The Impact of Opt-In versus Opt-Out Consent Process for the Use of Donor Human Milk on Feeding Practice and Growth Pattern in Preterm Neonates

Abstract

Objective

This study aimed to assess the impact of an opt-out consent process compared with the opt-in consent process for using pasteurized donor human milk (PDHM) on feeding practices and growth in preterm neonates.

Study Design

A retrospective study of 200 neonates ≤28 weeks’ gestation compared the effects of opt-in versus opt-out consent processes on time to first enteral feed, feeding types, growth trajectories, and prematurity-related morbidities. Descriptive statistics were used to compare the two groups.

Results

The opt-out process did not significantly alter the time to the first enteral feed (range opt-in: 0–12 days vs. opt-out 0–5 days, Mann–Whitney U; p = 0.295). Compared with the opt-in process, the opt-out process was associated with less formula use, less mother’s milk use, and more PDHM/mixed mom’s milk-PDHM use in the first 4 weeks of life (p < 0.01). Unadjusted analyses showed a greater weight gain in the opt-out group on days of life/DOLs 14 and 28 (2.49 [0.20, 4.78] and 3.94 [0.07, 7.80]%, p < 0.05), and following adjustment, it remained significant for DOL 14 (1.87 [0.02, 3.71]). Prematurity-related morbidities were similar between the two groups.

Conclusion

The opt-out consent process may impart short-term growth benefits in preterm infants. However, long-term outcomes require further study.

Exclusive breastfeeding is recommended by the American Academy of Pediatrics (AAP), the Centers for Disease Control and Prevention (CDC), and the World Health Organization (WHO) for the first 6 months of life.^1–3 For preterm neonates and those with very low birth weight at birth (VLBW), appropriately fortified mother’s own milk (MOM) is the preferred source. If this is insufficient or the mother cannot provide milk, the AAP has described pasteurized donor human milk (PDHM) as a safe alternative.^4,5

The use of PDHM has increased in recent years due to the recognition that human milk carries long-term benefits for neonates, especially preterm neonates.⁶ Compared with preterm formula, human milk reduces the rate of necrotizing enterocolitis (NEC) by up to 2.5-fold,⁷ reduces time to full enteral feed,⁸ decreases the incidence of sepsis,⁹ bronchopulmonary dysplasia (BPD),¹⁰ retinopathy of prematurity (ROP),¹¹ and improves neurodevelopmental outcomes.¹² Delays in initiating enteral feeds have been shown to increase intestinal inflammation, a risk factor for NEC.¹³ For these reasons, PDHM is often used as a bridge for enteral nutrition until MOM becomes available.¹⁴ Many neonatal intensive care units (NICU) feeding protocols use PDHM and MOM for the first 30 days of birth up to 34 weeks gestational age (GA), after which NEC incidence decreases.^15,16

Informed consent is a fundamental ethical requirement and a critical component of medical decision-making in any field, including neonatology. Since neonates cannot make healthcare decisions, consent is usually provided by parents or legal guardians through the opt-in or opt-out process.¹⁷ The approach to consent is typically tailored to the clinical contexts, for example, routine care, emergency or life-saving interventions, or participation in research or clinical trials.¹⁸ In the opt-in process, consent is required for study inclusion. The opposite is true for the opt-out process, where any eligible subjects are by default enrolled in a study or intervention, and consent is required for study exclusion.¹⁹ The opt-out process has been shown to increase participation and decrease sample bias.²⁰ While parents do not object to the opt-out process, more than half preferred the opt-in process.²¹

