Socioeconomic,Demographic,and Clinical Factors Associated with Postpartum Readmission

Introduction

The postpartum period is one of the increased vulnerabilities, as new mothers often have unmet physical, psychological, and social needs. ¹ The potential lack of structured medical care, along with new stressors, may result in patients receiving inadequate care. In some cases, this may lead to rehospitalization, which is considered a core measure of health care quality. ² Postpartum readmission is not only associated with increasing health care costs but may have a significant impact on a patient’s postpartum recovery. ³

Identification of risk factors associated with postpartum readmission may help guide interventions to reduce this rate. Readmission data from ‘State Inpatient Databases’ have shown that maternal comorbidities’ such as hypertension and diabetes have been associated with increased readmission rates. ⁴ In addition to comorbidities, further evaluation of these databases showed that racial and socioeconomic disparities also contribute to readmission rates. Specifically, Black patients, patients in the poorest quartile of median income, and those with public insurance were found to have higher rates of readmission. ⁵

Individual socioeconomic factors may only demonstrate one component of the racial disparities associated with postpartum readmission. For example, structural racism and neighborhood characteristics such as unemployment rates were shown to be associated with an increased rate of preterm deliveries. ⁶ A more detailed representation of individual socioeconomic status may be found by studying neighborhoods, which allows inclusion of certain factors, such as access to safe and adequate housing, education, and unemployment rates. Despite identifying certain individual social factors associated with postpartum readmission, there is limited knowledge of neighborhood characteristics linked to postpartum readmission. By evaluating these factors, also known as social determinants of health, we may be able to address gaps that may exist and provide interventions to reduce the risk of readmission.

The objective of this study was to determine which socioeconomic, demographic, and clinical characteristics are associated with postpartum readmission. Both individual patient-level and neighborhood-level factors were evaluated. The authors hypothesized that higher maternal comorbidity burden and socially vulnerable neighborhoods would be associated with an increased risk of postpartum readmission.

Materials and Methods

This was a retrospective cohort study of all pregnant patients who delivered between January 1, 2018 and March 1, 2020. Data were obtained from 7 hospitals within a large academic health system in New York. Four hospitals are in the New York City boroughs, and three are located in Long Island. Two of the hospitals are perinatal regional centers. All patients who delivered after 20 weeks of gestational age were evaluated for inclusion. Patients who presented to the emergency room postpartum but remained less than 23 hours and were not admitted were excluded from the study. Patients were identified by querying the system-wide inpatient electronic health record system (Sunrise Clinical Manager; Sunrise Clinical Manager, Allscripts Corp., Chicago, IL) for unique encounter numbers designated to each delivery hospitalization. The encounter number was then used to determine if a patient was readmitted within 42 days after discharge. Each delivery hospitalization was included for patients who had more than one delivery within the study period. Institutional Review Board approval was obtained prior to conducting the study.

Clinical data was obtained from the electronic medical record (EMR). Maternal and delivery characteristics included age, race, ethnicity, marital status, preferred language, health insurance type, body mass index (BMI), gestational weight gain, mode of delivery, chorioamnionitis, endometritis, blood product administration, intensive care unit admission during delivery hospitalization, and reason for readmission. Maternal comorbidities were identified by International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10) codes, and clinical documentation.

Individual neighborhood characteristics were determined by obtaining the ZIP code of residence for each patient and linking it to data from the United States Census Bureau’s American Community Survey. ⁷ Variables evaluated included mean persons per household, median household income, percentage of households receiving supplemental assistance, percentage of single-parent households, percentage of owner-occupied housing, unemployment rate, and percentage of adults with less than high school education.

The primary outcome was postpartum readmission which was defined as admission for 24 hours or more within 42 days of delivery. Demographic information, medical comorbidities, and delivery hospitalization characteristics were compared between patients who were readmitted and those who were not readmitted. The analysis also examined the characteristics associated with readmission, categorized by race and ethnicity. A secondary analysis was then conducted to assess specific characteristics associated with patients readmitted for hypertensive disorders and patients readmitted for other indications.

