Abstract
BACKGROUND:
The association between pregnant patients participating in obstetrical studies and pregnancy outcomes is understudied.
OBJECTIVE:
This study aimed to evaluate the association between participation in obstetrical studies and maternal and neonatal outcomes.
STUDY DESIGN:
This was a retrospective cohort study of all patients delivering at a single center from 2013 to 2018. Patients with pregnancy loss at <13 weeks of gestation or major fetal anomalies were excluded. Patients who enrolled in one or more obstetrical studies were categorized as “study participants” and were compared with patients who did not enroll in an obstetrical study, that is, “study nonparticipants.” The primary outcome was a composite of maternal morbidity diagnosed up to 6 weeks after delivery. The secondary outcomes included composite neonatal morbidity and other perinatal outcomes. Bivariate analyses compared baseline demographics and outcomes between groups. Multivariable logistic regression was used to estimate adjusted odds ratios with 95% confidence intervals. Subgroup analyses by study design (trial or observational) were planned.
RESULTS:
Of 19,569 patients included in this analysis, 3848 (19.7%) were study participants, and 15,721 (80.3%) were study nonparticipants. Among study participants, 3023 (78.6%) enrolled in a trial, and 825 (21.4%) enrolled in an observational study. The study participants had higher body mass index and were more likely to be younger, non-Hispanic Black, publicly insured, nulliparous, and undergo cesarean delivery than study nonparticipants. Compared with study nonparticipants, the study participants had significantly lower odds of composite maternal morbidity (9.2% vs 8.7%; adjusted odds ratio, 0.83; 95% confidence interval, 0.73 −0.95) and composite neonatal morbidity (27.5% vs 18.6%; adjusted odds ratio, 0.53; 95% confidence interval, 0.48–0.58). In addition, the odds of fetal death, 5-minute Apgar score of <5, neonatal death, maternal and neonatal intensive care unit admissions, and lengths of stay were lower for study participants than for study nonparticipants. In stratified analyses, maternal morbidity was only significantly decreased among trial participants; however, there was a significantly reduced odds of neonatal morbidity, regardless of study design (trial or observational vs no study).
CONCLUSION:
Participation in obstetrical studies was associated with decreased maternal and neonatal morbidities after adjusting for potential confounders. This underscored the importance of pregnant patients enrolling in obstetrical clinical studies and potentially benefiting from the additional surveillance. Further study of how study participation exerts this effect on pregnancy outcomes is warranted.
Keywords: cesarean delivery, fetal death, maternal death, maternal intensive care unit admission, neonatal death, observational study, obstetrical trial, postpartum readmission
Clinical research in the field of obstetrics is aimed ultimately to benefit the patient, the fetus and/or neonate, or both. Pregnant patients may be hesitant to participate in voluntary obstetrical clinical research for a variety of reasons, such as mistrust of the medical system, unknown short- or longterm effect of a particular treatment on the fetus or fertility, or other cultural or social beliefs, whereas others may strongly desire participation in a study believing that they are contributing to a scientific pursuit or obtaining care that could be advantageous for themselves or their fetus.1,2 Previous analyses and a systematic review of the effects of participation in research trials among the general population, including oncologic trials, have demonstrated an overall weak positive effect on patient outcomes compared with the effects of the clinical intervention being offered.3 However, in a Cochrane database review, the effect of participation in trials vs nonparticipation did not consistently show benefit in a broad population, including medical, surgical, oncologic, pediatric, psychiatric, and obstetrics and gynecologic patients.4
There are limited data among pregnant patients evaluating the impact of participation in obstetrical studies on overall perinatal morbidity outcomes. In a recent 2017 systematic review and meta-analysis of 21 randomized control trials in obstetrical (n=11) and gynecologic (n=10) patients, studies were included for analysis if they reported the same clinical outcomes for their participants and their comparable nonparticipants outside of the trial in the same publication. They demonstrated that participants experienced an average 25% increased odds of improved trial outcomes compared with nonparticipants, regardless of the efficacy of the trial therapy.5 Notably, that analysis was mixed in its composition of obstetrical and gynecologic studies and included only randomized controlled trials (RCTs) and not observational studies. In addition, given the heterogeneity of included studies, the outcomes evaluated were trial-specific outcome measures and not general morbidity. For example, if the primary trial assessed postpartum hemorrhage, study participants were compared with study nonparticipants concerning postpartum hemorrhage and not overall pregnancy benefit. Given the paucity of data on participation in clinical obstetrical studies, we aimed to evaluate the impact of obstetrical study participation on both maternal and neonatal outcomes. We hypothesized that obstetrical study participants will have improved maternal and neonatal outcomes compared with their contemporaneous study nonparticipants.
Materials and Methods
We conducted a retrospective cohort study of all patients delivering at our tertiary, university-based hospital between January 1, 2013, and December 31, 2018. Institutional review board (IRB) approval was obtained before study initiation (IRB-300007509). The Strengthening the Reporting of Observational Studies in Epidemiology guidelines for cohort studies were followed.
Patients who delivered a singleton or multiple pregnancy at ≥13 0/7 weeks estimated gestational age during the study period were included. If a patient had multiple pregnancies within the period, only the first pregnancy was included. Patients were excluded if they had a fetal loss before 13 weeks of gestation, missing estimated gestational age at delivery, or known major fetal anomalies.
