Abstract
Background
Gestational diabetes mellitus (GDM) and hypertensive disorders of pregnancy (HDP) mark elevated postpartum cardiometabolic risk, but less is known about how recorded nursing-process indicators relate to screening completion within institutional postpartum follow-up pathways.
Methods
This single-center retrospective cohort used an institutional postpartum follow-up analytic cohort from Zibo Central Hospital. Records of women delivered from January 1, 2021 to December 31, 2025 were reviewed through 84 days postpartum, and inclusion required sufficient follow-up, clinical covariate, nursing-process, and screening-completion information. The primary outcome was completion of an institutional cardiometabolic screening package comprising recorded postpartum blood pressure, postpartum BMI, glucose-related testing, and lipid profile. Modified Poisson regression estimated risk ratios (RRs) with sequential adjustment for delivery year, clinical covariates, and recorded nursing contact score.
Results
Of 1286 candidate records, 800 women were included: 376 controls, 214 GDM only, 139 HDP only, and 71 coexisting GDM/HDP. Screening completion was 63.8%, 77.1%, 77.0%, and 88.7%, respectively. After delivery-year and clinical adjustment, GDM only (RR 1.18, 95% CI 1.05–1.33), HDP only (RR 1.18, 95% CI 1.03–1.35), and coexisting GDM/HDP (RR 1.37, 95% CI 1.20–1.57) had higher screening completion than controls. After adding nursing contact score, group associations attenuated; nursing contact score remained associated with screening completion (RR 1.11 per point, 95% CI 1.06–1.17). Among screened women, coexisting GDM/HDP showed the highest abnormal blood pressure, abnormal glucose, and BMI ≥30.0 kg/m2 rates.
Conclusion
Women with coexisting GDM/HDP had the highest documented screening completion and substantial observed cardiometabolic abnormality among screened records. Recorded nursing-process indicators were associated with screening completion within this selected institutional cohort, supporting pathway-level interpretation rather than causal intervention claims.
Keywords: gestational diabetes mellitus, hypertensive disorders of pregnancy, postpartum screening, cardiometabolic risk, nursing-process indicators, retrospective cohort
Introduction
Gestational diabetes mellitus (GDM) and hypertensive disorders of pregnancy (HDP) are common pregnancy complications that signal increased cardiometabolic risk after delivery. Current diabetes and obstetric guidance emphasizes postpartum glucose testing after GDM and a broader transition from a single postpartum visit to an ongoing care process tailored to maternal risk.1,2 Cardiovascular prevention statements also recognize adverse pregnancy outcomes as early-life opportunities for risk identification and prevention.3 Cohort studies and meta-analyses have linked GDM with later type 2 diabetes and cardiovascular disease,4–6 and HDP with future cardiovascular morbidity.7,8 The coexistence of GDM and gestational hypertensive disorders may identify a smaller but clinically important group with overlapping metabolic and vascular risk signals.9,10
These risks have biologic plausibility beyond epidemiologic association. Diabetes-related dysglycemia may promote endothelial dysfunction, oxidative stress, vascular inflammation, and accelerated atherosclerotic processes, while obesity-related inflammatory and inflammation-resolution pathways help contextualize BMI and lipid markers as part of broader cardiometabolic risk clustering.11,12 In women with coexisting GDM/HDP, metabolic stress and hypertensive vascular maladaptation may therefore converge, supporting postpartum assessment of glucose, blood pressure, BMI, and lipid profile as early risk-recognition markers rather than isolated obstetric follow-up items.
Despite these recommendations and risk data, postpartum screening and follow-up after GDM or HDP remain inconsistent in routine care. A nursing-focused systematic review reported persistent disparities in postpartum follow-up and risk screening after these complications,13 and health-system studies have shown incomplete postpartum diabetes testing even where guideline-recommended screening is available.14,15 Similar gaps exist after HDP, where ambulatory postpartum follow-up is an opportunity for early cardiovascular risk-factor detection but is not uniformly achieved.16 These gaps make the postpartum transition a critical interval in which risk recognition must be linked to practical follow-up, reminders, health education, monitoring guidance, and referral.
Women with coexisting GDM and HDP are particularly relevant to postpartum follow-up pathways because this subgroup combines metabolic and vascular pregnancy complications while also requiring coordinated post-discharge care. Routine clinical datasets often capture diagnoses and laboratory values, whereas recorded nursing-process indicators that help women return for screening are less frequently analyzed. Systems-level approaches and reminder interventions have been evaluated to improve postpartum diabetes screening,17–21 and text-message remote monitoring has been used to improve postpartum blood pressure ascertainment after hypertensive pregnancy complications.22 These interventions address different outcomes: glucose-test adherence, blood pressure ascertainment, and broader cardiometabolic package completion should not be treated as interchangeable. Evaluating early post-discharge follow-up, review reminders, health education, monitoring guidance, specialty clinic contact, follow-up frequency, and first review timing can clarify how high-risk obstetric patients remain connected to postpartum cardiometabolic assessment.
This study evaluated postpartum cardiometabolic screening completion and observed marker abnormalities among women with no GDM/HDP, GDM only, HDP only, and coexisting GDM/HDP in a selected institutional follow-up analytic cohort. It also examined whether recorded nursing-process indicators were associated with documented screening completion. The study was designed to describe an institutional follow-up pathway and its association with screening adherence, not to infer that nursing contact independently caused improved outcomes.
Methods
Study Design and Setting
This single-center retrospective cohort study was conducted at Zibo Central Hospital and was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidance for observational research.23 Candidate records were identified from the institutional postpartum follow-up database among women who delivered between January 1, 2021 and December 31, 2025 and had at least one postpartum follow-up-related record in the institutional system. Postpartum nursing follow-up and cardiometabolic screening records were reviewed through 84 days after delivery, with final follow-up on March 25, 2026. The study used de-identified institutional data from delivery records, obstetric diagnosis records, postpartum nursing follow-up records, and the laboratory information system.
