- Email: [email protected]
Utility of Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) score in maternal admissions to the intensive care unit
American Journal of Obstetrics and Gynecology (2006) 194, e13–e15
www.ajog.org
Utility of Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) score in maternal admissions to the intensive care unit Tobey A. Stevens, MD,a Mary A. Carroll, MD,a Pamela A. Promecene, MD,a Marilyn Seibel, RN,b Manju Monga, MDa Department of Obstetrics, Gynecology and Reproductive Sciences, University of Texas Medical School Houstona and Memorial Hermann Hospital,b Houston, TX Received for publication September 7, 2005; revised January 6, 2006; accepted January 20, 2006
KEY WORDS Maternal death Intensive care unit APACHE III
Objective: A mean Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) score of O50 is associated with increased intensive care unit mortality rate in nonpregnant cardiac and trauma patients. The objective was to determine the usefulness of the APACHE III score in maternal admissions to an intensive care unit in a tertiary care center in an urban multicultural city. Study design: This was a retrospective review of all maternal admissions (O20 weeks of gestation or after delivery) to an intensive care unit between January 2002 and May 2004. Demographics, obstetric and medical history, and 20 physiologic variables that comprise the APACHE III were recorded. The minimum APACHE III score (lowest risk of death) is 0; maximum is 299. The association between APACHE III score and maternal death was assessed with Mann Whitney U test. Significance was assumed at a probability value of !.05. Results: Fifty-eight subjects met the study criteria. Thirty percent of these women were admitted antepartum (27 G 1.0 weeks of gestation); 31% of the women were admitted on the day of delivery; and 29% of the women were admitted after delivery. Mean maternal age was 27 G 6.7 years. Acute conditions that resulted in transfer to the intensive care unit included preeclampsia (24%), cardiorespiratory disease (21%), hemorrhage (16%), infection (12%), trauma (7%), and thromboembolism (3%). Fifty-five percent of the women had no previous underlying obstetric complications, and 98% of the women had no underlying chronic health condition. Fifty-eight percent of the women received care in a medical intensive care unit; 28% of the women received care in a surgical intensive care unit; 10% of the women received care in a cardiac intensive care unit, and 3% of the women received care in a neurologic intensive care unit. The mean intensive care unit stay was 3.7 G 4.6 days, and the mean hospital stay was 9.0 G 7 days. Three patients died; the rest of the patients went home in good condition. The median APACHE III score was 34 (range, 14-102) and was not correlated with maternal death. Conclusion: The APACHE III is not associated with risk of intensive care unit–related maternal death. Ó 2006 Mosby, Inc. All rights reserved.
Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.01.073
e14 Several studies have addressed the reasons for admission or maternal outcome in women admitted to the intensive care unit (ICU) during pregnancy.1-3 Some studies have described the usefulness of ICU scoring systems in pregnant women,4-11 but only 1 study has been based in the United States.11 Gilbert et al11 recently summarized the use of the Simplified Acute Physiologic Score (SAPS II) in 233 antepartum and postpartum patients who were admitted to a medical ICU in New Jersey from 1991 to 1998. They found that SAPS II predicted hospital mortality rates in obstetric patients who were admitted for medical reasons, but not for indications that were related to pregnancy and delivery. SAPS II is just 1 of several prognostic scoring systems that are available to predict an individual’s risk of dying in the hospital. These models are often derived over several years and may become obsolete because they no longer reflect current practice.12 Recently, The Acute Physiologic and Chronic Health Evaluation (APACHE) tool, which was developed in 1978, has been revised (APACHE III). This scoring system records 27 parameters that are related to diagnosis and physiologic abnormalities on admission to the ICU, age, and pre-existing medical problems.13 Although this has been validated in numerous studies in nonpregnant patients, no previous study has reported the use of the APACHE III scoring system in maternal admissions to the ICU. The objective of this study was to determine the usefulness of the APACHE III score in maternal admissions to an ICU in a tertiary care center.
