. 2024 May 24;11(1):535.

doi: 10.1038/s41597-024-03225-z.

BOLD: Blood-gas and Oximetry Linked Dataset

João Matos^{1

2

3

4}, Tristan Struja⁵, Jack Gallifant^{6

7}, Luis Nakayama^{6

8}, Marie-Laure Charpignon⁹, Xiaoli Liu¹⁰, Nicoleta Economou-Zavlanos¹¹, Jaime S Cardoso^{12

13}, Kimberly S Johnson¹⁴, Nrupen Bhavsar^{15

16}, Judy Gichoya¹⁷, Leo Anthony Celi^{6

18

19}, An-Kwok Ian Wong^{20

21}

Affiliations

¹ Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, MA, USA. jcmatos@mit.edu.
² Faculty of Engineering, University of Porto (FEUP), Porto, Portugal. jcmatos@mit.edu.
³ Institute for Systems and Computer Engineering, Technology and Science (INESCTEC), Porto, Portugal. jcmatos@mit.edu.
⁴ Duke University, Department of Medicine, Division of Pulmonary, Allergy, and Critical Care Medicine, Durham, NC, USA. jcmatos@mit.edu.
⁵ Medical University Clinic, Kantonsspital Aarau, Aarau, Switzerland.
⁶ Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, MA, USA.
⁷ Department of Critical Care, Guy's and St Thomas' NHS Trust, London, United Kingdom.
⁸ Department of Ophthalmology, São Paulo Federal University, São Paulo, SP, Brazil.
⁹ Institute for Data Systems and Society, Massachusetts Institute of Technology, Cambridge, MA, US.
¹⁰ Center for Artificial Intelligence in Medicine, The General Hospital of PLA, Beijing, China.
¹¹ Duke University, AI Health, Durham, NC, USA.
¹² Faculty of Engineering, University of Porto (FEUP), Porto, Portugal.
¹³ Institute for Systems and Computer Engineering, Technology and Science (INESCTEC), Porto, Portugal.
¹⁴ Duke University, Department of Medicine, Division of Geriatrics, Durham, NC, USA.
¹⁵ Duke University, Department of Biostatistics and Bioinformatics, Division of Translational Biomedical Informatics, Durham, NC, USA.
¹⁶ Duke University, Department of Surgery, Division of Surgical Sciences, Durham, NC, USA.
¹⁷ Emory University, Department of Radiology, Atlanta, GA, USA.
¹⁸ Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
¹⁹ Department of Medicine, Beth Israel Deaconess Medical Center, Boston, MA, USA.
²⁰ Duke University, Department of Medicine, Division of Pulmonary, Allergy, and Critical Care Medicine, Durham, NC, USA. med@aiwong.com.
²¹ Duke University, Department of Biostatistics and Biomedical Informatics, Division of Translational Biomedical Informatics, Durham, NC, USA. med@aiwong.com.

PMID: 38789452
PMCID: PMC11126612
DOI: 10.1038/s41597-024-03225-z

BOLD: Blood-gas and Oximetry Linked Dataset

João Matos et al. Sci Data. 2024.

. 2024 May 24;11(1):535.

doi: 10.1038/s41597-024-03225-z.

Authors

Affiliations

¹ Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, MA, USA. jcmatos@mit.edu.
² Faculty of Engineering, University of Porto (FEUP), Porto, Portugal. jcmatos@mit.edu.
³ Institute for Systems and Computer Engineering, Technology and Science (INESCTEC), Porto, Portugal. jcmatos@mit.edu.
⁴ Duke University, Department of Medicine, Division of Pulmonary, Allergy, and Critical Care Medicine, Durham, NC, USA. jcmatos@mit.edu.
⁵ Medical University Clinic, Kantonsspital Aarau, Aarau, Switzerland.
⁶ Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, MA, USA.
⁷ Department of Critical Care, Guy's and St Thomas' NHS Trust, London, United Kingdom.
⁸ Department of Ophthalmology, São Paulo Federal University, São Paulo, SP, Brazil.
⁹ Institute for Data Systems and Society, Massachusetts Institute of Technology, Cambridge, MA, US.
¹⁰ Center for Artificial Intelligence in Medicine, The General Hospital of PLA, Beijing, China.
¹¹ Duke University, AI Health, Durham, NC, USA.
¹² Faculty of Engineering, University of Porto (FEUP), Porto, Portugal.
¹³ Institute for Systems and Computer Engineering, Technology and Science (INESCTEC), Porto, Portugal.
¹⁴ Duke University, Department of Medicine, Division of Geriatrics, Durham, NC, USA.
¹⁵ Duke University, Department of Biostatistics and Bioinformatics, Division of Translational Biomedical Informatics, Durham, NC, USA.
¹⁶ Duke University, Department of Surgery, Division of Surgical Sciences, Durham, NC, USA.
¹⁷ Emory University, Department of Radiology, Atlanta, GA, USA.
¹⁸ Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, MA, USA.
¹⁹ Department of Medicine, Beth Israel Deaconess Medical Center, Boston, MA, USA.
²⁰ Duke University, Department of Medicine, Division of Pulmonary, Allergy, and Critical Care Medicine, Durham, NC, USA. med@aiwong.com.
²¹ Duke University, Department of Biostatistics and Biomedical Informatics, Division of Translational Biomedical Informatics, Durham, NC, USA. med@aiwong.com.