In many NICUs, the opt-in process remains the primary consenting method.²² However, in recent years, some centers have started using the opt-out process for some instances, such as delivery room management, emergency cases, and data linkage near or after discharge from the NICU.^23–25 This process provides equitable access while maintaining parental autonomy. The consent model for PDHM use largely relies on the opt-in process and is often bundled with the other consents upon admission to the NICU.²⁶ However, this process may be inadequate as parents may miss crucial information and lack the time to process it.^26,27 Given the well-documented advantages of PDHM for preterm infants and the widely used PDHM by NICUs nationwide, default enrollment with an option to decline (opt-out process) can be a viable alternative to the opt-in process. We hypothesized that the opt-out consent process would expedite the time to the first enteral feed (measured in days; primary outcome). The secondary outcomes included the rates of MOM versus formula usage and infant growth trajectory during the NICU stay. We will also explore prematurity-related morbidities: NEC (defined as per modified Bell criteria: stage II or greater),²⁸ early-onset sepsis/EOS (sepsis with onset before or equal to day of life/DOL 3),²⁹ late-onset sepsis/LOS (onset after DOL 3),²⁹ patent ductus arteriosus/PDA requiring intervention (composite of either pharmacotherapy [indomethacin, acetaminophen, or ibuprofen] device closure, or surgical ligation), intraventricular hemorrhage/IVH using ultrasonographic diagnosis (grades 1–3, and IVH with intraparenchymal hemorrhage),^30,31 and BPD as determined by the Jensen criteria (any oxygen requirement at 28 DOL or 36 weeks’ corrected age) were also recorded.^32,33

Materials and Methods

Until December 2018, our NICU utilized the traditional opt-in consent process to initiate enteral feeds using PDHM. Using this process, PDHM became available after mothers signed a consent form. However, several circumstances prohibited the opt-in consent in the immediate postnatal period, for example, when mothers are clinically unstable postdelivery or when delivery occurs at the community hospitals and neonates are transferred to our center, separated from their mothers. The inability to obtain consent immediately after birth often delays initiating the first enteral feeding. To minimize such delay, we adopted an alternative opt-out consent process for neonates born at ≤32 weeks’ gestation after getting approval from the risk management team at our institution. In this opt-out process, the discussion of PDHM usage was incorporated in the prenatal consult (for inborn neonates) or transport process (for outborn neonates). Parents’ verbal responses were recorded in the electronic medical record (EMR) system, allowing PDHM to be readily used in the postnatal period.

We conducted a retrospective EMR review of neonates between 23- and 28-week gestation born at Tufts Medical Center (TMC) NICU in Boston, MA, from November 2016 to April 2022. We also included preterm neonates born elsewhere and transferred to our center shortly after birth. Exclusion criteria included neonates born at < 23 weeks’ gestation based on our resuscitation practice guidelines at the time, those with scheduled admissions for surgical procedures, with known or suspected congenital malformations or chromosomal anomalies, and death before 48 hours of life. We also excluded subjects who had incomplete anthropometric or demographic data.

To evaluate our outcomes, we divided subjects into two epochs: November 2016 to December 2018 (opt-in) and January 2019 to April 2022 (opt-out). Staff education on the opt-out process took place between October 2018 and December 2018 through meeting announcements and email communications, and implementation of this new workflow commenced in January 2019 with no washout period. The prenatal consult template was modified to include a checkmark stating the verbal assent of the parents to administer PDHM as soon as clinically indicated. The Tufts Medical Center Risk Management department was involved in modifying the new process. The opt-out consent was obtained in person as part of the prenatal consultation and transfer process.

Demographic data included infant sex, GA, singleton or multiple gestation pregnancy, delivery type (vaginal delivery vs. cesarean/C-section), site of delivery (inborn vs. outborn), APGAR scores at 1 and 5 minutes, antenatal steroid use, length of stay, and anthropometric measurements at birth and discharge (birth weight, length, head circumference, and corresponding percentiles according to Fenton 2003). Due to the limitations of the EMR, race and ethnicity were not available or limited to a few subjects. Weight was recorded on DOL 7, 14, and 28, and weight percentage changes were compared with birth weight at these time points. Subjects received enteral feeding within 24 hours of life, unless they are deemed clinically unstable, such as hemodynamic instability, metabolic acidosis, need for vasoactive support, and severe respiratory distress (hypercarbia/hypoxemia). Guidelines for increasing enteral feeding volumes were followed according to the weight- and GA-based feeding pathway protocol in the NICU. Specifically, for neonates <32 weeks with birth weight between 1,001 and 1,250 g, enteral feeding is gradually advanced to a full feed over a 10-day period, and for birth weight < 1,000 g over a 12-day period. Neonates were considered to receive a full feed when they reached a total fluid of 150 mL/kg/day. The same feeding pathway protocols were used in either epoch, that is, before or after the opt-out consent process.

