Is it time to modify the Apgar score?

We read with interest the publication “Associations between provider-assigned Apgar score and neonatal race” by Edwards et al.1Edwards S.E. Wheatley C. Sutherland M. Class Q.A. Associations between provider-assigned Apgar score and neonatal race.Am J Obstet Gynecol. 2023; 228: 229.e1-229.e9Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar The authors reported that providers inaccurately assigned Black neonates significantly lower Apgar scores, which were driven by lower color scores. Full-term Black neonates were admitted to the neonatal intensive care unit (NICU) at higher rates than non-Black neonates when controlling for confounders (eg, umbilical artery blood gases and pH, gestational age, and maternal-fetal complications). Moreover, the authors concluded that the inaccuracies between Black neonates’ lower Apgar scores and objective findings, such as cord pH and blood gases, may be a factor in unnecessary admissions to the NICU of Black neonates, and they suggested that “colorism” and racial biases exist among healthcare providers in assigning lower Apgar scores to Black newborns. The findings of inappropriately lower Apgar scores among Black newborns may be of concern with a proposed rule that was published on August 4, 2022, in Section 1557 of the Patient Protection and Affordable Care Act (or “Obamacare”), which prohibits discrimination based on race, color, national origin, age, disability, or sex (including pregnancy, sexual orientation, gender identity, and sex characteristics).2Centers for Medicare and Medicaid ServicesNondiscrimination in health programs and activities. Federal Register.https://www.federalregister.gov/documents/2022/08/04/2022-16217/nondiscrimination-in-health-programs-and-activitiesDate: 2022Date accessed: January 11, 2023Google Scholar A new provision (§92.210) applies Section 1557’s nondiscrimination requirement to the use of clinical algorithms. Specifically, §92.210 explicitly prohibits discrimination in the use of clinical algorithms to support decision-making in covered health programs and activities. The Table shows the excerpts from §92.210. Once the proposed updated Section 1557 is adopted, hospitals will be responsible for discriminatory decisions based on clinical algorithms.TableExcerpts from §92.210Excerpts from “Use of clinical algorithms in decision-making (§ 92.210)”“[The] proposed § 92.210 states that a covered entity must not discriminate against any individual on the basis of race, color, national origin, sex, age, or disability through the use of clinical algorithms in its decision-making.”“The Department believes it is [crucial] to address this issue explicitly in this rulemaking given recent research demonstrating the prevalence of clinical algorithms that may result in discrimination.”“[The] OCR believes that [the] proposed § 92.210 would [place] covered entities on notice that they cannot use discriminatory clinical algorithms and may need to make reasonable modifications in their use of the algorithms.”“By overrelying on a clinical algorithm in their decision-making, such as by replacing or substituting their clinical judgment with a clinical algorithm, a covered entity may risk violating Section 1557 if their decision rests on or results in discrimination.”“Clinical algorithms are tools used to guide healthcare decision-making and can range in form from flowcharts and clinical guidelines to complex computer algorithms, decision support interventions, and models.”“Clinical algorithms are used for screening, risk prediction, diagnosis, prognosis, clinical decision-making, treatment planning, healthcare operations, and allocation of resources,[546] all of which affect the care that individuals receive. Recent studies have found that healthcare tools using clinical algorithms may create or contribute to discrimination on the bases protected by Section 1557 and, as a result of their use by covered entities in their health care decision-making, may lead to poorer health outcomes among members of historically marginalized communities.”“Clinical algorithms commonly include clinical and sociodemographic variables and measures of healthcare utilization.”“Race and ethnicity are often used as explicit input variables. Known as “race correction” or “race norming,” this practice adjusts an algorithm’s output on the basis of a patient's race or ethnicity.”