Methods This was a retrospective review of all maternal admissions to ICU between January 2002 and May 2004 at Memorial Hermann Hospital, a tertiary care center in Houston, Texas. Any chart that was coded for a gynecologic or obstetric diagnosis related group and intensive care admission was pulled. All women who were admitted during the study period who were R20 weeks of gestation or !6 weeks after delivery were included. Demographics, obstetric and medical history, and 20 physiologic variables that comprise the APACHE III were recorded. The minimum APACHE score (lowest risk of death) is 0; the maximum score is 299. A mean APACHE score of O50 is associated with increased ICU death in nonpregnant cardiac and trauma patients.13 The APACHE score was derived from nonpregnant patients and therefore does not include specific pregnancy-related diseases. The association between the APACHE III score and maternal death was assessed with the Mann Whitney U and Fisher exact tests. Significance was assumed at a probability value of !.05.
Stevens et al
Results Fifty-eight subjects met the study criteria. Thirty percent of the women were admitted antepartum (27 G 1.0 weeks); 31% of the women were admitted on the day of delivery; 29% of the women were admitted after delivery. The mean maternal age was 27 G 6.7 years. Acute conditions that resulted in transfer to the ICU included complications of preeclampsia (24%), cardiorespiratory disease (21%), hemorrhage (16%), infection (12%), trauma (7%), and thromboembolism (3%). Other reasons for admission included drug overdose, diabetic ketoacidosis, myasthenic crisis, anaphylaxis, sublingual hematoma, and altered mental status. Ninety-eight percent of the women had no underlying chronic health condition. The only patient with an underlying chronic health condition was a 33-year-old woman at 22 weeks of gestation with acquired immunodeficiency syndrome and presumed pneumocystis carinii pneumonia. Fifty-eight percent of the women received care in a medical ICU; 28% of the women received care in a surgical ICU; 10% of the women received care in a cardiac ICU, and 3% of the women received care in a neurologic ICU. The mean ICU stay was 3.7 G 4.6 days, and the mean hospital stay was 9.0 G 7 days. Three patients died; the rest of the patients went home in good condition. The first patient who died was a 23-year-old at 30 weeks of gestation who had been transferred to the medical ICU after an acetaminophen (Tylenol) overdose with liver failure, acidosis, fetal death, and subsequent multiorgan failure, cardiac arrest, and death. Her APACHE score was 102. The second patient was a 30-year-old pregnant women (G10P6) with a history of chronic hypertension who was morbidly obese. She had an induction of labor at term for chronic hypertension and experienced superimposed preeclampsia during labor. On the day of the delivery, she had acute shortness of breath, hypoxia, and cardiorespiratory arrest. She was transferred to the medical ICU and was noted to have pulmonary hypertension on echocardiogram there. She required extensive cardioventilatory support and at day 12 was noted to have no electrophysiologic activity on electroencephalography and was removed from ventilatory support. Her APACHE score was 33. The third patient was a 17-year-old pregnant woman at 29 weeks of gestation with a twin gestation. She was admitted with placental abruption, intrauterine fetal death of both twins, and severe preeclampsia. Her APACHE score was 57. The patient experienced disseminated intravascular coagulation and received replacement with fresh frozen plasma, packed red blood cells, and cryoprecipitate during labor. She was delivered vaginally but then had a hypertensive crisis and was transferred to the medical ICU immediately after delivery. Several hours later, a computed tomography scan confirmed herniation of her brainstem; after electroencephalography,
Stevens et al she was removed from ventilatory support on day 2 of admission. The median APACHE III score was 34 (range, 14102); 11 subjects had APACHE scores of O50. An APACHE score was not correlated with maternal death. An APACHE score O50 was associated with increased length of ICU stay and hospital stay (P ! .05).
