The obstetrics comorbidity index as a predictive tool for risk-appropriate maternal care Authors: Allison R McCarter, Regan N Theiler, Megan E Branda, Rebecca M Smith, Emily E Sharpe, Vanessa E Torbenson Published: 2024 ABSTRACT AbstractBackgroundThe aim of this study was to assess the correlation between obstetric co-morbidity index (OBCMI) and severe maternal morbidity (SMM) in antepartum obstetrics transfers. By utilizing a population of transfers to a level IV maternal care facility, we hope to demonstrate validity for the use of OBCMI in the triage of transfer to provide risk-appropriate maternal care.MethodsAntepartum obstetrics transfers to a single level IV maternal care facility from 1/1/2016 to 12/31/2020 that resulted in delivery during the same encounter were included in this retrospective study. The components of the OBCMI score were retrospectively collected by manual chart review of transfer and admission notes in the electronic medical record (EMR). SMM was determined via ICD-10 and CPT code extraction from time of transfer through six weeks postpartum and confirmed by the same reviewer. Mode of delivery, length of stay, ICU admission, readmission and reoperation were obtained via institutional databases and manual EMR review.ResultsAmong 561 transfers meeting the inclusion criteria, the median OBCMI was significantly higher for transfers with a maternal-only indication (n = 232) compared to fetal-only( n = 282) (median [IQR], 6 [4–8], 5 [4–6], and 1 [0–2] for maternal-only, maternal–fetal combined (n = 47), and fetal-only; p < 0.001). The prevalence of SMM was 16.8% (39/232), 27.7% (13/47), 2.1% (6/282), p < 0.0001 for those transferred for maternal, fetal and maternal, and fetal only indications respectively. The median (IQR) OBCMI score was 5 (4–8) and 3 (1–5) for those with versus without SMM. A cut-off OBCMI score of ≥ 4 was identified with 81% sensitivity (95% CI 68.6–90.1%) in predicting SMM (P = < .0001) and was noted to be significantly associated with operative delivery, blood transfusion, ICU admission, prolonged hospitalization, and reoperation. Using a cut-off OBCMI score of ≥ 4 on the population transferred for maternal and maternal–fetal combined indications only (279) yielded a specificity of 90.4% and sensitivity of 23.8% (p = 0.024).ConclusionOBCMI was demonstrated to discriminate for SMM in a population of obstetrics transfers to a Level IV maternal care facility. When stratifying for maternal indicated transfers, the ability of OBCMI as a predictive tool decreased. The obtained cutoff OBCMI value of ≥ 4 had high specificity but may miss a significant population that would benefit from transfer. Use of the OBCMI may be too crude of a measure to provide a comprehensive risk assessment to predict SMM and adverse obstetrics outcomes. Further studies that may include newer tools such as machine learning may be necessary to develop a more clinically useful tool.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12884-024-06992-0.Keywords: Severe maternal morbidity, Obstetrics comorbidity index, Risk assessment, Hemorrhage, Maternal death Background In 2015, the American College of Obstetricians and Gynecologists introduced Levels of Maternal Care (LoMC) to promote risk-appropriate obstetrics care, with the goal of reducing maternal morbidity and mortality [1]. Four levels of maternal care exist, ranging from basic care (level I), specialty care (level II), subspecialty care (level III), and regional perinatal health care centers (level IV). The leading causes of maternal mortality include preexisting conditions and obstetrics complications, such as postpartum hemorrhage or hypertensive disorders [2]. Studies suggest up to 60% of pregnancy-related deaths during 2013–2017 were potentially preventable [3]. Thus, those with near-miss mortality or severe maternal morbidity (SMM) events must be considered in the discussion for how to best reduce maternal mortality. The incidence of SMM increased by nearly 200% between 1993 and 2014 [4, 5], driven primarily by blood transfusion, though a 20% increased rate of SMM is still demonstrated when