“Covered entities should take steps to ensure that the use of clinical algorithms does not result in discrimination on the basis of race, color, national origin, sex, age, or disability in their health programs and activities.”“Given the increasing reliance on clinical algorithms to inform decision-making in the area of healthcare, and the reality that the implementation of these tools may be discriminatory under Section 1557, the Department proposes § 92.210 to make explicit that covered entities are prohibited from discriminating through the use of clinical algorithms on the basis of race, color, national origin, sex, age, or disability under Section 1557.”The Affordable Care Act (Section 1557) prohibits discrimination on the basis of race, color, national origin, sex, age, or disability in certain health programs and activities.OCR, Office for Civil Rights.Grünebaum. Modification of the Apgar score. Am J Obstet Gynecol 2023. Open table in a new tab The Affordable Care Act (Section 1557) prohibits discrimination on the basis of race, color, national origin, sex, age, or disability in certain health programs and activities. OCR, Office for Civil Rights. Grünebaum. Modification of the Apgar score. Am J Obstet Gynecol 2023. The Apgar score is a clinical algorithm commonly used for newborns at 1 and 5 minutes after birth mostly to report the status of the newborn immediately after birth and the response to resuscitation if needed.3Committee Opinion No. 644: the Apgar score.Obstet Gynecol. 2015; 126: e52-e55Crossref PubMed Scopus (0) Google Scholar It consists of 5 components, each with a score between 0 and 2 and a maximum score of 10. Apgar scoring was introduced by Virginia Apgar in 1953.4Apgar V. A proposal for a new method of evaluation of the newborn infant.Curr Res Anesth Analg. 1953; 32: 260-267Crossref PubMed Google Scholar Of note, 1 of the 5 components, skin color, scores the color of the skin: a score of 2 is given if the skin is “completely pink,” a score of 1 is given if the skin is “acrocyanotic,” and a score of 0 is given if the skin is “pale.” It is the only clinical algorithm in obstetrics and neonatology that includes skin color (which is also a proxy for race) as part of its scoring. There is growing concern that these kinds of algorithms may reproduce racial and gender disparities.1Edwards S.E. Wheatley C. Sutherland M. Class Q.A. Associations between provider-assigned Apgar score and neonatal race.Am J Obstet Gynecol. 2023; 228: 229.e1-229.e9Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar,5Obermeyer Z. Powers B. Vogeli C. Mullainathan S. Dissecting racial bias in an algorithm used to manage the health of populations.Science. 2019; 366: 447-453Crossref PubMed Scopus (1323) Google Scholar Clinicians often rely on prediction algorithms, such as the Apgar score, to identify and help patients with complex health needs and to implement interventions when Apgar scores are low. Royce et al6Royce C.S. Morgan H.K. Baecher-Lind L. et al.The time is now: addressing implicit bias in obstetrics and gynecology education.Am J Obstet Gynecol. 2022; ([Epub ahead or print])Google Scholar described strategies to address implicit bias, and they suggested crucial reviews of epidemiology and evidence-based medicine for underlying assumptions based on discriminatory practices or structural racism, which unintentionally reinforce stereotypes and bias. Edwards et al1Edwards S.E. Wheatley C. Sutherland M. Class Q.A. Associations between provider-assigned Apgar score and neonatal race.Am J Obstet Gynecol. 2023; 228: 229.e1-229.e9Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar and many others report that the Apgar score exhibits significant bias; lower Apgar scores are reported more frequently among newborns of color, and higher ones for White newborns, without a biologic explanation.6Royce C.S. Morgan H.K. Baecher-Lind L. et al.The time is now: addressing implicit bias in obstetrics and gynecology education.Am J Obstet Gynecol. 2022; ([Epub ahead or print])Google Scholar,7Grünebaum A. Bornstein E. Dudenhausen J.W. et al.Hidden in plain sight in the delivery room - the Apgar score is biased.J Perinat Med. 2023; ([Epub ahead of print])Crossref Scopus (0) Google Scholar With lower Apgar scores, more newborns of color require unnecessary interventions, such as NICU admissions, compared with White newborns.1Edwards S.E. Wheatley C. Sutherland M. Class Q.A. Associations between provider-assigned Apgar score and neonatal race.Am J Obstet Gynecol. 