The Initial Maternal Cost of Providing 100 mL of Human Milk for Very Low Birth Weight Infants in the Neonatal Intensive Care Unit

Abstract

Objectives

Human milk (HM) feeding is associated with lower incidence and severity of costly prematurity-specific morbidities compared to formula feeding in very low birth weight (VLBW; <1,500 g) infants. However, the costs of providing HM are not routinely reimbursed by payers and can be a significant barrier for mothers. This study determined the initial maternal cost of providing 100 mL of HM for VLBW infants during the early neonatal intensive care unit (NICU) stay.

Methods

This secondary analysis examined data from 111 mothers who provided HM for their VLBW infants during the early NICU stay. These data were collected during a multisite, randomized clinical trial where milk output and time spent pumping were recorded for every pumping session (n = 13,273). The cost analysis examined the cost of the breast pump rental, pump kit, and maternal opportunity cost (an estimate of the cost of maternal time).

Results

Mean daily milk output and time spent pumping were 558.2 mL (SD = 320.7; range = 0–2,024) and 98.7 minutes (SD = 38.6; range = 0–295), respectively. The mean cost of providing 100 mL of HM varied from $2.60 to $6.18 when maternal opportunity cost was included and from $0.95 to $1.55 when it was excluded. The cost per 100 mL of HM declined with every additional day of pumping and was most sensitive to the costs of the breast pump rental and pump kit.

Conclusions

These findings indicate that HM is reasonably inexpensive to provide and that the maternal cost of providing milk is mitigated by increasing milk output over the early NICU stay.

Introduction

Human milk (HM) feedings reduce the risk of costly and handicapping morbidities (e.g., sepsis, necrotizing enterocolitis) in a dose–response manner during the early post-birth period for very low birth weight (VLBW; <1,500 g) infants.^1–18 Thus, access to adequate amounts of HM from the infant's own mother should be a priority in the neonatal intensive care unit (NICU). However, the basic lactation equipment that these mothers need—a hospital-grade electric breast pump and a pump collection kit^19,20—is not routinely reimbursed by public and private third-party payers.²¹ Thus, these mothers must incur the cost for the breast pump and pump kit, an economic burden that translates into a barrier^20,22 to providing HM for a VLBW infant. In the immediate post-birth period, a mother who cannot access a breast pump and kit has two choices: (1) attempt to initiate lactation with less costly, but significantly less effective, manual or electric pumps, thereby risking lactation complications (e.g., mastitis, lactation failure); or (2) feed donor milk and/or commercial formula. The purpose of this study was to examine the initial maternal cost of providing 100 mL of HM for her VLBW infant during the early post-birth period and to compare this cost to the alternatives: donor milk and commercial formula.

Materials and Methods

Design

This study is a secondary analysis of an existing dataset of 132 mothers who provided HM for their VLBW infants immediately after birth between January 2002 and March 2004. These data were originally collected as part of a multisite, randomized clinical trial that tested the efficiency, efficacy, comfort, and convenience of three different suction patterns in a new hospital-grade electric breast pump (Symphony Pump; Medela, Inc., McHenry, IL).^23,24

Sample

The sample for this study was drawn from the records of the 132 mothers who participated in the randomized clinical trial. Inclusion criteria for the original study were as follows: birth of an infant(s) with birth weight <1,250 g and/or gestational age ≤32 weeks; infant admitted to NICU; mother provided HM; and written informed consent for study participation. For this secondary analysis, records for mothers enrolled in the original study were eligible for analysis if completed milk output logs were returned.^23,24 Of the 132 original records, 116 met this criterion and were eligible for inclusion in this analysis. Five records were excluded because no pumping times were documented, leaving a total study sample size of 111 records. The median number of days for study participation was 21 (range = 5–41; interquartile range = 17–24). This analysis was limited to the first 31 days of the original study because only 10 mothers provided data beyond 31 days. The demographic characteristics of the sample are displayed in Table 1. The original study was approved by the Institutional Review Boards at all four sites where data were collected, and the secondary analysis was approved by the Institutional Review Boards at Rush University Medical Center (Chicago, IL) and Saint Louis University (St. Louis, MO).