blood transfusion is excluded. Data suggests that obstetrics complications are more common at hospitals with low delivery volumes [6], and certain high-risk conditions have been demonstrated to have improved outcomes at larger institutions [7]. Obstetricians are tasked with identifying patients at greatest risk of maternal morbidity and mortality and initiating transfer to a center that meets a patient’s unique needs. Certain conditions clearly necessitate transfer to a level IV facility, such as complex maternal cardiac disease and placenta accreta spectrum [8] though patients with a constellation of lower risk conditions may also benefit from higher LOMC facilities to reduce maternal morbidity and mortality. As such, it is necessary to develop tools to risk stratify patients in a standardized fashion to deliver risk-appropriate care. The obstetric comorbidity index (OBCMI) is a tool that attempts to measure obstetrics risk, as determined by preexisting or obstetrics conditions [9, 10]. Patients are assigned points for their respective maternal and obstetrics comorbidities, where points are summed to generate a single numeric value. The OBCMI was originally designed as a tool for risk-adjustment in large database driven studies and has been historically used as a research tool to describe the acuity of a given study population [11]. Its use has been validated in retrospective cohort studies in distinct populations [10, 12] with a demonstrated relationship between increasing OBCMI and incidence of SMM in a prospective study [13]. The aim of this study was to evaluate the association between elevated OBCMI and incidence of SMM within a population of patients referred to a level IV center in a rural health system. We specifically examined this relationship within a transfer population to a level IV facility, with the goal of identifying an OBCMI cutoff that may be of utility in providing risk-appropriate care by revealing those at greatest risk of SMM. Methods The study population of this retrospective cohort study consisted of all antenatal obstetrics transfers to a single, level IV maternal care center from 1/1/2016 to 12/31/2020 that resulted in delivery during the same encounter. The level IV center is located in a small city surrounded by rural communities that support lower levels of obstetric care. Patients who had denied access to their medical records for research purposes were not included. The study was deemed exempt by the Mayo Clinic Institutional Review Board on 4/21/2021 and patient consent to participate was waived per the IRB. Patient records were accessed from 8/5/2021 through 6/22/2022 if the patient was eligible per Minnesota research authorization. The authors had access to protected health information and patient identity during the period of data collection. Antepartum transfers to Mayo Clinic Rochester from both internal and external facilities were identified by query of the centralized Mayo Clinic Admission and Transfer Center records. All patients who were transferred to the primary Obstetrics service or to an Intensive Care Unit (ICU) for co-management with Obstetrics were included. The list of transferred patients was then cross-referenced with an internal obstetrics delivery data mart. The electronic medical record of each patient who was transferred and had a delivery was then reviewed by the primary author to confirm that the delivery occurred during the same encounter as the transfer. Postpartum patients and those discharged prior to delivery at the level IV facility were excluded. The following patient and delivery characteristics were extracted from the obstetrics delivery data mart that is populated by electronic retrievals from the electronic medical record (EMR): age and date at delivery, prenatal body mass index, parity and gravidity prior to delivery, mode of delivery, birth outcome (living status), multiple gestation, gestational age, birth weight, Apgar scores, and neonatal intensive care (NICU) admission. Race and ethnicity at time of patient registration were obtained from an institutional database. Upon abstraction from the