This design was used to construct an analytic cohort for evaluating postpartum screening and recorded nursing-process indicators, rather than to estimate the population prevalence of GDM, HDP, or coexisting GDM/HDP among all deliveries. The final cohort should therefore be interpreted as an institutional postpartum follow-up analytic cohort rather than a prevalence-representative full delivery cohort.
Participants and Cohort Construction
A total of 1286 candidate postpartum follow-up records were initially identified from the institutional database. Before eligibility assessment, 139 records were excluded because of duplicate delivery or repeated records (n=32), delivery outside the study hospital or transfer before follow-up (n=41), missing essential delivery records (n=38), or an incomplete 84-day postpartum observation window (n=28). Among 1147 women eligible for detailed clinical assessment, 221 were excluded because of pre-existing diabetes mellitus (n=49), chronic hypertension before pregnancy or before 20 weeks of gestation (n=66), major cardiovascular, renal, hepatic, autoimmune, or malignant disease (n=37), missing or inconsistent GDM/HDP diagnostic information (n=31), or incomplete key clinical records (n=38).
Among 926 women with classifiable GDM/HDP status, a further 126 were excluded because of no usable postpartum nursing follow-up record (n=45), insufficient nursing-process variables (n=41), missing key covariates required for multivariable modeling (n=24), or inconsistent screening-completion status (n=16). The final analytic cohort included 800 women: 376 controls without GDM or HDP, 214 women with GDM only, 139 women with HDP only, and 71 women with coexisting GDM and HDP (Figure 1).
Figure 1.
Cohort construction flowchart. A total of 1286 candidate records were identified from the institutional postpartum follow-up database of Zibo Central Hospital between January 1, 2021 and December 31, 2025. After exclusions for duplicate or incomplete records, clinical ineligibility, missing GDM/HDP diagnostic information, insufficient nursing-process variables, missing key covariates, and inconsistent screening-completion status, 800 women were included in the final analytic cohort. The cohort should be interpreted as an institutional postpartum follow-up analytic cohort rather than a prevalence-representative full delivery cohort.
Patient-level baseline variables for records excluded before final analytic-cohort construction were not retained in the current analytic file; therefore, a full included-versus-excluded baseline comparison for these pre-analytic exclusions could not be calculated without unsupported reconstruction. Exclusion stages and baseline-comparability availability are summarized in Supplementary Table S1.
Definition of GDM and HDP Groups
Participants were categorized into four mutually exclusive groups according to documented obstetric diagnoses during the index pregnancy: control, GDM only, HDP only, and coexisting GDM/HDP. GDM was diagnosed in the institutional obstetric record based on a 75-g oral glucose tolerance test performed at 24–28 gestational weeks according to contemporaneous institutional criteria. The thresholds used in routine practice corresponded to one-step diagnostic criteria: fasting plasma glucose ≥5.1 mmol/L, 1-hour glucose ≥10.0 mmol/L, or 2-hour glucose ≥8.5 mmol/L.1,24 HDP was defined as new-onset hypertension after 20 weeks of gestation, including gestational hypertension and preeclampsia, according to contemporaneous obstetric diagnostic criteria.25 Women with pre-existing diabetes mellitus or chronic hypertension before pregnancy or before 20 gestational weeks were excluded. The analysis relied on recorded institutional diagnoses and did not centrally re-adjudicate GDM or HDP status.
Postpartum Nursing-Process Indicators
Seven recorded nursing-process indicators were analyzed: completion of telephone or SMS follow-up at 3–5 days after discharge, postpartum review reminder, health education, monitoring guidance for blood pressure, glucose, nutrition, activity, or weight management, specialty clinic or nurse-led follow-up contact, number of follow-up contacts, and days to first postpartum review. The nursing contact score was calculated as the sum of the first five binary nursing-process indicators and ranged from 0 to 5. It was treated as a pragmatic count of recorded nursing contacts, not as a validated psychometric scale, randomized intervention-intensity measure, or direct continuity-of-care construct. Follow-up count was analyzed separately as a contact-frequency variable. Days to first postpartum review was treated as a timing outcome because it is undefined for women without a recorded review.
Outcomes
The primary outcome was completion of an institutional postpartum cardiometabolic screening package within 84 days after delivery. In the analysis dataset, completed screening corresponded to recorded postpartum blood pressure, postpartum BMI assessment, glucose-related testing, and lipid profile. The source follow-up record captured these four fields as one scheduled postpartum review/laboratory bundle; the package was therefore recorded as complete when that bundle was documented and absent when no completed bundle was documented. Standalone partial testing outside the bundle was not separately captured in the analytic file. This package was an institutional follow-up outcome and should not be interpreted as direct replication of a single external guideline.
Secondary outcomes among women with completed screening were abnormal blood pressure, abnormal glucose metabolism, postpartum BMI ≥30.0 kg/m2, and continuous postpartum cardiometabolic markers. Abnormal blood pressure was defined as systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg. Abnormal glucose metabolism was defined as fasting plasma glucose ≥5.6 mmol/L, 2-hour oral glucose tolerance test glucose ≥7.8 mmol/L, or HbA1c ≥5.7%. The BMI outcome was labeled as postpartum obesity-level BMI and was defined as postpartum BMI ≥30.0 kg/m2; it should not be interpreted as the broader Chinese overweight threshold.
Covariates
Clinical covariates included delivery year, age, BMI at booking or the first prenatal visit, parity, gestational week at delivery, delivery mode, preterm birth, and multiple pregnancy. Delivery year was included as a categorical covariate to account for potential changes in postpartum follow-up systems from 2021 through 2025. Insulin use and severe HDP were summarized descriptively because they are closely related to exposure group and disease severity. The current analytic dataset contains a severe-HDP marker but does not separate gestational hypertension from preeclampsia; available severe-HDP distributions are reported in Supplementary Table S2.
Statistical Analysis
Continuous variables are presented as mean ± standard deviation for clinical interpretability. Group comparisons used Kruskal–Wallis tests for continuous variables because some variables were not normally distributed. Categorical variables are presented as n/N (%) and were compared using chi-square tests when feasible. Percentages used non-missing denominators. The missingness profile is reported in Supplementary Table S3.
