tag:theconversation.com,2011:/uk/topics/clinical-data-7709/articlesClinical data – The Conversation2023-05-26T18:01:24Ztag:theconversation.com,2011:article/2061682023-05-26T18:01:24Z2023-05-26T18:01:24ZIncluding race in clinical algorithms can both reduce and increase health inequities – it depends on what doctors use them for<figure><img src="https://images.theconversation.com/files/528403/original/file-20230525-15-2tu1k6.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C2121%2C1412&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">An increasing number of health care decisions rely on information from algorithms.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/doctors-discussing-patients-test-results-royalty-free-image/1062188494">Tom Werner/Digital Vision via Getty Images</a></span></figcaption></figure><p>Health practitioners are <a href="https://doi.org/10.1056/NEJMms2004740">increasingly concerned</a> that because race is a social construct, and the biological mechanisms of how race affects clinical outcomes are often unknown, including race in predictive algorithms for clinical decision-making may worsen inequities.</p>
<p>For example, to calculate an estimate of kidney function called the <a href="https://doi.org/10.7326%2F0003-4819-150-9-200905050-00006">estimated glomerular filtration rate, or eGFR</a>, health care providers use an algorithm based on age, biological sex, race (Black or non-Black) and serum creatinine, a waste product the kidneys release into the blood. A higher eGFR value means better kidney health. These eGFR predictions are used to <a href="https://optn.transplant.hrsa.gov/professionals/by-organ/kidney-pancreas/kidney-allocation-system/">allocate kidney transplants in the U.S.</a></p>
<p>Based on this algorithm, which was <a href="https://www.kidney.org/atoz/content/race-and-egfr-what-controversy">trained on actual GFR values from patients</a>, a Black patient would be assigned a higher eGFR than a non-Black patient of the same age, sex and serum creatinine level. This implies that some Black patients would be considered to have healthier kidneys than otherwise similar non-Black patients and less likely to be assigned a kidney transplant.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/1O7Ov1nxMc0?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">Biased clinical algorithms can lead to inaccurate diagnoses and delayed treatment.</span></figcaption>
</figure>
<p>In 2021, however, researchers found that excluding race in the original eGFR equations could <a href="https://doi.org/10.1056/NEJMoa2102953">lead to larger discrepancies</a> between estimated and actual GFR values for both Black and non-Black patients. They also found adding an additional biomarker called cystatin C can improve predictions. However, even with this biomarker, excluding race from the algorithm still led to elevated discrepanies across races.</p>
<p>I am a <a href="https://scholar.google.com/citations?user=AR72duAAAAAJ&hl=en">health economist and statistician</a> who studies how unobserved factors in data can result in biases that lead to inefficiencies, inequities and disparities in health care. My recently published research suggests that excluding race from certain diagnostic algorithms <a href="https://www.science.org/doi/10.1126/sciadv.add2704">could worsen health inequities</a>.</p>
<h2>Different approaches to fairness</h2>
<p>Researchers use <a href="https://plato.stanford.edu/entries/economic-justice/">different economic frameworks</a> to understand how society allocates resources. Two key frameworks are utilitarianism and equality of opportunity.</p>
<p>A purely <a href="https://doi.org/10.3386/w30700">utilitarian outlook</a> seeks to identify what features would get the most out of a positive outcome or reduce the harm from a negative one, ignoring who possesses those features. This approach allocates resources to those with the most opportunities to generate positive outcomes or mitigate negative ones.</p>
<p>A utilitarian approach would always include race and ethnicity to improve the prediction power and accuracy of algorithms, regardless of whether it’s fair. For example, utilitarian policies would aim to maximize overall survival among people seeking organ transplants. They would allocate organs to those who would survive the longest from transplantation, even if those who may not survive the longest due to circumstances outside their control and need the organs most would die sooner without the transplant.</p>
<p>Although utilitarian approaches do not take fairness into account, an approach that does would ask two questions: How do we define fairness? Are there conditions when maximizing an algorithm’s prediction power and accuracy would not conflict with fairness?</p>
<p>To answer these questions, I apply the <a href="https://www.jstor.org/stable/41106460">equality of opportunity</a> framework, which aims to allocate resources in a way that allows everyone the same chance of obtaining similar outcomes, without being disadvantaged by circumstances outside of their control. Researchers have used this framework in many contexts, such as <a href="https://www.jstor.org/stable/447264">political science</a>, <a href="https://press.uchicago.edu/ucp/books/book/chicago/E/bo22415931.html">economics</a> and <a href="https://www.jpe.ox.ac.uk/papers/what-makes-discrimination-wrong/">law</a>. The U.S. Supreme Court has also applied equality of opportunity in <a href="https://edeq.stanford.edu/sections/section-4-lawsuits/landmark-us-cases-related-equality-opportunity-k-12-education">several landmark rulings in education</a>.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Health care worker looking at tablet in an exam room" src="https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/528406/original/file-20230525-21-rcdl1v.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Including different variables in clinical algorithms can lead to very different results.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/unrecognizeable-person-using-digital-tablet-royalty-free-image/1421626437">SDI Productions/E+ via Getty Images</a></span>
</figcaption>
</figure>
<h2>Equality of opportunity</h2>
<p>There are two fundamental principles in equality of opportunity.</p>
<p>First, inequality of outcomes is unethical if it results from differences in circumstances that are outside of an individual’s own control, such as the income of a child’s parents, exposure to systemic racism or living in <a href="https://theconversation.com/black-mothers-trapped-in-unsafe-neighborhoods-signal-the-stressful-health-toll-of-gun-violence-in-the-u-s-203307">violent and unsafe environments</a>. This can be remedied by compensating individuals with disadvantaged circumstances in a way that allows them the same opportunity to obtain certain health outcomes as those who are not disadvantaged by their circumstances.</p>
<p>Second, inequality of outcomes for people in similar circumstances that result from differences in individual effort, such as practicing health-promoting behaviors like diet and exercise, is not unethical, and policymakers can reward those achieving better outcomes through such behaviors. However, differences in individual effort that occur because of circumstances, such as living in an area with <a href="https://theconversation.com/how-urban-planning-and-housing-policy-helped-create-food-apartheid-in-us-cities-154433">limited access to healthy food</a>, are not addressed under equality of opportunity. Keeping all circumstances the same, any differences in effort between individuals should be due to preferences, free will and perceived benefits and costs. This is called <a href="https://doi.org/10.1257/jel.20151206">accountable effort</a>. So, two individuals with the same circumstances should be rewarded according to their accountable efforts, and society should accept the resulting differences in outcomes.</p>