Patients who enrolled in at least 1 obstetrical study were categorized as “study participants,” whereas patients who did not enroll in any study were categorized as “study nonparticipants.” Obstetrical studies were characterized by type: trial or observational design (eg, prospective cohort). Trials included those with an intervention compared with no intervention, whereas observational studies had no intervention. Furthermore, studies were further classified by time point of patient enrollment: outpatient antepartum (prenatal clinics), inpatient antepartum (antepartum hospitalization), or intrapartum (at time of admission for delivery). Studies conducted solely in the postpartum period were excluded. Studies conducted from 2013 to 2018 at our institution and meeting inclusion criteria are presented in Supplemental Table 1 and categorized by study type and time point of patient enrollment. Participation in multiple concurrent studies was possible depending on each specific study enrollment criteria.
The primary outcome was a composite of maternal morbidity and mortality diagnosed up to 6 weeks after delivery. This composite included transfusion of at least 1 unit of packed red blood cells (pRBCs), clinically diagnosed endometritis, wound infection or complication, postpartum intensive care unit (ICU) admission, postpartum readmission, hysterectomy, or maternal death. A morbidity composite was chosen as the primary outcome as each morbidity may be rare. In addition, many of the items chosen for our composite represent outcomes that are core outcome measures used to standardly evaluate maternal care,6 and several of which are included in the Centers for Disease Control and Prevention (CDC) defined severe maternal morbidity outcomes.7 At our institution, obstetrical patients who consent to receive blood products typically receive a transfusion of PRBCs for a hemoglobin level of <7 g/dL in a postpartum setting or a hemoglobin level of <8 g/dL if the patient has symptomatic anemia or clinical suspicion for ongoing bleeding. Wound infection or complication included clinically diagnosed infections (cellulitis or other superficial wound infections, fasciitis or other deep wound infections, or intra-abdominal abscess) and noninfectious complications (wound hematoma, superficial or deep seroma, or any wound dehiscence, including superficial skin, fascial, or hysterotomy dehiscence).8
Our main secondary outcome was a composite of neonatal morbidity and mortality including fetal death, 5-minute Apgar score of <5, arterial pH of <7.0, neonatal ICU admission, and neonatal death. A neonatal composite, as may be customary, was chosen given the rarity of these adverse neonatal outcomes. Other secondary outcomes included estimated gestational age at delivery, maternal length of stay (in days), neonatal birthweight (in grams), and neonatal length of stay (in days). For patients with multiple pregnancies, a categorical adverse outcome was noted to be present if it occurred within any neonate, whereas continuous adverse outcomes were measured for the firstborn neonate in the multifetal birth.
The exposure of participation in one or more obstetrical studies was obtained through electronic abstraction of the Obstetrics and Gynecology Study Tracking System (STS), a continuously updated registry of all obstetrical and gynecologic research studies both completed and ongoing at our university. The classification of the study design and time point of enrollment was performed by 3 authors and confirmed by 2 research coordinators who are responsible for the administration and upkeep of the STS. The patients at our institution are routinely assessed for enrollment in clinical studies by both on-call physician staff and research nurses in a variety of settings, including outpatient clinics, the triage unit, labor and delivery, and the inpatient floors. Research nurses are generally available to assist with enrollment or data collection during business hours throughout the week and weekend; for inpatient studies, there is 24-hour research nurse coverage. The informed consent process is completed by IRB-trained resident or attending physicians or by certified research nurses depending on the individual study specifications. For maternal and pregnancy characteristics and outcome measures, an institutional database of all deliveries spanning from 2013 to 2018 (>24,000 cases) was created using previously validated methods.8 As such, outcomes were obtained through a combination of validated coded data abstraction and individual medical record review of all clinical, imaging, and operative notes (including available referral records).
Maternal demographics, pregnancy characteristics, and primary and secondary outcomes were compared between the study participants and nonparticipants using the chi-squared test of association or Fisher exact tests for categorical variables and the Wilcoxon rank-sum or Student t tests for continuous variables, as appropriate. For binary outcomes, crude odds ratios (ORs) with 95% confidence intervals (CIs) were estimated using logistic regression with study nonparticipants as the referent group. Multivariable logistic regression was used to adjust for confounding factors with adjusted ORs and 95% CIs reported for the primary and secondary outcomes, as appropriate. Baseline characteristics that differed by study participation with P<.001 were included as covariates in the multivariate models; collinearity was assessed to avoid overadjustment. Particularly, variables that differed by study participation with P<.001 that were not included in the final models were highly collinear with variables included in models to generate a more parsimonious adjusted model. The adjustments for multiple comparisons were not assessed. Additional logistic regression analyses were performed, stratifying by study type (trial or observational), enrollment time point (outpatient antepartum, inpatient antepartum, or intrapartum), and number of studies an individual participated in (0, 1, 2, or ≥3). In addition, we evaluated for effect modification by year of delivery. All primary analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC), and outcomes were evaluated at a .05 level of significance.
Results
Of 24,637 total deliveries at our center between 2013 and 2018, 19,569 met the final inclusion criteria (Figure). Of the 19,569 patients included, 3848 (19.7%) were study participants, and 15,721 (80.3%) were study nonparticipants. Among study participants, 3023 (78.6%) enrolled in a trial, and 825 (21.4%) enrolled in an observational study. By the time point of enrollment, 1752 study participants (44.5%) enrolled in the outpatient antepartum setting, 234 patients (6.1%) enrolled in the inpatient antepartum setting, and 1862 patients (48.4%) enrolled in the intrapartum setting. Of note, 3041 study participants (79.0%) enrolled in 1 study, 686 study participants (17.8%) enrolled in 2 studies, and 121 study participants (3.2%) enrolled in ≥3 studies. In context, during the study period, 77 studies conducted at our institution enrolled pregnant patients: 41 (53.4%) were trials and 36 (46.7%) were observational studies. Among the studies, 52 (67.5%) had outpatient antepartum enrollment, 12 (15.6%) had inpatient antepartum enrollment, and 13 (16.9%) had intrapartum enrollment.