No a priori sample-size or statistical power calculation was performed because this retrospective study used all eligible records available in the institutional analytic dataset. Precision was therefore interpreted in relation to the available subgroup sizes, particularly the coexisting GDM/HDP group.
Because screening completion was common, the primary analysis used modified Poisson regression with robust standard errors to estimate RRs, following established recommendations for binary outcomes where odds ratios may overstate relative effects.26 Three sequential models were fitted: Model 1 included exposure group only; Model 2 added delivery year and clinical covariates; Model 3 added nursing contact score. Model 2 was treated as the primary group-comparison model because recorded nursing-process indicators may be on the care pathway, reflect clinical prioritization, mark patient engagement or access, or operate as colliders rather than simple confounders. Model 3 was therefore interpreted as exploratory/pathway-adjusted rather than as a fully adjusted causal estimate. To clarify this interpretation, we added a directed acyclic graph that distinguishes baseline clinical covariates, GDM/HDP status, disease severity, recorded nursing-process indicators, screening completion, observed marker abnormalities, unmeasured healthcare-access or patient-engagement factors, and selection into the analytic follow-up cohort (Supplementary Figure S1). Supplementary analyses included the full Model 3 covariate table, model convergence and multicollinearity diagnostics, a categorical nursing contact score model, a model replacing the score with individual nursing indicators, and sensitivity analyses excluding preterm births, excluding severe HDP, restricting to GDM and/or HDP groups, and stratifying by delivery period (Supplementary Tables S4 and S5).
To address the construct concern for the nursing contact score, Cronbach alpha was calculated for the five binary nursing-process indicators.27 Because the score represented recorded process contacts rather than a latent psychometric construct, low internal consistency was interpreted as evidence for using the score as a count index and for emphasizing individual nursing indicators. A descriptive attenuation analysis compared group RRs before and after adding nursing contact score using the same complete-case dataset. This analysis was mechanism-supportive only and was not interpreted as causal mediation (Supplementary Table S5).
Missingness was handled according to the meaning of each variable. Regression models used complete cases for variables included in each model. Postpartum cardiometabolic marker fields were structurally unavailable for women without completed screening and were not imputed as normal values. Marker abnormality analyses were therefore restricted to screened women with available data. A screened-versus-unscreened comparison was performed to evaluate potential selection bias related to screening completion. Because the current analytic file records blood pressure, postpartum BMI, glucose testing, and lipid profile as a complete institutional review bundle, no partial discordant component-completion pattern was present. This all-or-none structure reflects the source follow-up record architecture and does not imply that partial clinical testing could not occur outside the captured bundle. All screened records had all four component categories recorded and all unscreened records lacked all four categories. Component-specific completion models would therefore duplicate the package-completion model rather than add independent information. Component availability and component-concordance patterns are summarized in Supplementary Table S6. Screening completion by delivery year and best-case and worst-case bounds for abnormal markers under missing-not-at-random assumptions are reported in Supplementary Tables S7 and S8.
Results
Cohort Characteristics
The analytic cohort contained 800 postpartum women: 376 controls (47.0%), 214 with GDM only (26.8%), 139 with HDP only (17.4%), and 71 with coexisting GDM and HDP (8.9%). Compared with controls, women in the coexisting GDM/HDP group were older (32.0 ± 4.6 vs 29.2 ± 4.1 years), had higher BMI at booking or the first prenatal visit (27.0 ± 3.6 vs 23.0 ± 3.2 kg/m2), delivered earlier (37.2 ± 1.8 vs 38.9 ± 1.2 weeks), and had higher proportions of cesarean delivery, insulin use, severe HDP, and preterm birth (Table 1).
Table 1.
Baseline Maternal and Pregnancy Characteristics According to GDM/HDP Status
| Characteristic | Control | GDM Only | HDP Only | GDM + HDP | P value |
|---|---|---|---|---|---|
| Age, years | 29.2 ± 4.1 (n=376) | 30.9 ± 3.8 (n=214) | 30.6 ± 3.4 (n=139) | 32.0 ± 4.6 (n=71) | <0.001 |
| BMI at booking or first prenatal visit, kg/m2 | 23.0 ± 3.2 (n=376) | 25.7 ± 3.7 (n=214) | 25.6 ± 3.3 (n=139) | 27.0 ± 3.6 (n=71) | <0.001 |
| Parity | 0.7 ± 0.7 (n=376) | 0.7 ± 0.7 (n=214) | 0.6 ± 0.7 (n=139) | 0.7 ± 0.7 (n=71) | 0.182 |
| Gestational week | 38.9 ± 1.2 (n=373) | 38.6 ± 1.3 (n=214) | 37.7 ± 1.6 (n=137) | 37.2 ± 1.8 (n=70) | <0.001 |
| Cesarean delivery | 121/371 (32.6%) | 102/207 (49.3%) | 90/136 (66.2%) | 57/71 (80.3%) | <0.001 |
| Birth weight, g | 3279.2 ± 415.9 (n=369) | 3397.9 ± 401.9 (n=208) | 3009.5 ± 424.7 (n=134) | 3067.0 ± 451.8 (n=69) | <0.001 |
| Insulin use | 0/376 (0.0%) | 27/214 (12.6%) | 0/139 (0.0%) | 13/71 (18.3%) | <0.001 |
| Severe HDP | 0/376 (0.0%) | 0/214 (0.0%) | 39/135 (28.9%) | 27/68 (39.7%) | <0.001 |
| Preterm birth | 20/370 (5.4%) | 19/212 (9.0%) | 31/138 (22.5%) | 22/71 (31.0%) | <0.001 |
| Multiple pregnancy | 14/376 (3.7%) | 4/214 (1.9%) | 2/139 (1.4%) | 2/71 (2.8%) | 0.416 |
Notes: Percentages use non-missing denominators. Continuous variables are mean ± SD.