<p>Equality of opportunity implies that if algorithms were to be used for clinical decision-making, then it is necessary to understand what causes variation in the predictions they make. </p>
<p>If variation in predictions results from differences in circumstances or biological conditions but not from individual accountable effort, then it is appropriate to use the algorithm for compensation, such as allocating kidneys so everyone has an equal opportunity to live the same length of life, but not for reward, such as allocating kidneys to those who would live the longest with the kidneys.</p>
<p>In contrast, if variation in predictions results from differences in individual accountable effort but not from their circumstances, then it is appropriate to use the algorithm for reward but not compensation.</p>
<h2>Evaluating clinical algorithms for fairness</h2>
<p>To hold machine learning and other artificial intelligence algorithms accountable to a standard of equity, I applied the principles of equality of opportunity to
<a href="https://www.science.org/doi/10.1126/sciadv.add2704">evaluate whether race should be included</a> in clinical algorithms. I ran simulations under both ideal data conditions, where all data on a person’s circumstances is available, and real data conditions, where some data on a person’s circumstances is missing.</p>
<p>In these simulations, I unequivocally assume that <a href="https://www.genome.gov/genetics-glossary/Race">race is a social and not biological construct</a>. Variables such as race and ethnicity are often <a href="https://www.ama-assn.org/press-center/press-releases/new-ama-policies-recognize-race-social-not-biological-construct">proxies for various circumstances</a> individuals face that are out of their control, such as systemic racism that contributes to health disparities.</p>
<figure>
<iframe width="440" height="260" src="https://www.youtube.com/embed/926PqQUOVOg?wmode=transparent&start=0" frameborder="0" allowfullscreen=""></iframe>
<figcaption><span class="caption">As a social construct, race is often a proxy for nonbiological circumstances.</span></figcaption>
</figure>
<p>I evaluated two categories of algorithms.</p>
<p>The first, diagnostic algorithms, makes predictions based on outcomes that have already occurred at the time of decision-making. For example, diagnostic algorithms are used to predict the presence of gallstones in patients with abdominal pain or urinary tract infections, or to detect breast cancer using radiologic imaging.</p>
<p>The second, prognostic algorithms, predicts future outcomes that have not yet occurred at the time of decision-making. For example, prognostic algorithms are used to predict whether a patient will live if they do or do not obtain a kidney transplant.</p>
<p>I found that, under an equality of opportunity approach, diagnostic models that do not take race into account would <a href="https://www.science.org/doi/10.1126/sciadv.add2704">increase systemic inequities and discrimination</a>. I found similar results for prognostic models intended to compensate for individual circumstances. For example, excluding race from algorithms that predict the future survival of patients with kidney failure would fail to identify those with underlying circumstances that make them more vulnerable.</p>
<p>Including race in prognostic models intended to reward individual efforts <a href="https://www.science.org/doi/10.1126/sciadv.add2704">can also increase disparities</a>. For example, including race in algorithms that predict how much longer a person would live after a kidney transplant may fail to account for individual circumstances that could limit how much longer they live.</p>
<h2>Unanswered questions and future work</h2>
<p>Better biomarkers may one day be able to better predict health outcomes than race and ethnicity. Until then, including race in certain clinical algorithms could help reduce disparities.</p>
<p>Although my study uses an equality of opportunity framework to measure how race and ethnicity affect the results of prediction algorithms, researchers don’t know whether other ways to approach fairness would lead to different recommendations. How to choose between different approaches to fairness also remains to be seen. Moreover, there are questions about how multiracial groups should be coded in health databases and algorithms.</p>
<p><a href="https://sop.washington.edu/choice/">My colleagues and I</a> are exploring many of these unanswered questions to reduce algorithmic discrimination. We believe our work will readily extend to other areas outside of health, including education, crime and labor markets.</p><img src="https://counter.theconversation.com/content/206168/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Anirban Basu received funding support from a consortium of ten biomedical companies to the University of Washington through an unrestricted gift. </span></em></p>Biased algorithms in health care can lead to inaccurate diagnoses and delayed treatment. Deciding which variables to include to achieve fair health outcomes depends on how you approach fairness.Anirban Basu, Professor of Health Economics, University of WashingtonLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/2029352023-04-05T12:25:12Z2023-04-05T12:25:12ZOne way to speed up clinical trials: Skip right to the data with electronic medical records<figure><img src="https://images.theconversation.com/files/519333/original/file-20230404-15-mnljji.jpg?ixlib=rb-1.1.0&rect=0%2C0%2C2463%2C1216&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">It takes around 17 years for medical research to translate into clinical practice.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/touching-base-with-the-medical-community-royalty-free-image/1167942823">shapecharge/E+ via Getty Images</a></span></figcaption></figure><p>Scientific knowledge, as measured by numbers of papers published, has been estimated to <a href="https://doi.org/10.1057/s41599-021-00903-w">double every 17.3 years</a>. However, it takes an <a href="https://doi.org/10.1258%2Fjrsm.2011.110180">average of about 17 years</a> for health and medical research – going from basic lab studies on cell cultures and animals to clinical trials in people – to result in actual changes patients see in the clinic.</p>
<p>The typical process of medical research is generally <a href="https://theconversation.com/90-of-drugs-fail-clinical-trials-heres-one-way-researchers-can-select-better-drug-candidates-174152">not well equipped</a> to respond effectively to quickly evolving pandemics. This has been especially evident for the COVID-19 pandemic, in part because the virus the causes COVID-19 mutates frequently. Scientists and public health officials are often left <a href="https://theconversation.com/18-months-of-the-covid-19-pandemic-a-retrospective-in-7-charts-166881">continually scrambling</a> to develop and test new treatments to match emerging variants. </p>
<p>Fortunately, scientists may be able to bypass the typical research timeline and study treatments and interventions as they are used in the clinic nearly in real time by leveraging a common source of existing data – electronic medical records, or EMRs.</p>
<p>We are a team composed of an <a href="https://scholar.google.com/citations?user=0BCX1qIAAAAJ&hl=en">epidemiologist</a>, <a href="https://scholar.google.com/citations?user=LNTsvI8AAAAJ&hl=en">pharmacist</a> and <a href="https://profiles.dom.pitt.edu/card/faculty_info.aspx/Marroquin5220">cardiologist</a> at the University of Pittsburgh Medical Center. During the COVID-19 pandemic, we realized the need to quickly study and disseminate accurate information on the most effective treatment approaches, especially for patients at high risk of hospitalization and death. In our <a href="https://doi.org/10.7326/M22-1286">recently published research</a>, we used EMR data to show that early treatment with one or more of five different monoclonal antibodies substantially reduced the risk of hospitalization or death compared with delayed or no treatment. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Two surgeons reviewing medical records in front of computer screens" src="https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/519338/original/file-20230404-18-reloqo.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">EMRs contain a wealth of clinical data that could be used for research.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/nurses-plan-surgical-paperwork-royalty-free-image/140175520">Reza Estakhrian/The Image Bank via Getty Images</a></span>
