Maternal Health after Stillbirth: Postpartum Hospital Readmission in California

Abstract

Objective

The aim of this study is to assess whether the risk of postpartum readmission within 6 weeks of giving birth differs for women who had stillbirths compared with live births.

Study Design

Using data from the Office of Statewide Health Planning and Development in California, we performed a population-based cohort study of 7,398,640 births between 1999 and 2011. We identified diagnoses and procedures associated with the first postpartum hospital readmission that occurred within 6 weeks after giving birth. We used log-binomial models to estimate relative risk (RR) of postpartum readmission for women who had stillbirth compared with live birth deliveries, adjusting for maternal demographic, prepregnancy, pregnancy, and delivery characteristics.

Results

The rate of postpartum readmission was higher among women who had stillbirths compared with women who had live births (206 and 96 per 10,000 births, respectively). After adjusting for maternal demographic and medical characteristics, the risk of postpartum readmission for women who had stillbirths was nearly 1.5 times greater (adjusted RR = 1.47, 95% confidence interval: 1.35–1.60) compared with live births. Among women with stillbirths, the most common indications at readmission were uterine infection or pelvic inflammatory disease, psychiatric conditions, hypertensive disorder, and urinary tract infection.

Conclusion

Based on our findings, women who have stillbirths are at higher risk of postpartum readmissions within 6 weeks of giving birth than women who have live births. Women who have stillbirths may benefit from additional monitoring and counseling after hospital discharge for potential postpartum medical and psychiatric complications.

Stillbirths, which occur in 6 per 1,000 births in the United States, are associated with major maternal morbidity.¹ In a recent cohort study of delivery hospitalizations in California, women who had a stillbirth had a more than fourfold higher risk of severe maternal morbidity than women who had a live birth.² However, maternal health after stillbirth is an understudied and underappreciated public health issue. After stillbirth delivery, women are at increased risk of mental health problems such as depression and anxiety.^3,4 Beyond postpartum mental health morbidity, there is a dearth of information about other major medical and obstetric-related postpartum complications.

Women who have a stillbirth are more likely to experience adverse health outcomes at delivery, including hemorrhage, infection, and peripartum hysterectomy.^2,5,6 Relative to women who have a live birth—those who have a stillbirth are more likely to be nonwhite—have lower socioeconomic status, smoke, obesity, and have prepregnancy conditions like diabetes and hypertension. Additionally, stillbirths are much more likely to be born preterm than live births.⁷ These factors are associated with worse maternal health outcomes and postpartum hospital readmission.^8–17

In light of this evidence, we hypothesize that comparing with women who have a live birth, women who have a stillbirth are at increased risk of postpartum readmission. To address this hypothesis, we estimated the risk of a maternal postpartum readmission within 6 weeks of giving birth for women who had stillbirths and women who had live births in a large cohort of women in California. In our secondary analysis, we examined whether diagnoses and procedures associated with postpartum readmission differed for women who had stillbirths compared with women who had live births.

Materials and Methods

We performed a population-based cohort study using data from the Office of Statewide Health Planning and Development, which is part of the California Health and Human Services Agency. For this study, we used information from vital statistics birth and fetal death files and maternal hospital discharge records (which were linked by a unique and encrypted alphanumeric code) for live birth and stillbirth deliveries in California between 1997 and 2011. We identified stillbirths from fetal death records and live births from birth records, keeping only singleton births and multiple births resulting in all live births or all stillbirths. We excluded deliveries with the following characteristics: gestational age that was missing or reported as less than 20 weeks or more than 45 weeks, unsuccessful linkages between vital records and maternal hospital discharge records, women identified as having died on the maternal delivery record, and women whose delivery hospitalization discharge date was before the birth of the child (likely a data entry error). Rates of linkage between birth records and maternal delivery records were lower for stillbirths than live births (79 vs. 97%), which was mainly attributable to unsuccessful linkage of stillbirths born at <24 weeks’ gestation (Supplementary Table S1 (available in the online version) for a comparison of characteristics of linked and unlinked stillbirths. After exclusions, the final cohort included 7,398,640 of the 7,959,376 (93%) deliveries identified in the data (Fig. 1).

Fig. 1.

Cohort selection.