Statistical analysis included Chi-squared and Fisher’s exact tests as appropriate for comparison of categorical variables and t-test or the Wilcoxon rank-sum test, as appropriate, for continuous variables. Mixed effects logistic regression was used to evaluate factors associated with postpartum readmission while adjusting for potential confounders: advanced maternal age, BMI, race, insurance type, language, marital status, delivery characteristics, medical comorbidities, and neighborhood characteristic. Fixed effects logistic regression was used to assess the association of independent factors with postpartum readmission while stratifying by race and ethnicity. Hysterectomy, venous thromboembolism (VTE) events, insurance type, ICU admission, chorioamnionitis, and endometritis were not evaluated due to low frequency and unreliable estimates. A multinomial logistic regression was used for the secondary analysis comparing patients admitted for hypertensive disorders to non-readmitted patients and patients admitted for non-hypertensive disorders to non-readmitted patients. Endometritis, hysterectomy, and VTE events were not evaluated due to low frequency. Adjusted odds ratios (aOR) and 95% confidence interval (CI) were calculated for each group. A p-value of <.05 was considered statistically significant. Statistical analyses were performed with SAS® 9.04 (SAS Inc, Cary, NC.).

Results

During the study period, 57,626 unique patients were identified. Among this group, 1,623 had two or more births, resulting in 59,251 deliveries. Within that total, 3,225 (5.4%) were associated with a postpartum emergency room visit (not admitted) and excluded from the study. In total, 57,507 deliveries were included for analysis, and of those, 1,481 (2.5%) postpartum readmission occurred. Baseline characteristics of the study population are presented in Table 1 and stratified by readmission status.

Patients who were readmitted were predominantly older, nulliparous, Black race, and unmarried. There were higher rates of maternal comorbidities among those readmitted including excessive weight gain, gestational and pre-gestational diabetes, chronic hypertension, hypertensive disorders of pregnancy, asthma, and psychiatric disorders. Patients who were readmitted more often had cesarean deliveries and longer length of stay during delivery hospitalization. Neighborhood characteristics with increased rates of readmission included higher rates of single-parent homes, unemployment rates, need for public assistance as well as lower income, lower educational attainment, and smaller household size.

After adjusting for differences between groups, factors associated with increased risk of postpartum readmission included Black race in comparison to non-Hispanic white race (aOR: 1.56, 95% CI: 1.30–1.86) and public insurance (aOR: 1.19, 95% CI: 1.05–1.35). Neighborhood characteristics were not associated with the increased rate of postpartum readmission (shown in Table 2).

Maternal characteristics associated with readmission included advanced maternal age (aOR: 1.24, 95% CI: 1.10 −1.40), nulliparity (aOR: 1.23, 95% CI: 1.08–1.40), chronic hypertension (aOR: 2.83, 95% CI: 2.33–3.44), hypertensive disorders of pregnancy (aOR: 1.91, 95% CI: 1.60–2.29), gestational diabetes (aOR: 1.38, 95% CI: 1.15 − 1.64) psychiatric disorders (aOR: 1.58, 95% CI: 1.24–2.06), asthma (aOR: 1.35, 95% CI: 1.07–1.70), BMI >25 kg/m² (aOR: 1.22, 95% CI: 1.05–1.42), and weight gain >40 lb during pregnancy (aOR: 1.19, 95% CI: 1.01–1.40).

Delivery and hospitalization factors associated with increased likelihood of postpartum readmission include endometritis (aOR: 2.37, 95% CI: 1.11–5.08) and administration of blood products during delivery admission (aOR: 2.18, 95% CI: 1.68–2.82) (Table 2). The readmission risk factors were further stratified by race and ethnicity (Table 3). Certain characteristics were found to be more closely linked to readmission for specific racial groups. For instance, particular neighborhood characteristics were connected to readmission among patients who identified as Black, Hispanic, and multiracial. Additionally, psychiatric disorder (aOR: 1.66, 95% CI: 1.17–2.38) was only associated with increased readmission among patients of White race.