Table 1.

Characteristics of the Mothers and Their Infants

Variable	Mean ± SD or n (%)
Race (n = 111)
Caucasian	75 (67.6)
African American	25 (22.5)
Asian	7 (6.3)
Native American	2 (1.8)
Other	2 (1.8)
Ethnicity (n = 111)
Hispanic	14 (12.6)
Not Hispanic	97 (87.4)
Highest level of education (n = 111)
Less than high school	3 (2.7)
High school graduate	18 (16.2)
Some college	23 (20.7)
College graduate	67 (60.4)
Total family income (n = 109)
Less than $5,000	5 (4.6)
$5,000–$49,999	34 (31.2)
$50,000 or more	65 (59.6)
Prefer not to answer/doesn't know	5 (4.6)
WIC eligible (n = 111)
Yes	35 (31.5)
No	71 (64.0)
Unsure	5 (4.5)
Maternal age in years (n = 111)	30.71 ± 6.37
Infant birth weight (g) (n = 123)b	985.49 ± 263.14
Infant gestational age (weeks) (n = 111)	27.40 ± 2.31

^aWIC, Women, Infants and Children.

^bInfant birth weight is the average birth weight of all infants born to study mothers; 146 infants were born to 111 mothers. Complete birth weight data were available for 123 infants.

Measures

Daily maternal milk output

Mothers used a standardized milk output log to record total milk output to the nearest 10 mL from each breast for each pumping session throughout the study. Total milk output for each pumping session was calculated by adding the milk output for the right and left breasts. Total daily milk output was calculated by adding the total milk output from each pumping session for each study day.

Daily maternal time spent pumping

Mothers used the milk output log to record the start and end time for each pumping session for each breast throughout the original study. Total time for each pumping session was calculated by subtracting the start time from the end time for each breast. If the mother pumped both breasts simultaneously for an equal amount of time, the total time was equal to the time for one breast. If the mother pumped longer on one breast, the total time was equal to the longer pumping time. If the mother did not pump both breasts simultaneously, the total time was calculated by adding the pumping times for both breasts. Total daily time was calculated by adding the total time from each pumping session for each study day.

HM provision costs

The maternal cost of providing HM was calculated using the cost for the minimum items that are universally required by all mothers of VLBW infants. These items included: the monthly breast pump rental fee (BPRF); the breast pump collection kit cost (BPCK); and the maternal opportunity cost (MOC). The cost of the containers to store HM is borne by the hospital to ensure that the containers meet hospital standards for food safety and infection control. The definitions for the universally required costs are below. All costs are presented in 2008 U.S. dollars.

The monthly BPRF was measured as the charge from the institution to the mother for rental of the hospital-grade electric breast pump (BPRF = $35). BPRF is typically offered on a monthly basis.²¹

The BPCK was measured as the institutional charge to the mother (BPCK = $57). This cost is a one-time fixed cost because kits do not become obsolete, ineffectual, or damaged in the first month of pumping.²⁵

The MOC measures the monetary value of the mother's time spent pumping and is traditionally defined as the maternal wage that the mother forgoes by choosing to do something other than paid work. MOC was calculated by multiplying the maternal hourly wage by the total daily hours spent pumping. Maternal hourly wage was estimated using the imputation model of Winkler et al.²⁶ The imputation model predicts wage using a regression equation where the independent predictor variables are maternal characteristics, including age, age squared, education, race, ethnicity, and geographic location.