</figcaption>
</figure>
<h2>Using EMR data for research</h2>
<p>In the U.S., health care systems typically use EMR systems for documenting patient care and for administrative purposes like billing. While data collection is not uniform, these systems typically contain <a href="https://www.cms.gov/medicare/e-health/ehealthrecords">detailed records</a> that can include sociodemographic information, medical history, test results, surgical and other procedures, prescriptions and billing charges.</p>
<p>Unlike <a href="https://worldpopulationreview.com/country-rankings/countries-with-single-payer">single-payer health care systems</a> that integrate data into a single EMR system, such as in the U.K. and in Scandinavian countries, many large health care systems in the U.S. collect patient data using <a href="https://www.definitivehc.com/blog/most-common-inpatient-ehr-systems">multiple EMR systems</a>. </p>
<p>Having multiple EMR systems adds a layer of complexity to using such data to conduct scientific research. To address this, the University of Pittsburgh Medical Center developed and maintains a clinical data warehouse that compiles and harmonizes data across the seven different EMR systems its 40 hospitals and outpatient clinics use.</p>
<h2>Emulating clinical trials</h2>
<p><a href="https://doi.org/10.1007%2Fs00392-016-1025-6">Using EMR data for research</a> is not new. More recently, researchers have been looking into ways to use these large health data systems to <a href="https://doi.org/10.1093/aje/kwv254">emulate randomized controlled trials</a>, which are considered the gold standard study design yet are often costly and take years to complete.</p>
<p>Using this emulation framework, our team used the EMR data infrastructure at our institution to <a href="https://doi.org/10.7326/M22-1286">evaluate five different monoclonal antibodies</a> for which the Food and Drug Administration granted emergency use authorization to treat COVID-19. Monoclonal antibodies are human-made proteins designed to prevent a pathogen – in this case the virus that causes COVID-19 – from entering human cells, replicating and causing serious illness. Initially the authorizations were based on clinical trial data. But as the virus mutated, subsequent evaluations based on cell culture studies suggested a loss of effectiveness. </p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="Close-up of health care provider accessing medical record on tablet" src="https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/519339/original/file-20230404-473-eyylv1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">EMR data could be used to confirm that the results of cell culture studies would apply in the clinic.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/nurse-using-portable-computer-royalty-free-image/104509052">Solskin/DigitalVision via Getty Images</a></span>
</figcaption>
</figure>
<p>We wanted to confirm that the findings of cell-based studies applied to actual patients. So we evaluated anonymous clinical data from 2,571 patients treated with these monoclonal antibodies within two days of COVID-19 infection, matching them with data from 5,135 patients with COVID-19 who were eligible for but either did not receive these treatments or received them three or more days after infection. </p>
<p>We found that overall, people who received monoclonal antibodies within two days of a positive COVID-19 test reduced their risk of hospitalization or death by 39% compared with those who did not receive the treatment or received delayed treatment. In addition, patients with compromised immune systems reduced their risk of hospitalization or death by 55%, regardless of their age.</p>
<p>Our near-real-time analysis of COVID-19 patients treated with monoclonal antibodies during the pandemic confirmed the findings of the cell culture studies. Our findings suggest that by using data in this way, researchers may be able to evaluate treatments in times of urgency without having to perform clinical trials.</p>
<h2>Appropriate EMR data use</h2>
<p>Many health care institutions have EMR systems that researchers can harness to rapidly answer important research questions as they arise. However, because this clinical data is not specifically collected for research purposes, researchers need to <a href="https://doi.org/10.1146/annurev-publhealth-032315-021353">carefully design their studies</a> and use rigorous data validation and analysis. They also need to take great care to harmonize data from different EMR systems, select appropriate patient samples and minimize all sources of potential bias. </p>
<p>New pandemics and significant public health challenges are likely to emerge abruptly and in unpredictable ways. Given the treasure trove of data routinely collected across U.S. health care systems, we believe that careful use of these data can help answer urgent health questions in ways that are representative of who’s actually receiving care.</p><img src="https://counter.theconversation.com/content/202935/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Erin McCreary has served on scientific advisory boards for Shionogi, Inc and Merck.</span></em></p><p class="fine-print"><em><span>Kevin Kip and Oscar Marroquin do not work for, consult, own shares in or receive funding from any company or organization that would benefit from this article, and have disclosed no relevant affiliations beyond their academic appointment.</span></em></p>In health care crises, researchers can avoid waiting for clinical trial results by using data from health care systems to analyze the effectiveness of treatments for COVID-19 and other illnesses.Kevin Kip, Vice President of Clinical Analytics, University of PittsburghErin McCreary, Clinical Assistant Professor of Medicine, University of PittsburghOscar Marroquin, Associate Professor of Medicine, Epidemiology and Clinical and Translational Science, University of PittsburghLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/1733962021-12-09T13:34:54Z2021-12-09T13:34:54ZFiguring out omicron – here’s what scientists are doing right now to understand the new coronavirus variant<figure><img src="https://images.theconversation.com/files/436465/original/file-20211208-137612-ikwh0c.jpg?ixlib=rb-1.1.0&rect=686%2C0%2C7210%2C5150&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">A researcher works with COVID-19 samples from patients.</span> <span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/laboratory-operator-handles-positive-covid-19-samples-to-be-news-photo/1237075524">Thomas Samson/AFP via Getty Images</a></span></figcaption></figure><p><em>Scientists around the world have been racing to learn more about the new omicron strain of SARS-CoV-2, first declared a <a href="https://www.who.int/news/item/26-11-2021-classification-of-omicron-(b.1.1.529)-sars-cov-2-variant-of-concern">“variant of concern” on Nov. 26, 2021</a> by the World Health Organization. Officials cautioned that it would take several weeks before they’d know whether the recently emerged coronavirus variant is more contagious and causes more or less serious COVID-19 than delta and other earlier variants, and whether current vaccines can ward it off.</em></p>
<p><em><a href="https://scholar.google.com/citations?user=OQ7vzu0AAAAJ&hl=en&oi=ao">Peter Kasson is a virologist and biophysicist</a> at the University of Virginia who studies how viruses such as SARS-CoV-2 enter cells and what can be done to stop them. Here he explains what lab-based scientists are doing to help answer the outstanding questions about omicron.</em></p>