database, data checks were conducted via descriptive and visual statistics where values outside of expected range were chart reviewed and validated. The primary indication for transfer was classified as maternal-only, fetal-only or combined maternal and fetal (Supplemental Table 1). The components of the OBCMI were retrospectively collected by manual chart review of transfer and admission notes in the EMR. Using the modified points system for each component reported by Easter et al. [13], points were summed to generate a single numeric OBCMI score for each transfer. The EMR of 30 patients were independently reviewed by two obstetrics providers to collect the components of the OBCMI. Items of discrepancy between reviewers were reviewed and a consensus was determined. The two reviewers then independently reviewed an additional 50 patients. The primary reviewer then manually reviewed the EMR of the remaining patients in the study cohort to collect the OBCMI components. Severe maternal morbidity from time of antenatal transfer through six weeks postpartum was identified using the 21 indicators and corresponding 10th Revision of the International Classification of Diseases (ICD-10) diagnosis and procedure codes determined by the Centers for Disease Control and Prevention [14]. Instead of using the recommended ICD-10 codes for blood product transfusions we utilized CPT code 36,430, as ICD-10 code extraction erroneously failed to identify any instances of blood transfusion. The presence of SMM was defined as the presence of at least one of the 21 indicators. The primary reviewer then reviewed and confirmed all identified SMM cases. In addition, the primary reviewer manually reviewed the EMR of 13 patients with a OBCMI score > 10 who did not have any SMM indicators and confirmed the absence of SMM. Lastly, additional adverse obstetrics outcomes including chorioamnionitis, endometritis, postpartum hemorrhage, maternal length of hospitalization, ICU admission, readmission and reoperation were identified using ICD-10 codes. The occurrence of chorioamnionitis and endometritis were identified using ICD-10 codes. Reoperation and blood transfusion were obtained using CPT codes. Maternal length of stay, ICU admission and readmission within the health system were electronically retrieved from the EMR using codes and internal databases and confirmed by primary reviewer. Data management and statistical analysis was performed using SAS version 9.4 software package (SAS Institute, Inc, Cary, NC). Variables were descriptively summarized using standard descriptive statistics. The distribution of the OBCMI score was compared between the three primary indications for transfer using the Kruskal–Wallis test. Likewise, the distribution of the OBCMI score was compared between those with versus without SMM using the Wilcoxon rank sum test. A logistic regression model was fit to estimate the odds of SMM per a one-point increase in the OBCMI score and to estimate the c-statistic. The c-statistic is a measure of how well the OBCMI score discriminates between patients with versus without SMM and its interpretation is analogous to the area under a receiver operating characteristic curve. A cutoff for the OBCMI index was determined from the cutoff maximizing Youden’s J statistic, defined as sensitivity + specificity -1. Using this cutoff, we then divided the cohort into two groups representing those below or at/above cutoff. Obstetrics outcomes were compared between the two groups using the chi-square test for categorical outcomes and the Wilcoxon rank test for gestational age and length of stay. All calculated p-values were two-sided and p-values less than 0.05 were considered statistically significant. Results The study included 558 women with 561 deliveries following transfer to a level IV maternal care center during the 4-year study period. The characteristics of the study cohort are presented in Table 1. Table 1Summary of maternal, delivery, and infant characteristicsCharacteristicValue (N = 561)Maternal Age at delivery (years), Mean (SD)29.4 (5.7)Race White449 (80.0%) Black or African American38 (6.8%) Asian22 (3.9%) American Indian/Alaskan Native4 (0.7%) Native Hawaiian/Other Pacific Islander3 (0.5%) Not Disclosed45 (8.0%)Ethnicity Hispanic or Latino63 (11.2%) Not Hispanic or Latino480 (85.6%) Not Disclosed18 (3.2%)Prenatal BMI (kg/m2) < 25141 (25.1%) 25–30154 (27.4%) 30–35107 (19.1%) 35–4069 (12.3%) 40 + 75 (13.4%) Not documented15 (2.7%)Parity at time of transfer 047 (8.4%) 1248 (44.2%) 2130 (23.2%) 363 (11.2%) 4 + 73 (13.0%)Gravidity at time of transfer 018 (3.2%) 1206 (36.7%) 2118 (21.0%) 381 (14.4%) 4 + 138 (24.6%)Delivery and postpartum Mode of Delivery Cesarean 281 (50.1%) Spontaneous vaginal260 (46.3%) Operative vaginal20 (3.6%)3rd or 4th degree laceration2 (0.4%)Chorioamnionitis50 (8.9%)Endometritis4 (0.7%)Maternal length of stay (days), Median (IQR)4 (3, 6)Readmission within 6 weeks after delivery21 (3.7%)ICU admit within 6 weeks after delivery22 (3.9%)Multiple gestation40 (7.2%)Gestational age (weeks), Median (IQR)34.0 (31.0, 35.0)Birth weight (grams), Median (IQR)a2200 (1610, 2760)Apgar 1 min, Median (IQR)a7 (5, 8)Apgar 5 min, Median (IQR)a9 (7, 9)Infant living status Singleton: Living502 (89.5%) Singleton: Fetal demise9 (1.6%) Singleton: Neonatal death10 (1.8%) Twins: Both living36 (6.4%) Twins: 1 living and 1 fetal demise1 (0.2%) Twins: 1 living and 1 neonatal death1 (0.2%) Twins: 1 fetal demise and 1 neonatal death1 (0.2%) Triplets: All 3 living1 (0.2%)NICU Admission Singleton: Not admitted200 (35.7%) Singleton: Admitted321 (57.2%) Twins: Neither admitted6 (1.1%) Twins: One admitted3 (0.5%) Twins: Both admitted30 (5.3%) Triplets: All 3 admitted1 (0.2%)Abbreviations: BMI body mass index, IQR interquartile range, ICU intensive care unit, NICU neonatal intensive care unitaFor deliveries involving multiple gestation, the reported birthweight is the highest among the multiples and the reported APGAR score is the average among the multiples Summary of maternal, delivery, and infant characteristics Abbreviations: BMI body mass index, IQR interquartile range, ICU intensive care unit, NICU neonatal intensive care unit aFor deliveries involving multiple gestation, the reported birthweight is the highest among the multiples and the reported APGAR score is the average among the multiples Among our transfer population, the median OBCMI score was 3 (interquartile range (IQR) 1–6) and ranged from 0–20. The prevalence of individual OBCMI components are presented in Table 2. Table 2OBCMI componentsOBCMI componentN (% of 561)PointsPreeclampsia with severe features or eclampsia144 (25.7%)5Preeclampsia without severe features/gestational/chronic hypertension107 (19.1%)2Congestive Heart Failure4 (0.7%)5Pulmonary hypertension0 (0.0%)4Ischemic heart disease or cardiac arrhythmia6 (1.1%)3Congenital heart and/or valvular disease16 (2.9%)4Multiple gestation40 (7.1%)2Fetal death11 (2.0%)2Placenta previa or suspected accreta or abruption45 (8.0%)4Previous cesarean delivery or myomectomy118 (21.0%)1Autoimmune disease or lupus17 (3.0%)2HIV or AIDS1 (0.2%)2Sickle cell disease/bleeding disorder/coagulopathy/anticoagulation15 (2.7%)3Epilepsy/cerebrovascular accident/neuromuscular disorder4 (0.7%)2Chronic renal disease5 (0.9%)1Asthma57 (10.2%)1Diabetes mellitus with insulin therapy61 (10.9%)1Maternal age, years > 440340–4415 (2.7%)235–3979 (14.1%)1Substance use disorder36 (6.4%)2Alcohol abuse5 (0.9%)1BMI, kg/m2BMI > 5034 (6.1%)3BMI 40–5082 (14.6%)2Abbreviations: OBCMI obstetric co-morbidity index, HIV human immunodeficiency virus, AIDS acquired immunodeficiency syndrome, BMI body mass index OBCMI components Abbreviations: OBCMI obstetric co-morbidity index, HIV human immunodeficiency virus, AIDS acquired immunodeficiency syndrome, BMI body mass index The primary indication was fetal-only for 282 (50.3%) transfers, maternal-only for 232 (41.4%) transfers and both maternal and fetal for 47 (8.4%) transfers. The median OBCMI was significantly higher for transfers with a maternal indication compared to fetal-only (median [IQR], 6 [4-8], 5 [4-6], and 1 [0–2] for maternal only, maternal–fetal combined, and fetal-only; p < 0.001). Additionally, the prevalence of SMM was 