Postpartum healthcare utilization and morbidity among diabetic versus nondiabetic individuals in remote hypertension program

Abstract

Introduction

Individuals with diabetes (gestational or pregestational) are at increased risk of adverse outcomes, particularly in the setting of hypertension (HTN) in pregnancy. Though remote self‐measured blood pressure (SMBP) monitoring programs may reduce rates of HTN‐related hospital readmission, emergency department (ED) presentations, or severe maternal morbidity (SMM), it is unclear whether the extent of this reduction differs between those with and without diabetes. We aimed to compare outcomes between postpartum patients in our remote SMBP program with versus without diabetes.

Methods

Since November 2022, postpartum patients with HTN at our tertiary care hospital have been offered enrollment in our remote SMBP program, which involves a nurse practitioner remotely managing blood pressure for 6 weeks postpartum. For this analysis, participants were stratified by the presence of pregestational or gestational diabetes. Participants with unknown diabetes status were excluded. The primary outcome was a composite of postpartum readmission or ED presentation for HTN within 30 days of delivery hospitalization. Secondary outcomes included HTN‐related SMM and remote medication initiation. A generalized linear model was used to estimate relative risks (RRs) after adjustment for differences in demographic, obstetric, and medical conditions.

Results

Among the first 2000 participants in the SMBP program, 17% had diabetes. Compared to those without diabetes, those with diabetes were older, less likely to identify as White, or have gestational hypertension. They were more likely to have chronic hypertension, be delivered via cesarean, and have a lower median gestational age at delivery. After adjusting for these factors, there was no difference in the risk of the composite outcome of HTN‐related postpartum readmission or ED presentation between those with and without diabetes (15.92% vs. 16.56%; adjusted RR, 0.92 [0.70, 1.21]). Similarly, there were no differences in secondary outcomes.

Conclusion

In our remote SMBP program for postpartum patients with HTN, unplanned HTN‐related healthcare utilization or SMM was similar between those with and without diabetes. While this study may have been underpowered to detect small differences, these findings suggest that at a minimum, the current monitoring strategy meets the standard of care for patients with HTN and diabetes.

1. INTRODUCTION

Hypertension (HTN) affecting pregnancy is one of the leading causes of maternal mortality in the United States, documented in up to 31.6% of maternal deaths that occurred during delivery from 2017 to 2019 [1]. HTN affects up to 10% of pregnancies, with a 25% increase in the past 20 years [2]. The prevalence of risk factors for HTN such as maternal age, diabetes, and obesity is also increasing. For example, from 2016 to 2020 alone, gestational diabetes (GDM) cases increased by 20%, affecting 7.8 per 100 births in 2020 [3].

The American College of Obstetricians and Gynecologists (ACOG) recommends blood pressure (BP) evaluation for HTN within 10 days postpartum, but acknowledges that in‐person evaluations may be a burden on patients. ACOG posits the use of other methods for monitoring to minimize this burden: text messaging, phone support, and remote BP monitoring [4]. Remote BP monitoring may help improve adherence to ACOG recommendations for monitoring postpartum while improving perinatal health equity.

In November 2022, our high‐volume academic standalone women's hospital (Women & Infants Hospital in Providence, Rhode Island) launched a remote self‐monitoring blood pressure (SMBP) program. In this opt‐out program, an automatic BP cuff is provided to all postpartum patients diagnosed with HTN prior to hospital discharge from delivery. The program is staffed by a Family Nurse Practitioner and a bilingual English‐ and Spanish‐speaking certified community health worker.

Other SMBP programs have been shown to reduce rates of HTN‐related hospital readmission or emergency department (ED) presentations within 6 weeks of delivery, though data are lacking on the effect on severe maternal morbidity (SMM) [5, 6]. However, patients with comorbid hypertension and diabetes have been shown to be at higher risk of SMM and postpartum readmission than those with hypertension alone [7]. Additionally, maternal morbidity and mortality disproportionately affect women with multimorbidity; thus, it is possible that those with comorbid diabetes in the setting of HTN may have worse outcomes than those with hypertension alone. Examining whether an SMBP program for patients with HTN is associated with additional benefits among patients with versus without comorbid diabetes may help close an important gap in maternal health equity and in reducing preventable HTN‐related complications in the highest risk populations. In this secondary analysis, to explore the impact of multimorbidity in an hypertensive disorders of pregnancy population, we aimed to compare outcomes between postpartum patients in our remote SMBP program with versus without diabetes.