The outcome of interest was postpartum hospital readmission. Readmissions were defined as inpatient hospital admissions within 6 weeks of giving birth. We examined the International Classification of Diseases (ICD)-9 diagnoses and procedures for the first postpartum readmission. Diagnoses and procedures at readmission were grouped into 17 diagnosis clusters and four procedure clusters based on methods used in previous studies.¹⁸ Codes used to define these clusters are presented in Supplementary Table S2 (available in the online version) in the supplemental materials. Women can have multiple diagnosis (up to 25) and procedure (up to 21) codes listed in one postpartum readmission; therefore, their first postpartum visit could result in them having diagnoses or procedures that are included in multiple predefined clusters. If diagnoses and procedure codes did not meet the criteria for inclusion in any of the predefined clusters, the reason for readmission was classified as “other.”

Risks of postpartum hospital readmission within 6 weeks, overall and by diagnosis and procedure cluster, were compared between women who had stillbirths and live births using log-linked binomial regression models. These models were sequentially adjusted, first for maternal demographic and prepregnancy characteristics, and then also adjusting for pregnancy and delivery characteristics. Maternal demographic information was obtained from the vital records: race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, Asian or Pacific Islander, other, and unknown), age (<20, 20–34, and 35 years), insurance type (Medi-Cal, private, other, or unknown), and education level (high school or less, some college, completed college, and unknown). Prepregnancy comorbidities included preexisting diabetes and preexisting hypertension, which were identified in the delivery hospitalization record using ICD-9 codes. Pregnancy and delivery characteristics included gestational hypertension, preeclampsia, and gestational diabetes, parity (nulliparous and multiparous), plurality (singleton and multiple), length of gestation (<28, 28–31, 32–36, 37–40, and 41þ weeks), mode of delivery (vaginal and cesarean), length of stay in hospital after delivery (<2 days, 2–3 days, and >3 days), and whether the woman had severe maternal morbidity. We identified severe maternal morbidity using diagnosis and procedure codes corresponding to a list of 18 indicators identified by the Centers for Disease Control and Prevention.^19–21 Codes used to define prepregnancy and pregnancy comorbidities, and severe maternal morbidity, are listed in Supplementary Table S2 (available in the online version) in the supplemental materials. We adjusted for this latter set of factors because they may serve as markers of mothers that have more complications (and in this interpretation serve as potential confounders).

We conducted three sets of sensitivity analyses to assess the robustness of our findings. As described above, in our primary analysis, we examined all diagnosis and procedure codes listed in the first postpartum readmission and categorized them into clusters. Other studies have used only the primary diagnosis to define reason for readmission.¹⁸ In our first sensitivity analysis, we examined whether the relative risk of readmission between women with stillbirths and women with live birth differed when considering only the primary diagnosis. Our primary analysis included all births between 20 and 45 weeks. Given the known co-occurrences of stillbirth, SMM, and preterm birth,^2,17 we conducted a second sensitivity analysis limiting our cohort to births at or after 37 weeks’ gestation. Lastly, we extended the postpartum period for readmission beyond 6 weeks. Recent studies indicate that readmissions for psychiatric conditions related to the stillbirth can occur up to 1 year after the birth.²² Our linked datasets include information on readmissions up to 9 months after the birth. Thus, in our final sensitivity analysis, we examined whether the risk and reason for readmission in 9 months after giving birth differed for women who had stillbirths versus live births.

All statistical tests were performed using SAS statistical software, version 9.4. Stanford University Institutional Review Board and the California State Committee for the Protection of Human Subjects reviewed and approved this study.

Results

The cohort included 7,398,640 births, of which 29,832 (0.4%) were stillbirths. Stillbirths were more common among women who had non-Hispanic Black race/ethnicity, were 35 years or older, did not attend college and had preexisting diabetes, preexisting hypertension, preeclampsia, and pregnancies with multiple gestations (Table 1). Compared with women having live births, women having stillbirths were more likely to deliver preterm, experience severe maternal morbidity, undergo vaginal delivery, and have shorter lengths of stay following birth.

Table 1.