<h2>Does prior immunity protect against omicron?</h2>
<p>These are the key lab results everyone is waiting for: How effective are the antibodies people already have at fighting off omicron? If you got the booster shot, are you protected? Or if you had COVID-19 and then were vaccinated?</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="artist's rendition of a virus with antibodies surrounding it" src="https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=338&fit=crop&dpr=1 600w, https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=338&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=338&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=424&fit=crop&dpr=1 754w, https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=424&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/436467/original/file-20211208-25-152atd7.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=424&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Will the antibodies people already have recognize and thwart omicron?</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/photo/antibodies-attacking-sars-cov-2-virus-corona-virus-royalty-free-image/1328466445">Dr_Microbe/iStock via Getty Images</a></span>
</figcaption>
</figure>
<p>The goal is to see how well antibodies from real people who have had COVID-19 or have been vaccinated against it can hold off omicron in petri dishes in the lab. Scientists expect that antibodies from people exposed to other variants won’t work as well against omicron because of its mutations, but they need to measure how much less well and whether it’s still enough to stop the virus. </p>
<p>To answer these questions, most researchers first make a version of the SARS-CoV-2 virus that can <a href="https://doi.org/10.3390/v12050513">enter cells but not reproduce</a>. A few specialized labs with <a href="https://theconversation.com/we-work-with-dangerous-pathogens-in-a-downtown-boston-biocontainment-lab-heres-why-you-can-feel-safe-about-our-research-163197">extra levels of biosecurity</a> use the actual virus. Scientists add antibodies from the blood of people vaccinated against or recovered from COVID-19 to the virus. They then mix this with human lung cells to see whether the antibodies can stop the virus from infecting the cells.</p>
<p>My laboratory performs this kind of work with <a href="https://doi.org/10.1038/s41541-021-00399-0">SARS-CoV-2</a> and other <a href="https://doi.org/10.1021/acscentsci.8b00494">emerging viruses</a>. Researchers have used these well-established techniques to test out <a href="https://doi.org/10.1038/s41586-021-03696-9">antibodies after COVID-19 recovery</a>, as well as different vaccines and <a href="https://doi.org/10.1056/NEJMc2113468">different variants</a>. </p>
<p>If antibodies people made against prior variants can’t stop omicron from infecting lung cells in the lab, then those antibodies probably won’t protect people out in the world either.</p>
<p>The very first early results are starting to come back, and it looks like <a href="https://www.ahri.org/wp-content/uploads/2021/12/MEDRXIV-2021-267417v1-Sigal.pdf">antibodies against earlier variants are less successful at blocking omicron</a>. Researchers took antibodies from six people who each had two doses of vaccine and from six other people who each had two doses of vaccine and had also recovered from an earlier COVID-19 infection. Antibodies from both groups of people were about 40 times worse at stopping omicron than original SARS-COV-2 strains, based on how much antibody was needed to prevent infection. But the people whose immune systems had seen the virus three times – that is, were doubly vaccinated and had also recovered from COVID-19 – had antibody levels that were high enough to still stop infection.</p>
<p>I’d expect people who have received booster vaccines will have similar or greater levels of immunity and will be at least moderately protected from omicron. But it will need to be tested. <a href="https://www.pfizer.com/news/press-release/press-release-detail/pfizer-and-biontech-provide-update-omicron-variant">Pfizer has said their early results agree with this prediction</a>, but the data is not yet publicly available. All of this work is not yet peer reviewed and still very preliminary.</p>
<p>Scientists will need to determine how a drop in “neutralization titer,” or how good antibodies are at blocking the virus in the lab, corresponds to a drop in “<a href="https://www.who.int/news-room/feature-stories/detail/vaccine-efficacy-effectiveness-and-protection">vaccine effectiveness</a>” or how likely a vaccinated person is to get COVID-19 compared to an unvaccinated one. Scientists know that <a href="https://doi.org/10.1038/s41591-021-01377-8">better antibodies correspond to more effective vaccines</a>, but the precise numerical relationships need to be determined.</p>
<p><iframe id="ikaxY" class="tc-infographic-datawrapper" src="https://datawrapper.dwcdn.net/ikaxY/1/" height="400px" width="100%" style="border: none" frameborder="0"></iframe></p>
<h2>How contagious is omicron compared to delta?</h2>
<p>The past pandemic year has shown that contagiousness, or transmissibility, has been the key factor in determining whether a coronavirus variant becomes dominant. Delta’s transmissibility has made it the current dominant variant because it simply outran others. But that situation may change with time.</p>
<p>The basic elements of the viral “life” cycle are getting into cells, making more virus, and getting out. Scientists can measure each of these stages in the lab and <a href="https://www.science.org/doi/10.1126/science.abl6184">report what aspects of a variant</a> make it more or less transmissible. In addition to binding to human cells better, some mutations enhance the packaging of new virus and the delivery of its genes once the virus gets into the cell.</p>
<p>While lab-based science can help people understand the biology behind just why a variant is more or less contagious, right now nature is doing a much bigger real-world experiment. Disease surveillance data from the <a href="https://twitter.com/_nickdavies/status/1466204363110633476?s=20">U.K.</a> and <a href="https://files.ssi.dk/covid19/omikron/statusrapport/rapport-omikronvarianten-07122021-1t6o">other countries</a> where delta has been dominant suggest that omicron is gaining share and may eventually displace delta.</p>
<p>Exactly how this plays out may differ from one country to another, depending on factors like the number of vaccinated people and which variants were previously in circulation, but this news about how good omicron is at spreading is concerning.</p>
<h2>Does omicron make people more or less sick?</h2>
<p>This is again a question that will be answered much more quickly by the thousands of people infected with omicron than by work in the lab. It’s important to remember, though, that nature’s experiments are not as carefully controlled as lab experiments. Precise lab work will help explain why omicron might be different, but the first answers here will come from hospitals.</p>
<p>Lab-based scientists will be working with hospitals to analyze what makes some patients more or less sick once they contract omicron. Some early numbers suggest that the <a href="https://www.samrc.ac.za/news/tshwane-district-omicron-variant-patient-profile-early-features">first omicron cases are mostly mild</a>, but public health officials urge caution: Most cases of all COVID-19 variants are mild, and <a href="https://www.samrc.ac.za/news/tshwane-district-omicron-variant-patient-profile-early-features">many of those infected so far with omicron are younger</a>. Hospitalization counts tend to increase somewhat after the initial increase in cases. So this question will take time to answer.</p>
<figure class="align-center zoomable">