Nursing-Process Indicators
Recorded nursing-process indicators differed substantially by group. The coexisting GDM/HDP group had the highest coverage of postpartum review reminders (78.6%), health education (92.6%), monitoring guidance (91.3%), and specialty clinic contact (64.3%). Three- to five-day follow-up was more common in the high-risk groups than in controls, with 69.0% coverage in the coexisting group. Mean nursing contact score increased from 2.4 ± 1.1 in controls to 4.0 ± 1.0 in the coexisting group. The first recorded postpartum review occurred earlier in the coexisting group than in controls (40.2 ± 9.3 vs 52.1 ± 9.4 days), and review within 42 days was more common (60.3% vs 16.7%) (Table 2).
Table 2.
Postpartum Nursing Follow-Up Indicators According to GDM/HDP Status
| Characteristic | Control | GDM Only | HDP Only | GDM + HDP | P value |
|---|---|---|---|---|---|
| 3–5 day follow-up | 169/372 (45.4%) | 137/210 (65.2%) | 94/137 (68.6%) | 49/71 (69.0%) | <0.001 |
| Postpartum review reminder | 225/374 (60.2%) | 158/210 (75.2%) | 100/139 (71.9%) | 55/70 (78.6%) | <0.001 |
| Health education | 216/369 (58.5%) | 152/210 (72.4%) | 92/136 (67.6%) | 63/68 (92.6%) | <0.001 |
| Monitoring guidance | 175/367 (47.7%) | 154/209 (73.7%) | 92/139 (66.2%) | 63/69 (91.3%) | <0.001 |
| Specialty clinic contact | 105/369 (28.5%) | 82/209 (39.2%) | 60/135 (44.4%) | 45/70 (64.3%) | <0.001 |
| Nursing contact score | 2.4 ± 1.1 (n=348) | 3.3 ± 1.0 (n=193) | 3.2 ± 1.0 (n=130) | 4.0 ± 1.0 (n=64) | <0.001 |
| Follow-up contacts, count | 2.2 ± 1.1 (n=372) | 3.0 ± 1.1 (n=210) | 3.0 ± 1.2 (n=136) | 3.5 ± 1.1 (n=71) | <0.001 |
| First review time, days | 52.1 ± 9.4 (n=240) | 49.4 ± 9.6 (n=165) | 45.4 ± 9.0 (n=107) | 40.2 ± 9.3 (n=63) | <0.001 |
| First review within 42 days | 40/240 (16.7%) | 34/165 (20.6%) | 44/107 (41.1%) | 38/63 (60.3%) | <0.001 |
| Nursing contact level | <0.001 | ||||
| Low (0–1) | 68/348 (19.5%) | 8/193 (4.1%) | 4/130 (3.1%) | 1/64 (1.6%) | |
| Moderate (2–3) | 224/348 (64.4%) | 103/193 (53.4%) | 75/130 (57.7%) | 18/64 (28.1%) | |
| High (4–5) | 56/348 (16.1%) | 82/193 (42.5%) | 51/130 (39.2%) | 45/64 (70.3%) |
Notes: Percentages use non-missing denominators. Continuous variables are mean ± SD. The P value for nursing contact level is the overall comparison across low, moderate, and high categories.
The nursing contact score was retained as a simple recorded-contact count. Cronbach alpha for the five binary nursing indicators was 0.158, supporting the decision not to interpret the score as a validated latent scale.
Screening Completion and Cardiometabolic Markers
Postpartum cardiometabolic screening completion was 63.8% in controls, 77.1% in the GDM-only group, 77.0% in the HDP-only group, and 88.7% in the coexisting GDM/HDP group (p<0.001; Figure 2). Among women with completed screening, the coexisting GDM/HDP group showed the highest observed rates of abnormal blood pressure (22.2%), abnormal glucose (36.5%), and postpartum BMI ≥30.0 kg/m2 (23.8%). These marker findings were conditional on screening completion and should not be interpreted as full-cohort prevalence estimates (Table 3).
Figure 2.
Postpartum screening completion by GDM/HDP group and nursing contact level. Bars show screening completion percentages in the final analytic cohort. Nursing contact level was categorized as low, moderate, and high using the five-item recorded-contact count.
Table 3.
Postpartum Cardiometabolic Screening Completion and Abnormal Marker Detection
| Characteristic | Control | GDM Only | HDP Only | GDM + HDP | P value |
|---|---|---|---|---|---|
| Screening complete in cohort | 240/376 (63.8%) | 165/214 (77.1%) | 107/139 (77.0%) | 63/71 (88.7%) | <0.001 |
| Abnormal blood pressure among screened | 7/240 (2.9%) | 4/165 (2.4%) | 19/107 (17.8%) | 14/63 (22.2%) | <0.001 |
| Abnormal glucose among screened | 1/240 (0.4%) | 36/165 (21.8%) | 7/107 (6.5%) | 23/63 (36.5%) | <0.001 |
| Postpartum BMI ≥30.0 kg/m2 among screened | 5/240 (2.1%) | 30/165 (18.2%) | 12/107 (11.2%) | 15/63 (23.8%) | <0.001 |
| Postpartum BMI, kg/m2 | 23.6 ± 3.3 (n=240) | 26.5 ± 3.7 (n=165) | 26.1 ± 3.4 (n=107) | 27.5 ± 3.7 (n=63) | <0.001 |
| Systolic blood pressure, mmHg | 116.5 ± 8.9 (n=240) | 120.7 ± 9.3 (n=165) | 128.4 ± 9.2 (n=107) | 130.2 ± 8.3 (n=63) | <0.001 |
| Diastolic blood pressure, mmHg | 75.5 ± 6.4 (n=240) | 76.9 ± 6.2 (n=165) | 81.1 ± 6.3 (n=107) | 82.1 ± 5.3 (n=63) | <0.001 |
| Fasting plasma glucose, mmol/L | 4.8 ± 0.3 (n=240) | 5.2 ± 0.3 (n=165) | 5.0 ± 0.3 (n=107) | 5.2 ± 0.3 (n=63) | <0.001 |
| 2-hour OGTT glucose, mmol/L | 6.0 ± 0.6 (n=240) | 7.0 ± 0.6 (n=165) | 6.3 ± 0.7 (n=107) | 7.1 ± 0.7 (n=63) | <0.001 |
| HbA1c, % | 5.2 ± 0.2 (n=240) | 5.4 ± 0.2 (n=165) | 5.3 ± 0.2 (n=107) | 5.5 ± 0.2 (n=63) | <0.001 |
| Total cholesterol, mmol/L | 4.6 ± 0.5 (n=240) | 4.9 ± 0.5 (n=165) | 5.2 ± 0.4 (n=107) | 5.3 ± 0.5 (n=63) | <0.001 |
| Triglycerides, mmol/L | 1.3 ± 0.3 (n=240) | 1.5 ± 0.4 (n=165) | 1.6 ± 0.3 (n=107) | 1.9 ± 0.4 (n=63) | <0.001 |
| LDL-C, mmol/L | 2.6 ± 0.4 (n=240) | 2.8 ± 0.4 (n=165) | 2.9 ± 0.4 (n=107) | 3.1 ± 0.4 (n=63) | <0.001 |
| HDL-C, mmol/L | 1.5 ± 0.2 (n=240) | 1.3 ± 0.2 (n=165) | 1.3 ± 0.1 (n=107) | 1.2 ± 0.2 (n=63) | <0.001 |
Notes: Percentages use non-missing denominators. Marker summaries are restricted to screened women with available values.