2023; 228: 229.e1-229.e9Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar The inaccuracy of the Apgar score becomes more evident with the addition of §92.210 to Section 1557 of the Affordable Care Act. Once §92.210 has been adopted, it would arguably create greater enforcement risk for hospitals that continue to use clinical algorithms, such as Apgar scores that include skin color. Embedding skin color, such as with the Apgar score, into basic data and decisions of healthcare may not comply with the new provision (§92.210) that applies Section 1557’s nondiscrimination requirement to the use of clinical algorithms. The American College of Obstetricians and Gynecologists has said that the Apgar score “is a tool for standardized assessment.”4Apgar V. A proposal for a new method of evaluation of the newborn infant.Curr Res Anesth Analg. 1953; 32: 260-267Crossref PubMed Google Scholar Its color component is not standardized and is not objective when skin color interpretation varies and people of color are consistently reported to have lower Apgar scores without a biologic explanation. Replacing the skin color score with another color component, as some have done, such as documenting the color of mucous membranes, is not sensitive or specific and is too unreliable because color interpretations are subjective and prone to errors, skin colors vary, and there is little evidence-based research to support it. Another component to replace skin color would be pulse oximetry, which determines the peripheral blood oxygen saturation. However, that too has been found to be unreliable in people of color.8Valbuena V.S.M. Merchant R.M. Hough C.L. Racial and ethnic bias in pulse oximetry and clinical outcomes.JAMA Intern Med. 2022; 182: 699-700Crossref PubMed Scopus (7) Google Scholar As part of complying with Section 1557 §92.210, independent of whether that section is being adopted or not, we should heed Virginia Apgar’s original comments made nearly 70 years ago.4Apgar V. A proposal for a new method of evaluation of the newborn infant.Curr Res Anesth Analg. 1953; 32: 260-267Crossref PubMed Google Scholar She said that the scoring for skin color is “by far the most unsatisfactory sign” because of “the inherited pigmentation of the skin of colored children.”4Apgar V. A proposal for a new method of evaluation of the newborn infant.Curr Res Anesth Analg. 1953; 32: 260-267Crossref PubMed Google Scholar Consequently, we should take note of the Office of Civil Rights’ recommendations that “covered entities cannot use discriminatory clinical algorithms and may need to make reasonable modifications in their use of the algorithms” and consider modifying the Apgar score by removing the skin color component so that the Apgar score has a range between 0 and 8 and a maximum score of 8. Further research is needed to assess the effect of this change, including whether the current cutoffs for a low Apgar score of <7 and <4 should remain the same or not. eyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiJkMzVlOTMyMWEyOGMxYjU1MjJhMzViZDQ0NWMyMTU0NyIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNjkwMTc0MjEyfQ.oRh184i-ugJftijRE2bKEoRf1jYSyMzD5LbVcFSiB68-lSSZ5fN-xI08vPMWVXGwwC-a60qABOD9ar4apcDE4PLCUbl1dVPNjyX04oQ9cjL296n1bbyvbWngtyQPyOwkABiMSDD-QuzAMH-jDRiI9LtVGINpnnXUsR3QF4w7PIZcKPXmwD_8B5ddlFlzmfvNsuvy1wtJueDxTuCp-MnkCxfNBLouJuqjH8rUvilOGkHqFjLJO27F-wScYJalJ9VXpex6vXNji5x8R3elM9q7e-gjwBj5pT6xZqawzYiQfGG0BJDaSrrKX3oL3zF-luNCeVfWzIKseIhSiDSH2K1dxg Download .mp4 (19.33 MB) Help with .mp4 files Video 1eyJraWQiOiI4ZjUxYWNhY2IzYjhiNjNlNzFlYmIzYWFmYTU5NmZmYyIsImFsZyI6IlJTMjU2In0.eyJzdWIiOiIzM2RjMWIyOWJkNjY4NWNkNjVjZDM5MDczODZkYjMyNSIsImtpZCI6IjhmNTFhY2FjYjNiOGI2M2U3MWViYjNhYWZhNTk2ZmZjIiwiZXhwIjoxNjkwMTc0MjEyfQ.KUAJ7MZ_ng27l2VslBRJCK2rNnhx0ZSSDm0k43svUwtnZScW-4B19tKWLmLTXzsrPYnbFfD1tB6QIKC6qPMt-cr03Hh43y6CNc9_M4iLf03f6ea-a5O9G-GZUHp-1Y2IYVMd2hUqy5oRtpy-n6VN-Qwfm5mdsmB-QMyjrjd-e97AhIkOqdnTh_vHilFJFQZkc2dWDY07BU9hihzeBxEUfMDq_dSQObVbKrr1FmxWsnkYRp1M_cpxdw5IZ0danZwo60l9muaBmrzmDSeuAkWZhLQO6xtUgu3blfjCfLoDLEp9-jP37otZGPZpKqEAVwgm2okJ1poo-GFRy9ByxFhTGg Download .mp4 (8.22 MB) Help with .mp4 files Video 2