16.8% (39/232), 27.7% (13/47) and 2.1% (6/282),, p < 0.0001 for those transferred for maternal, fetal and maternal, and fetal only indications respectively Table 3. Table 3Summary of the prevalence of specific severe maternal morbidity during the initial hospitalization and up to 6 weeks following deliverySMM by Transfer IndicationBOTH (N = 47)FETAL (N = 282)MATERNAL (N = 232)Total (N = 561)p valueSMM < 0.00011 No34 (72.3%)276 (97.9%)193 (83.2%)503 (89.7%) Yes13 (27.7%)6 (2.1%)39 (16.8%)58 (10.3%)(report generated on 13OCT2023)1Chi-Square Summary of the prevalence of specific severe maternal morbidity during the initial hospitalization and up to 6 weeks following delivery (report generated on 13OCT2023) 1Chi-Square A total of 58 patients (10.3%) met criteria for one or more SMM indicators from time of transfer through 6 weeks postpartum, with a majority occurring during the transfer/delivery hospitalization (52 of 58). As summarized in Table 4, the most common SMM events in our population were acute renal failure and blood transfusion. Table 4Summary of the prevalence of specific severe maternal morbidity during the initial hospitalization and up to 6 weeks following deliveryIndicatorN (% of 561)Acute Myocardial Infarction0Aneurysm0Acute Renal Failure19 (3.4%)Respiratory9 (1.6%)Amniotic Fluid Embolism0Cardiac Arrest0Disseminated Intravascular Coagulation5 (0.9%)Eclampsia2 (0.4%)Heart Failure/arrest during surgery0Puerperal Cerebrovascular Disorders3 (0.5%)Pulmonary edema/Acute Heart Failure7 (1.2%)Severe Anesthesia Complications0Sepsis8 (1.4%)Shock6 (1.1%)Sickle Cell Disease with Crisis0Air Embolism5 (0.9%)Blood Transfusion34 (6.1%) Summary of the prevalence of specific severe maternal morbidity during the initial hospitalization and up to 6 weeks following delivery Figure 1 compares the distribution of OBCMI scores between patients with and without SMM events. The median OBCMI score for those with SMM was 5 (IQR 4–8) compared to 3 (IQR 1–5) in those without SMM (p < 0.001). The c-statistic or area under the receive operative characteristic curve (AUC) for OBCMI as a predictor of SMM was 0.73 (95% CI 0.67–0.79). For every one-point increase in OBCMI, patients experienced at 1.24 increase in odds of SMM (95% CI 1.15 – 1.34). As depicted in Fig. 2, the proportion of patients experiencing a SMM event increased from 0.8% (1/123) for those with an OBCMI score of 0 to 25% (24/96) for those with OBCMI ≥ 9. Distribution of the OBCMI scores according to the presence of SMM(Red) or absence of SMM (Blue) Rate of SMM according to OBCMI score Among the 58 patients with SMM, 47 had an OBCMI of ≥ 4. Using a cut-off OBCMI of ≥ 4 yields a sensitivity of 81.0% (47/58, 95% CI 68.6.3 – 90.1%) and specificity of 57.6% (290/503, 95% CI 53.2 – 62.0%) with a p-value of < 0.0001. Robust obstetrics outcomes were then evaluated for those with an OBCMI above and below a cutoff of 4. The proportion of transfers with a cesarean delivery, blood transfusion, ICU admission, and reoperation was significantly higher in those with an OBCMI score ≥ 4 compared to < 4 (Table 5). In addition, the median length of stay was significantly longer for those with an OBCMI score ≥ 4. Table 5Comparison of Adverse Obstetrics Outcomes above versus below OBCMI cutoffOutcomeOBCMIP- value† < 4 (N = 301) ≥ 4 (N = 260)SMM11 (3.7%)47 (18.1%)< .0001Delivery and postpartum Mode of delivery< 0.001 Cesarean112 (37.2%)169 (65.0%) Spontaneous vaginal176 (58.5%)84 (32.3%) Operative vaginal13 (4.3%)7 (2.7%)Chorioamnionitis37 (12.3%)13 (5.0%)0.003Endometritis3 (1.0%)1 (0.4%)0.39Maternal length of Stay (days), Median (IQR)3.0 (2.0, 5.0)4.0 (3.0, 7.0)< 0.001Readmission within 6 weeks of delivery8 (2.7%)13 (5.0%)0.15ICU admission within 6 weeks of delivery6 (2.0%)16 (6.2%)0.01Reoperationa2 (0.7%)11 (4.2%)0.005Blood transfusion8 (2.7%)26 (10.0%)< 0.001Infant Gestational Age at Delivery (weeks), Median (IQR)34.0 (31.9, 35.9)34.0 (31.4, 35.6)0.11Abbreviations: SMM severe maternal, IQR Interquartile range†Comparisons between the two groups were evaluated using the chi-square test for categorical outcomes and the Wilcoxon rank test for gestational