Letters
Corresponding Author: Lesley Rathbun, MSN, CNM, FNP, American Although adverse neonatal outcomes to low-risk mothers
Association of Birth Centers, 3123 Gottschall Rd, Perkiomenville, PA 18074
(lesley@charlestonbirthplace.com).
Conflict of Interest Disclosures: Ms Rathbun has completed and submitted the
ICMJE Form for Disclosure of Potential Conflicts of Interest and reported being
the president of the American Association of Birth Centers. No other disclosures
were reported.
1. Woo VG, Milstein A, Platchek T. Hospital-affiliated outpatient birth centers:
a possible model for helping to achieve the triple aim in obstetrics. JAMA. 2016;
316(14):1441-1442.
2. Stapleton SR, Osborne C, Illuzzi J. Outcomes of care in birth centers:
demonstration of a durable model. J Midwifery Womens Health. 2013;58(1):
3-14.
3. Thornton P, McFarlin BL, Park C, et al. Cesarean outcomes in US birth
centers and collaborating hospitals: a cohort comparison. J Midwifery Womens
Health. 2016.
In Reply Our Viewpoint examined the concept of birth centers
and suggested a path to broader adoption in the United States
through a network of hospital-affiliated birth centers, a model
used in the United Kingdom.
The Birthplace in England national prospective cohort
study accounted for 95% of freestanding midwifery units
in the United Kingdom, collecting comprehensive data
across different delivery settings with high response rates.1
We thought the generalizability of these data were stronger
than available US studies. The outcomes were excellent for
low-risk women and their infants and provide a strong case
for hospital-affiliated outpatient birth centers.
By contrast, the most comprehensive study evalu-
ating outcomes among US birth centers comes from a vol-
untary registry run by the American Association of Birth
Centers2 into which birth centers voluntarily opt in or out.
The study 2 included 79 birth centers, which represent
approximately 25% of the freestanding birth centers in the
United States. Although these data demonstrate good mater-
nal and infant outcomes, the potential for selection bias
reduces generalizability.
We agree that accreditation of birth centers is essential for
enhancing safe care. However, approximately two-thirds of US
birth centers are not accredited, which undermines confi-
dence in the accreditation process.
Additionally, we advocate for a more robust collaboration
protocol than is outlined by the Commission for the Accredi-
tation of Birth Centers.3 Although standards by the American
Association of Birth Centers require a “transfer arrangement
with a hospital,” the description of such an agreement is
minimal. A prearranged plan for access to acute services
must exist, but no written agreement is required between the
birth center and the receiving facility. The birth center does
not have to provide practice protocols to collaborating physi-
cians or hospitals unless specifically asked. The only handoff
required before transfer is a call to the hospital. We envision a
more collaborative process in which hospital-based physi-
cians and birth center midwives agree and mutually create
practice protocols, including conditions necessitating patient
transfer and logistical arrangements for safe transfer. Such
processes could lead to significantly better care for patients
and enhanced hospital-based clinician support of a birth cen-
ter model.
646 JAMA February 14, 2017 Volume 317, Number 6 (Reprinted)
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/936050/ by a Washington University - St Louis User on 02/14/2017
are rare, they occur. One study found that over a 2-year
period in Oregon, planned out-of-hospital births were associ-
ated with higher rates of perinatal death than planned
in-hospital births. 4 Although these data are not directly
applicable to outpatient birth centers because they included
all planned out-of hospital births including home births, they
raise concerns regarding the risks of out-of-hospital births.
Certified nurse midwives are trained in neonatal re-
suscitation, but in their role as primary caregivers, their
focus should be on the delivering mother. Furthermore,
although many perinatal clinicians, including nurse mid-
wives and obstetricians, are trained in newborn care and
maintain Neonatal Resuscitation Program certification,
resuscitation skills significantly deteriorate 2 to 3 months
after certification.5 To mitigate skill loss, we advocate that
an advanced-practice clinician dedicated to newborn care,
such as a pediatrician or nurse practitioner, work in both
hospital and birth center settings to maintain their resusci-
tation skills. These clinicians should be immediately avail-
able for birth center deliveries needing emergent neonatal
resuscitation.
Victoria G. Woo, MD
Arnold Milstein, MD
Terry Platchek, MD
Author Affiliations: Clinical Excellence Research Center, Stanford University,
Stanford, California (Woo, Milstein); Department of Pediatrics, Lucile Packard
Children’s Hospital, Palo Alto, California (Platchek).
