Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Oct 1.
Published in final edited form as: Spine J. 2020 Jun 26;21(10):1649–1651. doi: 10.1016/j.spinee.2020.06.014

SMART on FHIR in spine: integrating clinical prediction models into electronic health records for precision medicine at the point of care

Aditya V Karhade 1, Joseph H Schwab 1, Guilherme Del Fiol 2, Kensaku Kawamoto 2
PMCID: PMC7762727  NIHMSID: NIHMS1607551  PMID: 32599144

Abstract

Recent applications of artificial intelligence have shown great promise for improving the quality and efficiency of clinical care. Numerous clinical decision support tools exist in today’s electronic health records (EHRs) such as medication dosing support, order facilitators (e.g., procedure specific order sets), and point of care alerts. However, less has been done to integrate artificial intelligence (AI)-enabled risk predictors into EHRs despite wide availability of validated risk prediction tools. An interoperability standard known as SMART on FHIR (Substitutable Medical Applications and Reusable Technologies on Fast Health Interoperability Resources) offers a promising path forward, enabling digital innovations to be seamlessly integrated with the EHR with regard to the user interface and patient data. For the next step in progress towards the goal of learning healthcare and informatics-enabled spine surgery, we propose the application of SMART on FHIR to integrate existing and new risk predictions tools in spine surgery through an EHR add-on-application

Keywords: artificial intelligence, clinical decision support, diagnosis, electronic medical records, integration, machine learning, natural language processing, prediction, SMART on FHIR, spine

Manuscript:

Recent applications of artificial intelligence have shown great promise for improving the quality and efficiency of clinical care.[1, 2] For example, Lundberg et al. developed algorithms for prediction of intraoperative hypoxemia that performed better than anesthesiologists and subsequently improved anesthesiologist performance when made available with real-time model explanations.[3] Similarly, Hollon et al. developed algorithms for automated intraoperative tumor diagnosis that performed at the level of trained pathologists in a fraction of the time (150 seconds versus 20 minutes).[4] In spine surgery, applications of machine learning have included prediction and diagnosis in spinal oncology, trauma, infections, degenerative conditions, and adult spinal deformity.[5-19]

In 1928, L.J. Henderson reported the first medical use of a pictographic tool known as a nomogram.[20] A nomogram is a two-dimensional, graphical calculator. In the years following Henderson’s report, nomograms were applied extensively for assisted clinical decision making.[21-27] In fact, nomograms and risk scores were ideal for clinicians when patient records existed on paper alone.

Today, after the 2009 Health Information Technology for Economic and Clinical Health (HITECH) Act, the majority of physicians in the United States use electronic health records (EHRs) for everyday clinical workflow.[28] EHRs bring a unique opportunity to integrate clinical decision support (CDS) tools into the clinical workflow. Numerous CDS tools exist in today’s EHRs such as medication dosing support, order facilitators (e.g., procedure specific order sets), and point of care alerts.[29, 30] However, less has been done to integrate AI-enabled risk predictors into EHRs despite wide availability of validated risk prediction tools.[1]

Most risk prediction tools today exist as risk scores in published manuscripts or as web-or smartphone-based digital calculators by professional societies, academic medical centers, research study groups, or aggregating-platforms.[5, 6, 31] Clinicians that seek to use these systems at the point of care are often forced to interrupt their workflow, navigate to these separate systems, manually look-up and input the required patient information, wait for the results, and then return to their workflow in order to integrate the decision support guidance. In addition, the rise of EHRs has led to concerns of increased documentation time, impaired patient-physician interactions, and physician burnout.[29, 32-34] In order to realize the potential of AI-enabled risk prediction models, these models must be seamlessly integrated into EHRs such that they not only improve the performance of clinicians but also make the clinical workflow easier and more efficient.

