Clarification of the methods and statistics in the study “Planned home birth and the association with neonatal hypoxic ischemic encephalopathy”

Shane W. Wasden; Stephen T. Chasen; Jeffrey M. Perlman; Jessica L. Illuzzi; Frank A. Chervenak; Amos Grunebaum; Heather S. Lipkind

doi:10.1515/jpm-2017-0080

Article Publicly Available

Clarification of the methods and statistics in the study “Planned home birth and the association with neonatal hypoxic ischemic encephalopathy”

Shane W. Wasden , Stephen T. Chasen , Jeffrey M. Perlman , Jessica L. Illuzzi , Frank A. Chervenak , Amos Grunebaum and Heather S. Lipkind

Published/Copyright: July 28, 2017

Published by

Become an author with De Gruyter Brill

Submit Manuscript Author Information Explore this Subject

From the journal Journal of Perinatal Medicine Volume 46 Issue 2

Dear Editor,

Thank you for the letter [1] and the opportunity to address the outlined concerns regarding our article [2]. Hopefully this response will clarify the areas of confusion, and address the statistical concerns.

In the abstract there is a typographical error, and the incorrect confidence interval was reported. The correct confidence intervals were reported within the results section of the paper. Thank you for drawing our attention to this error.

In regards to Table 3: All planned home births were attended by midwives, and all unplanned out of hospital births were unattended except for one which was attended by the emergency medical services (EMS). In this investigation we sought to evaluate the association between the location of delivery and hypoxic ischemic encephalopathy (HIE), and location of birth was analyzed as hospital birth or out of hospital birth. Out of hospital births were further stratified as unplanned out of hospital delivery, and planned home birth. Please note that this distinction is clearly shown in Table 4 as well as described in the results section. The association between HIE and birth attendant (or lack thereof) was not directly evaluated. However, neonates with HIE had an increased odds of unplanned out of hospital birth (most of which were unattended) and an increased odds of planned home birth (all of which were attended by midwives). A limitation of this investigation is that more detailed information regarding the midwives and their level of training is not available within this dataset.

As noted in our discussion, and reflected in your comments, location of delivery encompasses several important determinants of poor neonatal outcome, including quality of intrapartum fetal monitoring, expertise of maternity care provider and access to acute care services. Suboptimal circumstances, or care, in any of these domains could increase the risk for poor neonatal outcome.

As for the statistical concerns, when we perform a conditional regression on the match for planned home birth the adjusted odds ratio increases from 21.0 (1.7–256.4) to 38.2 (2.43–600.13). In Table 4, the list of variables included in the adjusted regression analysis are specifically listed in the footnote including diabetes, hypertension, prenatal care, race/ethnicity, maternal age, education, payer, parity, pre-pregnancy weight and mode of delivery.

In order to test whether there were too many variables in our regression models, we performed backwards-stepwise elimination removing one variable at a time to test whether the model was better or worse with inclusion in the model by examining the F statistic. In the end, the model was optimized with the inclusion of the variables mode of delivery and hypertension yielding a significant odds ratio of 23.9 (2.0–291.6). At the request of reviewers, many publications ultimately provide the odds ratios adjusted for all variables of interest rather than only those that turn out to be statistically significant confounders, if this does not change the conclusion of the paper. In this case, reducing the number of variables in the model does not change the conclusion, but it does narrow the confidence intervals, which is a benefit.

As authors, we do acknowledge that the wide confidence intervals in this study are a clear indication for further larger studies of this topic. This data set is unique in that it includes hospital data of infants with HIE from a large region but there is more work to be done to further explore the risks of home birth. Although we acknowledge that there are limitations to this investigation there are many strengths including our use of a linked dataset and detailed information on HIE. This study does raise concerns about the potential safety of planned home birth in this and other similar regions.

Corresponding author: Shane W. Wasden, MD, Department of Obstetrics and Gynecology, Division of Maternal-Fetal Medicine, New York Presbyterian Hospital, Weill Cornell Medical Center, 525 East 68th Street, Suite J-130, New York, NY 10065, USA, Tel.: (212) 746-1604

Author’s statement
Conflict of interest: Authors state no conflict of interest.
Informed Consent: Not applicable.
Ethical approval: Not applicable.

References

[1] Maimburg RK. Clarification of the methods and statistics in the study “Planned home birth and the association with neonatal hypoxic ischemic encephalopathy”. J Perinat Med. 2018;46:225–6.10.1515/jpm-2016-0412Search in Google Scholar PubMed

[2] Wasden SW, Chasen ST, Perlman JM, Illuzzi JL, Chervenak FA, Grunebaum A, et al. Planned home birth and the association with neonatal hypoxic ischemic encephalopathy. J Perinat Med. 2017;45:1055–60.10.1515/jpm-2016-0292Search in Google Scholar PubMed

Received: 2017-3-10

Accepted: 2017-4-28

Published Online: 2017-7-28

Published in Print: 2018-2-23

Articles in the same Issue

https://doi.org/10.1515/jpm-2017-0080