<a href="https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=1000&fit=clip"><img alt="nurse attends a COVID-19 patient on a hospital ward" src="https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/436468/original/file-20211208-25-5qjw44.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px"></a>
<figcaption>
<span class="caption">Epidemiological data about how real patients are faring will fill in the picture.</span>
<span class="attribution"><a class="source" href="https://www.gettyimages.com/detail/news-photo/registered-nurse-attends-a-patient-with-covid-19-at-the-news-photo/1235025034">Apu Gomes/AFP via Getty Images</a></span>
</figcaption>
</figure>
<h2>How are lab data and public health data complementary?</h2>
<p>Laboratories will provide the first results on immune protection against omicron, although this will be followed up with public health data that will likely confirm the lab results. Public health data will bring the first results on contagiousness and disease severity, which will then be explained by laboratory results.</p>
<p>Once the initial answers from public health data are in, laboratory results are still important to understand why these changes happened and to help predict what future variants will do. How do officials declare a variant of concern in the first place? It’s a combination of public health data and understanding from the lab.</p>
<h2>What do we know already?</h2>
<p>Variants of SARS-CoV-2 don’t change the laws of physics and biology. They cannot leap tall buildings in a single bound. Physical barriers like high-grade masks and good ventilation will still stop the virus. And, very likely, vaccines will continue to provide some amount of protection. The question is how much, and whether the world needs to <a href="https://theconversation.com/how-can-scientists-update-coronavirus-vaccines-for-omicron-a-microbiologist-answers-5-questions-about-how-moderna-and-pfizer-could-rapidly-adjust-mrna-vaccines-172943">change the current vaccines</a> or just provide more of them.</p>
<p>[<em>Research into coronavirus and other news from science</em> <a href="https://theconversation.com/us/newsletters/science-editors-picks-71/?utm_source=TCUS&utm_medium=inline-link&utm_campaign=newsletter-text&utm_content=science-corona-research">Subscribe to The Conversation’s new science newsletter</a>.]</p><img src="https://counter.theconversation.com/content/173396/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Peter Kasson has received funding from the National Institutes of Health, the National Science Foundation, the Knut and Alice Wallenberg Foundation, the Swedish Research Council, and TG Therapeutics. He is affiliated with the University of Virginia and Uppsala University. </span></em></p>Careful lab work will complement public health data as researchers worldwide focus on omicron, asking questions about contagiousness, severity of disease and whether vaccines hold up against it.Peter Kasson, Associate Professor of Molecular Physiology and Biomedical Engineering, University of VirginiaLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/1032412018-09-24T20:58:53Z2018-09-24T20:58:53ZBroke your arm? Exercise the other one to strengthen it…<figure><img src="https://images.theconversation.com/files/237194/original/file-20180919-146148-4sihbv.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">In a research study, students with an immobilized left arm who trained their opposite wrist completely preserved both the strength and muscle volume in the left arm.</span> <span class="attribution"><span class="source">(Shutterstock)</span></span></figcaption></figure><p>If you have ever broken an arm and had to wear a cast or splint for a few weeks, you will be familiar with the alarming loss of muscle and uneasy feeling of weakness experienced after removing your cast. </p>
<p>Most people do not do much exercise while a broken arm is healing and can struggle with this loss of muscle, known as “atrophy,” and weakness for many weeks after the injury. </p>
<p>A new study published recently in the <a href="https://www.physiology.org/doi/full/10.1152/japplphysiol.00971.2017"><em>Journal of Applied Physiology</em></a>, conducted in my lab by graduate student Justin Andrushko, suggests an effective strategy to offset muscle weakness might be to exercise the other arm.</p>
<p>We recruited a group of 16 college students to wear casts on their left wrists for four weeks. Half of these students exercised their right arm aggressively three days per week using a type of training known as “eccentric training” — which lengthens the muscle during contraction, and is quite effective for building muscle and enhancing strength. </p>
<p>Before and after the study period, we measured wrist strength in several different ways and quantified muscle volume using a Computed Tomography (CT) scan of the forearm. As expected, those students who did not train lost about 20 per cent of their strength and about three per cent of their muscle volume after four weeks. </p>
<p>Remarkably, the students who trained their opposite wrist completely preserved both the strength and muscle volume in the left, immobilized arm. This research has received a lot of attention.</p>
<h2>Possible ‘mirror’ contractions</h2>
<p>The phenomenon that creates the effect is known as “cross-education,” and has been documented for over a century, but the new study is one of just a handful to measure the effect when the opposite limb is immobilized. </p>
<p>We are the first to examine the effects using CT scans to measure muscle volume, and to measure the strength of multiple muscle groups in both arms (i.e. wrist flexors and extensors). </p>
<p>It turns out that the effect appears to be quite specific: training of the right wrist flexors preserved the left wrist flexors, but not the extensor muscles.</p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=400&fit=crop&dpr=1 600w, https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=400&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=400&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=503&fit=crop&dpr=1 754w, https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=503&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/237192/original/file-20180919-158240-c1i3k1.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=503&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">The loss of muscle after removing a cast from a broken limb can be alarming.</span>
<span class="attribution"><span class="source">(Shutterstock)</span></span>
</figcaption>
</figure>
<p>We do not fully understand what causes the effect. Most of the published work points towards changes in the nervous system relating to how the sides of the brain share information, or how they adapt together after training one arm. However, we are fascinated with the muscle size preservation effects. </p>
<p>Unfortunately, the study did not take detailed measures of anything inside the muscle. We suspect there could be some yet unknown connection between nervous system changes and the balance of muscle protein.</p>
<p>One theory is that there are small contractions, known as “mirror” contractions, under the cast while training the opposite side. We measured these contractions and they are very small — perhaps too small to preserve the muscle — but they are present. We need to do more research to understand the role of these small contractions in relation to prevention of atrophy.</p>
<h2>Consider training the opposite limb</h2>
<p>Although the results are exciting, we caution that the study was a controlled lab experiment involving young healthy volunteers without a real injury. </p>
<p>More work in clinical settings is needed before any changes to standard rehabilitation practices can be discussed. </p>
<p>There have been a few clinical studies already published — about wrist fracture and recovery from stroke and knee surgery — with promising results. The clinical studies seem more positive for fracture and stroke recovery and less so after knee surgeries. </p>