age and length of stayaReoperation included dilation and curettage, exploratory laparotomy and hysterectomy Comparison of Adverse Obstetrics Outcomes above versus below OBCMI cutoff Abbreviations: SMM severe maternal, IQR Interquartile range †Comparisons between the two groups were evaluated using the chi-square test for categorical outcomes and the Wilcoxon rank test for gestational age and length of stay aReoperation included dilation and curettage, exploratory laparotomy and hysterectomy Discussion We assessed OBCMI as a predictive tool for SMM in obstetrics transfers to a level IV maternal care center in a rural population. After distinguishing indication for transfer as maternal-only, fetal-only or both maternal and fetal, we affirmed the anticipated finding of an elevated OBCMI score in those requiring transfer for maternal or both maternal and fetal indications relative to fetal-only transfers as well as the resultant SMM in these cohorts. This serves as an internal control as only two components of OBCMI are related to the fetus. We redemonstrated the strong association between elevated OBCMI and an increased incidence of SMM as well as good discriminatory ability (c-statistic 0.73), consistent with previous studies [13, 15]. One study using Medicaid patients was able to confirm the likelihood of SMM with increasing OBCMI but was unable to specify a meaningful cut off [16]. Our study took this finding a step further by proposing an OBCMI cutoff score of ≥ 4, which yielded a sensitivity of 81.0% and specificity of 57.6% in predicting SMM in our population. This cutoff was identified to maximize Youden’s J statistic, while favoring sensitivity given the significance of an SMM event. Robust obstetrics outcomes were compared for those above and below the proposed OBCMI cutoff, where increased risk of operative delivery, prolonged hospitalization, ICU admission and reoperation were noted with an OBCMI score of ≥ 4. While we did not evaluate the SMM of those patients with similar OBCMI that were not transferred but remained in their initial hospital setting, the components of SMM that were experienced such as acute renal failure and hemorrhage are diagnoses that a larger institution has the resources to manage compared to small rural hospitals. Using an OBCMI cutoff score of ≥ 4, for the 279 patients transferred for maternal or maternal–fetal combined indications, the specificity of 90.4% and sensitivity of 23.8% (p = 0.024) indicate that in clinical practice this threshold may offer the high specificity desired for a tool used to make maternal transfer recommendations. Given the significance of SMM events, it is reasonable to consider that transfer to a level IV maternal care center may be of value given their ready access to subspecialty services, invasive diagnostic/therapeutics, availability of blood components and intensive care. Maximal clinical benefit of the OBCMI may be found in optimizing the specificity to determine which patients will clearly benefit form a higher level of care. Further studies are necessary to reproduce and improve upon our findings in a larger, more diverse sample and across multiple health systems to assess generalizability outside of our single level IV maternal care center. In addition, the OBCMI cutoff proposed herein remains unvalidated, where our findings simply suggest a strong association between OBCMI and SMM as well as adverse obstetrics outcomes. Future studies will be aimed at validation of the OBCMI cutoff with a true control population in a prospective fashion as patients are triaged for transfer to a higher level of maternal care. Lastly newer techniques such as machine learning may offer the opportunity to improve upon the OBCMI as a predictive tool. Implications As found in our study, the majority of transfers to higher level centers are due to fetal indications rather than maternal. Clear definitions have existed for levels of neonatal care since 2004 and there is data to show improved outcomes for infants born in the correct level facility in the intervening years [17]. The concept of levels of maternal care is