{"id":43,"date":"2017-09-22T22:58:33","date_gmt":"2017-09-22T22:58:33","guid":{"rendered":"http:\/\/blogs.harvard.edu\/shailin\/?p=43"},"modified":"2017-09-22T22:58:33","modified_gmt":"2017-09-22T22:58:33","slug":"democratized-diagnostics-why-medical-artificial-intelligence-needs-vetting","status":"publish","type":"post","link":"https:\/\/archive.blogs.harvard.edu\/shailin\/2017\/09\/22\/democratized-diagnostics-why-medical-artificial-intelligence-needs-vetting\/","title":{"rendered":"Democratized Diagnostics: Why Medical Artificial Intelligence Needs Vetting"},"content":{"rendered":"<p><i>Originally published on September 22, 2017, on the Petrie-Flom Center for Health Law Policy, Biotechnology, and Bioethics\u00a0<a href=\"http:\/\/t.umblr.com\/redirect?z=http%3A%2F%2Fblogs.harvard.edu%2Fbillofhealth%2F2016%2F09%2F12%2Fhospitals-should-think-before-performing-searches-for-law-enforcement%2F&amp;t=ZjY4MDlhMGY4ZDRmYzE3YWJiZTQxM2Y0OWE4OWRlYjViMTc5MjZkOCxxbFEyOUZWcg%3D%3D&amp;b=t%3APU6INWmFXOv29eWadaAWOA&amp;p=http%3A%2F%2Fshailinthomas.tumblr.com%2Fpost%2F165231436317%2Fhospitals-should-think-before-performing-searches&amp;m=1\">Bill of Health<\/a>\u00a0blog.<\/i><\/p>\n<p style=\"text-align: justify\">Pancreatic cancer is one of the deadliest illnesses out there.\u00a0 The five-year survival rate of patients with the disease is\u00a0<a href=\"http:\/\/pancreatic.org\/pancreatic-cancer\/about-the-pancreas\/prognosis\/\">only about 7%<\/a>.\u00a0 This is, in part, because few observable symptoms appear early enough for effective treatment.\u00a0 As a result, by the time many patients are diagnosed the prognosis is poor.\u00a0 There is an app, however, that is attempting to change that.\u00a0\u00a0<a href=\"https:\/\/ubicomplab.cs.washington.edu\/publications\/biliscreen\/\">BiliScreen<\/a>\u00a0was developed by researchers at the University of Washington, and it is designed to help users identify pancreatic cancer early with an algorithm that analyzes selfies.\u00a0 Users take photos of themselves, and the app\u2019s artificially intelligent\u00a0<a href=\"https:\/\/ubicomplab.cs.washington.edu\/publications\/biliscreen\/\">algorithm detects slight discolorations<\/a>\u00a0in the skin and eyes associated with early pancreatic cancer.<\/p>\n<p style=\"text-align: justify\">Diagnostic apps like BiliScreen represent a huge step forward for preventive health care.\u00a0 Imagine a world in which the vast majority of chronic diseases are caught early because each of us has the power to screen ourselves on a regular basis.\u00a0 One of the\u00a0<a href=\"https:\/\/www.cdc.gov\/healthcommunication\/toolstemplates\/entertainmented\/tips\/preventivehealth.html\">big challenges<\/a>\u00a0for the modern primary care physician is convincing patients to get screened regularly for diseases that have relatively good prognoses when caught early.<\/p>\n<p style=\"text-align: justify\">I\u2019ve written before about the\u00a0<a href=\"http:\/\/blogs.harvard.edu\/billofhealth\/2017\/01\/26\/artificial-intelligence-medical-malpractice-and-the-end-of-defensive-medicine\/\">possible impacts of artificial intelligence and algorithmic medicine<\/a>, arguing that both medicine and law will have to adapt as machine-learning algorithms surpass physicians in their ability to diagnose and treat disease.\u00a0 These pieces, however, primarily consider artificially intelligent algorithms licensed to and used by medical professionals in hospital or outpatient settings.\u00a0 They are about the relationship between a doctor and the sophisticated tools in her diagnostic toolbox \u2014 and about how relying on algorithms could decrease the pressure physicians feel to order unnecessary tests and procedures to avoid malpractice liability.\u00a0 There was an underlying assumption that these algorithms had already been evaluated and approved for use by the physician\u2019s institution, and that the physician had experience using them.\u00a0 BiliScreen does not fit this mold \u2014 the algorithm is not a piece of medical equipment used by hospitals, but rather part of an app that could be downloaded and used by anyone with a smartphone.\u00a0 Accordingly, apps like BiliScreen fall into a category of \u201cdemocratized\u201d diagnostic algorithms. While this democratization has the potential to drastically improve preventive care, it also has the potential to undermine the financial sustainability of the U.S. health care system.