia (6.4%)28 and England (6.6%).29 Data from Queensland30 and a recent study from the United States using hospital discharge survey data31 showed considerably lower rates of PPH. Caution must be exercised when comparing rates since definitions and recording practices may differ. Accurate documentation of PPH in the patient notes is of paramount importance. Improved documentation in medical records will increase the accuracy of routinely collected data. Use of the framework (see Table 1) will facilitate uniform reporting and monitoring of PPH at hospital, regional and State level and enable hospitals to fulfil the reporting and monitoring requirements of the NSW Department of Health’s PPH policy framework. Conclusions The findings from this study show that the ISC is a useful data source for monitoring the rate of PPH in NSW. However, as PPH is commonly underestimated, the quality of the data would be improved by better recognition and documentation starting from the clinical setting. The increased rate of PPH during the birth admission is concerning. Further research using linked birth and hospital discharge data would provide more detailed information about the impact of risk factors for PPH in the NSW population. With the recent dissemination of the NSW Department of Health’s PPH policy framework across the State, we would anticipate an increased awareness and reporting of PPH in the future. Acknowledgements The authors would like to thank the hospital staff who collect data for the Inpatient Statistics Collection. They would also like to acknowledge the NSW Department of Health staff who designed, implemented and maintain the HOIST data warehouse system. Christine Roberts is supported by a National Health and Medical Research Council (NHMRC) of Australia Public Health Practitioner Fellowship. Jane Ford is supported by the Health Evaluation and Research Outcomes Network (HERON), a NHMRC Capacity Building Grant in Population Health Research. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australian and New Zealand Journal of Public Health Wiley

Loading next page...
 
/lp/wiley/trends-in-postpartum-haemorrhage-VMVDgQxGX6

References (30)

Publisher
Wiley
Copyright
Copyright © 2006 Wiley Subscription Services, Inc., A Wiley Company
ISSN
1326-0200
eISSN
1753-6405
DOI
10.1111/j.1467-842X.2006.tb00109.x
Publisher site
See Article on Publisher Site