Procedures

The procedures for the original study were detailed in previous publications^23,24 and are briefly summarized here. Mothers were enrolled in the study within 48 hours of giving birth and were instructed to simultaneously pump both breasts six to eight times per day for a minimum of 15–20 minutes per session until the onset of lactogenesis II. Lactogenesis II is the point in time when the mother's milk “comes in.” After this time, mothers were instructed to continue pumping six or eight times per day and to pump until the steady flow of milk had stopped for 2 minutes. Demographic data were collected using a questionnaire. Time spent pumping and milk output were collected using a standardized milk output log for 3 weeks following study enrollment or until the mother met either of the following conditions: (1) mother discontinued pumping or (2) mother completed all study questionnaires and interviews.

Complete time and milk output data were available for 99.2% of the 13,273 pumping sessions. The procedures for missing data were as follows: If time spent pumping was recorded but not milk output, output was estimated using the mean of the following two estimates: the average milk output during that pumping session on the day before and the day after the missing pumping session and the average milk output for each session that study day; If milk output was recorded but not time spent pumping, time was estimated using the mean of the following two estimates: the average of the time spent pumping during that pumping session on the day before and the day after the missing pumping session and the average of the time spent pumping for each session that study day.

Analysis

Data were managed using SPSS for Windows Version 14 (SPSS, Inc., Chicago, IL) and Microsoft (Redmond, WA) Excel 2003. Frequencies and descriptive statistics were used to describe the data. Descriptive statistics were calculated for each day of the study period and for the entire study period.

Cost analysis

The cost of providing 100 mL of HM was calculated using the costing techniques outlined by Drummond et al.²⁷ In brief, these techniques consist of four steps: (1) define analysis assumptions; (2) identify and measure factors that influence cost; (3) conduct a sensitivity analysis; and (4) evaluate and compare the financial value of the next alternatives: Donor milk and commercial formula. All costs are presented in 2008 U.S. dollars.

The sources of cost for this study were the BPRF, the BPCK, and the MOC. Both the BPRF and BPCK were considered fixed costs because the mother pays the BPRF and BPCK cost regardless of whether or not she pumps for the whole month. MOC was a variable cost because the time pumped varied for each day throughout the study period. The total cost of providing HM was measured as: cost per 100 mL = ([BPRF + BPCK + MOC]/total milk output) × 100.

After the cost per 100 mL was calculated using the above equation, a one-way sensitivity analysis was conducted to determine which of the three components of the cost estimate (BPRF, BPCK, or MOC) most influenced the total cost of providing HM. Sensitivity analyses are performed by replacing the original cost estimates with alternative estimates obtained from market prices such as those listed by various suppliers and retailers. Published market prices from suppliers and retailers were used in this sensitivity analysis to be conservative in calculating the variation in the cost of providing HM. This conservative approach was used because published market prices are typically higher than an institution's negotiated price. Additionally, published market prices represent a more accurate determination of the maternal provision cost if she must incur these costs herself.

The sensitivity for the BPRF and BPCK was investigated using average market prices (± 2 SD). The mean BPRF and BPCK from the market review were $69.92 (SD = $10.45) and $50.08 (SD = $4.57), respectively.

The sensitivity of the cost estimate to MOC was examined using a planned wage variation. The variation was the 2008 federal minimum wage ($6.55)²⁸; this value is the minimum monetary value for legal employment in the United States regardless of state and local government mandates.

Finally, the cost of HM per 100 mL was compared to the next best alternatives (donor milk and commercial formula). Donor milk is purchased by the hospital to ensure that the milk meets hospital standards for food safety and infection control. The mean cost per 100 mL for donor milk was obtained by contacting several institutions that purchase donor milk and milk banks that sell donor milk (D.L. Spatz, N.M. Hurst, and D.J. Miracle, personal communications). The mean cost per 100 mL of donor milk was $13.59 per 100 mL (SD = $1.40).