FIGURE.
Flow diagram depicting patients identified for inclusion in the final analysis
EGA, estimated gestational age.
Study participants were younger, had higher body mass index (BMI), and were more likely to be non-Hispanic Black race and ethnicity, publicly insured, nonmarried, nulliparous, and use tobacco than study nonparticipants (Table 1). Although study participants had a higher incidence of chronic hypertension, the groups were similar concerning other maternal comorbidities. In addition, study participants were more likely to have a singleton pregnancy and develop gestational diabetes mellitus or preeclampsia (Table 1).
TABLE 1.
Baseline maternal demographics and pregnancy characteristics of patients delivered at a tertiary care center, 2013–2018
| Maternal and pregnancy characteristics | No participation in the study (n=15,721) | Participation in the study (n=3848) | P value |
|---|---|---|---|
| Maternal age (y) | 27.6±6.0 | 26.3±6.0 | <.001 |
| Prepregnancy BMI (kg/m2) | 29.3±8.0 | 31.8±8.9 | <.001 |
| Race and ethnicity | <.001 | ||
| Black, non-Hispanic | 6987 (44.4) | 2385 (62.0) | |
| White, non-Hispanic | 5582 (35.5) | 928 (24.1) | |
| Hispanic | 2459 (15.6) | 442 (11.5) | |
| Other | 692 (4.4) | 93 (2.4) | |
| Insurance status | <.001 | ||
| Private | 5539 (35.2) | 758 (19.7) | |
| Public | 9930 (63.2) | 3057 (79.4) | |
| Self-pay | 252 (1.6) | 33 (0.9) | |
| Married | 6127(39) | 935 (24.3) | <.001 |
| Nulliparous | 6023 (38.3) | 1981 (51.5) | <.001 |
| Tobacco use | 2061 (13.1) | 552 (14.3) | .04 |
| Chronic hypertension | 1214(7.7) | 432 (11.2) | <.001 |
| Pregestational diabetes mellitus | 551 (3.5) | 160(4.2) | .05 |
| Systemic lupus erythematosus | 99 (0.6) | 20 (0.5) | .43 |
| Chronic renal disease | 97 (0.6) | 14(0.4) | .06 |
| Singleton pregnancy | 15,071 (95.9) | 3777 (98.2) | <.001 |
| Gestational diabetes mellitus | 1091 (6.9) | 391 (10.2) | <.001 |
| Preeclampsia | 2336 (14.9) | 649 (16.9) | .002 |
Data are presented as mean±standard deviation or number (percentage), unless otherwise indicated.
BMI, body mass index.
With respect to mode of delivery, study participants were more likely to undergo a cesarean delivery than study nonparticipants (Table 2). Study participants were more likely to be induced rather than undergo spontaneous labor or no labor than their study nonparticipant counterparts. Moreover, the groups differed concerning the indication for cesarean delivery and mode of anesthesia (Table 2). After analyzing collinearity among these potential characteristics, only maternal age, BMI, insurance status, mode of delivery, and labor type remained in multivariate models. As such, race and ethnicity, marital status, parity, tobacco use, singleton pregnancy, chronic hypertension, preeclampsia, gestational diabetes mellitus, indication for cesarean delivery, and anesthetic type were highly collinear with one of our included variables and thus not included in the adjusted analysis to avoid overadjustment.
TABLE 2.
Intrapartum characteristics of patients delivered at a tertiary care center, 2013–2018
| Maternal and pregnancy characteristics | No participation in the study (n=15,721) | Participation in the study (n=3848) | P value |
|---|---|---|---|
| Mode of delivery | <.001 | ||
| Spontaneous vaginal delivery | 10,230 (65.1) | 2183 (56.7) | |
| Operative vaginal delivery | 514 (3.3) | 116 (3) | |
| Cesarean delivery | 4977 (31.7) | 1549 (40.3) | |
| Labor | <.001 | ||
| Spontaneous | 3276 (20.8) | 475 (12.3) | |
| Induction | 9764 (62.1) | 2781 (72.3) | |
| No labor | 2681 (17.1) | 592 (15.4) | |
| Chorioamnionitisa | 862 (5.5) | 226 (5.9) | .344 |
| Indication for cesarean delivery | <.001 | ||
| Malpresentation | 753 (15.4) | 151 (9.8) | |
| Nonreassuring fetal status | 1206 (24.7) | 461 (30) | |
| Any labor arrest | 832 (17.1) | 432 (28.1) | |
| Repeat cesarean delivery | 1632 (33.5) | 416 (27.1) | |
| Multiple pregnancies | 111 (2.3) | 10 (0.7) | |
| Other maternal indication | 344 (7.1) | 67 (4.4) | |
| Mode of anesthesia | <.001 | ||
| Neuraxial | 12,475 (79.6) | 3406 (88.6) | |
| GETA | 749 (4.8) | 154 (4) | |
| IV, pudendal, or none | 2451 (15.6) | 286 (7.4) |
Data are presented as number (percentage), unless otherwise indicated.