Comment Prognostic scoring systems have been used in intensive care settings for O20 years and have been used to compare actual and expected outcomes for comparable groups of patients, to describe heterogeneous populations in research trials, and to study the usefulness of ICUs.12-14 Because of evolving medical practices and populationspecific characteristics, prognostic scoring systems may become obsolete and may not be relevant in certain patient populations.12 The APACHE III prognostic scoring system was developed in 1991 with the use of physiologic data, age, chronic health, and disease components from a database of 17,440 unselected medical/surgical ICU admissions at 40 hospitals and was noted to have good predictive accuracy (receiver operated curve, 0.90) in that population.13 The results of our study suggest that APACHE III may not be predictive of maternal death after ICU admission. Possible reasons for this include the young age of our population (age !45 years is assigned of a score of 0 in APACHE III) and the lack of comorbid chronic underlying conditions (which may account for up to 23 points in APACHE III). However, the predicted mortality rate of an APACHE III score of 34 is 8% in other populations.13 The lack of statistical association between APACHE III and maternal death in our population may be due to the relatively small number of subjects in our study. Sample size determination with a significance level of .05 indicates that 1454 subjects would be required to show correlation between an APACHE score of O50 and a mortality rate with 80% power, given the low rate of death in this population. Several studies around the world have examined the usefulness of an earlier version of the ICU scoring system, APACHE II, in the prediction of maternal death after ICU admission.4-10 Although some researchers found that APACHE II accurately predicted maternal death,4,10 other researchers found that APACHE II overestimated maternal mortality rates5 or, conversely, that it underestimated maternal mortality rates because of obstetric problems.7 These differences may be due to higher mortality rates in some populations7,9 or to overutilization of ICUs in less acutely ill populations10 The only study in the United States to address this question used another scoring system, SAPS II. That study found that this system overestimated maternal mortality rates in all patients and specifically in patients who were
e15 admitted with obstetric indications but accurately predicted death in patients who were admitted for medical reasons.11 In a general obstetric population from military facilities in the United States, Novicoff et al15 developed a comprehensive risk-assessment model to predict maternal outcome. This model performed well in a population with low maternal mortality rates (0.01%; receiver operating characteristic curve, 0.75) but has not been tested in critically ill patients. It is possible that a scoring system that is derived from an obstetric population may have better usefulness in the prediction of maternal death. Our study is limited, as are all previous studies of prognostic scoring systems in obstetric populations, by its retrospective design. Future studies should attempt prospectively to compare predicted and observed maternal mortality rates with the use of previously validated ICU prognostic scoring tools.
References 1. Kilpatrick SJ, Matthay MA. Obstetric patients requiring critical care: a five-year review. Chest 1992;101:1407-12. 2. Panchal S, Arria AM, Harris AP. Intensive care utilization during hospital admission for delivery: prevalence, risk factors, and outcomes in a statewide population. Anesthesiology 2000;92:1537-44. 3. Zeeman GG, Wendel GD, Cunningham FG. A blueprint for obstetric critical care. Am J Obstet Gynecol 2003;188:532-6. 4. el-Solh AA, Grant BJ. A comparison of severity of illness scoring systems for critically ill obstetric patients. Chest 1996;110:1299-304. 5. Mahutte N, Murphy-Kaulbeck L, Le Q, Solomon V, Benjamin A, Boyd ME. Obstetric admissions to the intensive care unit. Obstet Gynecol 1999;94:263-6. 6. Cohen J, Singer P, Kogan A, Hod M, Bar J. Course and outcome of obstetric patients in a general intensive care unit. Acta Obstet Gynecol Scand 2000;79:846-50. 7. Karnad DR, Lapsia V, Krishnan A, Salvi VS. Prognostic factors in obstetric patients admitted to an Indian intensive care unit. Crit Care Med 2004;32:1294-9. 8. Heinonen S, Tyrvainen E, Saarikoski S, Ruokonen E. Need for maternal critical care in obstetrics: a population-based analysis. Int J Obstet Anesth 2002;11:260-4. 9. Demikiran O, Dikmen Y, Utku T, Urkmez S. Critically ill obstetric patients in the intensive care unit. Int J Obstet Anesth 2003;12:266-70. 10. Mirghani HM, Hamed M, Ezimokhai M, Weerasinghe DS. Pregnancy-related admissions to the intensive care unit: pregnancyrelated admissions to the intensive care unit. Int J Obstet Anesth 2004;13:82-5. 11. Gilbert TT, Smulian JC, Martin AA, Ananth CV, Scorza W, Scardella AT, et al. Obstetric admissions to the intensive care unit: outcomes and severity of illness. Obstet Gynecol 2003;102:897-903. 12. Herridge MS. Prognostication and intensive care unit outcome: the evolving role of scoring systems. Clin Chest Med 2003;24: 751-62. 13. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos BG, et al. The APACHE III prognostic system. Chest 1991; 100:1619-36. 14. Gunning K, Rowan K. Outcome data and scoring systems. BMJ 1999;319:241-4. 15. Novicoff WM, Wagner DP, Knaus WA, Kane EK, Cecere F, Draper E, et al. Initial development of a system-wide maternalfetal outcomes assessment program. Am J Obstet Gynecol 2000; 183:291-300.