The most common readmission diagnosis was hypertensive disorder, followed by infectious disorders and sepsis (Table 4). The mean and median days for readmission were 6.48 and 4.24 days, respectively. When evaluating which characteristic was associated with the patient’s admitted of hypertense disorder, compared to other causes, certain risk factors were identified (Table 5). Patients who had chronic hypertension (aOR: 4.16, 95% CI: 3.30–5.23), hypertensive disorders of pregnancy, (aOR: 2.56, 95% 2.06–3.18), advanced maternal age (aOR: 1.59, 95% 013.5–1.86) and BMI >25 kg/m² (aOR: 1.55 95% CI: 1.25–1.91) were more like to be admitted for hypertensive disorder postpartum than for other indications. In contrast, patients who delivered preterm (aOR: 1.64 95% CI: 1.28–2.10), had chorioamnionitis (aOR: 1.89 95% CI: 1.10–3.23), or received blood products (aOR: 2.87 95% CI: 2.03–4.07) were more likely to be readmitted for non-hypertensive-related indications.

Discussion

Overall, 2.5% of women were readmitted within 6 weeks after delivery during our study period, which was similar to prior reported rates. ^4,5 Several clinical and demographic factors including chronic hypertension, endometritis, blood product administration, excessive weight gain, and BMI 25> kg/m² were associated with postpartum readmission.

The postpartum period, sometimes referred to as the “fourth trimester,” includes the first 12 weeks after birth and is a critical time when patients are acclimating to new physical and environmental changes. ⁸ Similar to prior population studies, we found that maternal comorbidities such as diabetes, hypertensive disorders, and maternal characteristics including advanced maternal age were associated with increased risk of readmission. ^4,9 Additionally, the need for blood product administration and endometritis also contributed to higher readmission rates. ^10,11 However, contrary to past studies, we did not find a correlation between the mode of delivery and readmission rates. ¹² In addition, we did observe a higher rate of readmission among nulliparous women, which has not been previously reported.

Prior studies have also illustrated that Class III obesity (BMI >40 kg/mm²) is associated with higher rates of postpartum readmission. ¹³ We found risk of postpartum readmission began with BMI at or above 25 kg/mm² and the risk of readmission increased with increasing BMI. An additional new finding in our study was that excessive weight was also associated with readmission. As the obesity rates continue to climb, one would expect resulting elevation in pregnancy complications and postpartum readmission. Studies have demonstrated that implementing dietary and exercise changes can decrease weight gain and enhance pregnancy outcomes. ¹⁴ However, more research is necessary to investigate the effects of nutritional interventions on postpartum readmission.

There are certain factors that can increase the chance of postpartum readmission, such as race, public insurance, and lower income. ^{5,12,15 –17} Our study similarly identified black patients, and those with public insurance also had an increased rate of readmission. Not only do these disparities exist in readmission rates, but they also impact outcomes during the readmission hospitalization, such as an increased risk of pulmonary edema and heart failure. ¹⁸ This study sought to further expand our understanding by examining neighborhood characteristics associated with postpartum readmission. A neighborhood housing composition may reflect the patient’s social support. For example, women from single-family homes may lack the resources to dedicate time to their postpartum recovery, potentially compounding an already vulnerable state. Though our findings illustrated that racial disparity existed, we were not able to identify more detailed neighborhood characteristics associated with readmission among all patients. Chronic hypertension was a strong risk factor for hospital readmissions. Data from surveillance indicate that Black patients have higher rates of chronic hypertension compared to other racial groups. ¹⁹ To investigate if this may have contributed to the higher readmission rates among Black patients, the data were stratified by race and ethnicity. It was found that chronic hypertension was a significant risk factor for readmission across all racial groups.