Abstract

Carolyn A. Cameron Centre for Perinatal Health Services Research, University of Sydney, New South Wales Abstract Objective: To assess trends and outcomes of postpartum haemorrhage (PPH) in New South Wales (NSW). Methods: A population-based descriptive study of all 52,151 women who had a PPH either during the hospital stay for the birth of their baby or requiring a re-admission to hospital between 1994 and 2002. Data were obtained from the de-identified computerised census of NSW hospital in-patients and analysed to examine trends over time. The outcome measures included maternal death, hysterectomy, admission to intensive care unit (ICU), transfusion and major maternal morbidity, including procedures to reduce blood supply to the Christine L. Roberts Centre for Perinatal Health Services Research, University of Sydney, New South Wales, and Royal Prince Alfred Hospital, New South Wales Emily C. Olive, Jane B. Ford Centre for Perinatal Health Services Research, University of Sydney, New South Wales Wendy E. Fischer New South Wales Pregnancy and Newborn Services Network ostpartum haemorrhage (PPH) is the leading cause of direct maternal death in Australia.1 Eight deaths attributable to PPH were reported between 1997 and 1999, a rise from five in the previous triennium. Maternal morbidity resulting from PPH is harder to quantify, but it is estimated that between 5-15% of women giving birth in developed countries have excessive blood loss,2,3 with approximately 2% of postnatal women subsequently admitted to hospital with a PPH.4 Morbidity associated with PPH includes hysterectomy, other surgical procedures to reduce blood flow to the uterus, transfusion, renal failure and coagulation deficiencies. In 2000, a coronial inquest into a maternal death in New South Wales (NSW) sparked a review of hospital PPH policies. Deficiencies in existing clinical policies were identified and in November 2002, the NSW Department of Health issued a new evidence-based policy framework for the prevention, early recognition, management and monitoring of PPH. The policy recommends active management of the third stage of labour for all women, that is, the administration of a prophylactic oxytocic drug, early cord clamping and controlled cord traction to deliver the placenta. These measures have been shown to reduce the risk of PPH by more than 50%.5 The aim of this paper is to assess trends and outcomes of PPH prior to the introduction of the new policy framework. To date, there is little published data on the rate of PPH in Australia. uterus, acute renal failure and postpartum coagulation defects. Results: From 1994 to 2002 both the number and adjusted (for under-reporting) rate of PPH during the birth admission increased from 8.3% of deliveries to 10.7%. The rate of PPH adjusted for maternal age and mode of delivery was similar to the unadjusted rate. There was a sixfold increase in the rate of transfusions from 1.9% of women who haemorrhaged Methods Data source De-identified data were obtained from the NSW Department of Health’s computerised census of all in-patients treated in NSW hospitals – the Inpatient Statistics Collection (ISC). The ISC contains details of all inpatient separations (discharges, transfers and deaths), including information relating to the patient’s diagnosis and the procedures undertaken, from every public, private, and repatriation hospital, private day procedure centre and public nursing home in NSW.6 Each period of in-patient care generates a separate record regardless of the time between the date of separation and a to 11.7%. Hospital readmissions for PPH declined from 1.2% of deliveries to 0.9%. These were statistically significant changes. There were no significant changes in the rate of hysterectomies, procedures to reduce blood supply to the uterus, admissions to ICU, acute renal failure or coagulation defects. Conclusion: The increased rate of PPH during the birth admission is concerning. The increase in PPH could not be explained by increasing maternal age or caesarean sections. Linked birth and hospital discharge data could determine whether the increase in PPH is caused by other changes in obstetric practices or population. (Aust N Z J Public Health 2006; 30: 151-6) Correspondence to: Ms Carolyn Cameron, Centre for Perinatal Health Services Research, Building DO2 University of Sydney, New South Wales 2006. Fax: (02) 9351 7742; e-mail: ccameron@perinatal.usyd.edu.au Submitted: May 2005 Revision requested: July 2005 Accepted: September 2005 2006 VOL. 30 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Cameron et al. Article subsequent re-admission. The ISC is a financial year collection from 1 July through to 30 June of the following year. Herein, each financial year will be referred to by the second half of the financial year, so for example, the financial year 1993/94 will be known as 1994. The diagnostic and procedure information within the ISC is coded using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9)7 for the years to June 1998, and the International Classification of Diseases, Tenth Revision, Australian Modification (ICD-10)8 from July 1998 onwards. Each separation is classified using up to 40 diagnostic and 31 procedure variables. Study population PPH may occur during the birth admission or women may be admitted with a PPH subsequent to the birth. Thus the study population included all women discharged from a NSW hospital between 1 July 1993 and 30 June 2002 following a birth, where a diagnosis of PPH was recorded in the hospital separation data, or with a primary or stay diagnosis of delayed or secondary postpartum haemorrhage (a postpartum admission for PPH). Homebirths and births to NSW women in hospitals outside NSW were not available. major maternal morbidity and maternal death per 100 PPHs. We separately report the incidence