Maternal demographic, clinical, and delivery characteristics among women who had stillbirths and live births: California, 1997 to 2011

	Stillbirth (n = 29,832)	Live birth (n = 7,368,808)	Difference
	n (%)	n (%)	p-Value
Demographic and prepregnancy characteristics
Race/ethnicity
Non-Hispanic White	7,925 (26.6)	2,182,554 (29.6)	<0.001
Non-Hispanic Black	3,510 (11.8)	429,390 (5.8)
Hispanic	14,938 (50.1)	3,742,788 (12.2)
Asian/Pacific Islander	3,082 (10.3)	900,030 (12.2)
Other	135 (0.5)	36,093 (0.5)
Unknown	242 (0.8)	77,953 (1.1)
Mother’s age
< 20	3,335 (11.2)	718,355 (9.8)	<0.001
20–34	20,077 (67.3)	5,407,160 (73.4)
≥ 35	6,420 (21.5)	1,243,293 (16.9)
Payment type at birth
Medi-Cal	13,570 (45.5)	3,324,240 (45.1)	<0.001
Private	13,193 (44.2)	3,742,928 (50.8)
Other/unknown	3,069 (10.3)	301,640 (4.1)
Mother’s education level
High school or less	17,351 (58.2)	2,040,557 (54.6)	<0.001
Some college	6,430 (21.6)	1,505,700 (20.4)
Completed college	4,487 (15.0)	1,669,264 (22.7)
Unknown/missing	1,564 (5.2)	173,287 (2.4)
Nulliparous	11,512 (38.6)	2,846,231 (38.6)	0.900
Preexisting diabetes	807 (2.7)	56,635 (0.8)	<0.001
Preexisting hypertension	1,162 (3.9)	106,492 (1.5)	<0.001
Pregnancy characteristics
Multiple birth (nonsingleton)	833 (2.8)	114,048 (1.6)	<0.001
Gestational age at delivery
< 28 wk	11,315 (37.9)	38,131 (0.5)	<0.001
28–31 wk	4,258 (14.3)	61,193 (0.8)
32–36 wk	6,628 (22.2)	589,790 (8.0)
37–40 wk	6,489 (21.8)	5,620,524 (76.3)
41 + wk	1,142 (3.8)	1,059,170 (14.4)
Gestational diabetes	1,366 (4.6)	412,228 (5.6)	<0.001
Gestational hypertension	544 (1.8)	146,345 (2.0)	0.045
Preeclampsia	1,546 (5.2)	223,409 (3.0)	<0.001
Birth characteristics
Cesarean birth	4,524 (15.2)	2,061,398 (28.0)	<0.001
Severe maternal Morbiditya	1,650 (5.5)	69,547 (0.9)	<0.001
Length of stay in hospital after birth
< 2 d	20,124 (67.5)	2,291,406 (31.1)	<0.001
2–3 d	7,731 (25.9)	4,502,104 (61.1)
> 3 d	1,977 (6.63)	575,298 (7.8)

^aBased on Centers for Disease Control And Prevention Index.¹⁹

Among the 7,398,640 women with live births, 70,559 were readmitted within 6 weeks (0.96%), and among the 29,832 women with stillbirths, 614 were readmitted within 6 weeks (2.06%). The median time to readmission was 11 days after delivery for women with stillbirth and women with live birth. Fig. 2 illustrates the cumulative percentage of women with a postpartum readmission within 6 weeks of giving birth. As noted, just under 1% of women who had a live birth were readmitted within 6 weeks; the same percent of women who had a stillbirth were readmitted within 9 days of giving birth.

Fig. 2.

Cumulative percent of women with a postpartum readmission within 6 weeks of giving birth.

The risk of readmission was significantly higher among women who had stillbirths than women who had live births (RR = 2.17, 95% confidence interval [CI] : 2.01–2.36; Table 2). In our sequential models, the relative risk of readmission decreased to 1.94 (95% CI: 1.79–2.10) after adjusting for demographic and antepartum characteristics, and 1.47 (95% CI: 1.35–1.60) after additionally adjusting for pregnancy, and delivery characteristics.

Table 2.