<p>Lab-controlled studies like the one we conducted are important to understand the underlying mechanisms of the effect, and to maximize its potential in future clinical work.</p>
<p>While more work in clinical settings is certainly needed, we can still recommend that if you ever experience a limb fracture, you might consider training your opposite limb. As with many types of exercise training, the risk of this approach is quite low and could have important benefits.</p><img src="https://counter.theconversation.com/content/103241/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Jonathan Farthing receives funding from the Natural Sciences and Engineering Research Council of Canada, the Saskatchewan Health Research Foundation, and the Royal University Hospital Foundation.</span></em></p><p class="fine-print"><em><span>Justin Andrushko received funding from Canadian Institute of Health Research MSc Scholarship.</span></em></p>A research study shows that training the other limb can actually help preserve muscle in a broken and immobilized one.Jonathan Farthing, Associate Professor, College of Kinesiology, University of SaskatchewanJustin Andrushko, PhD Student, University of SaskatchewanLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/923172018-02-27T17:06:36Z2018-02-27T17:06:36ZThe key to treating multiple sclerosis could be inside sufferers’ own bodies<figure><img src="https://images.theconversation.com/files/208110/original/file-20180227-36686-1dyi84f.jpg?ixlib=rb-1.1.0&rect=0%2C29%2C5000%2C3218&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">
</span> <span class="attribution"><a class="source" href="https://www.shutterstock.com/image-vector/nerve-cell-anatomy-detailed-illustration-on-645801253">Tefi/Shutterstock</a></span></figcaption></figure><p>Fat often gets a bad press, but if it didn’t coat the cables that connect our neurons, we’d be in a lot of trouble. Sufferers of multiple sclerosis and a host of other nervous system diseases have first-hand experience of this, with few safe and effective treatment options available. Only now are new treatments appearing on the horizon that might just make a big difference.</p>
<p>In order for us to think, feel and move, information must move around the brain accurately and rapidly. Vital in this process are long wire-like structures called axons, which conduct the electrical currents that encode our thoughts from neuron to neuron.</p>
<p>Most of our axons are sheathed in a fatty substance called <a href="https://www.nationalmssociety.org/What-is-MS/Definition-of-MS/Myelin">myelin</a> which, like the plastic coating on a wire, provides insulation for efficient conduction and protects the axon from damage.</p>
<p>Unfortunately, many diseases damage these myelin sheaths. For example, in <a href="https://theconversation.com/explainer-multiple-sclerosis-32662">multiple sclerosis</a> (MS), the immune system – usually our body’s defence against disease – attacks its own myelin in the brain and spinal cord, leaving the underlying axons exposed. Like a worn-down phone charger, these bare axons can no longer conduct electricity effectively, and are vulnerable to damage. Depending on which cables are damaged, this can cause tingling, weakness, visual problems, and eventually difficulty moving, speaking and swallowing.</p>
<figure> <img src="https://upload.wikimedia.org/wikipedia/commons/4/48/Saltatory_Conduction.gif"><figcaption> An unmyelinated axon and a myelinated axon, side-by-side. Source: www.docjana.com</figcaption></figure>
<p><a href="https://www.mssociety.org.uk/dmts">Most current therapies</a> for MS attempt to stop the immune system from attacking the myelin sheaths. This can reduce damage, but it can’t reverse it. So the condition of many patients deteriorates even while on these drugs. Stem cell transplantation therapy has shown recent promise in treating MS, but such treatments are aggressive and can <a href="https://theconversation.com/can-stem-cell-therapy-really-treat-multiple-sclerosis-63162">seriously endanger patients’ health</a>, requiring chemotherapy to almost completely eliminate the patient’s immune system before attempting to reboot it to an earlier, more healthy stage.</p>
<p>Now, a different kind of stem cell offers exciting potential for a raft of new treatments that could reverse symptoms of MS and other myelin diseases, rather than just slow them – and without the need for transplantation.</p>
<h2>A new hope</h2>
<p>After myelin damage, stem cells called <a href="https://en.wikipedia.org/wiki/Oligodendrocyte_progenitor_cell">OPCs</a> can create specialised brain cells called oligodendrocytes, which send octopus-like arms to wrap new myelin around damaged axons. OPCs are already scattered throughout the brains of MS sufferers, but <a href="https://www.ncbi.nlm.nih.gov/pubmed/23595275">only in some people</a> do they produce enough of the specialised brain cells that regenerate myelin, and therefore <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5006855/">reduce symptoms</a>.</p>
<p>Recent years have seen <a href="https://www.nature.com/articles/nrn.2017.136">great advances</a> in our understanding of how to influence OPC stem cells to respond properly to myelin damage. We can now grow them in hundreds of tiny artificial wells, each containing a different drug and several microscopic axon-mimicking cables, and examine which drugs best kick-start the OPCs into re-myelinating action. <a href="http://www.msdiscovery.org/news/new_findings/12139-novel-remyelination-assay-allows-high-throughput-drug-screening">This innovative lab technique</a> is helping researchers to fast identify the most promising concoctions to take to clinical trials.</p>
<p>Surprisingly, recent discoveries also show that the same immune system responsible for attacking and damaging myelin can also play a beneficial role in regenerating it. For example, immune cells called microglia can gobble up the debris of the old myelin sheaths, clearing the way for new myelin to regenerate. Drugs targeting this process have already <a href="https://www.ncbi.nlm.nih.gov/pubmed/25609628">helped mice to regenerate mylein</a> and will likely be seen in clinical trials soon. What’s more, new <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5006855/">medical imaging technologies</a> will allow us to monitor how well all of these new drugs regenerate myelin inside patients in real time.</p>
<p>The next few years will be an exciting time, as we begin to see clinical data on how these new drugs can help people living with MS. After years of struggle to find an effective treatment, we may just find that the key was inside our bodies all along.</p><img src="https://counter.theconversation.com/content/92317/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Chris McMurran receives funding from MedImmune, and the Jean Shanks Foundation. </span></em></p>All multiple sclerosis sufferers have stem cells with the potential to heal them, but scientists are only just figuring out how to kick them into action.Chris McMurran, MB/PhD Candidate, University of CambridgeLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/637822016-08-12T15:33:59Z2016-08-12T15:33:59ZAI can excel at medical diagnosis, but the harder task is to win hearts and minds first<figure><img src="https://images.theconversation.com/files/133962/original/image-20160812-14381-spu2im.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption"></span> <span class="attribution"><span class="source">angellodeco/shutterstock.com</span></span></figcaption></figure><p>Scientists in Japan reportedly saved a woman’s life by applying artificial intelligence to help them diagnose a rare form of cancer. Faced with a 60-year-old woman whose cancer diagnosis was unresponsive to treatment, they supplied an AI system with huge amounts of clinical cancer case data, and it <a href="http://futurism.com/ai-saves-womans-life-by-identifying-her-disease-when-other-methods-humans-failed/">diagnosed the rare leukemia</a> that had stumped the clinicians in just ten minutes.</p>