2. MATERIALS AND METHODS

Postpartum patients with any HTN diagnosis (including gestational hypertension, preeclampsia with and without severe features, chronic hypertension, chronic hypertension with superimposed preeclampsia, eclampsia) are approached in their postpartum hospital room at Women & Infants Hospital (Providence, RI) by the SMBP team (D.S. and M.M.C.) where they are provided with a calibrated automatic BP cuff, instructed on use, and given contact information. Emergent signs and symptoms of HTN, and reasons for ED evaluation are provided. Participants take their BP once daily for 6 weeks, which are entered via MyChart directly to their electronic medical record, during which the BPs are monitored by the SMBP team and antihypertensive medications are initiated and titrated by the Family Nurse Practitioner (D.S.) remotely to maintain BPs below 140/90 mmHg. If persistently severe range BPs are noted, or patients contact the SMBP staff with concerns for severe features, the FNP instructs that patient to present to the ED for evaluation. After 6 weeks postpartum, patients are discharged back to the care of their OB/GYN or primary care provider.

As part of a quality improvement initiative, data from the first 2000 participants in our SMBP program were abstracted from the medical record by two trained research assistants (N.K., W.H.) and manually entered into a REDCap database. A total of 10% of charts were randomly selected for auditing by a trained epidemiologist (T.J.) to confirm accuracy. For this planned analysis, all SMBP participants with a confirmed presence or absence of diabetes diagnosis were included, and stratified by pregestational or gestational. Diabetes diagnoses were defined by having an international classification of diseases‐10 code associated with both pregestational diabetes (E08‐E13) and gestational diabetes (O24.4) or a diagnosis of pregestational diabetes or gestational diabetes described in the admission or discharge notes. Due to the small number of pregestational diabetic patients included in our analytic population, pregestational and gestational diabetes patients were intentionally combined for analysis. Baseline characteristics (age, race and ethnicity, parity, gestational age at delivery, and education levels, among others) were compared. Continuous variables were compared using the t‐test or Mann–Whitney U test. Categorical variables were compared using Fisher's exact tests. The primary outcome was a composite of postpartum readmission or ED presentation for HTN within 30 days of delivery hospitalization, which was defined as having an elevated BP or preeclampsia symptom as the leading indication for readmission or presentation, respectively. Secondary outcomes included individual components of the primary composite outcome and HTN‐related SMM (stroke, seizure/eclampsia, acute fatty liver of pregnancy, posterior reversible encephalopathy syndrome, pulmonary edema, heart failure, hemolysis, elevated liver enzymes and low platelets syndrome, placental abruption, postpartum hemorrhage, acute kidney injury, and transaminitis) during or within 30 days of delivery hospitalization. A generalized linear model was used to estimate relative risks (RRs) after adjustment for differences in baseline characteristics. This study was approved by the Care New England—Women & Infants Hospital Institutional Review Board (#2187816).

3. RESULTS

Among the first 2000 people in the SMBP program, 1965 people had data on the presence or absence of concurrent diagnoses of diabetes and were included in this analysis. In total, 337 (17%) had diabetes (70 pregestational and 267 gestational). Compared to those without diabetes, those with diabetes were older (median 33, IQR [29, 37] vs. 31 IQR [27, 35], p < 0.001), less likely to identify as White (52.2% vs. 58.3%, p = 0.006), more likely to be delivered via cesarean (60.78% vs. 47.53%, p < 0.001), and had a lower median gestational age at delivery (37.7 vs. 38.1, p < 0.001) (Table 1). There were also differences in type of HTN diagnosis: those with diabetes were less likely to have gestational hypertension (19.6% vs. 30.3%, p < 0.001) but more likely to have chronic hypertension (22% vs. 12.5%, p < 0.001) as well as chronic hypertension with superimposed preeclampsia (11.0% vs. 5.3%, p < 0.001) compared to those without diabetes. Rates of preeclampsia with and without severe features and eclampsia were the same in both groups.

TABLE 1.

Demographics.