Additionally, we determined the type of first enteral feeds at these time points, time to regain birth weight, type and duration of use of central lines (composite of umbilical arterial and venous catheters, peripherally inserted central catheters, and surgical lines), length of stay, and disposition at discharge. Diagnoses of NEC, EOS, LOS, PDA requiring intervention (composite of either pharmacotherapy [indomethacin, acetaminophen, or ibuprofen], device closure, or surgical ligation), IVH (grades 1–3, and IVH with intraparenchymal hemorrhage), and BPD as aforementioned were also recorded.

Comparisons were made between the two groups. Descriptive statistics in each group were calculated, including mean and standard deviation (SD) for normally distributed variables, median and interquartile range (IQR) for nonnormally distributed variables, and percentages for categorical variables. For demographic variables and outcomes, two-sample t-tests were used to compare means, the Mann–Whitney U test was used to compare medians, and chi-square tests or Fisher’s exact tests were used to compare categorical variables between opt-in and opt-out groups. A p-value <0.05 was considered statistically significant. We performed linear regression analyses to account for potential confounding of the association between the consent process and time to first feed and growth measurements. We used prior knowledge, availability of data, and a univariate screen (p < 0.2) to identify possible covariates. With these, we performed backward selection using Akaike’s Information Criterion (AIC) for each outcome and took the union of selected predictors to be used for all outcomes. The following set of potential confounders was included in all regression models: GA, multiple pregnancy, mode of delivery, outborn status, birth weight, and APGAR scores at 1 and 5 minutes. We additionally included length of stay and squared length of stay in the adjusted model with growth at discharge as the outcome. For all models, we inspected residuals and calculated variance inflation factors to assess issues of model fit and multicollinearity. For the outcome of day-to-first enteral feed, we additionally fit a Poisson regression with robust standard error estimation as a sensitivity analysis to account for the skewed distribution of this outcome.

For the primary outcome of time/day to the first enteral feed, we conducted a post hoc power analysis for a difference in means. We had 80% power to detect Cohen’s effect size of 0.41, which is considered a small-to-medium effect size.³⁴ For day to first enteral feed, which we observed to have an SD of 1.29, this corresponds to detecting a difference in means of 0.52 days. To be conservative, if the true SD is 20% larger, we could detect a difference in means of 0.63 days. From this, we believe we have adequate power to detect sufficiently small, clinically meaningful differences for our primary outcome. For growth outcomes, with SDs of 6.5, 8.2, 13.9, and 146% growth for DOL 7, 14, 28, and discharge, respectively, we expected to have the power to detect differences of 3.2, 4.0, 6.8, and 71% growth for DOL 7, 14, 28, and discharge, assuming conservatively that the true SD is 20% greater than our estimates.

Results

Study Population and Clinical Data

As shown in the flowchart in ►Supplementary Fig. S1 (available in the online version only), we screened 224 EMRs per the inclusion criteria (GA: 23–28 weeks gestation). Per the exclusion criteria, we retained 200 neonates in the final analysis (80 in the opt-in group and 120 in the opt-out group). There was a greater rate of C-section delivery in the opt-out group (n = 88 [53.8%] vs. n = 43 [73.3%], p = 0.007). Otherwise, the two groups were comparable (►Table 1). No significant differences were found in the median length of stay (opt-in 79.5 [56.5, 95.0] vs. opt-out 81.0 [58.0, 112.3] days, p = 0.157) or median central line duration (opt-in 11.0 [9.0, 16.0] vs. opt-out 12.0 [9.0, 15.0] days, p = 0.664).