of PPH during the birth admission and PPH requiring a subsequent admission to hospital. PPH is known to be under-reported by 41%.10 We determined PPH rates adjusted for under-reporting. To assess any contribution of increases in both maternal age and the caesarean section rate, we generated rates of PPH adjusted for maternal age and mode of delivery. We used the direct method of standardisation with 2002 as the standard population. Changes in other outcomes over time were investigated using χ² for trend statistic. The level of significance for trend was set at p<0.01 because of the large number of births and statistical comparisons. Analyses were conducted using the SAS System for Windows via the NSW Department of Health’s HOIST (Health Outcomes Information and Statistical Toolkit) data warehouse system. The study was approved by the University of Sydney Human Research Ethics Committee. Results A total of 775,073 women gave birth in a NSW hospital over the nine-year study period and 52,151 women had a postpartum haemorrhage recorded in their NSW hospital separation data (ISC record). Of these, 43,505 women (5.6% of those giving birth) had a PPH during the birth admission and 8,646 women (1.1% of those giving birth) were readmitted to hospital with a PPH. Definitions and outcomes Birth admissions were identified by a diagnosis of delivery or if the Australian National Diagnosis Related Group code (a classification system used to assist with resource allocation) indicated a delivery (see Table 1 for ANDRG codes). PPH is defined in the NSW Department of Health’s policy Framework For Prevention, Early Recognition And Management Of Postpartum Haemorrhage as any blood loss where a clinical diagnosis of PPH is recorded, or if blood loss is reported as greater than 500 ml following a vaginal delivery or a post-caesarean blood loss of more than 750 ml.9 This is consistent with both ICD-9 and ICD-10 definitions for PPH7,8 (see Table 1 for ICD-10 codes). Maternal outcomes among women with PPH included transfusion (including whole blood, red cells, platelets, coagulation factors, plasma and blood expanders), dilatation and curettage (D&C), hysterectomy, admission to an intensive care unit (ICU), procedures to reduce blood supply to the uterus (uterine artery or internal iliac artery ligation or embolisation), acute renal failure, disseminated intravascular coagulopathy (DIC), postpartum coagulation disorders and maternal death. We defined major maternal morbidity as any procedure to reduce blood supply to the uterus, acute renal failure following labour and delivery, DIC or a postpartum coagulation disorder. Other maternal factors available for analysis were age (categorised as <20, 20-34 and ≥35) and delivery (categorised as vaginal and caesarean section). Table 1: Framework for identifying and reporting PPH and associated morbidities using hospital discharge data (ICD-10 codes).8 PPH during birth admission 1. Select Delivery Z37.0 – Z37.9 (in any of the diagnosis fields) or ANDRG V4.2: O01A – D, O02Z, O60A – D and 2. PPH O72 (in any of the diagnosis fields) PPH during the birth admission and associated morbidity 1. As above and 2. Hysterectomy 35653 (in any of the procedure fields) or 3. Major maternal morbidity (in any of the procedure/diagnosis fields) a. Procedures to reduce blood supply to the uterus 34103-12, 35321-00 or b. DIC/Postpartum coagulation defects D65, O72.3 or c. Acute renal failure following delivery O90.4 or 4. Transfusion 13706, 92060-92064 (in any of the procedure fields) Readmission for PPH 1. PPH O72.2 (as the principal [icd10d1] or stay diagnosis [icd10d2]) Readmission for PPH and associated morbidity 1. As above and 2. Postpartum evacuation of uterus (dilatation and curettage/ suction curettage) 16564 (in any of the procedure fields) or 3. Transfusion 13706, 92060-92064 (in any of the procedure fields) Analysis We estimated the annual incidence of PPH per 100 deliveries and the incidence of hysterectomy, transfusion, admission to ICU, AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2006 VOL. 30 NO. 2 Vulnerable People Trends in postpartum haemorrhage The number and rate of PPHs occurring during the birth admission increased significantly over time from a crude rate of 4.9% in 1994 to 6.3% in 2002 (see Table 2, Figure 1). After adjusting for under-reporting, PPHs during the birth admission increased from 8.3% in 1994 to 10.7% in 2002 (see Table 2). Postpartum haemorrhage rates adjusted for maternal age and mode of delivery were similar to the unadjusted PPH rates (see Figure 1). There was a significant increase in transfusion associated with PPH in both the birth admission and subsequent hospitalisation. In 1994, less than 1.9% of women who haemorrhaged following delivery received a transfusion. This proportion rose more than sixfold to 11.7% in 2002 (see Table 2). A total of 226 women had a hysterectomy, ranging between 19 and 31 per annum (2-4 per 1,000 deliveries). Among women with a PPH the rate of admission to intensive care increased slightly over the study period but did not reach statistical significance (p=0.04). There was no change in the proportion of women who suffered major maternal morbidity associated with PPH. Twenty-four women underwent procedures to reduce blood flow to the uterus between 1998 (when these codes were introduced) and 2002, 51 women (range 2-10 per annum) were reported to have acute renal failure following delivery and 334 women (range 21-47 per annum) were reported to have DIC or postpartum coagulation defects. There was a significant decrease in the number and proportion of women readmitted to hospital with a PPH over the study period (see Table 2). Among the women readmitted, the proportion undergoing a D&C also decreased from 74% in 1994 to 54% in 2002. Among women who had a PPH either during the birth