	With diabetes (n = 337)	Without diabetes (n = 1,628)	p value
Age	33 (29, 37)	31 (27, 35)	<0.001
Primary insurance type			0.55
Private	170 (50.45%)	868 (53.38%)
Medicaid/Medicare	165 (48.96%)	751 (46.19%)
None	2 (0.59)	7 (0.43%)
Race			0.006
White	176 (52.23%)	949 (58.29%)
Black	58 (17.21%)	305 (18.73%)
Multiracial or other	100 (29.67%)	342 (21.01%)
Unknown	3 (0.89%)	32 (1.97%)
Ethnicity			0.67
Hispanic	117 (34.72%)	526 (32.31%)
Non‐Hispanic	218 (64.69%)	1090 (55.47%)
Unknown	2 (0.59%)	12 (0.74%)
HDP diagnosis prior to delivery
Gestational hypertension	66 (19.58%)	494 (30.34%)	<0.001
Preeclampsia with severe features	57 (16.91%)	329 (20.21%)	0.18
Preeclampsia without severe features	47 (13.95%)	203 (12.47%)	0.47
Chronic hypertension with SIPE	37 (10.98%)	86 (5.28%)	<0.001
Eclampsia	0 (0.00%)	5 (0.31%)	0.59
Chronic Hypertension	74 (21.96%)	207 (12.71%)	<0.001
Gestational age (weeks) at delivery	37.7 (36.5, 39.0)	38.1 (36.9, 39.3)	<0.001
Mode of delivery			<0.001
Vaginal	131 (39.2%)	849 (52.49%)
Cesarean	203 (60.78%)	769 (47.53%)
Postpartum day of discharge	3 (2, 4)	3 (2, 4)	<0.001
Any BP submitted during the study period			0.76
Yes	194 (57.5%)	922 (56.6%)
No	143 (42.4%)	706 (43.4%)

Abbreviations: BP, blood pressure; HDP, hypertensive disorders of pregnancy; SIPE, superimposed preeclampsia.

Bold values statistically significance p values <0.05.

After adjusting for differences identified between study groups, there was no difference in the risk of the composite outcome of HTN‐related postpartum readmission or ED presentation between those with and without diabetes (15.9% vs. 16.6%; adjusted RR [aRR], 0.92 [0.70, 1.21]). Similarly, there were no differences in secondary outcomes: rates were similar between groups for ED visits due to hypertension (15.9% vs. 16.2%; aRR, 0.94 [0.72, 1.25]), hospitalization due to hypertension rates (9.1% vs. 10.0%; aRR, 0.85 [0.58, 1.24]), and SMM during or after delivery hospitalization (11.9% vs. 11.7%; aRR, 0.92 [0.64–1.31]) (Table 2). Rates of SMBP program participation, defined as any BP measurement entered into the patient's electronic medical record during the first 6 weeks postpartum, were assessed for effect measure modification and were similar between those with and without diabetes (57.6% vs. 56.6%, p = 0.76).

TABLE 2.

Outcomes.

	With diabetes (n = 337)	Without diabetes (n = 1,628)	p value	Relative risk (95% CI)	Adjusted RR (95% CI) a
Primary outcomes
ED visit or hospitalization for hypertension	53 (15.92%)	266 (16.56%)	0.81	0.96 (.73‐1.26)	0.92 (0.70, 1.21)
Secondary outcomes
ED visit due to hypertension	53 (15.92%)	261 (16.24%)	0.93	0.98 (0.75, 1.28)	0.94 (0.72, 1.25)
Hospitalization due to hypertension	30 (9.06%)	161 (10.04%)	0.69	0.90 (0.63, 1.31)	0.85 (0.58, 1.24)
Severe maternal morbidity (any) b	40 (11.87%)	191 (11.73%)	0.93	1.01 (0.74, 1.39)	0.92 (0.64–1.31)

Note: Data presented as n (%) or median (interquartile range) unless otherwise specified.

Abbreviations: CI, confidence interval; ED, emergency department; RR, relative risk.

^a Adjusted for age, race, and hypertensive diagnosis at discharge.

^b Includes stroke, seizure/eclampsia, acute fatty liver of pregnancy (AFLP), posterior reversible encephalopathy syndrome (PRES), pulmonary edema, heart failure, hemolysis, elevated liver enzymes and low platelets (HELLP) syndrome, placental abruption, postpartum hemorrhage, acute kidney injury, and transaminitis.