The mean cost per 100 mL for commercial formula was obtained using a three-step procedure. First, budgetary reports documenting the unit price paid by the hospital for formula products for VLBW infants at one of the study institutions were examined (M. Fogleman, Rush University Medical Center departmental budget for formula, 2009). Next, formulas were classified into three categories: powdered preterm formula (Enfacare Lipil 22^® [Mead Johnson, Evansville, IN], Neosure^® [Abbott Nutrition, Columbus, OH]), powdered specialty formula (Nutramigen^® [Mead Johnson], Portagen^® [Mead Johnson], Pregestimil^® [Mead Johnson], Elecare^® [Abbott Nutrition], Neocate^® [Nutricia North America, Rockville, MD]), and liquid ready-to-feed preterm formula (Specialcare^® [Abbott Nutrition] ready-to-feed). Next, the price per 100 fluid mL was calculated by a registered NICU dietitian for each powdered formula using the two most common caloric dilutions: 20–22 Cal/oz and 24–27 Cal/oz (M. Fogleman, Rush University Medical Center institutional formulary for mixing formula, 2009). The cost per 100 fluid mL for each dilution was calculated by converting the cost per fluid ounce to the cost per milliliter multiplied by 100 mL. Lastly, a cost estimate for 100 mL of sterile water to mix the formula was added to each powdered formula estimate to obtain the final cost per 100 mL of powdered formula (Rush University Medical Center, purchasing department electronic price list for sterile water, 2009). Table 2 displays the cost per 100 mL of formula by category and caloric dilution. The cost per 100 mL ranged from $0.63 for regular 20–22 Cal/oz powdered preterm formula to $2.97 per 100 mL for liquid ready-to-feed 20–22 Cal preterm formula.

Table 2.

The Average Cost of Preterm Formula by Formula Category and Caloric Dilution (2008 U.S. Dollars)

	Average cost for 100 mL of formula	Cost for 100 mL sterile water	Total cost per 100 mL
20–22 Cal/oz
Preterm formulaa	$0.58	$0.05	$0.63
Specialty formulab	$0.84	$0.05	$0.89
Liquid ready-to-feed preterm formula	$2.97	NA	$2.97
24–27 Cal/oz
Preterm formulaa	$0.69	$0.05	$0.73
Specialty formulab	$1.06	$0.05	$1.11

NA, not applicable.

^aPreterm formula includes Enfacare Lipil 22 and Neosure.

^bSpecialty formula includes Nutramigen, Portagen, Pregestimil, Elecare, and Neocate.

Results

The mean daily milk output and time spent pumping were 558.2 mL (SD = 320.7; range = 0–2,024) and 98.7 minutes (SD = 38.6; range = 0–295), respectively. These measures varied throughout the study period (Fig. 1). Mean daily milk output increased steadily during the first 8 study days and then remained relatively constant through study day 24. Mean daily time spent pumping increased sharply during the first 4 study days and then trended downward through study day 26. Greater variability in milk output and time spent pumping during the last week of this study was probably due to the declining sample size (75% of the sample had completed the original study by day 24).

FIG. 1.

The mean daily milk output (in mL) and the mean daily time spent pumping (in minutes) for study participants (n = 111).

The mean cost of providing 100 mL of HM was calculated with and without MOC (Table 3). When MOC was included, the mean cost of providing 100 mL of HM varied from $2.60 to $6.18. When MOC was excluded, the mean cost of providing 100 mL of HM varied from $0.95 to $1.55.

Table 3.

Mean Maternal Cost of Providing 100 mL of HM (n = 111) Calculated Using Actual Hospital Charges and Estimated Market Prices (2008 U.S. Dollars)

	Actual hospital charges		Mean market price
	Pump only ($35)	Pump and kit ($92)	Pump ($69.92) and kit ($50.08)	Lower limit pump ($49.02) and kit ($40.94)	Upper limit pump ($90.82) and kit ($59.22)
Calculated without MOC	$0.36	$0.95	$1.24	$0.93	$1.55
Calculated with MOC using
Federal minimum wage ($6.55)	$2.60	$3.19	$3.48	$3.17	$3.79
Imputed wage (mean = $13.44)	$5.00	$5.58	$5.87	$5.56	$6.18

Figure 2 is a cost curve that depicts the average cost of providing 100 mL of HM for each day throughout the study period. The cost per 100 mL declined with every additional day of pumping. The curve is stable through day 24 of the study and becomes less stable thereafter, which was probably related to the declining sample size during the last week of the study.