Primary Modified Poisson Models
In the group-only modified Poisson model, GDM only (RR 1.21, 95% CI 1.08–1.34), HDP only (RR 1.21, 95% CI 1.07–1.36), and coexisting GDM/HDP (RR 1.39, 95% CI 1.24–1.56) were associated with higher screening completion compared with controls. After adjustment for delivery year and clinical covariates, the associations remained positive: GDM only (RR 1.18, 95% CI 1.05–1.33), HDP only (RR 1.18, 95% CI 1.03–1.35), and coexisting GDM/HDP (RR 1.37, 95% CI 1.20–1.57).
After adding nursing contact score in the exploratory/pathway-adjusted model, group associations attenuated and no longer reached conventional statistical significance: GDM only (RR 1.07, 95% CI 0.94–1.21), HDP only (RR 1.07, 95% CI 0.93–1.23), and coexisting GDM/HDP (RR 1.15, 95% CI 0.98–1.35). Nursing contact score remained associated with screening completion (RR 1.11 per one-point increase, 95% CI 1.06–1.17) (Table 4 and Figure 3). In the same complete-case dataset, adding nursing contact score attenuated the log-RR for coexisting GDM/HDP by approximately 54.4%, supporting a care-pathway interpretation without establishing causal mediation. Model 2 and Model 3 converged, and maximum column-level variance inflation factors were 2.99 and 3.00, respectively (Supplementary Table S4).
Table 4.
Sequential Modified Poisson Models for Postpartum Cardiometabolic Screening Completion
| Model | Term | Adjusted RR (95% CI) | P value |
|---|---|---|---|
| Model 1: group only | GDM only vs control | 1.21 (1.08–1.34) | <0.001 |
| Model 1: group only | HDP only vs control | 1.21 (1.07–1.36) | 0.002 |
| Model 1: group only | GDM + HDP vs control | 1.39 (1.24–1.56) | <0.001 |
| Model 2: year + clinical covariates | GDM only vs control | 1.18 (1.05–1.33) | 0.005 |
| Model 2: year + clinical covariates | HDP only vs control | 1.18 (1.03–1.35) | 0.014 |
| Model 2: year + clinical covariates | GDM + HDP vs control | 1.37 (1.20–1.57) | <0.001 |
| Model 3: model 2 + nursing contact score | GDM only vs control | 1.07 (0.94–1.21) | 0.301 |
| Model 3: model 2 + nursing contact score | HDP only vs control | 1.07 (0.93–1.23) | 0.330 |
| Model 3: model 2 + nursing contact score | GDM + HDP vs control | 1.15 (0.98–1.35) | 0.077 |
| Model 3: model 2 + nursing contact score | Nursing contact score, per 1-point increase | 1.11 (1.06–1.17) | <0.001 |
Notes: Models use robust standard errors. Model 2 adjusts for delivery year, age, BMI at booking or the first prenatal visit, parity, gestational week, delivery mode, preterm birth, and multiple pregnancy. Model 3 adds nursing contact score.
Figure 3.
Adjusted risk ratios for postpartum screening completion. Risk ratios were estimated using modified Poisson regression with robust standard errors. Group-effect estimates are shown from the sequential group-only, clinical-adjustment, and exploratory/pathway-adjusted models; the nursing contact score estimate is from the model that additionally included the recorded-contact count.
Missingness and Sensitivity Analyses
All 575 screened women had recorded postpartum blood pressure, glucose-related testing, lipid markers, and postpartum BMI values, while these markers were unavailable for 225 unscreened women. Compared with unscreened women, screened women had higher BMI at booking or the first prenatal visit (24.8 ± 3.6 vs 23.7 ± 3.7 kg/m2), higher nursing contact scores (3.1 ± 1.1 vs 2.5 ± 1.1), more follow-up contacts (2.8 ± 1.2 vs 2.5 ± 1.1), and higher coverage of 3–5 day follow-up, reminders, health education, and monitoring guidance (Table 5). This pattern supports non-random marker availability and screening-related selection. Screening completion varied by delivery year from 65.4% to 75.3% (Supplementary Table S7). Under extreme missing-not-at-random bounds, full-cohort abnormality estimates ranged from 5.5% to 33.6% for abnormal blood pressure, 8.4% to 36.5% for abnormal glucose metabolism, and 7.8% to 35.9% for postpartum BMI ≥30.0 kg/m2 (Supplementary Table S8).
Table 5.