4. DISCUSSION

In this analysis of participants in our postpartum SMBP program, rates of unplanned HTN‐related postpartum readmission, HTN‐related ED presentation, and SMM were similar between those with versus without concurrent diagnoses of diabetes. We also confirmed via formal testing for effect modification that the effect of SMBP in terms of participation did not vary by diabetes status. While prior studies had found a decrease in unplanned healthcare utilization and SMM when comparing those who did versus did not participate in an SMBP program [5, 6], they did not specifically analyze whether the effect of these programs differed among those with comorbid conditions such as diabetes. Because the presence of both GDM and HTN has been shown to increase SMM and readmission rates postpartum [7, 8], our study provides insight into the fact that concurrent diabetes and HTN diagnoses do not appear to worsen rates of HTN‐related healthcare utilization and SMM in patients in SMBP programs.

Our findings differ from prior studies, which suggest that postpartum people with comorbid conditions such as diabetes and HTN are at higher likelihood of readmission [7] after discharge from childbirth hospitalization. This difference is likely because prior studies on this topic were not based on populations of SMBP program participants. Indeed, though uncontrolled HTN has been identified as the most common etiology of postpartum readmissions [9], a recent meta‐analysis concluded that SMBP program participation increases rates of recommended BP ascertainment after hospital discharge [5], allowing providers to detect mild elevations in BP and act on them prior to patient progression to severe, uncontrolled HTN. Our findings suggest that additional monitoring among patients with concurrent diabetes and HTN is not needed beyond what is offered in standard SMBP programs to prevent hospital readmissions or ED presentations related to HTN.

This study lays the foundation for additional investigations into the effect of SMBP programs on perinatal outcomes. Like many SMBP programs [6], our program stops 6 weeks after delivery, which prevents us from examining the effect that SMBP programs may have on long‐term outcomes such as cardiovascular risk among patients with versus without diabetes. Indeed, our outcomes were assessed within the first 30 days of childbirth. It is possible that SMBP program benefits may extend beyond 6 weeks postpartum. In addition, it is possible that SMBP participation improves postpartum glycemic control due to the increased number of contacts that SMBP participants have after hospital discharge with clinical staff. We were also unable to assess postpartum glucose control because in our hospital system, outpatient perinatal care is provided by more than 10 distinct prenatal care clinics that do not utilize a central electronic medical record. Given the association between SMBP participation and increased engagement with recommended postpartum care, it is also possible that engagement with SMBP is associated with improved glycemic control. Examining this association is beyond the scope of this project; thus, additional research is needed to clarify whether SMBP programs among pregestational diabetics alter glycemic control after delivery.

This study has many strengths. First, our study population reflects that of our opt‐out SMBP program, which enrolls a diverse participant population, with a high percentage of non‐White and Hispanic‐identifying individuals, and patients with public insurance. Therefore, because HTN and GDM disproportionately affect racial minorities and those of lower socioeconomic status [10], findings from this study are generalizable to many similar populations. Second, our hospital delivers more than 90% of patients with perinatal HTN in our state. Thus, findings from this study can be considered as statewide data. Third, our data were collected and analyzed prospectively, which adds robustness to our findings.

However, this study is not without limitations. First, we did not stratify by type of diabetes in our cohort. This intentional decision was because of the small number of patients with pregestational diabetes in the overall SMBP program. It is possible that there are differences between unplanned healthcare utilization or SMM between pregestational diabetes with versus without HTN that were masked by the larger population of patients with gestational diabetes. Thus, additional research specifically examining the association between comorbid pregestational diabetes and HTN may elucidate the association between SMBP and rates of healthcare utilization or SMM. Second, data on ED presentations directly related to diabetic complications were not abstracted as our study focused on HTN‐related outcomes. Lastly, unplanned healthcare utilization at other hospitals was not captured, which may underestimate our outcome.

In conclusion, in our remote SMBP program for postpartum patients with HTN, unplanned HTN‐related healthcare utilization or SMM was similar between those with and without diabetes. These findings provide reassurance that patients with comorbid conditions do not appear to have worse HTN‐related outcomes if they participate in a remote SMBP program.

FUNDING INFORMATION

The authors received no specific funding for this work.