admission or a subsequent admission for PPH, there were five maternal deaths recorded during the nine-year period, four during the birth admission and one during a readmission for PPH. Table 1 gives ICD-10 codes used to identify maternal morbidity associated with PPH that could be used at hospital, local, State and national levels to ensure uniform reporting. Discussion We report trends in PPH and our findings show that the rate of PPH during the birth admission increased between 1994 and 2002 to more than 6% of all deliveries, prior to the introduction of the NSW Department of Health’s policy framework on PPH. Possible explanations for the increase may be attributed to changes in three main areas – the ascertainment and reporting of PPH, variations in obstetric practices, and changes in maternal characteristics. Determining the quantity of blood loss at delivery is almost always a subjective measure relying primarily on visual estimation by caregivers and is underestimated11 and consequently underreported. We know of no reason to expect a change in PPH reporting during the study period, although it has probably been systematically under-reported. A recent validation study of the ISC for birth admission diagnostic codes compared with patient notes found that although most reported PPH cases were true cases (positive predictive value 94%), PPH was under-numerated (sensitivity 58.6%, [95% CI 38.9-76.5]), indicating that these rates are an underestimate.10 (This is consistent with a 1990 validation of population birth data (sensitivity 66%)).12 Although the administrative coding system used by hospitals was updated to ICD-10 in July 1998, the definition of PPH did not alter and a trend of increasing PPH during the birth admission was evident Table 2: Maternal morbidity and mortality associated with PPH between 1 July 1993 and 30 June 2002. Obstetric event or procedure PPH at delivery (n) (% of deliveries) PPH rate adjusted for age and mode of delivery (% of deliveries) PPH rate adjusted for under-reporting (% of deliveries) Hysterectomy (n) (% of PPH) Transfusion (n) (% of PPH) Admission to ICU (n) (% of PPH) Major maternal morbidity (n) (% of PPH) Readmission for PPH (n) (% of deliveries) Transfusion (n) (% of PPH) D&Cb (n) (% of PPH) 4,281 (4.9) 4.8 8.3 20 (0.5) 81 (1.9) 45 (1.1) 47 (1.10) 1,067 (1.2) 7 (0.7) 793 (74.3) 4,106 (4.6) 4.6 8.1 22 (0.5) 121 (3.0) 45 (1.1) 40 (0.97) 975 (1.1) 7 (0.7) 700 (71.8) 4,377 (5.1) 5.0 8.8 24 (0.6) 211 (4.8) 67 (1.5) 45 (1.03) 1,062 (1.3) 20 (1.9) 728 (68.6) 4,786 (5.5) 5.2 9.4 26 (0.5) 314 (6.6) 46 (1.0) 36 (0.75) 1,009 (1.2) 24 (2.4) 717 (71.1) 4,800 (5.6) 5.5 9.5 31 (0.7) 429 (8.9) 55 (1.2) 51 (1.06) 998 (1.2) 25 (2.5) 698 (69.9) 5,104 (5.9) 5.8 10.1 31 (0.6) 491 (9.6) 53 (1.0) 56 (1.10) 991 (1.2) 26 (2.6) 581 (58.6) 5,374 (6.2) 6.1 10.6 27 (0.5) 548 (10.2) 57 (1.1) 29 (0.54) 893 (1.0) 29 (3.3) 570 (63.8) 5,378 (6.4) 6.2 10.9 19 (0.4) 560 (10.4) 59 (1.1) 42 (0.78) 929 (1.1) 38 (4.1) 530 (57.1) 5,299 (6.3) 6.3 10.7 26 (0.5) 619 (11.7) 65 (1.2) 56 (1.06) 722 (0.9) 44 (6.9) 388 (53.7) p valuea <0.001 0.43 <0.001 0.04 0.21 <0.001 <0.001 <0.001 Notes: (a) p value for χ²trend (b) Postpartum evacuation of uterus (dilatation and curettage/suction curettage). 2006 VOL. 30 NO. 2 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH Cameron et al. Article prior to the introduction of ICD-10. The introduction of the new policy framework and focus on PPH may increase PPH reporting even if practices to reduce PPH are more rigorously instituted. Changes in obstetric practice that have an impact on known risk factors for PPH, such as alterations to the therapeutic regimen and techniques used in induction, delivery and management of the third stage of labour, have occurred during the study period. Although we do not have information on practices among individual clinicians or hospitals (and do not know if practice reflects the available evidence), a Cochrane review first published in 1997 and updated in 2000 found active management of the third stage of labour more effective than expectant management in terms of blood loss, postpartum haemorrhage and other serious complications.5 A systematic review comparing prophylactic syntometrine and oxytocin for delivery of the placenta found syntometrine was associated with less blood loss (when blood loss was moderate) but more side-effects (nausea, vomiting and hypertension) than oxytocin.13 The Royal Australian and New Zealand College of Obstetricians and Gynaecologists (RANZCOG) issued a statement in 2003 indicating that third-stage management may be either active or expectant but recommending active management, although clinicians are free to practice at their own discretion.14 Publication of major clinical trials has been shown to influence treatment practices,15 however we do not know what impact the Cochrane reviews and RANZCOG statement had on third-stage management and the ISC does not include data on medications used in labour. Other changes in obstetric practice may have occurred that have an effect on PPH risk factors, such as an increase in inductions of labour.16,17 Figure 1: Trends in PPH rates in NSW, 1994-2002. Increasing maternal age has been one of the most significant changes in maternal characteristics over the past decade.16,17 In 1994, 13% of women giving birth were 35 years or older and by 2002 this proportion had risen to more than 18%.18,19 Older women are at increased risk of adverse maternal and perinatal outcomes including multiple gestation (associated with advances in assisted reproductive technologies), breech presentation and caesarean section.20-22 Caesarean sections have increased from 17.3% of all births in NSW in 1994 to 24.9% in 2002.18,19,23 We assessed the contribution of these factors to the PPH rate. However, the similarity between the adjusted and unadjusted rates