Screened versus Unscreened Comparison
| Characteristic | Screened | Unscreened | P value |
|---|---|---|---|
| Age, years | 30.2 ± 4.2 (n=575) | 30.1 ± 3.7 (n=225) | 0.870 |
| BMI at booking or first prenatal visit, kg/m2 | 24.8 ± 3.6 (n=575) | 23.7 ± 3.7 (n=225) | <0.001 |
| Parity | 0.7 ± 0.7 (n=575) | 0.7 ± 0.7 (n=225) | 0.310 |
| Gestational week | 38.4 ± 1.5 (n=569) | 38.6 ± 1.4 (n=225) | 0.143 |
| Birth weight, g | 3236.9 ± 448.7 (n=559) | 3268.1 ± 412.2 (n=221) | 0.355 |
| Nursing contact score | 3.1 ± 1.1 (n=532) | 2.5 ± 1.1 (n=203) | <0.001 |
| Follow-up contacts, count | 2.8 ± 1.2 (n=567) | 2.5 ± 1.1 (n=222) | <0.001 |
| Insulin use | 34/575 (5.9%) | 6/225 (2.7%) | 0.087 |
| Severe HDP | 51/568 (9.0%) | 15/225 (6.7%) | 0.358 |
| Preterm birth | 71/568 (12.5%) | 21/223 (9.4%) | 0.274 |
| Multiple pregnancy | 12/575 (2.1%) | 10/225 (4.4%) | 0.111 |
| 3–5 day follow-up | 344/569 (60.5%) | 105/221 (47.5%) | 0.001 |
| Postpartum review reminder | 409/572 (71.5%) | 129/221 (58.4%) | <0.001 |
| Health education | 404/565 (71.5%) | 119/218 (54.6%) | <0.001 |
| Monitoring guidance | 373/564 (66.1%) | 111/220 (50.5%) | <0.001 |
| Specialty clinic contact | 218/561 (38.9%) | 74/222 (33.3%) | 0.174 |
Notes: Percentages use non-missing denominators. Continuous variables are mean ± SD.
The categorical nursing score model showed higher screening completion for moderate contact (RR 1.31, 95% CI 1.04–1.64) and high contact (RR 1.51, 95% CI 1.20–1.91) compared with low contact. In the model with individual nursing indicators, 3–5 day follow-up (RR 1.11, 95% CI 1.01–1.23), postpartum review reminder (RR 1.16, 95% CI 1.04–1.29), health education (RR 1.18, 95% CI 1.06–1.32), and monitoring guidance (RR 1.14, 95% CI 1.03–1.27) were associated with screening completion, while specialty clinic contact was not (RR 0.99, 95% CI 0.90–1.09). Among screened women classified with abnormal glucose metabolism, 2-hour OGTT most commonly drove the classification (31/67, 46.3%), followed by HbA1c (26/67, 38.8%) and fasting plasma glucose (16/67, 23.9%). Sensitivity analyses excluding preterm births, excluding severe HDP, restricting to GDM and/or HDP groups, and stratifying by delivery period produced directionally consistent positive associations between nursing contact score and screening completion (Supplementary Table S5).
Discussion
In this single-center retrospective cohort based on an institutional postpartum follow-up database, women with coexisting GDM and HDP had the highest documented postpartum cardiometabolic screening completion and the highest observed marker abnormality among screened records. They also had more documented nursing follow-up contacts and earlier first postpartum review. These findings support the clinical relevance of coexisting GDM/HDP as a postpartum cardiometabolic risk and follow-up pathway priority, while avoiding the stronger claim that nursing contact independently caused screening completion or improved cardiometabolic outcomes.
The clinical screening package has a biologic rationale but remains a screening construct. Glucose abnormalities may identify residual dysglycemia and diabetes-related endothelial injury, blood-pressure ascertainment captures vascular dysfunction after HDP, and BMI/lipid measures reflect adiposity-related inflammatory and atherogenic risk clustering.11,12 Coexisting GDM/HDP may concentrate these pathways, but the present study measured early postpartum screening completion and observed marker abnormalities only; it did not measure endothelial biomarkers or long-term atherosclerotic events.
The revised cohort construction is important for interpretation. The study did not start from the full delivery registry to estimate population prevalence. Instead, it evaluated candidate postpartum follow-up records with sufficient clinical, nursing-process, covariate, and screening-completion information. This framing directly limits generalizability but is better aligned with the research question, which concerns how documented nursing-process indicators are linked to screening completion within an institutional follow-up pathway. Because patient-level records for excluded women were not available in the current analytic file, a full included-versus-excluded comparison could not be calculated without unsupported reconstruction. We therefore summarized exclusion stages and baseline-comparability availability rather than imputing or reconstructing excluded-record characteristics (Supplementary Table S1).
The sequential modified Poisson models are central to interpretation. Group associations were positive after adjustment for delivery year and clinical covariates in Model 2, which is the primary group-comparison model. Those associations attenuated after nursing contact score was added in Model 3, which is best interpreted as exploratory and pathway-adjusted. The directed acyclic graph clarifies why Model 2 was retained as the primary group-comparison model and why Model 3 was treated as pathway-adjusted rather than as a fully confounder-adjusted causal model. This pattern is consistent with several explanations: higher-risk women may have been more actively prioritized for follow-up, documented contacts may reflect patient engagement or access, and some nursing-process indicators may lie on the pathway between clinical risk recognition and screening completion. The descriptive attenuation analysis supports a care-pathway interpretation but should not be described as causal mediation because nursing contact was not randomized and may reflect clinician prioritization, disease severity, patient engagement, and follow-up-system access.
The score-level and individual-indicator analyses give more actionable nursing information than the total score alone. Health education, postpartum review reminders, monitoring guidance, and 3–5 day follow-up each remained positively associated with screening completion when modeled together, whereas specialty clinic contact did not. This suggests that broad, repeatable nursing-process actions that reach many women may be more strongly linked to screening completion than specialty-clinic contact, which may occur later or be reserved for selected patients. These associations remain observational and should be treated as hypotheses for prospective pathway evaluation.
The screened-versus-unscreened comparison clarifies the missing-data problem. Marker fields were structurally unavailable among unscreened women, and screened women differed from unscreened women in BMI and documented nursing-process indicators. Marker abnormality rates may be overestimated if clinically high-risk women were more likely to return for screening, but they may also be underestimated if high-risk women without screening were lost to follow-up. The best-case and worst-case bounds show how wide the possible full-cohort range could be under extreme assumptions. For this reason, abnormal blood pressure, glucose, and postpartum BMI ≥30.0 kg/m2 results should be interpreted as observed abnormality among screened records, not population prevalence in the full analytic cohort.