<\/p>\n<p style=\"text-align: justify\">Democratized diagnostic algorithms should be a source of financial concern for the health care system because of the malpractice risks they will create for physicians who disagree with them. A study published in\u00a0<a href=\"http:\/\/jamanetwork.com\/journals\/jamaoncology\/fullarticle\/2108852\">JAMA Oncology in 2015<\/a>\u00a0found that patients demand specific medical interventions in 8.7% of encounters. \u00a0As democratized diagnostic apps proliferate in app stores, patients with \u201cdiagnoses\u201d may begin to appear in doctors\u2019 offices \u2014 smartphones in hand \u2014 demanding tests and procedures more frequently. These apps will likely be wrapped in legal language disclaiming all diagnostic accuracy \u2014 telling users to consult a physician for an actual evaluation \u2014 but a jury may find a \u201cdiagnosis\u201d from such an app sufficient to establish that a physician who failed to pursue further testing or treatment was negligent.\u00a0 Thus, if a patient presents with a cancer warning from an app, a physician may feel obligated to run a barrage of tests to confirm or refute the algorithm\u2019s determination \u2014 even if the patient exhibits none of the symptoms typically associated with the diagnosis.<\/p>\n<p style=\"text-align: justify\">The potential influx of patients will be exacerbated by the liability concerns faced by the apps themselves.\u00a0 It\u2019s a common joke that patient information websites like\u00a0<a href=\"http:\/\/www.webmd.com\/\">WebMD<\/a>\u00a0will identify any set of symptoms as a potential indicator of a terminal illness.\u00a0\u00a0 But this makes sense given the liability landscape.\u00a0 WebMD does not want to risk discouraging patients from seeking medical attention \u2014 or pursuing more traditional preventive screenings \u2014 because of assurances received on the website.\u00a0 Diagnostic apps will want to avoid similar liability, and as a result they may be designed to either over-diagnose or tell patients to consult a physician regardless of the algorithm\u2019s analysis.<\/p>\n<p style=\"text-align: justify\">The systemic harm of this perfect storm will be its impact on defensive medicine.\u00a0 As previously mentioned, defensive medicine is when a physician orders more diagnostic tests and procedures than a patient\u2019s condition warrants, and it puts an immense burden on the U.S. health care system.\u00a0 In a 2013 survey of private-sector physicians,\u00a0<a href=\"https:\/\/www.forbes.com\/sites\/realspin\/2013\/08\/27\/defensive-medicine-a-cure-worse-than-the-disease\/#72c295f37c95\">75 percent<\/a>\u00a0admitted to using defensive medicine to avoid lawsuits, resulting in an estimated\u00a0<a href=\"https:\/\/www.forbes.com\/sites\/realspin\/2013\/08\/27\/defensive-medicine-a-cure-worse-than-the-disease\/#72c295f37c95\">$650 billion<\/a>\u00a0spent annually on unnecessary care.\u00a0 As democratized diagnostic apps proliferate, it is easy to see how defensive medicine will increase commensurately.\u00a0 Regardless of a patient\u2019s symptoms, no physician will want to disregard an app diagnosis only find herself in front of a jury trying to explain why she ignored the algorithm\u2019s warning.<\/p>\n<p style=\"text-align: justify\">The underlying issue here is fundamentally one of quality control.\u00a0 If these artificially intelligent algorithms were universally superior to physicians in their diagnostic capabilities, they could decrease defensive medicine by limiting expensive testing to cases where the likelihood of a positive finding is very high.\u00a0 However, when any coder can spin up a \u201cdiagnostic\u201d algorithm and throw it on an app store, it will be difficult for physicians to determine which apps to trust and which to ignore.\u00a0 Some of these apps, including BiliScreen, may be\u00a0<a href=\"http:\/\/www.nature.com\/nature\/journal\/v542\/n7639\/full\/nature21056.html?foxtrotcallback=true\">able to diagnose diseases<\/a>\u00a0earlier and more accurately than physicians with decades of experience, but it\u2019s unlikely that all medical apps available from an app store for $0.99 will be so good. \u00a0Indeed\u00a0<a href=\"https:\/\/www.ncbi.nlm.nih.gov\/pmc\/articles\/PMC4286553\/\">a 2014 study<\/a>\u00a0published in Translational Behavioral Medicine reported an \u201cenormous range of quality among [mobile health] apps.