GETA, general endotracheal anesthesia; IV, intravenous.
Characteristic assessed only among patients undergoing spontaneous or induced labor.
The primary outcome of composite maternal morbidity or mortality occurred in 8.7% of study participants compared with 9.2% of study nonparticipants, which was not significantly different (OR, 0.94; 95% CI, 0.83–1.06) (Table 3). However, after adjustment for covariates, the study participants had significantly lower odds of maternal morbidity (adjusted OR [aOR], 0.83; 95% CI, 0.73–0.95). Of the composite components, the study participants had lower odds of pRBC transfusion (aOR, 0.73; 95% CI, 0.60–0.90) and lower odds of postpartum ICU admission (aOR, 0.56; 95% CI, 0.34–0.95). Furthermore, there was a shorter maternal length of stay among study participants than among study nonparticipants (3.9±3.0 vs 4.1±4.9 days; P=.03) (Table 3), although these small differences may not be clinically significant.
TABLE 3.
Primary and secondary outcomes by participation in any obstetrical study
| Outcome | No participation in the study (n=15,721) | Participation in the study (n=3848) | OR (95% CI) | aOR (95% CI)a |
|---|---|---|---|---|
| Composite maternal outcome | 1443 (9.2) | 333 (8.7) | 0.94 (0.83–1.06) | 0.83 (0.73 –0.95) |
| pRBC transfusion | 574 (3.7) | 118 (3.1) | 0.83 (0.68–1.02) | 0.73 (0.60–0.90) |
| Endometritis | 80 (0.5) | 27 (0.7) | 1.38 (0.89–2.14) | 1.26 (0.80–1.99) |
| Wound infection and complication | 355 (2.3) | 114 (3) | 1.32 (1.07–1.64) | 0.93 (0.74–1.17) |
| Postpartum ICU admission | 137 (0.9) | 17 (0.4) | 0.5 (0.30–0.84) | 0.42 (0.25–0.69) |
| Postpartum readmission | 693 (4.4) | 159 (4.1) | 0.93 (0.78–1.11) | 0.91 (0.76–1.09) |
| Hysterectomy | 71 (0.5) | 9 (0.2) | 0.52 (0.26–1.04) | 0.52 (0.26–1.06) |
| Maternal death | 11 (0.1) | 0 (0) | .10 | — |
| Maternal length of stay (d) | 4.1±4.9 | 3.9±3.9 | .03 | <.001 |
| Composite neonatal outcome | 4322 (27.5) | 715 (18.6) | 0.6 (0.55–0.66) | 0.53 (0.48–0.58) |
| Miscarriage or fetal death | 409 (2.6) | 19 (0.5) | 0.19 (0.12–0.29) | 0.20 (0.13–0.32) |
| 5-min Apgar score<5 | 943 (6) | 106 (2.8) | 0.44 (0.36–0.54) | 0.44 (0.35–0.54) |
| Arterial pH<7.0 | 117 (0.9) | 34 (1) | 1.12 (0.76–1.64) | 0.95 (0.64–1.41) |
| NICU admission | 3759 (24) | 665 (17.3) | 0.66 (0.60–0.73) | 0.57 (0.52–0.63) |
| Neonatal death | 218 (1.4) | 19 (0.5) | 0.35 (0.22–0.56) | 0.42 (0.26–0.68) |
| Gestational age at delivery (wk) | 37.1±4.6 | 38.2±3.1 | <.001 | <.001 |
| Neonatal birthweight (g) | 2878.2±918.3 | 3058.1±710.0 | <.001 | <.001 |
| Neonatal length of stay (d) | 9.9±23.0 | 6.5±16.4 | <.001 | <.001 |
Data are presented as number (percentage) or mean±standard deviation, unless otherwise indicated. Unadjusted and adjusted P values were presented in cases of a continuous variable; adjusted P values were not presented for insufficient observations.
aOR, adjusted odds ratio; CI, confidence interval; ICU, intensive care unit; NICU, neonatal intensive care unit; OR, odds ratio; pRBC, packed red blood cell.
Variables included in the regression analysis: maternal age, body mass index, insurance status, mode of delivery, and labor type.
The main secondary outcome of composite neonatal morbidity or mortality occurred in 18.6% of study participants compared with 27.5% of study nonparticipants, corresponding to a 40% decreased odds of morbidity among participants (OR, 0.60; 95% CI, 0.55–0.66) (Table 3). This remained significant after adjustment for covariates (aOR, 0.53; 95% CI, 0.48–0.58). Study participants had significantly lower odds of miscarriage or fetal death (aOR, 0.20; 95% CI, 0.13–0.32), 5-minute Apgar score of <5 (aOR, 0.44; 95% CI, 0.35–0.54), neonatal ICU admission (aOR, 0.57; 95% CI, 0.52–0.63), and neonatal death (aOR, 0.42; 95% CI, 0.26–0.68). There was no difference in cord arterial pH of <7.0. Study participants, compared with study nonparticipants, had a significantly later estimated gestational age at delivery (38.2±3.1 vs 37.1±4.6 weeks; adjusted P≤.001), a higher birthweight (3058.1±710.0 vs 2878±918.3 g; adjusted P≤.001), and a shorter neonatal length of stay (6.5±16.4 vs 9.9±23.0 days; adjusted P≤.001) (Table 3).