Table 1.

Demographic and clinical characteristics by the consent process

Variables	Opt-in (n = 80)	Opt-out (n = 120)	p-Valuea
Gestational age (wk)	26.9 [25.5, 27.9]	27.3 [25.9, 28.0]	0.447
Male	37 (46.2)	61 (50.8)	0.624
Multiple pregnancies	16 (20.0)	39 (32.5)	0.075
Prenatal steroids	74 (92.5)	107 (89.2)	0.588
Outborn	12 (15.0)	29 (24.2)	0.163
C-section	43 (53.8)	88 (73.3)	0.007
Birth weight (g)	912.29 (232.49)	884.37 (241.90)	0.414
Birth weight percentile	48.20 [30.32, 68.97]	44.60 [21.58, 63.30]	0.139
Birth head circumference (cm)	24.19 (1.74)	24.03 (2.13)	0.571
Birth head circumference percentile	45.60 [21.50, 64.12]	35.55 [17.60, 56.97]	0.105
Birth length (cm)	34.50 (2.75)	34.20 (3.35)	0.482
Birth length percentile	37.10 [22.55, 60.60]	29.45 [12.75, 60.37]	0.160
1-min APGAR	5.00 [3.00, 6.50]	5.00 [2.00, 6.00]	0.143
5-min APGAR	8.00 [7.00, 9.00]	7.00 [5.00, 8.00]	0.067

^aWelch two-sample t-test (normally distributed continuous measures); Wilcoxon rank sum test (nonnormally distributed continuous measures); Fisher’s exact test (categorical measures with expected cell counts < 5); Pearson’s chi-squared test (categorical measures with expected cell counts ≥ 5). Data are presented as mean (standard deviation) for normally distributed continuous measures, median [interquartile ranges] for nonnormally distributed continuous measures, or n (%) for categorical measures.

Enteral Feeding Characteristics

For our primary outcome, the median time to first enteral feed did not differ significantly between the two groups in unadjusted or adjusted analyses (opt-in 1 [1,2] vs. opt-out 1 [1,2] days, p = 0.295, unadjusted p = 0.295, linear regression p = 0.270, Poisson regression p = 0.228). Although a broader time range was seen in the opt-in than opt-out group (0–12 days vs. 0–5 days), no difference was found by adjusting for variances of the times of first feeds using Levine’s test (p = 0.39). The median time to achieve full enteral feeds was also similar between the two groups: 13 [11, 18] days vs. 14 [11, 17] days, p = 0.92.

The types of enteral feeds were compared between the two groups across various time points: at first enteral feed, at regain birth weight, on DOLs 7, 14, 28, and discharge day (►Table 2). Within the first 28 days, the distributions of feeding types differed significantly between the groups (p < 0.05) at all time points except DOL 14. Compared with the opt-in group, the opt-out group had a consistent pattern of less formula/mixed-use, less MOM use, and more PDHM or mixed MOM/PDHM use. At discharge, feeding types were distributed similarly between the groups (p = 0.51). Exclusive breast milk use was similar in the two groups at <20%, and formula and mixed breast milk/formula were the primary feed of choice for both groups at discharge (82.5% in the opt-in group vs. 83.3% in the opt-out group).

Table 2.

Types of feeding at various time points by the consent process

Timing	First feed		Regain BW		DOL 7		DOL 14		DOL 28		DC
Consent; feed types	Opt-in	Opt-out	Opt-in	Opt-out	Opt-in	Opt-out	Opt-in	Opt-out	Opt-in	Opt-out	Opt-in	Opt-out
MOM	47.5	10.0	62.5	47.5	61.3	29.2	62.5	56.7	62.5	57.5	17.5	16.7
PDHM	33.7	60.0	15.0	20.0	12.5	28.3	11.2	15.8	1.3	17.5	0.0	0.0
MOM/PDHM	17.5	30.0	11.3	29.2	13.7	29.2	10.0	19.2	10.0	10.8	0.0	0.0
Formula	0.0	0.0	0.0	0.0	0.0	0.0	1.3	0.0	18.7	0.8	52.5	60.0
Mixed	1.3	0.0	1.2	0.0	0.0	0.0	1.3	0.0	3.0	4.2	30.0	23.3
NPO	0.0	0.0	10.0	3.3	12.5	13.3	13.7	8.3	3.7	9.2	0.0	0.0
p-Valuea	<0.0001		0.005		<0.0001		0.160		<0.0001		0.519