newer and not yet substantiated with similar volumes of research. While the aim of transfer is to improve outcomes, it often separates pregnant persons from their support structures which can be a significant emotional and financial burden. Additionally, unnecessary transfer of patients away from a local hospital, results in loss of revenue for the originating facility which has implications for rural centers which have lower volumes that make it difficult to support fixed operating costs [18]. It is imperative therefore to identify those patients most likely to benefit from a maternal transfer and further validate OBCMI or other tools as a decision aid for clinicians. Strengths and limitations Strengths of our study include robust internal databases to capture all antenatal transfers across four years’ time, and reliable access to information at originating and transfer facility given common EMR and manual chart review by a single, validated obstetrics provider. Limitations include patient selection of those transferred to our single level IV maternal care facility, consisting of a relatively homogenous and non-Hispanic White population. Inability to retrospectively calculate SMM and OBCMI of all patients that remained in the primary hospital setting to evaluate if their outcome was similar. However, given that patients who were transferred for fetal only indication had a lower OBCMI and lower SMM serves as an internal control of sorts highlighting the ability of OBCMI to discriminate those at risk of SMM. Patients had to be stable enough to transfer in the antepartum period, where postpartum transfers were excluded from the study. Data acquisition included retrospective chart review of the EMR, along with ICD-10 and CPT code extraction and electronic retrieval from institutional databases. As such, our study is limited by the integrity of documentation and coding between the originating and transfer facilities. Conclusions The OBCMI is a research and clinical tool aimed to quantify maternal risk in a single numeric value. Our study demonstrates the possible value of OBCMI to distinguish for SMM in a novel patient population of obstetrics transfers. A cut off OBCMI score of ≥ 4 is proposed and was further demonstrated to be associated with increased risk of operative delivery, blood transfusion, ICU admission, prolonged hospitalization, and reoperation. After further validation with other transfer populations, our results may be applied clinically to identify those at greatest risk of SMM and adverse obstetrics outcomes to aid in the triage of obstetrics patients to a level IV maternal care facility. Supplementary Information Supplementary Material 1. Acknowledgements We would like to acknowledge Amy L Weaver MS who was a significant contributor to this work before she retired. Authors’ contributions AM, RT and VET conceived the study. AM, RT, MB, RS, ES and VET designed the study. AM and VET gathered the data. MB performed the statistical calculations. AM, RT, MB, RS, ES and VET analyzed the data. AM drafted the manuscript. All authors have revised and approved the submitted version and have agreed to be responsible for its contents. Funding No funding support. Data availability Primary dataset generated during the current study are available on Zenodo and can be accessed by requestor on reasonable request to Vanessa Torbenson at Torbenson.vanessa@mayo.edu DOI https://zenodo.org/doi/10.5281/zenodo.10525171. Declarations Ethics approval and consent to participate The study was deemed exempt on 4/21/2021 by the Mayo Clinic Institutional Review Board in accordance with the Declaration of Helsinki and patient consent to participate was waived. Charts were accessed if the patient was eligible per Minnesota research authorization. Consent for publication NA. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1. Data Availability Statement Primary dataset generated during the current study are available on Zenodo and can be accessed by requestor on reasonable request to Vanessa Torbenson at Torbenson.vanessa@mayo.edu DOI https://zenodo.org/doi/10.5281/zenodo.10525171.