Corresponding Author: Victoria G. Woo, MD, Clinical Excellence Research
Center, Stanford University, 75 Alta Rd, Stanford, CA 94305
(vwoo1@stanford.edu).
Conflict of Interest Disclosures: The authors have completed and submitted
the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were
reported.
1. Hollowell J, Puddicombe D, Rowe R, et al. The Birthplace national prospective
cohort study: perinatal and maternal outcomes by planned place of birth.
Birthplace in England research programme. Final report part 4. http://www.nets
.nihr.ac.uk/__data/assets/pdf_file/0006/84948/SDO_FR4_-08-1604-140_V04
.pdf. Accessed December 2, 2016.
2. Stapleton SR, Osborne C, Illuzzi J. Outcomes of care in birth centers:
demonstration of a durable model. J Midwifery Womens Health. 2013;58(1):3-14.
3. Commission for the Accreditation of Birth Centers. Stay Current with
the CABC Indicators R.Ed v.1.1 (effective 6/15/2016). https://www
.birthcenteraccreditation.org/go/get-cabc-indicators/. Accessed December 21,
2016.
4. Snowden JM, Tilden EL, Snyder J, Quigley B, Caughey AB, Cheng YW.
Planned out-of-hospital birth and birth outcomes. N Engl J Med. 2015;373(27):
2642-2653.
5. Patel J, Posencheg M, Ades A. Proficiency and retention of neonatal
resuscitation skills by pediatric residents. Pediatrics. 2012;130(3):515-521.
Oxygen Supplementation Among Patients
in the Intensive Care Unit
To the Editor In the randomized clinical trial among critically
ill patients comparing conservative (partial arterial oxygen
pressure [PaO2] of 70-100 mm Hg or arterial oxyhemoglobin
saturation [Sp O 2 ] of 94%-98%) with conventional (Pa O 2
of <150 mm Hg or Sp O 2 of 97%-100%) oxygen therapy,
Dr Girardis and colleagues1 observed an absolute decrease in
jama.com

intensive care unit (ICU) mortality of 8.6% in the conserva-
tive vs conventional therapy groups. The study has several
shortcomings.
First, adherence to oxygenation targets could not be reli-
ably assessed. Oxygenation and oxygen exposure were re-
ported as time-weighted PaO2 and fraction of inspired oxygen
(FiO2). However, arterial blood samples were only analyzed on
clinical indication, at times just once a day. The frequency of
FiO2 measurements was not presented, and SpO2 values, al-
though measured continuously, were not reported.
Second, it seems unlikely that an absolute time-weighted
FiO2 difference of 3% between groups (36% in conservative
group vs 39% in conventional group) would generate a rela-
tive reduction in ICU mortality of 43%. Such an effect is not
supported by previous studies with substantially larger dif-
ferences in FiO2 of 5% to 10%, such as a pilot randomized clini-
cal trial2 that found no difference in ICU mortality and a before-
and-after trial with similar ICU mortality and a relative
reduction of hospital mortality of 11%.3 This suggests that other
factors, such as imbalances in baseline characteristics, might
have contributed to the survival difference.
Third, despite the well-known rapidly increasing risk of
pulmonary damage at FiO2 levels above 60%,4 the protocol did
not describe an upper FiO2 limit to attain the high PaO2 tar-
gets in the conventional group. Only median FiO2 levels were
shown; range or interquartile range was not reported. In our
opinion, for the safety of the conventional group, the proto-
col should have included an upper limit of FiO2 as has been done
in other trials.2 We therefore believe that the results of the
Oxygen-ICU trial should be interpreted with caution.
Angelique M. E. Spoelstra-de Man, MD, PhD
Heleen M. Oudemans-van Straaten, MD, PhD
Armand R. J. Girbes, MD, PhD
Author Affiliations: Department of Intensive Care, Vrije Universiteit Medical
Center, Amsterdam, the Netherlands.
Corresponding Author: Angelique Spoelstra-de Man, MD, PhD, Vrije
Universiteit Medical Center, Department of Intensive Care, Amsterdam,
De Boelelaan 1117, 1081 HV, Amsterdam, the Netherlands (am.spoelstra@vumc.nl).
Conflict of Interest Disclosures: The authors have completed and submitted
the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported
performing a similar study with funding from the Netherlands Organization of
Health Research and Development and the European Society of Intensive Care
Medicine.
1. Girardis M, Busani S, Damiani E, et al. Effect of conservative vs conventional
oxygen therapy on mortality among patients in an intensive care unit: the
Oxygen-ICU randomized clinical trial. JAMA. 2016;316(15):1583-1589.