<p>The <a href="http://www.ibm.com/watson/what-is-watson.html">Watson AI system from IBM</a> matched the patient’s symptoms against 20m clinical oncology studies uploaded by a team headed by <a href="http://www.ims.u-tokyo.ac.jp/imsut/en/lab/advancedclinicalresearch/">Arinobu Tojo</a> at the University of Tokyo’s Institute of Medical Science that included symptoms, treatment and response. The <a href="https://www.mskcc.org/blog/msk-trains-ibm-watson-help-doctors-make-better-treatment-choices">Memorial Sloan Kettering Cancer Center</a> in New York has carried out similar work, where teams of clinicians and data analysts trained Watson’s machine learning capabilities with oncological data in order to focus its predictive and analytic capabilities on diagnosing cancers.</p>
<p>IBM Watson first became famous when it <a href="http://www.techrepublic.com/article/ibm-watson-the-inside-story-of-how-the-jeopardy-winning-supercomputer-was-born-and-what-it-wants-to-do-next/">won the US television game show Jeopardy</a> in 2011. And IBM’s previous generation AI, Deep Blue, became <a href="http://www-03.ibm.com/ibm/history/ibm100/us/en/icons/deepblue/">the first AI to best a world champion at chess</a> when it beat Garry Kasparov in a game in 1996 and the entire match when they met again the following year. From a perspective of technological determinism, it may seem inevitable that AI has moved from chess to cancer in 20 years. Of course, it has taken a lot of hard work to get it there.</p>
<p>But efforts to use artificial intelligence, machine learning and big data in healthcare contexts have not been uncontroversial. On the one hand there is wild enthusiasm – lives saved by data, new medical breakthroughs, and a world of personalised medicine tailored to meet our needs by deep learning algorithms fed by smartphones and FitBit wearables. On the other there’s considerable scepticism – a lack of trust in machines, the importance of individuals over statistics, privacy concerns over patient records and medical confidentiality, and generalised fears of a Brave New World. Too often the debate dissolves into anecdote rather than science, or focuses on the breakthrough rather than the hard slog that led to it. Of course the reality will be somewhere in the middle. </p>
<figure class="align-center ">
<img alt="" src="https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&fit=clip" srcset="https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=600&h=393&fit=crop&dpr=1 600w, https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=600&h=393&fit=crop&dpr=2 1200w, https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=600&h=393&fit=crop&dpr=3 1800w, https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=754&h=494&fit=crop&dpr=1 754w, https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=30&auto=format&w=754&h=494&fit=crop&dpr=2 1508w, https://images.theconversation.com/files/133964/original/image-20160812-16364-k5qazy.jpg?ixlib=rb-1.1.0&q=15&auto=format&w=754&h=494&fit=crop&dpr=3 2262w" sizes="(min-width: 1466px) 754px, (max-width: 599px) 100vw, (min-width: 600px) 600px, 237px">
<figcaption>
<span class="caption">Not everyone agrees on how big data can be put to work.</span>
<span class="attribution"><span class="source">wk1003mike/shutterstock.com</span></span>
</figcaption>
</figure>
<h2>There’s not just a technical battle to win</h2>
<p>In fact, it may surprise you to learn that the world’s first computerised clinical decision-support system, <a href="http://www.openclinical.org/dss.html#firstGen">AAPhelp</a>, was developed in the UK way back in 1972 by <a href="https://books.google.co.uk/books?id=gtR-vtbqhwIC&pg=PA1&dq=tim+de+dombal&hl=en&sa=X&ved=0ahUKEwjT2KTly7bOAhVLKcAKHf3VC0kQ6AEIJDAB#v=onepage&q=tim%20de%20dombal&f=false">Tim De Dombal</a> and one of my colleagues, <a href="https://medhealth.leeds.ac.uk/profile/651/398/dr_susan_clamp/publications">Susan Clamp</a>.</p>
<p>This early precursor to the genius AI of today used a <a href="http://www.saedsayad.com/naive_bayesian.htm">naive Bayesian algorithm</a> to compute the likely cause of acute abdominal pain based on patient symptoms. Feeding the system with more symptoms and diagnosis helped it to become more accurate over time and, by 1974, De Dombal’s team had trained the system to the point where it was more accurate at diagnosis than junior doctors, and almost as accurate as the most senior consultants. It took AAPhelp overnight to give a diagnosis, but this was on 1970s computer hardware.</p>
<p>The bad news is that 40 years on, AAPhelp is still not in routine use. </p>
<p>This is the reality check for the most ardent advocates of applying technology to healthcare: to get technology such as predictive AIs into clinical settings where they can save lives means tackling all those negative connotations and fears. AI challenges people and their attitudes: the professionals that the machine can outperform, and the patients that are reduced to statistical probabilities to be fed into complex algorithms. Innovation in healthcare can take decades.</p>
<p>Nevertheless, while decades apart both AAPHelp and IBM Watson’s achievements demonstrate that computers can save lives. But the use of big data in healthcare implies that patient records, healthcare statistics, and all manner of other personal details might be used by researchers to train the AIs to make diagnoses. People are increasingly sensitive to the way personal data is used and, quite rightly, expect the highest standards of ethics, governance, privacy and security to be applied. The revelations that one NHS trust had given <a href="https://www.newscientist.com/article/2086454-revealed-google-ai-has-access-to-huge-haul-of-nhs-patient-data/">access to 1.6m identiable patient records</a> to Google’s DeepMind AI laboratory <a href="http://www.dailymail.co.uk/news/article-3571433/Google-s-artificial-intelligence-access-private-medical-records-1-6million-NHS-patients-five-years-agreed-data-sharing-deal.html">didn’t go down well</a> when reported a few months ago. </p>
<p>The hard slog is not creating the algorithms, but the patience and determination required to conduct careful work within the restrictions of applying the highest standards of data protection and scientific rigour. At the University of Leed’s <a href="http://lida.leeds.ac.uk">Institute for Data Analytics</a> we recently used IBM Watson Content Analytics software to analyse 50m pathology and radiology reports from the UK. Recognising the sensitivities, we brought IBM Watson to the data rather than passing the data to IBM. </p>
<p>Using natural language processing of the text reports we double-checked diagnoses such as brain metastases, HER-2-positive breast cancers and <a href="http://www.nhs.uk/Conditions/Hydronephrosis/Pages/Introduction.aspx">renal hydronephrosis</a> (swollen kidneys) with accuracy rates already over 90%. Over the next two years we’ll be developing these methods in order to embed these machine learning techniques into routine clinical care, at a scale that benefits the whole of the NHS. </p>