FIG. 2.

The average daily cost of 100 mL of HM for pumping supply costs (BPRF and BPCK): actual hospital charges and estimated market price (2008 U.S. dollars).

The sensitivity analysis (Table 3) demonstrates that the cost per 100 mL of HM is sensitive to all three of the cost components: BPRF, BPCK, and MOC. MOC was the largest contributor to total cost per 100 mL of HM, with its addition tripling the cost per 100 mL. However, variation in maternal wage had less effect on total cost compared to variation in the cost of the BPRF and BPCK. For example, increasing wage from the federal minimum wage to the imputed wage, a 96.2% increase, resulted in only a 92.3% increase in the cost per 100 mL. In contrast, the cost per 100 mL changed proportionately with variations in the cost of the BPRF and BPCK. Changing the cost of the BPRF and BPCK from the actual hospital charges to the higher limit mean market cost, a 63.2% increase, resulted in a 63.1% increase in the cost per 100 mL.

The cost per 100 mL of HM ($0.95–$6.18) was less expensive than the cost per 100 mL of donor milk ($13.59) regardless of the variations in the cost of the BPRF, BPCK, and MOC. The cost per 100 mL of HM including MOC was less expensive than liquid ready-to-feed preterm formula ($2.97) when MOC was valued at the federal minimum wage ($2.60). In contrast, the cost per 100 mL of HM including MOC was more expensive than all formulas for all higher variations in MOC. The cost per 100 mL of HM excluding MOC ($0.95–$1.55) was less expensive than the 24–27 Cal/oz specialty formula ($1.06) and the liquid ready-to-feed preterm formula ($2.97). The cost per 100 mL of HM, excluding MOC, was more expensive than the 20–22 Cal/oz and 24–27 Cal/oz preterm formula and the 20–22 Cal/oz specialty formula.

Discussion

This is the first study to estimate the initial maternal cost of providing 100 mL of HM for VLBW infants and included only the required components that all mothers need to provide their HM: The BPRF, the BPCK, and the maternal time. The findings from this study suggest that institutions and payers may actually realize a cost savings by reimbursing costs for the BPRF and the BPCK. Our data reveal that the cost of HM from the infant's own mother ($0.95–$1.55) is less expensive than donor milk ($13.59), specialty formula ($1.06), and ready-to-feed formula ($2.97) that are commonly used in the NICU for this population.

This analysis also demonstrated that the cost per 100 mL of HM declined with each additional day of pumping (Fig. 2). This cost reduction is a predictable outcome because maternal milk output increases significantly during the early post-birth period²⁹ without an increase in pumping time or fixed costs. The increasing daily milk output rapidly mitigates the initial costs of providing milk, with a low mean cost per 100 mL of $1.03 on study day 22. Although not examined in this study, the cost per 100 mL in the second month post-birth would likely be lower because the fixed cost of the BPCK is removed and milk output should remain stable or increase.²⁹ Removing the cost of the BPCK in month 1 supports this speculation because the mean cost per 100 mL of HM decreases to $0.36 (Table 3). These findings suggest a return on investment for the BPRF and BPCK cost during the first weeks of pumping. Moreover, this trend suggests that the return on investment would remain stable or potentially increase with additional months of pumping. This trend also suggests that HM from the infants' own mother is less expensive than donor milk and commercial formula over the entire NICU stay.