PMID: 38789452
PMCID: PMC11126612
DOI: 10.1038/s41597-024-03225-z

Abstract

Pulse oximeters measure peripheral arterial oxygen saturation (SpO₂) noninvasively, while the gold standard (SaO₂) involves arterial blood gas measurement. There are known racial and ethnic disparities in their performance. BOLD is a dataset that aims to underscore the importance of addressing biases in pulse oximetry accuracy, which disproportionately affect darker-skinned patients. The dataset was created by harmonizing three Electronic Health Record databases (MIMIC-III, MIMIC-IV, eICU-CRD) comprising Intensive Care Unit stays of US patients. Paired SpO₂ and SaO₂ measurements were time-aligned and combined with various other sociodemographic and parameters to provide a detailed representation of each patient. BOLD includes 49,099 paired measurements, within a 5-minute window and with oxygen saturation levels between 70-100%. Minority racial and ethnic groups account for ~25% of the data - a proportion seldom achieved in previous studies. The codebase is publicly available. Given the prevalent use of pulse oximeters in the hospital and at home, we hope that BOLD will be leveraged to develop debiasing algorithms that can result in more equitable healthcare solutions.

PubMed Disclaimer

Conflict of interest statement

A.I.W. holds equity and management roles in Ataia Medical. All other authors report no conflicts of interest.

Figures

**Fig. 1**
Rationale and variables included in the dataset.

**Fig. 2**
Pipeline created to curate and merge the datasets.

**Fig. 3**
Oxyhemoglobin dissociation curve, per database, with the pH shift highlighted, on the extended dataset.

**Fig. 4**
Modified Bland-Altman plots, across race and ethnicity (White compared with racial and ethnic group), and across 2 time windows, on the extended dataset.

**Fig. 5**
RMSE and number of pairs with varying window between SaO₂ and SpO₂, per race and ethnicity, on the extended dataset.

**Fig. 6**
a. Flow diagram for MIMIC-III depicting cohort selection. b. Flow diagram for MIMIC-IV depicting cohort selection. c. Flow diagram for eICU-CRD depicting cohort selection.

**Fig. 7**
Flow diagram for the merged dataset.

**Fig. 8**
Completeness of the aligned covariates in the preprocessed dataset.

See this image and copyright information in PMC

Update of

BOLD: Blood-gas and Oximetry Linked Dataset - Open Source Research.
Matos J, Struja T, Gallifant J, Nakayama L, Charpignon ML, Liu X, Economou-Zavlanos N, Cardoso JS, Johnson KS, Bhavsar N, Gichoya J, Celi LA, Wong AI. Matos J, et al. medRxiv [Preprint]. 2023 Oct 3:2023.10.03.23296485. doi: 10.1101/2023.10.03.23296485. medRxiv. 2023. Update in: Sci Data. 2024 May 24;11(1):535. doi: 10.1038/s41597-024-03225-z. PMID: 37873343 Free PMC article. Updated. Preprint.

References

1. Sjoding MW, Dickson RP, Iwashyna TJ, Gay SE, Valley TS. Racial Bias in Pulse Oximetry Measurement. N. Engl. J. Med. 2020;383:2477–2478. doi: 10.1056/NEJMc2029240. - DOI - PMC - PubMed
1. Wong, A. I. et al. Analysis of discrepancies between pulse oximetry and arterial oxygen saturation measurements by Race/Ethnicity and association with organ dysfunction and mortality. JAMA Network Open, 10.1001/jamanetworkopen.2021.31674 (2021). - PMC - PubMed
1. Valbuena VSM, et al. Racial bias and reproducibility in pulse oximetry among medical and surgical inpatients in general care in the Veterans Health Administration 2013-19: multicenter, retrospective cohort study. BMJ. 2022;378:e069775. doi: 10.1136/bmj-2021-069775. - DOI - PMC - PubMed
1. Gottlieb, E. R., Ziegler, J., Morley, K., Rush, B. & Celi, L. A. Assessment of Racial and Ethnic Differences in Oxygen Supplementation Among Patients in the Intensive Care Unit. JAMA Internal Medicine vol. 182 849 10.1001/jamainternmed.2022.2587 (2022). - PMC - PubMed
1. Pollard TJ, Johnson AEW, Raffa J, Badawi O. The eICU Collaborative Research Database. physionet.org. 2017 doi: 10.13026/C2WM1R. - DOI - PMC - PubMed

Publication types

Actions

MeSH terms

Actions
Actions
Actions
Actions
Actions
Actions
Actions

Substances

Actions

Grants and funding

5U54MD012530/U.S. Department of Health & Human Services | NIH | National Institute on Minority Health and Health Disparities (NIMHD)

LinkOut - more resources

Full Text Sources

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

BOLD: Blood-gas and Oximetry Linked Dataset

Affiliations

BOLD: Blood-gas and Oximetry Linked Dataset

Authors

Affiliations

Abstract

Conflict of interest statement

Figures

Update of

References

Publication types

MeSH terms

Substances

Grants and funding

LinkOut - more resources

Full Text Sources