Hash and colleagues at Vanderbilt University Medical Center found that the SimulConsult diagnostic software saved time and was perceived by clinicians as useful in improving diagnosis.
Kulchak Rahm and colleagues combined Natural Language Processing of Electronic Health Record (EHR) data with SimulConsult’s ability to discern what information is useful. The result was a system that identifies useful information in the EHR, and uses it for accurate diagnosis.
The Centers for Disease Control and Prevention’s latest Covid guidelines have many Americans confused. Vaccinated people are supposed to resume wearing masks, lest they contract and spread the virus. Yet unvaccinated people are still strongly urged to get the shots, which are said to be highly effective. How can both these claims be true?
In a research paper, geneticists at Vanderbilt University Medical Center discussed their protocol for using SimulConsult in the diagnosis process. They suggested “In addition to increased communication, programs such as OMIM and SimulConsult can be used to analyze combinations of signs and symptoms to identify and prioritize differential diagnoses.”
The SimulConsult software runs just about everywhere, so we checked whether it could run on a Peloton exercise bicycle. Using the guidance (see hyperlink) for how to find the web browser on the Peloton, we found that indeed the software runs fine.
The SimulConsult software is most conveniently run on a regular computer or a tablet such as an iPad mini. But with the release of the iPhone 12 mini that fits in your pocket conveniently, we’ve made sure the software even runs with that small screen size so you use it whenever you need it. Even the genome analysis fits on the screen, but that is not something anyone will expect you to do on the go.
The SimulConsult Genome-Phenome Analyzer supports all HGNC gene symbols in use since 1 January 2010, even if the symbol has been changed. The hyperlinked article details how HGNC is systematically changing all the symbols that Excel turns into dates in a spreadsheet. But even if you use the old symbols such as SEPT12 in a variant table, or search for it in the software, SimulConsult recognizes that as being the HGNC symbol SEPTIN12.
In the 81 cases (with 216 individuals) the gene abnormality was recognized in 100%, and it was ranked #1 in 94% of cases. Large CNVs could be analyzed in an integrated analysis, performed in 24 of the cases. The process is rapid enough to allow for periodic reanalysis of unsolved cases.
“In all families (84 pregnancies) pertinence was > 95% between signs/symptoms and molecular cause”, meaning that not only was the correct diagnosis ranked as #1 but no other diagnoses were close.
Published in the Wall Street Journal on May 1, 2019 by Michael Segal, MD PhD
They can appear to be the proximate cause of a condition when they have nothing to do with the ultimate cause. Why are people afraid of vaccines?
See letter in response on 13 May.
SimulConsult releases new mobile version of its Phenome Analyzer and Genome-Phenome analyzer software
Released April 1, 2019
SimulConsult announces a new version of its diagnostic decision support tool. The tool combines the power of
curated human expertise and computational artificial intelligence (AI) to empower clinicians in diagnosis and
workup of patients. It is used today in 118 countries. The new version has a completely new interface that allows it to run on mobile devices as well as on computers. It is fast to use and puts the diagnostic power into a clinician’s hands whenever needed.
SimulConsult’s newly-formed research partner, PhenoSolve, LLC., received an SBIR I from the National Human Genome Research Institute (NHGRI) of the NIH beginning September 19, 2018 with Michael M. Segal MD PhD as Primary Investigator
The rationale is that the current state of genome analysis and clinical use is lagging in part due to a situation of siloed information and capabilities, a situation that Clayton Christensen characterizes as requiring end-to-end integration to set the stage for a subsequent environment of interoperable standards. While analogous to the radiology PACS, the G-PACS will also have advanced analysis capabilities and sophisticated handling of 2 types of information – patient findings and annotated genomic variants to enable genome-phenome analysis by clinicians and laboratorians using these components.
Experience with Integrating Diagnostic Decision Support Software with Electronic Health Records: Benefits versus Risks of Information Sharing
Published in The Journal for Electronic Health Data and Methods on December 6, 2017 by Michael M. Segal, Alanna K. Rahm, Nathan C. Hulse, Grant Wood, Janet L. Williams, Lynn Feldman, Gregory J. Moore, David Gehrum, Michelle Yefko, Steven Mayernick, Roger Gildersleeve, Margie C. Sunderland, Steven B. Bleyl, Peter Haug, Marc S. Williams
The article describes the result of an NIH-sponsored study of integrating SimulConsult’s diagnostic decision support system with Electronic Health Record systems.
Impact of a Patient-Facing Enhanced Genomic Results Report to Improve Understanding, Engagement, and Communication
Published in the Journal of Genetic Counselors in April 2018 (open access) by Williams JL, Rahm AK, Zallen DT, Stuckey H, Fultz K, Fan AL, Bonhag M, Feldman L, Segal MM, Williams MS.