An interoperability standard known as SMART on FHIR (Substitutable Medical Applications and Reusable Technologies on Fast Health Interoperability Resources) offers a promising path forward, enabling digital innovations to be seamlessly integrated with the EHR with regard to the user interface and patient data.[35-39] Prior work by two of the authors has shown that an EHR add-on-application for neonatology using SMART on FHIR resulted in time savings, high clinician usability ratings, and more clinically appropriate interventions.[35] This app also provided guidance on the likelihood of rebound hyperbilirubinemia following phototherapy.[40]

For the next step in progress towards the goal of learning healthcare and informatics-enabled spine surgery, we propose the application of SMART on FHIR to integrate existing and new risk predictions tools in spine surgery through an EHR add-on-application. We propose the following steps to accomplish this goal: (1) determination of the capacity for native EHR approaches to meet clinician needs, (2) selection of algorithms for prediction in spine surgery with demonstrated evidence of generalizability on external validation, (3) automating collection of real-time data required for model prediction using FHIR data interfaces, (4) development of visualization interfaces to allow for individual patient-level predictions and model explanations, (5) pilot studies to assess the impact of a SMART on FHIR EHR add-on-application on structure, process and patient outcomes, (6) deployment of the add-on-application at multiple institutions, (7) and prospective multi-center trials to determine the impact of the add-on-applications on time savings, user satisfaction, clinician performance, and patient outcomes.

Acknowledgments

Funding: KK and GDF were supported by Grant No. U24CA204800 from the Informatics Technology for Cancer Research program of the US National Cancer Institute

Footnotes

Disclosures:

AVK: none

JHS: Consultant Stryker

GDF: none

KK: KK reports honoraria, consulting, royalties, or sponsored research outside the submitted work from the past three years with McKesson InterQual, Hitachi, Pfizer, Klesis Healthcare, RTI International, Mayo Clinic, Vanderbilt University, the University of Washington, the University of California at San Francisco, MD Aware, and the U.S. Office of the National Coordinator for Health IT (via ESAC and Security Risk Solutions) in the area of health information technology. KK is also an unpaid board member of the non-profit Health Level Seven International health IT standard development organization, and he has helped develop a number of health IT tools which may be commercialized to enable wider impact. None of these relationships have direct relevance to the manuscript but are reported in the interest of full disclosure.

Ethics Statement: This study did not require institutional review board approval

Conflicts of Interest: The authors have no directly relevant conflicts of interest. Other disclosures are noted above.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References:

  • 1.Matheny M, Israni ST, Ahmed M, Whicher D. Artificial intelligence in health care: The hope, the hype, the promise, the peril. National Academy of Medicine, prepub. 2020:94–7. [Google Scholar]
  • 2.Yu K-H, Beam AL, Kohane IS. Artificial intelligence in healthcare. Nature biomedical engineering. 2018;2(10):719–31. [DOI] [PubMed] [Google Scholar]
  • 3.Lundberg SM, Nair B, Vavilala MS, Horibe M, Eisses MJ, Adams T, et al. Explainable machine-learning predictions for the prevention of hypoxaemia during surgery. Nature biomedical engineering. 2018;2(10):749–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Hollon TC, Pandian B, Adapa AR, Urias E, Save AV, Khalsa SSS, et al. Near real-time intraoperative brain tumor diagnosis using stimulated raman histology and deep neural networks. Nature Medicine. 2020:1–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Karhade AV, Ogink PT, Thio QC, Broekman ML, Cha TD, Hershman SH, et al. Machine learning for prediction of sustained opioid prescription after anterior cervical discectomy and fusion. The Spine Journal. 2019;19(6):976–83. [DOI] [PubMed] [Google Scholar]
  • 6.Karhade AV, Thio QC, Ogink PT, Bono CM, Ferrone ML, Oh KS, et al. Predicting 90-day and 1-year mortality in spinal metastatic disease: Development and internal validation. Neurosurgery. 2019;85(4):E671–E81. [DOI] [PubMed] [Google Scholar]
  • 7.Karhade AV, Thio QC, Ogink PT, Shah AA, Bono CM, Oh KS, et al. Development of machine learning algorithms for prediction of 30-day mortality after surgery for spinal metastasis. Neurosurgery. 2019;85(1):E83–E91. [DOI] [PubMed] [Google Scholar]
  • 8.Karhade AV, Ogink PT, Thio QC, Cha TD, Gormley WB, Hershman SH, et al. Development of machine learning algorithms for prediction of prolonged opioid prescription after surgery for lumbar disc herniation. The Spine Journal. 2019;19(11):1764–71. [DOI] [PubMed] [Google Scholar]
  • 9.Karhade AV, Bongers ME, Groot OQ, Kazarian ER, Cha TD, Fogel HA, et al. Natural language processing for automated detection of incidental durotomy. The Spine Journal. 2019. [DOI] [PubMed] [Google Scholar]
  • 10.Karhade AV, Bongers ME, Groot OQ, Cha TD, Doorly TP, Fogel HA, et al. Development of machine learning and natural language processing algorithms for preoperative prediction and automated identification of intraoperative vascular injury in anterior lumbar spine surgery. The Spine Journal. 2020. [DOI] [PubMed] [Google Scholar]
  • 11.Karhade AV, Thio Q, Ogink P, Kim J, Lozano-Calderon S, Raskin K, et al. Development of machine learning algorithms for prediction of 5-year spinal chordoma survival. World neurosurgery. 2018;119:e842–e7. [DOI] [PubMed] [Google Scholar]
  • 12.Karhade AV, Ahmed AK, Pennington Z, Chara A, Schilling A, Thio QC, et al. External validation of the sorg 90-day and 1-year machine learning algorithms for survival in spinal metastatic disease. The Spine Journal. 2020;20(1):14–21. [DOI] [PubMed] [Google Scholar]
  • 13.Karhade AV, Ogink P, Thio Q, Broekman M, Cha T, Gormley WB, et al. Development of machine learning algorithms for prediction of discharge disposition after elective inpatient surgery for lumbar degenerative disc disorders. Neurosurgical focus. 2018;45(5):E6. [DOI] [PubMed] [Google Scholar]
  • 14.Karhade AV, Shah AA, Bono CM, Ferrone ML, Nelson SB, Schoenfeld AJ, et al. Development of machine learning algorithms for prediction of mortality in spinal epidural abscess. The Spine Journal. 2019;19(12):1950–9. [DOI] [PubMed] [Google Scholar]
  • 15.Shah AA, Karhade AV, Bono CM, Harris MB, Nelson SB, Schwab JH. Development of a machine learning algorithm for prediction of failure of nonoperative management in spinal epidural abscess. The Spine Journal. 2019;19(10):1657–65. [DOI] [PubMed] [Google Scholar]
  • 16.Ames CP, Smith JS, Pellise F, Kelly M, Gum JL, Alanay A, et al. Development of predictive models for all individual questions of srs-22r after adult spinal deformity surgery: A step toward individualized medicine. Eur Spine J. 2019;28(9):1998–2011. [DOI] [PubMed] [Google Scholar]
  • 17.Lu J-T, Pedemonte S, Bizzo B, Doyle S, Andriole KP, Michalski MH, et al. Deepspine: Automated lumbar vertebral segmentation, disc-level designation, and spinal stenosis grading using deep learning. arXiv preprint arXiv:180710215. 2018. [Google Scholar]
  • 18.Bongers MER, Karhade AV, Villavieja J, Groot OQ, Bilsky MH, Laufer I, et al. Does the sorg algorithm generalize to a contemporary cohort of patients with spinal metastases on external validation? Spine J. 2020. [DOI] [PubMed] [Google Scholar]
  • 19.DeVries Z, Hoda M, Rivers CS, Maher A, Wai E, Moravek D, et al. Development of an unsupervised machine learning algorithm for the prognostication of walking ability in spinal cord injury patients. Spine J. 2020;20(2):213–24. [DOI] [PubMed] [Google Scholar]