Rate and relative risk of postpartum hospital readmission within 6 weeks of birth among women who had stillbirths versus live births, overall and by specific maternal diagnosis/procedure clusters at readmission; California, 1997 to 2011

Postpartum readmission	Stillbirth (n = 29,832)	Live birth (n = 7,368,808)	Relative risk (95% confidence interval)
	n (ratea)	n (ratea)	Unadjusted	Model 1b	Model 2c
All	614 (206)	70,559 (96)	2.17 (2.01–2.36)	1.94 (1.79–2.10)	1.47 (1.35–1.60)
Specific diagnosis/procedure clusters
Uterine infection/pelvic inflammatory disease	142 (48)	13,435 (18)	2.61 (2.22–3.09)	2.45 (2.08–2.90)	1.79 (1.50–2.15)
Psychiatric condition	113 (38)	5,433 (7)	5.15 (4.28–6.21)	3.83 (3.17–4.62)	2.29 (1.86–2.82)
Hypertensive disorder	100 (34)	8,040 (11)	3.08 (2.53–3.75)	1.93 (1.58–2.35)	1.49 (1.20–1.84)
Urinary tract infection	88 (29)	9,433 (13)	2.31 (1.87–2.85)	2.01 (1.63–2.48)	1.51 (1.21–1.89)
Nonspecific postpartum diagnosis	79 (26)	13,495 (18)	1.44 (1.16–1.81)	1.25 (1.01–1.57)	1.06 (0.85–1.35)
Dilation and curettage	72 (24)	4,729 (6)	3.77 (2.98–4.76)	3.71 (2.94–4.69)	1.84 (1.41–2.42)
Hemorrhage and/or retained products of conception	68 (23)	6,436 (9)	2.61 (2.06–3.32)	2.52 (1.98–3.21)	1.31 (1.01–1.71)
Transfusion	57 (19)	3,813 (5)	3.70 (2.85–4.80)	3.11 (2.39–4.05)	1.64 (1.23–2.17)
Gallbladder disease	51 (17)	11,853 (16)	1.06 (0.81–1.40)	1.06 (0.80–1.40)	1.10 (0.82–1.46)
Thrombotic event	51 (17)	2,521 (3)	5.00 (3.79–6.60)	4.38 (3.31–5.79)	2.71 (2.00–3.68)
Wound infection and/or breakdown	49 (16)	10,025 (14)	1.21 (0.91–1.60)	0.97 (0.73–1.29)	1.17 (0.87–1.55)
Acute cardiovascular disease	19 (6)	2,029 (3)	2.31 (1.47–2.64)	1.50 (0.95–2.36)	1.59 (1.00–2.55)
Upper respiratory infection	19 (6)	1,813 (2)	2.59 (1.65–4.07)	1.89 (1.20–2.98)	1.54 (0.96–2.48)
Pancreatitis	<15	3,671 (5)	-	-	-
Mastitis, breast abscess	<15	3,436 (5)	-	-	-
Appendicitis	<15	2,228 (3)	-	-	-
Anesthesia complications	<15	722 (1)	-	-	-
Bacteremia, sepsis	<15	606 (1)	-	-	-
Acute cerebrovascular disease	<15	588 (1)	-	-	-
Laparotomy	<15	581 (1)	-	-	-
Hysterectomy	<15	469 (1)	-	-	-
Otherd	46 (15)	6,694 (9)	1.70 (1.27–2.27)	1.57 (1.17–2.09)	0.95 (0.70–1.29)

Note: women can be included in multiple clusters.

^aPer 10,000 births.

^bAdjusted for maternal race/ethnicity, age at delivery, payment type at delivery, education level, parity, preexisting diabetes, and preexisting hypertension.

^cAdjusted for maternal race/ethnicity, age at delivery, payment type at delivery, education level, parity, preexisting diabetes, preexisting hypertension, multiple gestation, gestational age, gestational diabetes, gestational hypertension, preeclampsia, mode of delivery, severe maternal morbidity at delivery, and length of stay in hospital after birth.

^dNo diagnoses or procedures for this readmission fell into one of the 21 predefined clusters.