www.ajog.org
Utility of Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) score in maternal admissions to the intensive care unit Tobey A. Stevens, MD,a Mary A. Carroll, MD,a Pamela A. Promecene, MD,a Marilyn Seibel, RN,b Manju Monga, MDa Department of Obstetrics, Gynecology and Reproductive Sciences, University of Texas Medical School Houstona and Memorial Hermann Hospital,b Houston, TX Received for publication September 7, 2005; revised January 6, 2006; accepted January 20, 2006
KEY WORDS Maternal death Intensive care unit APACHE III
Objective: A mean Acute Physiology, Age, and Chronic Health Evaluation (APACHE III) score of O50 is associated with increased intensive care unit mortality rate in nonpregnant cardiac and trauma patients. The objective was to determine the usefulness of the APACHE III score in maternal admissions to an intensive care unit in a tertiary care center in an urban multicultural city. Study design: This was a retrospective review of all maternal admissions (O20 weeks of gestation or after delivery) to an intensive care unit between January 2002 and May 2004. Demographics, obstetric and medical history, and 20 physiologic variables that comprise the APACHE III were recorded. The minimum APACHE III score (lowest risk of death) is 0; maximum is 299. The association between APACHE III score and maternal death was assessed with Mann Whitney U test. Significance was assumed at a probability value of !.05. Results: Fifty-eight subjects met the study criteria. Thirty percent of these women were admitted antepartum (27 G 1.0 weeks of gestation); 31% of the women were admitted on the day of delivery; and 29% of the women were admitted after delivery. Mean maternal age was 27 G 6.7 years. Acute conditions that resulted in transfer to the intensive care unit included preeclampsia (24%), cardiorespiratory disease (21%), hemorrhage (16%), infection (12%), trauma (7%), and thromboembolism (3%). Fifty-five percent of the women had no previous underlying obstetric complications, and 98% of the women had no underlying chronic health condition. Fifty-eight percent of the women received care in a medical intensive care unit; 28% of the women received care in a surgical intensive care unit; 10% of the women received care in a cardiac intensive care unit, and 3% of the women received care in a neurologic intensive care unit. The mean intensive care unit stay was 3.7 G 4.6 days, and the mean hospital stay was 9.0 G 7 days. Three patients died; the rest of the patients went home in good condition. The median APACHE III score was 34 (range, 14-102) and was not correlated with maternal death. Conclusion: The APACHE III is not associated with risk of intensive care unit–related maternal death. Ó 2006 Mosby, Inc. All rights reserved.
Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2006 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2006.01.073
e14 Several studies have addressed the reasons for admission or maternal outcome in women admitted to the intensive care unit (ICU) during pregnancy.1-3 Some studies have described the usefulness of ICU scoring systems in pregnant women,4-11 but only 1 study has been based in the United States.11 Gilbert et al11 recently summarized the use of the Simplified Acute Physiologic Score (SAPS II) in 233 antepartum and postpartum patients who were admitted to a medical ICU in New Jersey from 1991 to 1998. They found that SAPS II predicted hospital mortality rates in obstetric patients who were admitted for medical reasons, but not for indications that were related to pregnancy and delivery. SAPS II is just 1 of several prognostic scoring systems that are available to predict an individual’s risk of dying in the hospital. These models are often derived over several years and may become obsolete because they no longer reflect current practice.12 Recently, The Acute Physiologic and Chronic Health Evaluation (APACHE) tool, which was developed in 1978, has been revised (APACHE III). This scoring system records 27 parameters that are related to diagnosis and physiologic abnormalities on admission to the ICU, age, and pre-existing medical problems.13 Although this has been validated in numerous studies in nonpregnant patients, no previous study has reported the use of the APACHE III scoring system in maternal admissions to the ICU. The objective of this study was to determine the usefulness of the APACHE III score in maternal admissions to an ICU in a tertiary care center.
Methods This was a retrospective review of all maternal admissions to ICU between January 2002 and May 2004 at Memorial Hermann Hospital, a tertiary care center in Houston, Texas. Any chart that was coded for a gynecologic or obstetric diagnosis related group and intensive care admission was pulled. All women who were admitted during the study period who were R20 weeks of gestation or !6 weeks after delivery were included. Demographics, obstetric and medical history, and 20 physiologic variables that comprise the APACHE III were recorded. The minimum APACHE score (lowest risk of death) is 0; the maximum score is 299. A mean APACHE score of O50 is associated with increased ICU death in nonpregnant cardiac and trauma patients.13 The APACHE score was derived from nonpregnant patients and therefore does not include specific pregnancy-related diseases. The association between the APACHE III score and maternal death was assessed with the Mann Whitney U and Fisher exact tests. Significance was assumed at a probability value of !.05.