<p>While we’ve had £12m investment for our facilities and the work we’re doing, we’re not claiming to have saved lives yet. The hard battle is first to win hearts and minds – and on that front there’s still a lot more work to be done.</p><img src="https://counter.theconversation.com/content/63782/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Owen A Johnson receives research funding from MRC, EPSRC, NIHR, the NHS and InnovateUK. He is a director of X-Lab Ltd., an e-health software company focused on disruptive innovation in healthcare. </span></em></p>Computer-aided decision-making has been shown to help in clinical contexts. But winning over doctors and patients is a different matter.Owen A Johnson, Senior Fellow, University of LeedsLicensed as Creative Commons – attribution, no derivatives.tag:theconversation.com,2011:article/197552013-11-01T04:14:23Z2013-11-01T04:14:23ZMaking all clinical data public is vital for better medical care<figure><img src="https://images.theconversation.com/files/34215/original/c2f4jd93-1383278228.jpg?ixlib=rb-1.1.0&q=45&auto=format&w=496&fit=clip" /><figcaption><span class="caption">A large proportion of drug trial data never gets published, skewing our picture of drugs' effectiveness and safety. </span> <span class="attribution"><span class="source">opensource.com</span></span></figcaption></figure><p><a href="http://www.bmj.com/content/347/bmj.f6104">An article</a> published in the journal of the British Medical Association, BMJ, earlier this week illustrates a devastating problem with the “evidence base” in the academic medical literature. </p>
<p>A large proportion of drug trials, particularly those sponsored by pharmaceutical companies, never get published, skewing our picture of drugs’ effectiveness and safety. </p>
<p>But it’s not all bad news; <a href="https://theconversation.com/register-all-trials-report-all-results-its-long-overdue-11603">a UK initiative called AllTrials</a> is seeking to remedy the situation by calling for the results of all clinical trials to be published.</p>
<h2>A festering problem</h2>
<p><a href="http://www.trialsjournal.com/content/11/1/37">Research published in 2010</a> showed results unfavourable to sponsored drugs are less likely to be published, or selectively published to put a favourable “spin” on poor results.</p>
<p>And internal pharmaceutical industry documents released from court cases show concealment of data is a widespread practice.</p>
<p>A colleague and I <a href="http://link.springer.com/article/10.1007%2Fs11673-010-9208-8">assessed such documents</a> about psychiatric medications from five pharmaceutical companies. The <a href="http://www.healthyskepticism.org/global/news/int/hsin2009-12">papers suggested</a> widespread overstatement of benefits and understatement of adverse effects. </p>
<p>Other researchers have found <a href="http://annals.org/article.aspx?articleid=727539">similar problems</a> with different drugs.</p>
<h2>Journals respond</h2>
<p>In response, some medical journals voluntarily agreed to publish only studies registered on a website of the US National Institutes of Health, <a href="http://clinicaltrials.gov/">ClinicalTrials.gov</a>. </p>
<p>At least studies with unfavourable results would not be “buried” by drug companies. But the BMJ article confirms that many registered studies still don’t get published.</p>
<p>Chief editors of major medical journals have condemned the status quo. Marcia Angell of the New England Journal of Medicine has called industry-sponsored research “<a href="http://jama.jamanetwork.com/article.aspx?articleid=182478">a broken system</a>”; Richard Horton of The Lancet answered the rhetorical question of his editorial “<a href="http://www.thelancet.com/journals/lancet/article/PIIS0140-6736%2802%2908198-9/fulltext">How tainted has medicine become?</a>” with “heavily, and damagingly so”; and the BMJ’s Fiona Godlee <a href="http://1boringoldman.com/index.php/2013/07/11/a-sticky-wicket/">told a British parliamentary committee</a> that “drug companies should not be allowed to evaluate their own products.” </p>
<p>The previous chief-editor of the BMJ, Richard Smith has <a href="http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.0020138">offered a radical solution</a> – journals should stop publishing drug trials and instead provide expert critiques of the full data and methodology of trials posted on regulated websites.</p>
<h2>The AllTrials initiative</h2>
<p>The <a href="http://www.alltrials.net/">AllTrials initiative</a> aims to make that vision a reality. <a href="http://www.alltrials.net/all-trials/">It intends</a> to make the (de-identified) results and methodology of drug trials available to independent researchers so journals can publish in-depth articles based on all of the full data.</p>
<p>In 11 months, the campaign’s <a href="http://www.alltrials.net/supporters/">petition has gathered</a> over 59,000 individual signatories and over 400 medical and health-care organisations. These include many British medical colleges and learned academic medical science institutions, such as the <a href="http://www.cochrane.org/">Cochrane Collaboration</a> and the British <a href="http://www.nice.org.uk/">National Institute for Health and Clinical Excellence</a> (NICE).</p>
<p>Although the campaign is progressing slower outside the United Kingdom, it is managing to get some traction internationally. </p>
<p>The World Association of Medical Editors (WAME), the South African Medical Research Council and the Canadian Agency for Drugs and Technology in Health have signed. </p>
<p>Perhaps with an eye to the future, the International Federation of Medical Students Associations, the American Medical Students Association and the Australian Medical Students Association have gazumped their senior colleagues and also signed.</p>
<h2>Trouble in Europe</h2>
<p>The European Medicines Agency (EMA) – equivalent of America’s Food and Drug Administration (FDA) or Australia’s Therapeutic Goods Administration (TGA) – adopted a similar policy in 2010, and decided to <a href="http://www.ema.europa.eu/ema/index.jsp?curl=pages/news_and_events/news/2013/10/news_detail_001926.jsp&mid=WC0b01ac058004d5c1">release much of the drug trial data</a> in its possession.</p>
<p>The EMA’s policy is currently stalled by a European Court injunction because of a <a href="http://www.prweb.com/releases/2013/9/prweb11120572.htm">lawsuit by two pharmaceutical companies</a>. And Europe’s pharmaceutical industry body <a href="http://www.alltrials.net/2013/5-things-you-can-do-today-to-help-us-with-alltrials/">has threatened</a> a “series of lawsuits”. </p>
<p>The ethical issue here has more to do with health professionals, journals, and governments; pharmaceutical companies need to make a profit. Nonetheless, GlaxoSmithKline has signed the AllTrials petition.</p>
<h2>The way forward</h2>
<p>Ultimately, people who practice and teach medicine, surgery, pharmacy, and other health-care professions, as well as regulatory and government all rely on good evidence to make the best decisions for our collective health. </p>
<p>As the chief editors of prestigious medical journals are now telling us, the current system is broken and unable to deliver evidence-based medicine. The implications of this are profoundly negative for both our health and financial well-being. </p>
<p>The <a href="http://www.alltrials.net/supporters-list/">AllTrials petition</a> has garnered impressive support in just under a year. But more international support is needed for its aims to be fully implemented. A healthier, better-informed future depends on its success. </p>
<p><em>This article has been amended. In its previous version, it stated that LEO Pharma and GlaxoSmithKline had signed the AllTrials petition, but the text has been corrected to reflect that only GlaxoSmithKline has signed.</em> </p>
<p><em>According to the <a href="http://www.alltrials.net/2013/another-pharma-company-commits-to-greater-transparency/">AllTrials website</a>, LEO Pharma has committed to greater transparency.</em></p><img src="https://counter.theconversation.com/content/19755/count.gif" alt="The Conversation" width="1" height="1" />
<p class="fine-print"><em><span>Peter Parry is affiliated with Healthy Scepticism (<a href="http://www.healthyscepticism.org">www.healthyscepticism.org</a>).</span></em></p>An article published in the journal of the British Medical Association, BMJ, earlier this week illustrates a devastating problem with the “evidence base” in the academic medical literature. A large proportion…Peter Parry, Child and adolescent psychiatrist & senior lecturer, The University of QueenslandLicensed as Creative Commons – attribution, no derivatives.