2. Panwar R, Hardie M, Bellomo R, et al; CLOSE Study Investigators; ANZICS
Clinical Trials Group. Conservative versus liberal oxygenation targets for
mechanically ventilated patients: a pilot multicenter randomized controlled
trial. Am J Respir Crit Care Med. 2016;193(1):43-51.
3. Helmerhorst HJ, Schultz MJ, van der Voort PH, et al. Effectiveness and
clinical outcomes of a two-step implementation of conservative oxygenation
targets in critically ill patients: a before and after trial. Crit Care Med. 2016;44
(3):554-563.
4. Kallet RH, Matthay MA. Hyperoxic acute lung injury. Respir Care. 2013;58(1):
123-141.
In Reply We agree with Dr Spoelstra-de Man and colleagues that
the results of the Oxygen-ICU trial should be interpreted with
jama.com
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/936050/ by a Washington University - St Louis User on 02/14/2017
Letters
caution. The design of the study (ie, single center, open label)
and early termination may have exaggerated the results and
an appropriate confirmatory trial is needed.1 Moreover, the
mechanisms by which mild hyperoxia increased bacteremia,
shock, and liver failure in the conventional group remain un-
clear and should be the objective of specific studies.
Nevertheless, the mortality reduction observed in our
study (absolute risk reduction [ARR], 8.3%, relative risk
reduction [RRR], 43%) is similar to data reported in the pilot
randomized trial by Panwar and colleagues2 in a selected
population. In this trial, the overall analysis indicated no
difference in mortality between conservative and liberal
oxygen strategies among patients with an expected length
of mechanical ventilation more than 24 hours. However, in
the prespecified subgroup of patients with respiratory fail-
ure (ie, PaO2/FiO2 < 300 mm Hg), including 67 of 103 ran-
domized patients, ICU mortality was lower with conserva-
tive vs liberal oxygen therapy with an ARR of 11% and a RRR
of 39%, although the difference was not significant due to
the small number of patients. In our trial, respiratory failure
was present in 58% of patients at enrollment and developed
in 6.5% more during their ICU stay. The difference in mean
PaO2 values between the 2 groups in the trial by Panwar and
colleagues was 23 mm Hg, close to the 15 mm Hg difference
in median PaO2 observed in our trial. Moreover, PaO2 data
reported by Panwar and colleagues2 suggest that they com-
pared a normoxic with a mild hypoxic strategy rather than a
normoxic with a mild hyperoxic strategy, as in our trial. The
mean PaO2 values were 92 mm Hg (95% CI, 82-96) in the lib-
eral oxygen therapy group and 70 mm Hg (95% CI, 68-73) in
the conservative group, with normal PaO2 values ranging
between 80 and 90 mm Hg.3 Therefore, the results of our
trial should be compared with caution to results from the
trial by Panwar and colleagues.
Regarding the adherence to oxygenation targets, our pro-
tocol used a pragmatic, easy-to-apply nurse order set, and we
did not plan to strictly verify the application by continuous re-
cordings of SpO2 and FiO2. On the case report forms, PaO2, FiO2,
and SpO2 were recorded when an arterial blood gas analysis was
performed. This approach limited the exact assessment of O2
exposure, particularly in the conventional group that re-
ceived an FiO2 of 1.0 during intubation, airway suction, or hos-
pital transfer. In this setting, SpO2 is redundant to PaO2, and
the time-weighted PaO2 and FiO2 values during the ICU stay
seem to be the most appropriate parameters for evaluating O2
exposure in our patients.
Indication of an upper limit of FiO2 in the protocol might
have been appropriate. Nevertheless, the exclusion of pa-
tients with acute respiratory distress syndrome from the study
and an upper limit of PaO2 of 150 mm Hg in the conventional
group made the use of FiO2 greater than 60% redundant. In
fact, median FiO2 values were 0.39 in the conventional group
and 0.36 in the conservative group.
Massimo Girardis, MD
Stefano Busani, MD
Author Affiliations: Department of Anesthesiology and Intensive Care,
University Hospital of Modena, Modena, Italy.
(Reprinted) JAMA February 14, 2017 Volume 317, Number 6 647
 Letters
Corresponding Author: Massimo Girardis, MD, Cattedra di Anestesia e
Rianimazione, Azienda Ospedaliera Universitaria di Modena, L.go del Pozzo 71,
41100 Modena, Italy (girardis.massimo@unimo.it).
Conflict of Interest Disclosures: Both authors have completed and submitted
the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were
reported.
