IFC Commentary on Bartick and Reinhold 2011
Related Materials
This commentary is in response to the article by Bartick and Reinhold, "The Burden of Suboptimal Breastfeeding in the United States: A Pediatric Cost Analysis." 1 The International Formula Council* supports the position of the World Health Organization, the American Academy of Pediatrics and other organizations that breastfeeding is ideal and offers specific child and maternal benefits. However, we have several major concerns about the methodology and assumptions underpinning this analysis:
1) The cost estimate of suboptimal breastfeeding is heavily dependent (73%) on what we believe is an inaccurate determination of value lost due to preventable infant deaths.
2) The authors relied on only one report for data on the reduction in disease risks associated with breastfeeding (a report by the Agency for Healthcare Research and Quality, i.e. AHRQ), and did not comment on its limitations nor comment on the findings of other researchers/ authoritative bodies and subsequent studies that have differed from results presented in the AHRQreport.
3) We question the basis for the selection of a 90% exclusive breastfeeding benchmark—this rate seems arbitrary and substantially exceeds current U.S. government public health breastfeeding goals.
Concerns regarding the cost estimate of suboptimal breastfeeding
Bartick and Reinhold assigned a value of $10.5 million to each of the calculated "excess deaths." Every life is of value and, although we strongly disagree with the view of assigning a dollar value to each human life, this was the methodology Bartick and Reinhold employed for their analysis. The resulting $9.5 billion figure calculated due to "excess death" contributes 73% of the total $13 billion "savings to the U.S." that is referenced in the conclusion. To test the plausibility of these values, we estimated that with a life expectancy of 77.7 years (U.S. average in 2007) and with a 3% annual rate of return, $10.5 million accumulated in a lifetime would yield a constant annuity stream of $342,000 per year. This is 7.9 times the mean real gross domestic product (GDP) per capita in 2007 (Bureau of Labor Statistics, 2009). Thus the revealed preference job risk approach that the authors chose values each year of life saved almost eight times the value of per capita annual product or consumption. In addition, their estimate did not incorporate the costs of resources (food, shelter, environmental impact, etc.) needed for an infant to become a productive adult. Given these generous assumptions and the lack of consideration for the costs of resources needed for each human life, we question the application of a $10.5 million value to each "excess death."
Concerns regarding the data source for reduction in disease risk
Bartick and Reinhold relied solely on the 2007 breastfeeding report from AHRQ for relative risk for the various diseases considered. 2 The AHRQ report, as an analysis of meta-analyses, is fundamentally limited by the value of the underlying observational data that were used. There are studies designed to show cause and effect between breastfeeding and certain prevalent health measures, for example the randomized prospective trial PROBIT,3 that the AHRQ report acknowledged but did not include. PROBIT enrolled more than 17,000 mother infant pairs and was focused on prevalent health measures. However, even PROBIT lacked sufficient power to determine a causal relationship between breastfeeding and excess deaths.
In the Bartick and Reinhold study, the three quantitatively most important diseases in contributing to the overall cost estimate to the U.S. were SIDS death, necrotizing enterocolitis (NEC) hospital costs and deaths, and deaths from lower respiratory infection (LRTI).
PROBIT reported a non-statistically significant difference in SIDS between treatment groups, (p = 0.12). Additionally, PROBIT found no difference in the risk of respiratory tract infections between control and breastfeeding promotion groups. These findings were not included in the AHRQ analysis nor in Bartick and Reinhold’s analysis.
Similarly, other authoritative bodies including American Academy of Pediatrics (AAP) have reached different conclusions than those reached in the AHRQ report. Regarding SIDS, it is notable that in 2005, the AAP Task Force on SIDS concluded "Although breastfeeding is beneficial and should be promoted for many reasons, the task force believes that the evidence is insufficient to recommend breastfeeding as a strategy to reduce SIDS."4 The differences in conclusions between AHRQ and AAP could be related to which original studies were considered in each analysis. Use of Bartick and Reinhold’s study as a basis of public policy must consider the selected nature of the science on breastfeeding on which the AHRQ report rests.