\u201d While accuracy rate could be a helpful metric, physicians will still face the difficulty of determining how low the accuracy rate must be for an app\u2019s diagnosis to be safely ignored.\u00a0 Would a jury find that disregarding an algorithm with 75% accuracy constitutes negligence?\u00a0 What about 60% or 45% accuracy?\u00a0 This will be almost impossible for any individual physician to predict.\u00a0 As a result, medical providers may feel pressure to order confirmatory tests in cases involving all but the least accurate apps.<\/p>\n<p style=\"text-align: justify\">One possible solution is vetting of diagnostic algorithms by a trusted third party. Having a curated set of accurate democratized diagnostic apps would serve two functions critical for reaping the benefits of these innovative technologies while avoiding their pitfalls.\u00a0 First, it would give physicians \u2014 and juries \u2014 guidelines for determining which algorithms warrant diagnostic deference.\u00a0 A physician could point to a poor rating by a vetting organization as a justification for her decision not to pursue additional testing for a patient with no clinically relevant symptoms.\u00a0 Second, a curated set of reliable apps would help guide patients to the most accurate diagnostic algorithms available, decreasing the likelihood of erroneous diagnoses and expensive physician-ordered tests to refute them.\u00a0 This quality control would maximize the ability of democratized diagnostic apps to serve both individual patients and the health care system as a whole \u2014 accurately identifying diseases early and decreasing the prevalence of defensive medicine by taking some diagnostic liability away from health care providers.<\/p>\n<p style=\"text-align: justify\">There are many ways this vetting could be accomplished.\u00a0 It could be done by a non-profit or university that reviews the evidence supporting diagnostic algorithms and rates them according to their accuracy, or it could be taken on by the platforms purveying these apps to users, such as the iOS App Store or the Google Play store.\u00a0 These tech companies already serve as gatekeepers,\u00a0<a href=\"https:\/\/developer.apple.com\/app-store\/review\/guidelines\/\">vetting both code and content<\/a>\u00a0before an app will be offered for download, so it would not be outlandish for them to require a certain degree of diagnostic accuracy prior to allowing a medical app on their platforms.<\/p>\n<p style=\"text-align: justify\">The promise of democratized diagnostic algorithms is immense \u2014 but as with all technological advancements, where there is promise there is also peril.\u00a0 If not vetted and implemented properly, each $0.99 app has the potential burden the health care system with additional unnecessary diagnostic testing.\u00a0 Empowering patients to conduct their own preventive screening could save countless lives \u2014 so long as it doesn\u2019t cause the already bloated health care system to collapse under its own weight.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Originally published on September 22, 2017, on the Petrie-Flom Center for Health Law Policy, Biotechnology, and Bioethics\u00a0Bill of Health\u00a0blog. Pancreatic cancer is one of the deadliest illnesses out there.\u00a0 The five-year survival rate of patients with the disease is\u00a0only about 7%.\u00a0 This is, in part, because few observable symptoms appear early enough for effective treatment.\u00a0 [&hellip;]<\/p>\n","protected":false},"author":8887,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-43","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"jetpack_featured_media_url":"","_links":{"self":[{"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/posts\/43","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/users\/8887"}],"replies":[{"embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/comments?post=43"}],"version-history":[{"count":1,"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/posts\/43\/revisions"}],"predecessor-version":[{"id":44,"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/posts\/43\/revisions\/44"}],"wp:attachment":[{"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/media?parent=43"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/categories?post=43"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/archive.blogs.harvard.edu\/shailin\/wp-json\/wp\/v2\/tags?post=43"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}