Stevens et al
Results Fifty-eight subjects met the study criteria. Thirty percent of the women were admitted antepartum (27 G 1.0 weeks); 31% of the women were admitted on the day of delivery; 29% of the women were admitted after delivery. The mean maternal age was 27 G 6.7 years. Acute conditions that resulted in transfer to the ICU included complications of preeclampsia (24%), cardiorespiratory disease (21%), hemorrhage (16%), infection (12%), trauma (7%), and thromboembolism (3%). Other reasons for admission included drug overdose, diabetic ketoacidosis, myasthenic crisis, anaphylaxis, sublingual hematoma, and altered mental status. Ninety-eight percent of the women had no underlying chronic health condition. The only patient with an underlying chronic health condition was a 33-year-old woman at 22 weeks of gestation with acquired immunodeficiency syndrome and presumed pneumocystis carinii pneumonia. Fifty-eight percent of the women received care in a medical ICU; 28% of the women received care in a surgical ICU; 10% of the women received care in a cardiac ICU, and 3% of the women received care in a neurologic ICU. The mean ICU stay was 3.7 G 4.6 days, and the mean hospital stay was 9.0 G 7 days. Three patients died; the rest of the patients went home in good condition. The first patient who died was a 23-year-old at 30 weeks of gestation who had been transferred to the medical ICU after an acetaminophen (Tylenol) overdose with liver failure, acidosis, fetal death, and subsequent multiorgan failure, cardiac arrest, and death. Her APACHE score was 102. The second patient was a 30-year-old pregnant women (G10P6) with a history of chronic hypertension who was morbidly obese. She had an induction of labor at term for chronic hypertension and experienced superimposed preeclampsia during labor. On the day of the delivery, she had acute shortness of breath, hypoxia, and cardiorespiratory arrest. She was transferred to the medical ICU and was noted to have pulmonary hypertension on echocardiogram there. She required extensive cardioventilatory support and at day 12 was noted to have no electrophysiologic activity on electroencephalography and was removed from ventilatory support. Her APACHE score was 33. The third patient was a 17-year-old pregnant woman at 29 weeks of gestation with a twin gestation. She was admitted with placental abruption, intrauterine fetal death of both twins, and severe preeclampsia. Her APACHE score was 57. The patient experienced disseminated intravascular coagulation and received replacement with fresh frozen plasma, packed red blood cells, and cryoprecipitate during labor. She was delivered vaginally but then had a hypertensive crisis and was transferred to the medical ICU immediately after delivery. Several hours later, a computed tomography scan confirmed herniation of her brainstem; after electroencephalography,
Stevens et al she was removed from ventilatory support on day 2 of admission. The median APACHE III score was 34 (range, 14102); 11 subjects had APACHE scores of O50. An APACHE score was not correlated with maternal death. An APACHE score O50 was associated with increased length of ICU stay and hospital stay (P ! .05).