These findings have important implications for clinical practice and health policy. To initiate lactation for a VLBW infant, mothers must pay between $90 and $150 for the first month's BPRF and the purchase of the BPCK. Most payers and institutions do not reimburse or pay these costs, so they are borne by the mothers.^20–24 This initial cost can be prohibitive because preterm birth means other unexpected expenses (e.g., increased insurance co-payments, hospital parking fees, lost wages). Programs that provide the BPRF and BPCK to mothers have reported high rates of HM feedings, even among low-income families.^19,20,22 These successes dispel the myth that mothers do not initiate lactation because they are disinterested or unmotivated.^20,22 Rather, the success of these programs implies that mothers will provide their milk when the cost barriers associated with lactation for a VLBW infant are removed or reduced.

In this analysis, MOC was the largest overall contributor to the cost of providing 100 mL of HM, and removing this cost substantially reduced the cost per 100 mL (Table 3). This finding raises the fundamental question of how MOC should be treated in economic analyses of breastfeeding or providing HM. From an economic perspective, opportunity cost is a standard measure in cost analyses because it captures the value of a person's time when she or he chooses one activity, such as pumping, over other activities, such as paid work or leisure time.

In economic analyses of breastfeeding or providing HM, it can be argued that MOC should be excluded from cost analyses because most mothers are not medically eligible to return to paid work during the first weeks post-birth. However, other experts contend that including the value of maternal time allows society to examine the economic contribution that mothers make when they choose to breastfeed or provide HM.³⁰ Because both arguments are legitimate, we have chosen to report the cost of providing 100 mL of HM with and without MOC, an approach that seems pragmatic for subsequent economic studies of breastfeeding and providing HM.

Finally, this study examines the initial maternal cost of providing HM and only analyzed the costs for items that are universally required by all mothers of VLBW infants to initiate HM feeding, including the BPRF, the BPCK, and maternal time. This study is a necessary first step in determining the cost of HM feedings because if a mother does not initiate pumping, then other potential HM-associated costs, including other maternal and institutional costs,^19,20 are not relevant. Future analyses should examine additional potential sources of maternal cost (e.g., breastfeeding bras, pads, creams, pillows, additional dietary needs, prescription pharmaceuticals) and institutional costs (e.g., HM storage containers, lactation specialist care, hospital freezer space) that are associated with HM feeding.^19,20

Conclusions

In conclusion, the findings from this study estimate that the initial maternal cost of providing her milk for her VLBW infant is less than $1.00 per 100 mL and that her milk costs less than donor milk and many types of commercial formula used in the NICU. Although the cost of providing HM is relatively small, it remains an economic barrier for many mothers. Removal of this economic barrier is the first step toward ensuring that access to effective lactation equipment becomes a standard of patient care for all mothers of VLBW infants. Providers, payers, and Women, Infants and Children (WIC) clinics can use these findings to justify the cost of providing BPRF and BPCK for mothers of VLBW infants.

Acknowledgments

This article was funded by grants AHRQ 1 R36 HS016012 and NIH 1 R01 NR010009. The original clinical trial that provided the dataset for this analysis was funded by Medela, Inc. The authors would like to thank the following individuals for their contribution to this manuscript: Mary Fogleman, M.S., NICU dietitian at Rush University Medical Center; Nancy M. Hurst, Ph.D., Assistant Director, Lactation Program and Mother's Own Milk Bank, Texas Children's Hospital and Assistant Professor of Pediatrics, Baylor College of Medicine; Donna J. Miracle, Ph.D., Director of Research and Clinical Operations, Indiana Mothers' Milk Bank, Inc.; and Diane L. Spatz, Ph.D., Helen M. Shearer Term Associate Professor of Nutrition and Associate Professor of Healthcare of Women and Childbearing Nursing, Penn and Clinical Nurse Specialist at the Children's Hospital of Philadelphia.

Disclosure Statement

B.J.J. has received research funding from Medela, Inc. P.M. has received research funding from Medela, Inc. and Prolacta Bioscience, has received speaking honoraria from Medela, Inc. and Abbott Nutritionals, and has provided consultation to special projects for Medela, Inc. J.L.E. has received research funding from Medela, Inc. T.M. has no conflict of interest to declare for this article.