From a nursing practice perspective, the results support an integrated postpartum pathway rather than a single isolated intervention (Figure 4). A feasible pathway would identify GDM/HDP status before discharge, provide 3–5 day post-discharge contact, send explicit postpartum screening reminders, reinforce home blood pressure, glucose, weight, and lifestyle monitoring, and connect women with abnormal findings to diabetes, hypertension, midwifery, or nurse-led postpartum clinics. Glucose screening adherence, blood pressure ascertainment, and complete cardiometabolic package completion may respond to different implementation strategies, so future pathway evaluations should measure these components separately. In the current analytic file, however, component availability was concordant with package completion because the source data captured the four marker categories through one completed institutional review bundle: no record showed completion of only blood pressure, only glucose testing, only postpartum BMI assessment, or only lipid testing. The most defensible conclusion is that recorded nursing-process indicators can help identify whether high-risk women remain connected to screening, while prospective studies are needed to test causal effects on long-term cardiometabolic outcomes.
Figure 4.
Conceptual postpartum care pathway for cardiometabolic screening. The pathway summarizes risk identification before discharge, early follow-up, reminders, health education, monitoring guidance, screening completion, abnormal-marker recognition, and referral or continued management. It is a conceptual interpretation informed by the study findings and prior postpartum follow-up literature, not an empirically validated causal pathway.
Limitations
Several limitations should be emphasized. First, the final cohort was derived from candidate records in an institutional postpartum follow-up database rather than the full delivery registry; therefore, the group distribution should not be interpreted as the natural prevalence of GDM, HDP, or coexisting GDM/HDP among all deliveries. Second, exclusions related to follow-up record availability, nursing-process variable completeness, covariate availability, and screening-completion consistency may introduce selection bias and limit generalizability, and patient-level excluded-record data were not available for a full included-versus-excluded comparison. Third, this was a single-center retrospective study and requires external validation in other healthcare settings. Fourth, nursing contact was not randomized and was likely influenced by baseline risk, clinician prioritization, patient engagement, healthcare access, education, insurance, and other unmeasured social determinants of follow-up. Fifth, marker fields were missing by design among unscreened women, so marker abnormality analyses were conditional on screening completion and susceptible to missing-not-at-random mechanisms. Sixth, the screening package had no partial component discordance in the analytic file because the source records captured one completed review bundle rather than standalone component orders; therefore, the study cannot distinguish women who completed only glucose testing, only blood pressure ascertainment, only postpartum BMI assessment, or only lipid testing. Seventh, the nursing contact score was an unvalidated count index with low internal consistency, so individual nursing indicators should be considered alongside the score. Eighth, postpartum BMI ≥30.0 kg/m2 identifies obesity-level BMI and does not capture the broader overweight threshold used in some Chinese adult criteria. Ninth, the coexisting GDM/HDP subgroup included 71 women, so some subgroup estimates may be imprecise. Finally, long-term cardiovascular events were not available, so the study cannot evaluate whether documented nursing follow-up reduces later cardiovascular disease.
Conclusion
In this institutional postpartum follow-up analytic cohort, women with coexisting GDM and HDP had the highest documented screening completion and the highest observed cardiometabolic marker abnormality among screened records. Nursing contact score and several recorded nursing-process indicators were associated with documented screening completion and may help identify modifiable points in postpartum care pathways. Prospective pathway evaluations in broader populations are required to determine causal effects and long-term cardiometabolic benefits.
Acknowledgments
Xiang Zhao and Yan Wei are co-first authors, contributed equally to this work.
Funding Statement
The work was not supported by any funding.
Data Sharing Statement
The study used de-identified institutional clinical and nursing follow-up data. Individual-level data are not publicly shared because they derive from hospital records. De-identified analytic code and aggregate tables may be made available from the corresponding author subject to institutional approval.
Ethics Approval and Consent to Participate
This retrospective study was approved by the Ethics Committee of Zibo Central Hospital (Approval No. YXLL2025R041). The requirement for individual informed consent was waived because the study used de-identified retrospective clinical and nursing follow-up data.
Disclosure
The authors declared that they have no conflicts of interest regarding this work.