Concerning the analysis by study design, trial participants had a lower risk of composite maternal morbidity (aOR, 0.76; 95% CI, 0.66–0.88) than study nonparticipants (Table 4). However, among observational study participants, there was no statistically significant difference found in composite maternal morbidity compared with nonparticipants (aOR, 0.90; 95% CI, 0.70–1.16), likely because of the reduced sample size. Consistent with the primary analysis, improvement in neonatal outcomes among study participants persisted regardless of study design (trial aOR, 0.48 [95% CI, 0.43–0.53]; observational aOR: 0.55 [95% CI, 0.45–0.60]) (Table 4).
TABLE 4.
Maternal and neonatal morbidity composite outcomes by type of study design and time point of study enrollment
| Composite outcomes | Study type | Study time point | ||||
|---|---|---|---|---|---|---|
| No participation (n=15,721) | Trial (n=3023) | Observational study (n=825) | Outpatient antepartum (n=1752) | Inpatient antepartum (n=234) | Intrapartum (n=1862) | |
| Maternal | 9.2 (Ref) | 8.7 (0.76 [0.66–0.88]) | 8.5 (0.90 [0.70–1.16]) | 8.9 (0.87 [0.73–1.04]) | 11.5 (1.13 [0.75–1.70]) | 8.1 (0.69 [0.57–0.82]) |
| Neonatal | 27.5 (Ref) | 18.9 (0.48 [0.43–0.53]) | 17.6 (0.55 [0.45–0.60]) | 17.2 (0.48 [0.42–0.55]) | 39.7 (1.55 [1.17–2.03]) | 17.2 (0.41 [0.36–0.47]) |
Data are presented as percentage (adjusted odds ratio [95% confidence interval]). No study participation is the Ref group. Variables Included In the regression analysis are as follows: maternal age, body mass index, insurance status, mode of delivery, and labor type. Ref, referent.
In the analysis by time point of enrollment, there was no difference in composite maternal morbidity by time point of study enrollment (Table 4). However, neonatal morbidity was decreased among patients enrolled in the outpatient antepartum setting (aOR, 0.48; 95% CI, 0.42–0.55) and intrapartum setting (aOR, 0.41; 95% CI, 0.36–0.47) but not in those enrolled in the inpatient antepartum setting (aOR, 1.55; 95% CI, 1.17–2.03) (Table 4).
In the analysis by the number of studies in which study participant was enrolled, maternal morbidity was decreased for participants in 1 study (aOR, 0.78; 95% CI, 0.68–0.90) but not statistically significantly decreased among those enrolled in 2 studies (aOR, 0.80; 95% CI, 0.61–1.05) or ≥3 studies (aOR, 0.82; 95% CI, 0.46–1.44) (Supplemental Table 2). This is likely due to the small sample size of patients participating in 2 or ≥3 studies. Consistent with the primary analysis, improvement in neonatal outcomes among study participants persisted regardless of study design (Supplemental Table 2).
There was no effect modification by year of delivery for the maternal composite (P=.86), suggesting that the lower odds of adverse maternal outcomes with study participation were similar across the study years. There was a differential effect in the neonatal composite (P≤.0001) by year of delivery such that the association between study participation and adverse neonatal outcome was weaker as time progressed (ie, 2013: aOR, 0.36 [95% CI, 0.28–0.47]; 2018: aOR, 0.54 [95% CI, 0.43–0.67]; the remainder of the results were not shown).
Discussion
Principal findings
In a large retrospective analysis comparing the impact of participation in obstetrical studies vs nonparticipation at a single academic center, study participants had 17% lower odds of composite maternal morbidity and improvements in select maternal outcomes including pRBC transfusion, ICU admission, and postpartum length of stay than study nonparticipants. In addition, there was a 47% lower odds of composite neonatal morbidity, miscarriage or fetal death, 5-minute Apgar score of <5, neonatal ICU admission, neonatal death, and shorter lengths of stay among study participants than among study nonparticipants. These findings of improvement in outcomes were similar for neonates regardless of the study type or time point of enrollment, whereas findings of improvement in maternal outcomes were most pronounced among trial participants.
Results
Studies assessing the impact of participation in trials vs nonparticipation in a broad nonpregnant population are mixed.3,9,10 A Cochrane database review evaluated the impact of participation in both trials and retrospective analyses among a general patient population, including medical, surgical, oncologic, pediatric, psychiatric, and obstetrics and gynecologic patients, on outcomes, including mortality, morbidity, and other clinically important outcomes, for example, pain scores and surgical complications (which were reported in the analysis of each publication). The authors concluded that study participation did not consistently show an improvement in outcomes. Most studies (84.6%) in their analysis had no difference in outcomes, 8.1% of studies had improvement in outcomes, and 7.3% of studies had worse outcomes with study participation.4 Similarly, a recent review of the “trial effect” evaluated among clinical trial participants in a nonpregnant oncologic population found no overall clinically significant harm or benefit, concerning trial endpoints, overall survival, disease-free survival, period-specific survival, and recurrence.11
Literature regarding the “trial effect” specifically within obstetrical studies is limited. A recent systematic review and meta-analysis evaluated the impact of trial participation vs nonparticipation among 21 clinical obstetrical and gynecologic RCTs.5 They demonstrated that trial participants experienced a nearly 25% improvement in the individual author-reported trial outcomes compared with matched nonparticipants. In addition, the benefit of trial participation was study specific, in that the benefits were greater for participants in high-quality trials than in low-quality trials. The benefit of trial participation was also greater when the intervention was unavailable to nonparticipants compared with when an intervention was available to nonparticipants. Sensitivity analysis demonstrated that the benefits of trial participation remained regardless of whether the trial intervention was effective. Although 11 studies were obstetrical and 10 studies were gynecologic, there was no subgroup analysis evaluating trials only in pregnant patients, and the authors acknowledge that there would be an insufficient number of studies to analyze a specific disease or complication. Our study has added to the existing literature by focusing on the impact of study participation solely among obstetrical patients in obstetrical studies. Although the aforementioned analysis used reported trial outcome measures to evaluate the positive or negative impact of trial participation, our clinical outcomes were not study specific and were chosen to assess the overall effect on maternal or fetal morbidity. Furthermore, our analysis was not limited to RCTs and included all types of trials (RCT and non-RCT) and observational studies as well, broadening the impact of the investigation.