Abbreviations: DC, discharge; DOL, day of life, regain BW = at the time infants regain their birth weight; MOM, mother’s own milk, breast milk; NPO, nil per os; PDHM, pasteurized donor human milk, mixed = breast milk and formula.

Note: All values are reported in percentages (%). Italicized p-values indicate significance.

^aFisher’s exact test (categorical measures with expected cell counts < 5); Pearson’s chi-squared test (categorical measures with expected cell counts ≥ 5).

Anthropometrics and Growth Data

The two groups were comparable in their anthropometrics at birth (weight, length, head circumference, and corresponding percentiles; ►Table 1). The median time needed to regain birth weight was similar between groups (10 [8, 13] days vs. 11 [9, 15] days, p = 0.08). Comparison of weights at various time points (DOLs 7, 14, 28, and discharge) showed that the opt-out group had greater weight gain percentages on DOLs 14 and 28. Weight changes were not different between groups at DOL 7 and at discharge (p < 0.05 for DOLs 14 and 28, ►Fig. 1). When adjusting for covariates, the difference between groups in growth at DOL 28 was no longer statistically significant (►Supplementary Table S1, available in the online version only).

Fig. 1.

Growth trajectory between groups during study standpoints. Compared with infants in the opt-in consent group, those in the opt-out consent group had more weight gain on DOLs 14 and 28. *p< 0.05.

Prematurity-Related Comorbidities

►Table 3 showed a comparable incidence of prematurity-related comorbidities in the two groups: NEC, EOS, LOS, PDA requiring intervention (pharmacotherapy or surgical ligation), IVH (various grades), and BPD (p-values > 0.05).

Table 3.

Prematurity-related comorbidities by the consent process

Demographics	Opt-in (n = 80)	Opt-out (n = 120)	p-Valuea
NEC	10 (12.5)	18 (15.0)	0.771
EOS	9 (11.2)	8 (6.7)	0.379
LOS	11 (13.8)	15 (12.5)	0.966
PDA with intervention	16 (20.0)	30 (25.0)	0.410
IVH			0.877
Grade 1	11 (13.8)	16 (13.0)
Grade 2	8 (10.0)	10 (8.3)
Grade 3	2 (2.5)	3 (2.5)
Grade 4 (IVH/IPH)	5 (6.2)	4(3.3)
BPD	36 (45.0)	54 (45.0)	1.00

Abbreviations: BPD, bronchopulmonary dysplasia; EOS, early-onset sepsis; IPH, intraparenchymal hemorrhage; IVH, intraventricular hemorrhage; LOS, late-onset sepsis; NEC, necrotizing enterocolitis; PDA, patent ductus arteriosus.

Note: All values are reported in percentages (%). Italicized p-values indicate significance.

^aFisher’s exact test (categorical measures with expected cell counts < 5); Pearson’s chi-squared test (categorical measures with expected cell counts ≥ 5).