We examined diagnoses and procedures at the first postpartum readmission based on 21 predefined clusters. Since we included all diagnoses present at first readmission, women could be included in multiple clusters. In both groups of women, approximately 50% were included in one cluster, 26% were included in two clusters, 10% were included in three clusters, and 4% were included in four or more clusters in their first postpartum hospitalization (Supplementary Table S3 (available in the online version) in supplemental materials. Among women with stillbirths, the most common diagnoses at the first postpartum readmission were uterine infection or pelvic inflammatory disease (48 per 10,000 stillbirths), psychiatric conditions (38 per 10,000 stillbirths), and hypertensive disorders (34 per 10,000 stillbirths; Table 2). Among women with live births, the most common diagnoses at first postpartum readmission were uterine infection or pelvic inflammatory disease (18 per 10,000 live births), nonspecific postpartum diagnosis (18 per 10,000 live births), and gallbladder disease (16 per 10,000 live births). For most clusters, relative risks were attenuated after adjusting for maternal demographic and prepregnancy characteristics, and further attenuated after adjusting for pregnancy and delivery characteristics.

After full adjustment, the risk of readmission for seven clusters was significantly higher for women with stillbirth deliveries; there were no clusters for which risk of readmission was higher among women with live births. The adjusted relative risk of readmission were greatest for thrombotic events (aRR = 2.71; 95% CI: 2.00–3.68) and psychiatric conditions (aRR = 2.29; 95% CI: 1.86–2.82).

Sensitivity Analyses

In the main analysis, diagnosis and procedure clusters were defined using all diagnoses and procedures listed in the first postpartum readmission. In our sensitivity analysis, we defined diagnosis clusters based only on the primary diagnosis (i.e., one diagnosis per woman). The 10 most common diagnosis-based clusters remained the same in both analyses, but the rankings varied. For example, using all diagnoses, the top three indications at readmission among women who had stillbirths were uterine infection or pelvic inflammatory disease, psychiatric conditions, and hypertensive disorders; using only primary diagnosis, the top three indications at readmission were uterine infection or pelvic inflammatory disease, hemorrhage and/or retained products of conception, and psychiatric conditions (Supplementary Table S4 [available in the online version]). The relative risks of readmission for each diagnosis cluster remained similar to the main analysis (Table 3).

Table 3.

Adjusted relative risk of postpartum hospital readmission among women who had stillbirths versus live births, by sensitivity analysis; California, 1997 to 2011

Postpartum readmission	Adjusteda relative risk (95% confidence interval)
	Main analysis	Sensitivity analysis 1: only consider primary diagnosis	Sensitivity analysis 2: subcohort of births at or after 37 weeks	Sensitivity analysis 3: readmissions within 9 months of giving birth
All	1.47 (1.35–1.60)	1.47 (1.35–1.60)	1.89 (1.60–2.23)	1.51 (1.43–1.59)
Specific diagnosis/procedure clusters
Uterine infection/pelvic inflammatory disease	1.79 (1.50–2.15)	1.69 (1.36–2.09)	2.06 (1.43–2.95)	1.65 (1.40–1.95)
Psychiatric condition	2.29 (1.86–2.82)	2.34 (1.69–3.25)	4.64 (3.23–6.69)	1.70 (1.49–1.93)
Hypertensive disorder	1.49 (1.20–1.84)	1.25 (0.89–1.75)	b	1.59 (1.36–1.87)
Urinary tract infection	1.51 (1.21–1.89)	1.58 (1.09–2.29)	1.95 (1.27–3.01)	1.47 (1.26–1.71)
Nonspecific postpartum diagnosis	1.06 (0.85–1.35)	1.04 (0.75–1.44)	1.11 (0.69–1.79)	1.05 (0.94–1.32)
Dilation and curettage	1.84 (1.41–2.42)	c	b	2.11 (1.74–2.54)
Hemorrhage and/or retained products of conception	1.31 (1.01–1.71)	1.16 (0.85–1.59)	2.36 (1.50–3.72)	1.44 (1.14–1.81)
Transfusion	1.64 (1.23–2.17)	c	b	1.69 (1.37–2.09)
Gallbladder disease	1.10 (0.82–1.46)	1.40 (0.97–2.04)	1.33 (0.82–2.18)	1.15 (0.99–1.33)
Thrombotic event	2.71 (2.00–3.68)	3.33 (2.37–4.69)	b	2.55 (1.93–3.35)
Wound infection and/or breakdown	1.17 (0.87–1.55)	1.07 (0.74–1.54)	1.64 (1.02–2.65)	1.19 (0.93–1.52)
Acute cardiovascular disease	1.59 (1.00–2.55)	b	b	1.19 (0.81–1.76)
Upper respiratory Infection	1.54 (0.96–2.48)	b	b	1.42 (1.04–1.95)
Otherd	0.95 (0.70–1.29)	1.65 (1.41–1.94)	2.28 (1.42–3.69)	1.63 (1.48–1.80)

Note: frequencies for each sensitivity analysis can be found in Supplementary Tables S4–S6 (available in the online version).