Regarding NEC, a concern about Bartick and Reinhold’s study is that each NEC death is estimated to have excess direct costs to the U.S. of $10.5 million (the same value estimate assumed for all other "excess deaths" addressed in the study) due to lost contribution of earnings from premature deaths of infants. As addressed earlier in this commentary, we question the basis for this assumption. Further, the population of infants susceptible to NEC, a disease that can occur in term infants, although it is more commonly seen in premature infants often have complications and thus poorer neurological and developmental outcomes than healthy full-term infants. Given their medical history, we question whether all NEC infants will have the same potential earnings as healthy term infants; however, this was the assumption made in Bartick and Reinhold’s analysis. And finally, Bartick and Reinhold used an odds ratio to determine excess costs due to NEC calculated based on exclusive breastfeeding rates at 3 months (32.1%) instead of initial breastfeeding rates (74.1%), even though NEC usually occurs in the first few weeks of life (most typically between 30-32 weeks post-conceptional age). Applying the odds ratio to an initial rate of breastfeeding decreases Bartick and Reinhold’s excess costs by more than half, a correction that applies independently from the other factors discussed above.
The third major contributor to Bartick and Reinhold’s excess cost estimate is LRTI mortality. Bartick and Reinhold used the crude OR from the AHRQ report, which based its conclusions on a meta-analysis5 in which the only confounders controlled for were either smoking or socioeconomic status. The meta-analysis did not control for male gender, low birthweight, third or later child, low APGAR score, low maternal education, and young maternal age, all identified by the Centers for Disease Control and Prevention (CDC) as risk factors for LRTI deaths,6 that co-vary with breastfeeding. Therefore, there is a high likelihood of residual confounding that was included in Bartick’ and Reinhold’s analysis, potentially over-estimating the relationship between breastfeeding and LRTI. Bartick and Reinhold also did not take into account the large, randomized prospective trial (PROBIT) for breastfeeding promotion that found no difference in the risk of respiratory tract infections between control and breastfeeding promotion groups.3
Concerns regarding selection of a 90% exclusive breastfeeding rate
Finally, it is important to remember that Bartick and Reinhold’s calculations and what has been promoted in the popular press, are based on achieving a rate of 90% exclusive breastfeeding for six months. There was no substantiation or detail given on selecting the 90% exclusive breastfeeding rates for six months number. Indeed, this choice seems arbitrary and is not in line with any U.S. government public health goals for breastfeeding. The 90% rate is far higher than current Healthy People 2020 Goals and it is also far beyond the increased rates of breastfeeding that the U.S. Preventative Health Service Task Force reports as actually achieved through a variety of breastfeeding promotion efforts.7
We do not question that breastfeeding provides optimal nutrition for infants and we agree that there may be real public health value to increased breastfeeding support. New mothers are already extraordinarily attuned to doing all they can to benefit the health and development of their infants. Bartick and Reinhold suggest that actions to improve breastfeeding rates, duration, and exclusivity could be cost effective. However to truly assess the cost-effectiveness of increased breastfeeding support, both the costs and the effectiveness need stronger evidence than the Bartick and Reinhold analysis provides.
*The International Formula Council is an association of manufacturers and marketers of formulated nutrition products, e.g., infant formulas and adult nutritionals, whose members are based predominantly in North America. IFC members are: Abbott Nutrition; Mead Johnson Nutrition; Nestlé Infant Nutrition; PBM Products, LLC, A Perrigo Company; and Pfizer Nutrition.
References Cited:
1. Bartick M, Reinhold A. The burden of suboptimal breastfeeding in the United States: a pediatric cost analysis. Pediatrics;125 (5):e1048-56.
2. Ip S, Chung M, Raman G, et al. Breastfeeding and Maternal and Infant Health Outcomes in Developed Countries. Rockville, MD: Agency for Healthcare Research and Quality; 2007. Evidence report/technology assessment No. 153.
3. Kramer, M. S., B. Chalmers, et al. "Promotion of Breastfeeding Intervention Trial (PROBIT): A Randomized Trial in the Republic of Belarus." JAMA; 2001 285(4): 413-420.
4. American Academy of Pediatrics, Task Force on Sudden Infant Death Syndrome. The Changing Concept of Sudden Infant Death Syndrome: Diagnostic Coding Shifts, Controversies Regarding the Sleeping Environment, and New Variables to Consider in Reducing Risk Pediatrics 2005;116:1245–1255.
5. Bachrach VR, Schwarz E, Bachrach LR. Breastfeeding and the risk of hospitalization for respiratory disease in infancy: a meta-analysis. Arch Pediatr Adolesc Med 2003;157(3):237-43.
6. Singleton et al, Risk Factors for Lower Respiratory Tract Infection Death Among Infants in the United States, 1999 –2004 Pediatrics 2009;124:e768–e776.
7. Chung M, Raman G, Trikalinos T, Lau J, Ip S. Interventions in Primary Care to Promote Breastfeeding: An Evidence Review for the U.S. Preventive Services Task Force. Ann Intern Med. October 21, 2008 2008;149(8):565-582.