Discussion

In this retrospective study, we evaluated the effect of an opt-out (alternative) consent process on the time to first enteral feed compared with the previously practiced opt-in consent process. The rationale for this predetermined primary outcome is based on the perception that the traditional consent process often delays an infant’s first feed due to the challenges in obtaining the signed consent form from mothers. Although the median time to the first enteral feed did not differ between the two groups, neonates in the opt-out group received their first feeds within the first 5 days after delivery, as opposed to 12 days in the traditional group. Important to note that due to the retrospective nature of our study and the limitations of EMR capabilities, we could only determine the time to first feed using the DOL and not precisely the specific hour of life. This limitation may falsely overestimate or underestimate the time to first enteral feeds in either group. Further, the retrospective nature of this study also limits our understanding of potential reasons for a wider time frame or a later initiation in the traditional group. However, we recognize that this difference was not significant. While it is not feasible to repeat the study in this setting due to the already adopted practice change and the absence of traditional practice, it could be possible in another institution with different consent processes or workflow using a prospective model and measuring feedings down to the hour of life. Moreover, future studies should investigate factors that may influence the initiation of the first enteral feed (e.g., neonatal, maternal, or unit-based factors).

Examination of the weight changes showed that compared with those in the opt-in group, neonates in the opt-out group had a greater weight gain during their NICU stay (►Supplementary Table S1, available in the online version only). Even after adjusting for various covariates, our results suggest a more favorable growth pattern in the opt-out than in the opt-in group, with a significantly greater weight gain seen by DOL 14. Multiple studies have found that early introduction of feedings and favorable catch-up growth correlate with better neurodevelopmental outcomes,³⁵ with the opposite being associated with higher risks of cognitive delay at 2 years of life.³⁶

The types of feeding at each time point are also noteworthy. As shown in ►Table 2, there were more MOM and formula/mixed uses and less PDHM use in the traditional opt-in than in the opt-out group. Given that our lactation support has remained the same throughout the study, we speculate that the lower MOM use may be related to the significantly greater proportion of C-sections in this group compared with the opt-in group (►Table 1). It is possible that there were greater incidences of pregnancy-related health issues necessitating C-section deliveries (e.g., preeclampsia, hypertension, and placental insufficiency), which in turn reduced the rate of breastfeeding initiation, increased breastfeeding difficulties, and earlier breastfeeding cessation.³⁷ It is also likely that mothers changed their initial plan to breastfeed depending on the reasons or types of C-sections (emergent vs. planned).

While it is not likely that the use of PDHM rather than MOM or formula contributes to the more favorable growth pattern seen in the opt-out group, the availability of PDHM as a supplement may provide assurance and a sense of well-being for some mothers, which in turn fosters positive care experiences, improves breastfeeding rates and helps promote growth in their infants.^38,39 Despite the greater use of PDHM in the opt-out group up to DOL 28, our data showed the rate of MOM use at discharge was similarly low, 17.5 and 16.7% in the opt-in and opt-out groups, respectively. This result suggests that while the practice change increased the adoption of PDHM use as a “bridge” to MOM use, this trend was not sustained beyond the first month of NICU stay and did not increase MOM use at discharge. One of the targets of the Healthy People 2030 (HP2030) Initiative is to ensure that 42.4% of infants exclusively receive breast milk.⁴⁰ As PDHM is a vital effort and investment to achieve exclusive human milk feeding, there is a critical need to understand barriers to translating higher PDHM use to MOM use at discharge so we can ensure the success of the HP2030 target goal.⁴¹ Future studies would benefit from gathering additional characteristics that may affect mother’s milk supply and ability to sustain milk supply, such as types of delivery (vaginal vs. cesarean), pregnancy-related health issues surrounding the delivery process, first-time versus recurrent breastfeeding, gravida and parity status, pre- or post-delivery physical or mental health issues that may prevent mother’s availability to pump, duration or number of attempts at breast milk pumping, number of lactation visits, and socioeconomic challenges that may hinder mother’s ability to continue pumping or breastfeeding.

The practice change was not linked with any difference in prematurity-related comorbidities (NEC, EOS, LOS, PDA, IVH, and BPD). This may highlight the complexity of premature births and that the PDHM consent practice change may not directly affect or modify these outcomes. Of interest was the incidence of NEC in the current study, which was 12.5 and 15.0% in the opt-in and opt-out groups, respectively. These percentages are higher than the average NEC incidence in VLBW infants in the United States in the last few decades (7% in 1993, 13% in 2008, and 9% in 2012).⁴² It is unclear if the high incidence of NEC was related to the unit-based practices at our center or may stem from potential errors when the diagnosis was coded, entered, or extracted. Nevertheless, future studies should continue examining the rate of NEC at our center and the factors contributing to or ameliorating this disease incidence.