^aAdjusted for maternal race/ethnicity, age at delivery, payment type at delivery, education level, parity, preexisting diabetes, preexisting hypertension, multiple gestation, gestational age, gestational diabetes, gestational hypertension, preeclampsia, mode of delivery, severe maternal morbidity at delivery, and length of stay in hospital after birth.

^bMissing because of small cell sizes.

^cMissing because this sensitivity analysis only considered diagnoses.

^dNo diagnoses or procedures for this readmission fell into one of the 21 predefined clusters.

In our second sensitivity analysis, we limited the cohort to live births and stillbirths that occurred at or after 37 weeks’ gestation. In this subcohort of 6,687,325 births, 7,631 (0.1% of births) were stillbirths; for both stillbirths and live births, the rate of readmission was slightly lower than what was seen in the primary analysis (Supplementary Table S5 [available in the online version]). However, the adjusted risk of readmission was slightly higher than in our primary analysis; women who had stillbirths had a 1.89 (95% CI: 1.60–2.23) times higher risk of readmission in the 6 weeks’ postpartum relative to women who had live births (Table 3).

In our third sensitivity analysis, we examined the risk of readmission over a 9-month period after delivery. Among the 7,398,640 women with live births, 187,276 (2.5%) were readmitted and among the 29,832 women with stillbirths, 1,709 (5.7%) were readmitted within 9 months. The most common indications for readmission for women with stillbirths were psychiatric conditions, hypertensive disorder, and gallbladder disease, and the most common indications for readmission among women who had live births were gallbladder disease, psychiatric conditions, and uterine infection/pelvic inflammatory disease (Supplementary Table S6 [available in the online version]). The adjusted relative risk of remained very similar (aRR = 1.51, 95% CI: 1.43–1.59) to the adjusted relative risk of readmission within 6 weeks of giving birth (Table 3).

Discussion

Principal Findings of the Study

Our findings indicate that more than 2% of women who have stillbirths were readmitted within 6 weeks of giving birth, compared with just under 1% of women who have live births. After adjusting for demographic and clinical characteristics, the risk of postpartum readmission was 46% higher among women who have stillbirths compared with women who have live births.

Results in the Context of Prior Evidence

Previous studies indicate that women who have stillbirths are at greater risk of morbidity during the delivery hospitalization.^2,5 Our study finds that higher risk of morbidity among women who have stillbirths continues after hospital discharge. At delivery, the most common complications among women who had stillbirths were blood transfusion and disseminated intravascular coagulation,² whereas in the 6 weeks’ postpartum, the most common complications at readmission among women who had stillbirths were uterine infection and pelvic inflammatory disease, and psychiatric conditions. Additionally, as with previous findings, we found that women who had stillbirths had shorter lengths of stay in hospital than women who had live births.²³ More than 65% of women who had a stillbirth were discharged less than 2 days after giving birth, whereas only 31% of women who had a live birth were discharged in this time period. This difference could be partially attributed to fact that cesarean births are less common among women who have stillbirths; however, risk of readmission remained higher even after adjusting for mode of delivery and length of stay.

Clinical Implications

Our findings highlight the need for closer postpartum follow-up after stillbirth. Current stillbirth management guidelines focus on identifying risk factors, examining the fetus, and providing appropriate counseling, including recurrence counseling.²⁴ We found that women who have stillbirths are at increased risk of postpartum readmissions for a variety of physical and mental health conditions. Recent obstetric guidelines highlight the need to optimize postpartum care, recommending a shift from a single postpartum visit at 6 weeks, to a comprehensive postpartum process.²⁵ These guidelines do not indicate specific needs and circumstances for women who had stillbirths other than to provide emotional support and bereavement counseling. Psychiatric conditions were the second most common indication at readmission, suggesting additional bereavement and mental health support to be provided at and after discharge. Compared with live births, the condition associated with the highest risk of readmission after stillbirth was a thrombotic event. Further studies are needed to examine whether differential etiologies explain the increased risk and to examine whether adherence to American College of Obstetrics and Gynecology guidelines for venous thromboembolism prevention are similar after a stillbirth compared with a live birth.