Clinical implications
Several theories exist as to the mechanism by which study participation is associated with improvement in patient outcomes. Patients enrolled in clinical studies receive care under prespecified and preapproved study protocols.12 As such, this may afford focus on current or evidence-based practice. A systematic review published in 2011 evaluating the impact of participation in nonobstetrical clinical studies found that physicians and other practitioners involved in the research studies were more likely to choose guideline-based therapies than their counterparts not involved in studies.9 In addition, patients who participate in clinical research may also receive additional oversight or safety monitoring by clinicians, nurses, or other ancillary research staff. As human subjects research must be conducted following approval of the IRB, it is subject to several regulations aimed at protecting the rights and welfare of study participants, including a group review process to assess research protocols and related materials (eg, informed consent documents and investigator brochures).12 Furthermore, certain trials undergo additional review by data monitoring committees, designed to access interim safety and efficacy data in an unblinded and confidential fashion, to further promote patient protection.13,14 It is plausible that these ethically regulated measures contribute to improvement in patient outcomes.
Other possibilities as to the mechanism for improvement in outcomes relates to the Hawthorne effect—where patients who are being observed change their behavior and improve their performance because they are being observed. Furthermore, the improvement in outcomes may be due to unmeasured confounding that study participants may have underlying differences compared with study nonparticipants. In congruence with this, study participants and nonparticipants had several differences in baseline characteristics (Tables 1 and 2). In our patient population, we found that those with study participation tended to be nulliparous, non-Hispanic Black race, publicly insured, and with several comorbidities, including chronic hypertension. It may be that there is something unique to that population who chooses study participation and warrants further analysis beyond adjustment in multivariable models. It may also be that certain patient populations with particular baseline characteristics had higher study participants because those characteristics qualified them for enrollment in a particular analysis—for instance, the ARRIVE trial (ClinicalTrials.gov number, NCT01990612) conducted at our institution enrolling nulliparous participants15 or the CHAP trial (ClinicalTrials.gov number, NCT02299414) enrolling patients with chronic hypertension.16 In addition, it is suggested that study participation increases the interface with medical providers and requires a basic level of prenatal care. As consistent prenatal care is associated with positive pregnancy outcomes and as inadequate prenatal care is associated with poor outcomes,17–19 it may be that trial participation is a proxy for prenatal care. In our analysis, there was a positive impact on maternal outcomes in trial participants; however, this was not demonstrated in subgroup analysis by time point of enrollment. Fetal outcomes were noted to be improved in patients enrolled not only in the outpatient antepartum setting but also in the intrapartum setting, challenging the notion that the impact of study participation is only through increased exposure to prenatal care. Further mechanistic studies are indicated to determine how study participation exerts its positive impact and how we as a community can extend these benefits of study participation to study nonparticipants. We hope that patients will be encouraged to participate in obstetrical clinical studies, and as such, the results of studies evaluating particular interventions and their impact on pregnancy outcomes can be applied out of a study setting to improve the care of the general pregnant population. In addition, our findings should be confirmed across other centers with different research infrastructures and patient demographics to confirm the applicability of these results.
Research implications
Hesitancy to conduct clinical obstetrical research in some instances stems from concerns regarding the potential for harm to fetuses and infants, including unforeseen teratogenic effects and adverse pregnancy outcomes.1 Among pregnant patients themselves, there may be a reluctance to enroll because of concerns over the well-being of their fetus, a perceived lack of benefit or research relevance, or mistrust of the medical system.1 Several additional factors influencing decisions to participate may include perceived interpersonal community support,20 partner reluctance, time constraints, and a desire to avoid additional examinations.21 In contrast, patients may choose to participate in research feeling that it enhances their understanding of their health or improves their care.2 Our analysis demonstrated that participation by gravidas in obstetrical studies was associated with improvement not only in maternal but also in fetal and neonatal outcomes. This association with improvement in neonatal outcomes was present when evaluating all obstetrical study designs and time points of enrollment. Our data can be used to aid in the informed consent process supporting the safety and potentially the benefit of participating in obstetrical research and addressing maternal fears regarding unanticipated risks to the fetus.
Strengths and limitations
There were multiple strengths in our study. We utilized a large, racially diverse cohort of 19,569 patients. Patients had the opportunity to be involved in both nationally based and institution-specific studies, increasing the applicability of our results. The use of the standardized and well-maintained Obstetrics and Gynecology STS and individual detailed medical record review by trained data reviewers minimized the risk of misclassification. The CDC enumerates several severe maternal morbidities, including transfusion of blood products, sepsis, shock, and hysterectomy,7 which were evaluated in our primary outcome and made this analysis relevant to providers nationally and worldwide. Adjustment for baseline maternal and pregnancy characteristics further reduced the risk of confounding inherent to all observational studies.