The article describes the result of a PCORI-sponsored study of providing information about genetic testing to patients, including a display driven by the highly granular onset information in SimulConsult’s diagnostic decision support system.
SimulConsult receives SBIR II from the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the NIH of the NIH beginning August 1, 2017 with Michael M. Segal MD PhD as Primary Investigator
This research aims to ease the shortage of pediatric rheumatologists by using decision support software to improve the ability of generalist physicians to make rheumatologic diagnoses. It extends to actual clinical practice our previous work that demonstrated large reductions in diagnostic error when tested using written case summaries, using a randomized controlled study. The research will be done after implementing improvements to the diagnostic tool suggested by the earlier pilot study.
Published by the Journal of Pediatric Rheumatology Online on December 13, 2016 (open access) by Michael M. Segal, Balu Athreya, Mary Beth F. Son, Irit Tirosh, Jonathan S. Hausmann, Elizabeth Y. N. Ang, David Zurakowski, Lynn K. Feldman, and Robert P. Sundel
The article describes the result of an NIH-sponsored study of adding rheumatology information to the SimulConsult diagnostic decision support system. The 26 clinicians demonstrated a significant reduction in diagnostic errors following introduction of the software, from 28% errors while unaided to 15% using decision support (p < 0.0001). Improvement was greatest for emergency medicine physicians (p = 0.013) and clinicians in practice for less than 10 years (p = 0.012). This error reduction occurred despite the fact that testers employed an “open book” approach to generate their initial lists of potential diagnoses, spending an average of 8.6 min using printed and electronic sources of medical information before using the diagnostic software.
SimulConsult, Inc. of Chestnut Hill, MA has been awarded START Stage I funding for their work on the SimulConsult® diagnostic decision support system. Lynn Feldman, CEO of SimulConsult, Inc. said “the START program fills a critical gap between the SBIR research program and achieving a revenue-producing product – investment in commercialization activities. The START program will allow us to migrate our technology to a new platform, which means we can price our product more competitively, improving our financial prospects and rate of growth.”
Published in the American Journal of Medical Genetics in May 2016 (open access) by Williams JL, Rahm AK, Stuckey H, Green J, Feldman L, Zallen DT, Bonhag M, Segal MM, Fan AL, Williams MS.
The article describes the responses by clinicians to SimulConsult’s prognosis table, as used in genomic reports. One sample: “I love it—no clicks, detailed, comprehensive enough that I didn’t feel I needed another source—everything a pediatrician would think about.”
Published in Applied Translational Genomics, in July 22, 2015 by Michael M. Segal.
In a brain MRI report, the following words often appear: “clinical correlation is recommended”. These words signify that inadequate clinical information was provided, or that an unexpected finding on the MRI should be assessed clinically. “Clinical correlation is recommended” is less common in a report about a single gene or simple gene panel. This is because the very act of ordering the test conveys much of what is important about the clinical situation, and only rarely is further information needed.
Genetics labs are moving into new territory as they adopt next-generation genomic sequencing. When moving beyond single gene tests and simple panels, more clinical correlation is needed. The complexity of interpretation becomes similar to a brain MRI, only more so.
In an exome, thousands of variants are found. Even after comparing to other family members, and using estimates of variant pathogenicity, many genes must be considered. Sometimes clinical correlation can be as simple as using the key clinical finding, assuming that you know which finding is key. But sometimes the situation is more complicated: variants are found in a gene that hadn’t been considered clinically, or two genes are needed to explain the clinical picture, and more clinical correlation is needed.
Clinical Pertinence Metric Enables Hypothesis-Independent Genome-Phenome Analysis for Neurologic Diagnosis
Published in the Journal of Child Neurology in June of 2015 by Michael M. Segal, MD, PhD, Mostafa Abdellateef, Ayman W. El-Hattab, MD, Brian S. Hilbush, PhD, Francisco M. De La Vega, PhD, Gerard Tromp, PhD, Marc S. Williams, MD, Rebecca A. Betensky, PhD, and Joseph Gleeson, MD
We describe an “integrated genome-phenome analysis” that combines both genomic sequence data and clinical information for genomic diagnosis. It is novel in that it uses robust diagnostic decision support and combines the clinical differential diagnosis and the genomic variants using a “pertinence” metric. This allows the analysis to be hypothesis-independent, not requiring assumptions about mode of inheritance, number of genes involved, or which clinical findings are most relevant. Using 20 genomic trios with neurologic disease, we find that pertinence scores averaging 99.9% identify the causative variant under conditions in which a genomic trio is analyzed and family-aware variant calling is done. The analysis takes seconds, and pertinence scores can be improved by clinicians adding more findings. The core conclusion is that automated genome-phenome analysis can be accurate, rapid, and efficient. We also conclude that an automated process offers a methodology for quality improvement of many components of genomic analysis.