Currently, only 60% of patients present to postpartum visits. ⁸ These appointments are typically 6 weeks after delivery. Beyond this singular appointment, outpatient care is otherwise limited, which may prompt women to utilize emergency rooms more frequently, leading to higher rehospitalization rates. ¹² On average, most of the readmission occurred within 6 days after discharge. Increasing access to health care earlier could potentially improve rates of readmission. For example, telehealth monitoring of patients with hypertensive disorders reduced the rate of readmission. ²⁰ Patients are routinely scheduled for a 72-hour postpartum blood pressure check if the pregnancy was complicated with hypertension. However, the data on how many hypertensive patients had a 72-hour follow-up after discharge and whether there were differences based on insurance type was not available. In addition, our study also illustrated that patients readmitted for hypertensive disorders have different risk factors compared to patients admitted for other indications. Future studies should assess if earlier or more frequent postpartum visits reduce readmission rates for high-risk patients and if follow-up can be tailored based on risk factors for type of readmission. For example, in addition to patients with hypertensive disorders of pregnancy, patients who are of advanced maternal age or have elevated BMI should also have earlier appointments for blood pressure checks. Future studies can also determine if racial disparities exist in postpartum outpatient follow-up and if those differences impact rates of readmission.

While hospital readmission is an important quality measure, the Society of Maternal Fetal Medicine advises against using postpartum readmission rates as a way to encourage health care providers to avoid readmitting patients when necessary. This is because the postpartum readmission rate is relatively low (between 1 and 2%), and there are complex risk factors associated with postpartum readmission. ²¹ Additionally, there is concern about unintended adverse outcomes. It seems that individual characteristics, and not hospital factors, contribute to higher readmission rates. ²² A holistic approach to patient care, which includes attention to medical and psychological comorbidities, as well as socioeconomic determinants of health, may indicate certain factors that leave patients prone to poor outcomes. By acknowledging these risk factors, specific interventions can be offered to mitigate postpartum readmission rates.

Our study has several strengths. Multiple hospitals within a large health system were evaluated. Each site serves a unique patient population. The subjects reside in both urban and suburban communities and span the full socioeconomic spectrum. Our study collected more detailed information directly from clinical records in the EMR, as well as using billing data. This approach differs from some previous research that relied on large databases based on International Classification of Disease (ICD) codes, which may have certain biases or inaccuracies. Using a shared EMR system with standardized data collection instruments minimized data extraction errors and misclassification.

Limitations of this study include its retrospective nature and inability to definitively determine causal factors and potential unadjusted confounders. We did not evaluate patient management in the emergency department and how any particular actions there affected the decision to admit versus discharge the patient. Future studies should also evaluate if patients who presented to the emergency room only had different risk factors than those readmitted. We also did not have access to data regarding postpartum follow-up visits including if and how many outpatient postpartum visits patients attended. The number of outpatient encounters for each patient after the delivery hospitalization, if any, is unknown. Neighborhood characteristics were assessed using the ZIP code. In highly populated areas, there may be wide variation in socioeconomic metrics. Thus, ZIP code-associated socioeconomic data is not always representative of those living in that area. Smaller unit areas such as census tracts may provide more accurate neighborhood socioeconomic data. Patient home addresses were not consistently available for this study; therefore, census tracking could not be performed. It could also not be confirmed if patients accurately represented their home addresses. Individual patient-level socioeconomic determinants of health such as income, employment, and education were not consistently available and could not be evaluated. It is not known if these patient characteristics differed from neighborhood characteristics. The data may also not be generalizable to other regions.

Our results demonstrate that individual clinical and social risk factors are associated with an increased likelihood of postpartum readmission. The neighborhood characteristics evaluated in this study were not associated with this outcome. Maternal comorbidities and pregnancy complications associated with postpartum readmission include chronic hypertension, elevated BMI, excessive gestational weight gain, endometritis, and peripartum blood product administration. Different risk factors existed among patients who were admitted for hypertensive disorders compared to patients admitted for other indications. Maternal social risk factors include Black race and public health insurance. Offering tailored medical and social support to those at higher risk may help lower the rate of postpartum readmission.