Comment Prognostic scoring systems have been used in intensive care settings for O20 years and have been used to compare actual and expected outcomes for comparable groups of patients, to describe heterogeneous populations in research trials, and to study the usefulness of ICUs.12-14 Because of evolving medical practices and populationspecific characteristics, prognostic scoring systems may become obsolete and may not be relevant in certain patient populations.12 The APACHE III prognostic scoring system was developed in 1991 with the use of physiologic data, age, chronic health, and disease components from a database of 17,440 unselected medical/surgical ICU admissions at 40 hospitals and was noted to have good predictive accuracy (receiver operated curve, 0.90) in that population.13 The results of our study suggest that APACHE III may not be predictive of maternal death after ICU admission. Possible reasons for this include the young age of our population (age !45 years is assigned of a score of 0 in APACHE III) and the lack of comorbid chronic underlying conditions (which may account for up to 23 points in APACHE III). However, the predicted mortality rate of an APACHE III score of 34 is 8% in other populations.13 The lack of statistical association between APACHE III and maternal death in our population may be due to the relatively small number of subjects in our study. Sample size determination with a significance level of .05 indicates that 1454 subjects would be required to show correlation between an APACHE score of O50 and a mortality rate with 80% power, given the low rate of death in this population. Several studies around the world have examined the usefulness of an earlier version of the ICU scoring system, APACHE II, in the prediction of maternal death after ICU admission.4-10 Although some researchers found that APACHE II accurately predicted maternal death,4,10 other researchers found that APACHE II overestimated maternal mortality rates5 or, conversely, that it underestimated maternal mortality rates because of obstetric problems.7 These differences may be due to higher mortality rates in some populations7,9 or to overutilization of ICUs in less acutely ill populations10 The only study in the United States to address this question used another scoring system, SAPS II. That study found that this system overestimated maternal mortality rates in all patients and specifically in patients who were
e15 admitted with obstetric indications but accurately predicted death in patients who were admitted for medical reasons.11 In a general obstetric population from military facilities in the United States, Novicoff et al15 developed a comprehensive risk-assessment model to predict maternal outcome. This model performed well in a population with low maternal mortality rates (0.01%; receiver operating characteristic curve, 0.75) but has not been tested in critically ill patients. It is possible that a scoring system that is derived from an obstetric population may have better usefulness in the prediction of maternal death. Our study is limited, as are all previous studies of prognostic scoring systems in obstetric populations, by its retrospective design. Future studies should attempt prospectively to compare predicted and observed maternal mortality rates with the use of previously validated ICU prognostic scoring tools.
References 1. Kilpatrick SJ, Matthay MA. Obstetric patients requiring critical care: a five-year review. Chest 1992;101:1407-12. 2. Panchal S, Arria AM, Harris AP. Intensive care utilization during hospital admission for delivery: prevalence, risk factors, and outcomes in a statewide population. Anesthesiology 2000;92:1537-44. 3. Zeeman GG, Wendel GD, Cunningham FG. A blueprint for obstetric critical care. Am J Obstet Gynecol 2003;188:532-6. 4. el-Solh AA, Grant BJ. A comparison of severity of illness scoring systems for critically ill obstetric patients. Chest 1996;110:1299-304. 5. Mahutte N, Murphy-Kaulbeck L, Le Q, Solomon V, Benjamin A, Boyd ME. Obstetric admissions to the intensive care unit. Obstet Gynecol 1999;94:263-6. 6. Cohen J, Singer P, Kogan A, Hod M, Bar J. Course and outcome of obstetric patients in a general intensive care unit. Acta Obstet Gynecol Scand 2000;79:846-50. 7. Karnad DR, Lapsia V, Krishnan A, Salvi VS. Prognostic factors in obstetric patients admitted to an Indian intensive care unit. Crit Care Med 2004;32:1294-9. 8. Heinonen S, Tyrvainen E, Saarikoski S, Ruokonen E. Need for maternal critical care in obstetrics: a population-based analysis. Int J Obstet Anesth 2002;11:260-4. 9. Demikiran O, Dikmen Y, Utku T, Urkmez S. Critically ill obstetric patients in the intensive care unit. Int J Obstet Anesth 2003;12:266-70. 10. Mirghani HM, Hamed M, Ezimokhai M, Weerasinghe DS. Pregnancy-related admissions to the intensive care unit: pregnancyrelated admissions to the intensive care unit. Int J Obstet Anesth 2004;13:82-5. 11. Gilbert TT, Smulian JC, Martin AA, Ananth CV, Scorza W, Scardella AT, et al. Obstetric admissions to the intensive care unit: outcomes and severity of illness. Obstet Gynecol 2003;102:897-903. 12. Herridge MS. Prognostication and intensive care unit outcome: the evolving role of scoring systems. Clin Chest Med 2003;24: 751-62. 13. Knaus WA, Wagner DP, Draper EA, Zimmerman JE, Bergner M, Bastos BG, et al. The APACHE III prognostic system. Chest 1991; 100:1619-36. 14. Gunning K, Rowan K. Outcome data and scoring systems. BMJ 1999;319:241-4. 15. Novicoff WM, Wagner DP, Knaus WA, Kane EK, Cecere F, Draper E, et al. Initial development of a system-wide maternalfetal outcomes assessment program. Am J Obstet Gynecol 2000; 183:291-300.