The optimal postpartum care practice for stillbirth is unknown; however, public health initiatives to monitor and address maternal health during and after delivery could improve outcomes for women who have stillbirths. This study provides sufficient evidence to justify why complications related to stillbirth as well as live birth delivery should be highlighted in future maternal morbidity and mortality guidelines. Future studies should investigate whether early postdischarge follow-up by primary care providers can reduce the readmission risk for women who deliver stillbirths.

Research Implications

The linkage of birth/fetal death records and maternal hospitalization records allowed us to obtain accurate information on postpartum readmissions among women who had live births and stillbirths. However, there were specific questions regarding postpartum care of women with stillbirths that these data could not answer. For example, we were unable to examine whether women with stillbirths received adequate postpartum care from their primary care provider or the timing of that care. Additionally, we could not determine reasons for the shorter hospital stay in women with stillbirths compared with live births. Women with stillbirths may want to be discharged sooner to avoid being in a maternity ward with other women who have live births, which could be difficult after experiencing such a traumatic loss. Finally, we examined postpartum readmission within 6 weeks. Future studies could extend this time-frame to examine readmissions among women with stillbirths within the first year of giving birth.

Strengths and Limitations

The large dataset with linkages between birth records and postpartum hospitalizations is a significant strength of our study. These data allowed for a detailed examination of diagnoses and procedures at postpartum readmission among women who had stillbirths, both of which are rare. Our study has several limitations. First, linkage between birth records and maternal delivery records was lower for stillbirths than live births (79 vs. 97%). Our concerns about selection bias due to the exclusion of unlinked stillbirths is minimal as the distribution of maternal sociodemographic characteristics was similar among stillbirth cases that were and were not linked to discharge records. However, it is unknown whether clinical factors that predispose women to very early stillbirth, which were the most likely to have unsuccessful linkage, would result in a higher or lower risk of readmission. Second, we do not have detailed information on preexisting conditions. We identified all prepregnancy and pregnancy conditions using diagnosis codes in the birth hospitalization record, which likely underestimates the true prevalence of diabetes and hypertension in these time periods.²⁶ Additionally, we do not have information on preexisting mental health conditions. Previous studies have found that women exposed to antipsychotic medication during pregnancy are at a twofold increased risk of stillbirth,²⁷ and women with preexisting mental health conditions are more likely to experience postpartum mental illness.²⁸ The lack of information on preexisting mental health conditions could lead to residual confounding; however, it is still useful to know the actual risk of readmission related to psychiatric conditions. We examined postpartum readmission within 6 weeks of giving birth; since women are at risk of being readmitted only after they are discharged from their birth hospitalization, length of follow-up varies across the cohort. However, more than 92% of women in the cohort (regardless of whether they had stillbirths or live births) were discharged within 3 days of giving birth. Additionally, we are unable to identify whether a stillbirth occurred intrapartum or antepartum. Given that risk factors for antepartum and intrapartum stillbirth are heterogeneous,²⁹ the risk of postpartum readmission may also vary. Lastly, our study included deliveries in California only between 1997 and 2011. Although care practices for women with stillbirths are unlikely to have dramatically changed in subsequent years, our results may not be generalizable outside of this time period and location.

Conclusion

Our study suggests that the risk of postpartum readmission within 6 weeks of delivery issubstantially higher among women who had stillbirth compared with women who had live births. Given the risk of adverse outcomes, both at delivery and after, women who have stillbirths may benefit from additional monitoring and counseling after hospital discharge for potential postpartum medical and psychiatric complications.

Key Points.

• Women who have stillbirths are at nearly 1.5 times greater risk of postpartum readmission than women who have live births.

• Uterine infections and pelvic inflammatory disease, and psychiatric conditions are the most common reasons for readmission among women who had a stillbirth.

• Women who have stillbirths may benefit from additional monitoring and counseling after hospital discharge for potential postpartum medical and psychiatric complications.