  • 20.Henderson LJ. Blood: A study in general physiology: Yale University Press; 1928. [Google Scholar]
  • 21.Raschke RA, Reilly BM, Guidry JR, Fontana JR, Srinivas S. The weight-based heparin dosing nomogram compared with a standard care nomogram: A randomized controlled trial. Annals of internal medicine. 1993;119(9):874–81. [DOI] [PubMed] [Google Scholar]
  • 22.KATTAN MW, REUTER V, MOTZER RJ, KATZ J, RUSSO P. A postoperative prognostic nomogram for renal cell carcinoma. The Journal of urology. 2001;166(1):63–7. [PubMed] [Google Scholar]
  • 23.Kattan MW, Leung DH, Brennan MF. Postoperative nomogram for 12-year sarcoma-specific death. Journal of clinical oncology. 2002;20(3):791–6. [DOI] [PubMed] [Google Scholar]
  • 24.Pereira NRP, Janssen SJ, van Dijk E, Harris MB, Hornicek FJ, Ferrone ML, et al. Development of a prognostic survival algorithm for patients with metastatic spine disease. JBJS. 2016;98(21):1767–76. [DOI] [PubMed] [Google Scholar]
  • 25.Devin CJ, Bydon M, Alvi MA, Kerezoudis P, Khan I, Sivaganesan A, et al. A predictive model and nomogram for predicting return to work at 3 months after cervical spine surgery: An analysis from the quality outcomes database. Neurosurgical focus. 2018;45(5):E9. [DOI] [PubMed] [Google Scholar]
  • 26.Radford EP Jr, Ferris BG Jr, Kriete BC. Clinical use of a nomogram to estimate proper ventilation during artificial respiration. New England Journal of Medicine. 1954;251(22):877–84. [DOI] [PubMed] [Google Scholar]
  • 27.Balachandran VP, Gonen M, Smith JJ, DeMatteo RP. Nomograms in oncology: More than meets the eye. The lancet oncology. 2015;16(4):e173–e80. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Adler-Milstein J, Jha AK. Hitech act drove large gains in hospital electronic health record adoption. Health Affairs. 2017;36(8):1416–22. [DOI] [PubMed] [Google Scholar]
  • 29.Bright TJ, Wong A, Dhurjati R, Bristow E, Bastian L, Coeytaux RR, et al. Effect of clinical decision-support systems: A systematic review. Ann Intern Med. 2012;157(1):29–43. [DOI] [PubMed] [Google Scholar]
  • 30.Wright A, Sittig DF, Ash JS, Feblowitz J, Meltzer S, McMullen C, et al. Development and evaluation of a comprehensive clinical decision support taxonomy: Comparison of front-end tools in commercial and internally developed electronic health record systems. J Am Med Inform Assoc. 2011;18(3):232–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Khor S, Lavallee D, Cizik AM, Bellabarba C, Chapman JR, Howe CR, et al. Development and validation of a prediction model for pain and functional outcomes after lumbar spine surgery. JAMA surgery. 2018;153(7):634–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Baumann LA, Baker J, Elshaug AG. The impact of electronic health record systems on clinical documentation times: A systematic review. Health Policy. 2018;122(8):827–36. [DOI] [PubMed] [Google Scholar]
  • 33.Pelland KD, Baier RR, Gardner RL. ‘It is like texting at the dinner table’: A qualitative analysis of the impact of electronic health records on patient–physician interaction in hospitals. BMJ Health & Care Informatics. 2017;24(2):216–23. [DOI] [PubMed] [Google Scholar]
  • 34.Adler-Milstein J, Zhao W, Willard-Grace R, Knox M, Grumbach K. Electronic health records and burnout: Time spent on the electronic health record after hours and message volume associated with exhaustion but not with cynicism among primary care clinicians. Journal of the American Medical Informatics Association. 2020;27(4):531–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Kawamoto K, Kukhareva P, Shakib JH, Kramer H, Rodriguez S, Warner PB, et al. Association of an electronic health record add-on app for neonatal bilirubin management with physician efficiency and care quality. JAMA network open. 2019;2(11):e1915343–e. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Esteva A, Robicquet A, Ramsundar B, Kuleshov V, DePristo M, Chou K, et al. A guide to deep learning in healthcare. Nature medicine. 2019;25(1):24–9. [DOI] [PubMed] [Google Scholar]
  • 37.Mandel JC, Kreda DA, Mandl KD, Kohane IS, Ramoni RB. Smart on fhir: A standards-based, interoperable apps platform for electronic health records. Journal of the American Medical Informatics Association. 2016;23(5):899–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Mandl KD, Kohane IS. No small change for the health information economy. N Engl J Med. 2009;360(13):1278–81. [DOI] [PubMed] [Google Scholar]
  • 39.Mandl KD, Mandel JC, Murphy SN, Bernstam EV, Ramoni RL, Kreda DA, et al. The smart platform: Early experience enabling substitutable applications for electronic health records. J Am Med Inform Assoc. 2012;19(4):597–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Chang PW, Kuzniewicz MW, McCulloch CE, Newman TB. A clinical prediction rule for rebound hyperbilirubinemia following inpatient phototherapy. Pediatrics. 2017;139(3). [DOI] [PubMed] [Google Scholar]

RESOURCES