1. Ferguson ND. Oxygen in the ICU: too much of a good thing? JAMA. 2016;316
(15):1553-1554.
2. Panwar R, Hardie M, Bellomo R, et al; CLOSE Study Investigators; ANZICS
Clinical Trials Group. Conservative versus liberal oxygenation targets for
mechanically ventilated patients: a pilot multicenter randomized controlled
trial. Am J Respir Crit Care Med. 2016;193(1):43-51.
3. Vincent JL, Abraham JL, Kochanek P, Vincent JL, et al. Textbook of Critical
Care 6th ed. Philadelphia, PA: Elsevier/Saunders, 2011:297-298.
CORRECTION
Omission of Data: In the Original Investigation entitled “Effect of Pritelivir Com-
pared With Valacyclovir on Genital HSV-2 Shedding in Patients With Frequent Re-
currences: A Randomized Clinical Trial”1 published in the December 20, 2016, issue
of JAMA, data were omitted. The Design, Setting, and Participants section of the
Abstract should have reported that “45 participants were randomized to receive prite-
livir.” In the Results section of the Abstract, the relative risk during valacyclovir treat-
ment should have been reported as “0.42.” In Table 1, the upper bound of the con-
fidence intervals in the “HSV DNA log10 copies/mL” subsection should have been
reported as “0.5,” “−0.1,” and “1.0,” respectively. The second and third row stubs in
the “Clinical End Points” section should have read “Recurrence or incidence rate” and
“Days with pain,” respectively. This article has been corrected online.
1. Wald A, Timmler B, Magaret A, et al. Effect of pritelivir compared with
valacyclovir on genital HSV-2 shedding in patients with frequent recurrences:
a randomized clinical trial. JAMA. 2016;316(23):2495-2503.
Incorrect Values and Excluded Units of Measure: In the Original Investigation
entitled “Global Burden of Hypertension and Systolic Blood Pressure of at Least
110 to 115 mm Hg, 1990-2015”1 published in the January 10, 2017, issue of JAMA,
incorrect values for changes in loss of disability-adjusted life-years (DALYs) were
reported in the Results section of the Abstract. For loss of DALYs associated with
systolic blood pressure of 140 mm Hg or higher, the loss increased from 95.9 mil-
648 JAMA February 14, 2017 Volume 317, Number 6 (Reprinted)
Copyright 2017 American Medical Association. All rights reserved.
Downloaded From: http://jama.jamanetwork.com/pdfaccess.ashx?url=/data/journals/jama/936050/ by a Washington University - St Louis User on 02/14/2017
lion (95% uncertainty interval [UI], 87.0-104.9 million) to 143.0 million (95% UI,
130.2-157.0 million). Additionally, some of the units of measure for rows in Table 1
were excluded. The rows that were labeled to report values for “Individuals” and
“Deaths” have been corrected to read: “Individuals, thousands” and “Deaths, thou-
sands.” This article was corrected online.
1. Forouzanfar MH, Liu P, Roth GA, et al. Global burden of hypertension and
systolic blood pressure of at least 110 to 115 mm Hg, 1990-2015. JAMA. 2016;317
(2):165-182. doi:10.1001/jama.2016.19043
Guidelines for Letters
Letters discussing a recent JAMA article should be submitted within 4
weeks of the article's publication in print. Letters received after 4 weeks
will rarely be considered. Letters should not exceed 400 words of text
and 5 references and may have no more than 3 authors. Letters report-
ing original research should not exceed 600 words of text and 6 refer-
ences and may have no more than 7 authors. They may include up to 2
tables or figures but online supplementary material is not allowed. All
letters should include a word count. Letters must not duplicate other ma-
terial published or submitted for publication. Letters not meeting these
specifications are generally not considered. Letters being considered for
publication ordinarily will be sent to the authors of the JAMA article, who
will be given the opportunity to reply. Letters will be published at the
discretion of the editors and are subject to abridgement and editing. Fur-
ther instructions can be found at http://jamanetwork.com/journals/jama
/pages/instructions-for-authors. A signed statement for authorship cri-
teria and responsibility, financial disclosure, copyright transfer, and
acknowledgment and the ICMJE Form for Disclosure of Potential Con-
flicts of Interest are required before publication. Letters should be sub-
mitted via the JAMA online submission and review system at https:
//manuscripts.jama.com. For technical assistance, please contact
jama-letters@jamanetwork.org.
Section Editor: Jody W. Zylke, MD, Deputy Editor.
jama.com