References
- 1.Bajaj M, McCoy RG, Balapattabi K.; American Diabetes Association Professional Practice Committee for Diabetes. 15. Management of diabetes in pregnancy: standards of care in diabetes-2026. Diab Care. 2026;49(suppl 1):S321–13. doi: 10.2337/dc26-S015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.American College of Obstetricians and Gynecologists. ACOG committee opinion No. 736: optimizing postpartum care. Obstet Gynecol. 2018;131(5):e140–e150. doi: 10.1097/AOG.0000000000002633 [DOI] [PubMed] [Google Scholar]
- 3.Parikh NI, Gonzalez JM, Anderson CAM, et al. Adverse pregnancy outcomes and cardiovascular disease risk: unique opportunities for cardiovascular disease prevention in women: a scientific statement from the American heart association. Circulation. 2021;143(18):e902–e916. doi: 10.1161/CIR.0000000000000961 [DOI] [PubMed] [Google Scholar]
- 4.Kramer CK, Campbell S, Retnakaran R. Gestational diabetes and the risk of cardiovascular disease in women: a systematic review and meta-analysis. Diabetologia. 2019;62(6):905–914. doi: 10.1007/s00125-019-4840-2 [DOI] [PubMed] [Google Scholar]
- 5.Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes mellitus after gestational diabetes: a systematic review and meta-analysis. Lancet. 2009;373(9677):1773–1779. doi: 10.1016/S0140-6736(09)60731-5 [DOI] [PubMed] [Google Scholar]
- 6.Feig DS, Zinman B, Wang X, Hux JE. Risk of development of diabetes mellitus after diagnosis of gestational diabetes. CMAJ. 2008;179(3):229–234. doi: 10.1503/cmaj.080012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Wu P, Haththotuwa R, Kwok CS, et al. Preeclampsia and future cardiovascular health: a systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes. 2017;10(2):e003497. doi: 10.1161/CIRCOUTCOMES.116.003497 [DOI] [PubMed] [Google Scholar]
- 8.Lo CCW, Lo ACQ, Leow SH, et al. Future cardiovascular disease risk for women with gestational hypertension: a systematic review and meta-analysis. J Am Heart Assoc. 2020;9(13):e013991. doi: 10.1161/JAHA.119.013991 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Echouffo Tcheugui JB, Guan J, Fu L, Retnakaran R, Shah BR. Association of concomitant gestational hypertensive disorders and gestational diabetes with cardiovascular disease. JAMA Network Open. 2022;5(11):e2243618. doi: 10.1001/jamanetworkopen.2022.43618 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ackerman-Banks CM, Lipkind HS, Palmsten K, Ahrens KA. Association between hypertensive disorders of pregnancy and cardiovascular diseases within 24 months after delivery. Am J Obstet Gynecol. 2023;229(1):65.e1–65.e15. doi: 10.1016/j.ajog.2023.04.006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Karakasis P, Theofilis P, Patoulias D, Vlachakis PK, Antoniadis AP, Fragakis N. Diabetes-driven atherosclerosis: updated mechanistic insights and novel therapeutic strategies. Int J Mol Sci. 2025;26(5):2196. doi: 10.3390/ijms26052196 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Karakasis P, Stachteas P, Iliakis P, et al. Inflammation and resolution in obesity-related cardiovascular disease. Int J Mol Sci. 2026;27(1):535. doi: 10.3390/ijms27010535 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Jones EJ, Hernandez TL, Edmonds JK, Ferranti EP. Continued disparities in postpartum follow-up and screening among women with gestational diabetes and hypertensive disorders of pregnancy: a systematic review. J Perinat Neonatal Nurs. 2019;33(2):136–148. doi: 10.1097/JPN.0000000000000399 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ferrara A, Peng T, Kim C. Trends in postpartum diabetes screening and subsequent diabetes and impaired fasting glucose among women with histories of gestational diabetes mellitus: a report from the translating research into action for diabetes (TRIAD) study. Diab Care. 2009;32(2):269–274. doi: 10.2337/dc08-1184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Brown SD, Hedderson MM, Zhu Y, et al. Uptake of guideline-recommended postpartum diabetes screening among diverse women with gestational diabetes: associations with patient factors in an integrated health system in USA. BMJ Open Diabetes Res Care. 2022;10(3):e002726. doi: 10.1136/bmjdrc-2021-002726 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lewey J, Levine LD, Yang L, Triebwasser JE, Groeneveld PW. Patterns of postpartum ambulatory care follow-up care among women with hypertensive disorders of pregnancy. J Am Heart Assoc. 2020;9(17):e016357. doi: 10.1161/JAHA.120.016357 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Yarrington C, Zera C. Health systems approaches to diabetes screening and prevention in women with a history of gestational diabetes. Curr Diab Rep. 2015;15(12):114. doi: 10.1007/s11892-015-0687-1 [DOI] [PubMed] [Google Scholar]
- 18.Clark HD, Graham ID, Karovitch A, Keely EJ. Do postal reminders increase postpartum screening of diabetes mellitus in women with gestational diabetes mellitus? A randomized controlled trial. Am J Obstet Gynecol. 2009;200(6):634.e1–634.e7. doi: 10.1016/j.ajog.2009.01.003 [DOI] [PubMed] [Google Scholar]
- 19.Stasenko M, Liddell J, Cheng YW, et al. Patient counseling increases postpartum follow-up in women with gestational diabetes mellitus. Am J Obstet Gynecol. 2011;204(6):522.e1–522.e6. doi: 10.1016/j.ajog.2011.01.057 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Khalifeh A, Khosla J, Cantor E, et al. A reminder system for postpartum diabetes screening after gestational diabetes: a randomized clinical trial. J Matern Fetal Neonatal Med. 2022;35(25):7518–7522. doi: 10.1080/14767058.2021.1951212 [DOI] [PubMed] [Google Scholar]
- 21.Sarmiento AJ, Bernardo DC, Isip-Tan IT. A randomized controlled trial on the effectiveness of short message service (SMS) reminders in improving postpartum follow-up among gestational diabetes mellitus patients. J ASEAN Fed Endocr Soc. 2019;34(1):62–72. doi: 10.15605/jafes.034.01.10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hirshberg A, Sammel MD, Srinivas SK. Text message remote monitoring reduced racial disparities in postpartum blood pressure ascertainment. Am J Obstet Gynecol. 2019;221(3):283–285. doi: 10.1016/j.ajog.2019.05.011 [DOI] [PubMed] [Google Scholar]
- 23.von Elm E, Altman DG, Egger M, et al; STROBE Initiative. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Bull World Health Organ. 2007;85(11):867–872. doi: 10.2471/BLT.07.045120 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Metzger BE, Gabbe SG, Persson B; International Association of Diabetes and Pregnancy Study Groups Consensus Panel. International Association of Diabetes and Pregnancy Study Groups recommendations on the diagnosis and classification of hyperglycemia in pregnancy. Diab Care. 2010;33(3):676–682. doi: 10.2337/dc09-1848 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.American College of Obstetricians and Gynecologists. Gestational hypertension and preeclampsia: ACOG Practice Bulletin, Number 222. Obstet Gynecol. 2020;135(6):e237–e260. doi: 10.1097/AOG.0000000000003891 [DOI] [PubMed] [Google Scholar]
- 26.Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702–706. doi: 10.1093/aje/kwh090 [DOI] [PubMed] [Google Scholar]
- 27.Cronbach LJ. Coefficient alpha and the internal structure of tests. Psychometrika. 1951;16(3):297–334. doi: 10.1007/BF02310555 [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The study used de-identified institutional clinical and nursing follow-up data. Individual-level data are not publicly shared because they derive from hospital records. De-identified analytic code and aggregate tables may be made available from the corresponding author subject to institutional approval.