As previously mentioned, our study is limited by its retrospective design, which limits the information that can be extracted, such as the ability to pinpoint the exact time of feed initiation, the specific reasons for the delay in the initiation of feed, refusal or acceptance rate of each consent process, etc. The use of EMR for our data gathering may also introduce human or system errors, either during data entry, extraction, or recording. A well-known limitation of retrospective studies is the introduction of selection and recall biases.⁴³ The retrospective review also limits our ability to follow up these infants longitudinally.⁴³ Our findings may result from unit-based or personnel practices that were not accounted for. For instance, the higher use of formula in the opt-in group may be due to the clinical team’s time constraints in obtaining consent. Mothers’ health conditions may be such that the team members must wait or try several times to obtain consent, yet such information is unavailable through the EMR review. The already implemented practice change also limits the replicability of this study, and the single-center context also limits the generalizability of our findings. Finally, the small-to-medium effect size likely limits the statistical significance of our findings, especially when data are adjusted for several covariates of interest. Nevertheless, our study identifies some findings that may be clinically important and worth replicating in a larger sample size.

Despite these weaknesses, to the best of our knowledge, our study is the first to report the impact of an opt-out consent process for using PDHM in the extremely premature population on short-term feeding practices and growth patterns. The opportunity to conduct a pre- and post-intervention by comparing two different consent processes and leveraging a natural or historical unit-based context without creating the intervention de novo enables us to assess infant health outcomes in a time- and cost-efficient fashion. While we did not evaluate parental views on the opt-in versus opt-out consent process, the opt-out consent process may provide the family with more time to weigh their options, gather more information about the use of PDHM, and even reconsider their consent.²⁷ Even though the benefits of PDHM are widely known, there remains some misperception about the safety of PDHM that warrants better communication and strategies to address parental concerns.⁴⁴ In clinical trials or research studies, the process of consenting in critically ill neonates before delivery has been shown to provide families with ample time to understand the costs and benefits of the intervention and make an informed decision.^45,46 A study examining the practice of an opt-out consent process in withholding versus continuing feeds around transfusion in preterm infants showed that clinicians perceived an opt-out consent as a feasible and acceptable method to provide informed choice.⁴⁷ Similarly, families also believed that an opt-out consent process allowed them to exercise their right to decide and participate in the study.⁴⁷ Therefore, discussing the use of PDHM during the antenatal consult may provide families with more time to ask questions and study the safety and benefits of PDHM.

Conclusion

In conclusion, the opt-out process is an alternative consent model that may be associated with a more favorable growth trajectory and less use of formula in the first month of life in preterm neonates. Although our results do not ascertain a causal relationship, as the first to report the utility of the opt-out consent process in using PDHM, our study may serve as a foundation for future studies. Other units still using an opt-in consent process can replicate and utilize a prospective, rather than retrospective, study design to circumvent the existing weaknesses in the current study. While EMR access limitation did not allow us to collect baseline opt-in and opt-out rates, such information allows a better understanding of the feasibility and success of each process in improving infant health outcomes. Future studies should also incorporate family perspectives in adapting the opt-out consent process, focus on understanding the barriers to achieving sustained MOM use at discharge, and explore the impact of PDHM on long-term neonatal outcomes.

Key Points.

• Opt-out is an alternative to the traditional opt-in consent for donor milk use in preterm infants.

• It is associated with better growth by day 14.

• Less formula use, and more donor/mixed milk use by day 28.

• Mother’s milk use at discharge and short-term outcomes are similar for both consent types.

• Future studies will focus on the long-term effects and benefits of the opt-out consent process.