Potential limitations of this analysis included the retrospective nature of this study. The reason for a patient’s nonparticipation in a specific study is not available and may have included being ineligible for a study, which could introduce selection bias. However, given the number and breadth of studies conducted at our institution, we expect that patients would have been eligible for at least 1 study throughout their pregnancy and therefore would have had the opportunity to participate if they had desired. Unfortunately, we did not have the retrospective data to match participants with nonparticipants concerning a specific study in a case-control manner. In addition, we were unable to assess the impact of enrollment in positive studies (ie, the primary outcome was statistically significant) vs enrollment in negative studies (ie, the primary outcome was not statistically significant). It is plausible that some of the effects on maternal or neonatal outcomes may be due to the trial intervention, and a better comparison could be of the placebo group to nonparticipants. However, the choice to participate in the study and the randomization to a particular treatment group was made before any results or publication, and therefore, the comparison between participation and nonparticipation may be more practical on the patient level. Furthermore, we did not adjust for nulliparity or multiple pregnancies because they were highly collinear with other factors in our adjusted models—this was done to avoid overfitting our models. Future studies could evaluate the independent effects of study participation in these select groups of patients.
Study participants seemed to have a higher rate of cesarean delivery than study nonparticipants; however, this does not necessarily imply causality. Several studies (both published and ongoing), which were conducted at our institution during our study time frame, enrolled only patients undergoing cesarean delivery,22,23 which could account for this difference. As our analysis spans 5 years, evidence-based practice guidelines have changed during this time frame, which could have affected the study outcomes. For instance, some studies were conducted before or after the implementation of the use of adjunctive azithromycin for cesarean delivery prophylaxis22 or use of antenatal late preterm steroids in select patients at risk of preterm delivery,24 which could affect composite maternal and neonatal morbidities. However, there was no effect modification by year of delivery on maternal composite, suggesting that the improvement in maternal outcomes was similar in each year throughout the study time frame. Although there was effect modification by study year whereby the association between study participation and the neonatal composite outcome was weaker over time, study participation was associated with lower odds of adverse neonatal outcomes in all years. The differential effect may have been due to the implementation of new maternal interventions or new advances in neonatal resuscitation as time progressed.
Conclusions
Elective participation in obstetrical studies was associated with decreased maternal and neonatal morbidities. This held true in a subgroup analysis of obstetrical trials, which tend to have a direct patient care impact. Our analysis supported the safety and importance of pregnant patients participating in obstetrical clinical studies. Further studies are needed to better understand why obstetrical study participants have improved outcomes, acknowledging the possibility of selection bias, and how they potentially benefit from additional surveillance.
Supplementary Material
AJOG MFM at a Glance.
Why was this study conducted?
There are mixed data regarding the impact of participation in clinical trials and observational studies on outcomes in the general medical and nonpregnant population. There are limited data on the effect of pregnant patients participating in obstetrical studies and pregnancy outcomes.
Key findings
Elective participation in any type of obstetrical study, including trial and observational designs, was associated with a decrease in composite maternal and neonatal morbidities. The decrease in neonatal morbidity with study participation was present regardless of the type of study, timing of enrollment, number of studies, or year of delivery.
What does this add to what is known?
Participation in obstetrical studies may improve pregnancy outcomes—both maternal and neonatal. After providing informed consent, pregnant patients should be encouraged to participate in obstetrical studies.
ACKNOWLEDGMENTS
The authors would like to thank Nancy Saxon, Lisa Dimperio, and Rachel LeDuke for their maintenance of the Obstetrics and Gynecology Study Tracking System database and their assistance in study categorization.
A.N.B. was supported by a grant (grant number K23HD103875) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development during the study. This study received no other external source of funding.
Footnotes
The authors report no conflict of interest.
Supplementary materials
Supplementary material associated with this article can be found in the online version, at doi:10.1016/j.ajogmf.2022.100729.
Contributor Information
Gabriella D. Cozzi, Center for Women’s Reproductive Health, University of Alabama at Birmingham, Birmingham, AL; Departments of Obstetrics and Gynecology.
Victoria C. Jauk, Center for Women’s Reproductive Health, University of Alabama at Birmingham, Birmingham, AL.
Jeff M. Szychowski, Center for Women’s Reproductive Health, University of Alabama at Birmingham, Birmingham, AL; Biostatistics, University of Alabama at Birmingham, Birmingham, AL..
Alan T. Tita, Center for Women’s Reproductive Health, University of Alabama at Birmingham, Birmingham, AL; Departments of Obstetrics and Gynecology.
Ashley N. Battarbee, Center for Women’s Reproductive Health, University of Alabama at Birmingham, Birmingham, AL; Departments of Obstetrics and Gynecology.
Akila Subramaniam, Center for Women’s Reproductive Health, University of Alabama at Birmingham, Birmingham, AL; Departments of Obstetrics and Gynecology.