suggests that the observed increase in PPHs could not be attributed to the changes in maternal age and caesarean section rates during the study period. Descriptive maternal data including information on parity, length and type of labour, and medical or obstetrical complications, which would enable the increase in PPH rates to be investigated, are not included in the ISC. The usual source of this information is the NSW Midwives Data Collection (MDC), however PPH has not been recorded on the MDC since 1998, thus necessitating the use of the ISC to determine PPH rates during the birth admission. In any case, postpartum haemorrhages requiring a readmission to hospital would have to be ascertained from the ISC as they are separate to the admission for delivery. The use of linked birth and hospital discharge data could determine whether the increase in PPH is caused by changes in obstetric practices or population. There was no significant increase in the rates of hysterectomy and major maternal morbidity over the study period. Surgical procedures such as internal iliac artery ligation and hysterectomy PPH as a % of all births 0 1994 1995 1996 1997 1998 Year PPH at delivery PPH rate adjusted for under-reporting Readmission for PPH PPH rate adjusted for age and mode of delivery 1999 2000 2001 2002 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH 2006 VOL. 30 NO. 2 Vulnerable People Trends in postpartum haemorrhage are used to manage severe PPH. Uterine artery embolisation is a relatively new procedure requiring the use of angiographic facilities, which are not available in all NSW hospitals. We found no significant change in the rate of admission to intensive care, although this may be an underestimation of the severity of PPH. A recent descriptive study from Scotland found that only about one-third of women with severe maternal morbidity are transferred to intensive care units – suggesting that as many maternity units provide high-dependency care there are fewer admissions to ICU24 and consequently the severity of the morbidity may be underestimated. The rates of acute renal failure and postpartum coagulation defects are among all women who had a PPH. One of the limitations of cross-sectional data is the inability to establish causation. Thus we can only determine that these outcomes occurred among women who had a PPH and not whether PPH, preeclampsia or multiple factors contributed to these outcomes. The significant increase in the rate of transfusion among both women who haemorrhaged during the birth admission and those subsequently admitted with a PPH may indicate that blood loss is increasing or there may be a lowering of the threshold to transfuse. Alternatively, it may reflect an increase in confidence both caregivers and patients have in the quality and safety of blood and blood products. Possible explanations for the reduction in readmissions for PPH may include changes in admission policy, with a higher threshold for readmission. Our findings suggest that women were more likely to receive treatment such as transfusion or D&C and less likely to be admitted for general care and monitoring. However, the number of women readmitted to hospital with a PPH may be overestimated as the reported figures for subsequent admissions are based on admissions, not on patients. Therefore a woman may have more than one admission for the same complication and would be counted every time she was admitted to hospital. There is considerable variation between jurisdictions in the definition, ascertainment and reporting of PPH.25 The NSW rate (6.3% during the birth admission in 2002) is lower than the recently reported rates in Victoria (9.2%)26 and New Zealand (8.7%),27 but comparable to rates in South Australia (6.4%)28 and England (6.6%).29 Data from Queensland30 and a recent study from the United States using hospital discharge survey data31 showed considerably lower rates of PPH. Caution must be exercised when comparing rates since definitions and recording practices may differ. Accurate documentation of PPH in the patient notes is of paramount importance. Improved documentation in medical records will increase the accuracy of routinely collected data. Use of the framework (see Table 1) will facilitate uniform reporting and monitoring of PPH at hospital, regional and State level and enable hospitals to fulfil the reporting and monitoring requirements of the NSW Department of Health’s PPH policy framework. Conclusions The findings from this study show that the ISC is a useful data source for monitoring the rate of PPH in NSW. However, as PPH is commonly underestimated, the quality of the data would be improved by better recognition and documentation starting from the clinical setting. The increased rate of PPH during the birth admission is concerning. Further research using linked birth and hospital discharge data would provide more detailed information about the impact of risk factors for PPH in the NSW population. With the recent dissemination of the NSW Department of Health’s PPH policy framework across the State, we would anticipate an increased awareness and reporting of PPH in the future. Acknowledgements The authors would like to thank the hospital staff who collect data for the Inpatient Statistics Collection. They would also like to acknowledge the NSW Department of Health staff who designed, implemented and maintain the HOIST data warehouse system. Christine Roberts is supported by a National Health and Medical Research Council (NHMRC) of Australia Public Health Practitioner Fellowship. Jane Ford is supported by the Health Evaluation and Research Outcomes Network (HERON), a NHMRC Capacity Building Grant in Population Health Research.

Journal

Australian and New Zealand Journal of Public HealthWiley

Published: Apr 1, 2006

There are no references for this article.