References
- 1.Frew PM, Saint-Victor DS, Isaacs MB, et al. Recruitment and retention of pregnant women into clinical research trials: an overview of challenges, facilitators, and best practices. Clin Infect Dis 2014;59(Suppl7):S400–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Purdy S, Finkelstein JA, Fletcher R, Christiansen C, Inui TS. Patient participation in research in the managed care environment: key perceptions of members in an HMO. J Gen Intern Med 2000;15:492–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Braunholtz DA, Edwards SJ, Lilford RJ. Are randomized clinical trials good for us (in the short term)? Evidence for a “trial effect. J Clin Epidemiol 2001;54:217–24. [DOI] [PubMed] [Google Scholar]
- 4.Vist GE, Bryant D, Somerville L, Birminghem T, Oxman AD. Outcomes of patients who participate in randomized controlled trials compared to similar patients receiving similar interventions who do not participate. Cochrane Database Syst Rev 2008;3:MR000009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Nijjar SK, D’Amico MI, Wimalaweera NA, Cooper N, Zamora J, Khan KS. Participation in clinical trials improves outcomes in women’s health: a systematic review and meta-analysis. BJOG 2017;124:863–71. [DOI] [PubMed] [Google Scholar]
- 6.Devane D, Begley CM, Clarke M, Horey D, OBoyle C. Evaluating maternity care: a core set of outcome measures. Birth 2007;34:164–72. [DOI] [PubMed] [Google Scholar]
- 7.Centers for Disease Control and Prevention. How does CDC identify severe maternal morbidity? 2019. Available at: https://www.cdc.gov/reproductivehealth/maternalinfanthealth/smm/severe-morbidity-ICD.htm#:~:text=To%20identify%20delivery%20hospitalizations%20with,ICD)%20diagnosis%20and%20procedure%20codes. Accessed July 2, 2022.
- 8.Lu MY, Blanchard CT, Ausbeck EB, et al. Evaluation of a risk-stratified, heparin-based, obstetric thromboprophylaxis protocol. Obstet Gynecol 2021;138:530–8. [DOI] [PubMed] [Google Scholar]
- 9.Clarke M, Loudon K. Effects on patients of their healthcare practitioner’s or institution’s participation in clinical trials: a systematic review. Trials 2011;12:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Peppercorn JM, Weeks JC, Cook EF, Joffe S. Comparison of outcomes in cancer patients treated within and outside clinical trials: conceptual framework and structured review. Lancet 2004;363:263–70. [DOI] [PubMed] [Google Scholar]
- 11.Fernandes N, Bryant D, Griffith L, et al. Outcomes for patients with the same disease treated inside and outside of randomized trials: a systematic review and meta-analysis. CMAJ 2014;186:E596–609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.US Food and Drug Administration. Institutional review boards (IRBs) and protection of human subjects in clinical trials. Available at: https://www.fda.gov/about-fda/center-drug-evaluation-and-research-cder/institutional-review-boards-irbs-and-protection-human-subjects-clinical-trials. Accessed January 4, 2022.
- 13.DAMOCLES Study Group. NHS Health Technology Assessment Programme. A proposed charter for clinical trial data monitoring committees: helping them to do their job well. Lancet 2005;365:711–22. [DOI] [PubMed] [Google Scholar]
- 14.Fleming TR, Hennekens CH, Pfeffer MA, DeMets DL. Enhancing trial integrity by protecting the independence of data monitoring committees in clinical trials. J Biopharm Stat 2014;24:968–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Grobman WA, Rice MM, Reddy UM, et al. Labor induction versus expectant management in low-risk nulliparous women. N Engl J Med 2018;379:513–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Tita AT, Szychowski JM, Boggess K, et al. Treatment for mild chronic hypertension during pregnancy. N Engl J Med 2022;386:1781–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Heaman MI, Martens PJ, Brownell MD, Chartier MJ, Derksen SA, Helewa ME. The association of inadequate and intensive prenatal care with maternal, fetal, and infant outcomes: a population-based study in Manitoba. Canada. J Obstet Gynaecol Can 2019;41:947–59. [DOI] [PubMed] [Google Scholar]
- 18.Holcomb DS, Pengetnze Y, Steele A, Karam A, Spong C, Nelson DB. Geographic barriers to prenatal care access and their consequences. Am J Obstet Gynecol MFM 2021;3:100442. [DOI] [PubMed] [Google Scholar]
- 19.Nussey L, Hunter A, Krueger S, et al. Sociodemographic characteristics and clinical outcomes of people receiving inadequate prenatal care: a retrospective cohort study. J Obstet Gynaecol Can 2020;42:591–600. [DOI] [PubMed] [Google Scholar]
- 20.Frew PM, Saint-Victor DS, Owens LE, Omer SB. Socioecological and message framing factors influencing maternal influenza immunization among minority women. Vaccine 2014;32:1736–44. [DOI] [PubMed] [Google Scholar]
- 21.van Delft K, Schwertner-Tiepelmann N, Thakar R, Sultan AH. Recruitment of pregnant women in research. J Obstet Gynaecol 2013;33:442–6. [DOI] [PubMed] [Google Scholar]
- 22.Tita AT, Szychowski JM, Boggess K, et al. Adjunctive azithromycin prophylaxis for cesarean delivery. N Engl J Med 2016;375:1231–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Tuuli MG, Liu J, Tita ATN, et al. Effect of prophylactic negative pressure wound therapy vs standard wound dressing on surgical-site infection in obese women after cesarean delivery: a randomized clinical trial. JAMA 2020;324:1180–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gyamfi-Bannerman C, Thom EA, Blackwell SC, et al. Antenatal betamethasone for women at risk for late preterm delivery. N Engl J Med 2016;374:1311–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.