Algorithms Can Treat Physician Fallibility
I’ve learned by assembling the following information that the politically touted “best health care in the world” at primary and acute care levels is outdated and at least 10 to 20 years behind the times.
And that’s a conservative estimate, based on my 2015 experience with the Desert Orthopedic Center (DOC) at the Eisenhower Medical Center in Rancho Mirage, California. There, I underwent cervical spine fusion surgery where a 3-ring binder was eventually filled with the paper trail of my diagnosis and treatment. It bulged with paper and followed me into what was touted as state-of-the-art surgery. Even in the medical profession, the three-inch thick paper file of a patient’s history symbolizes a healthcare care information system of 50 years ago. Sharing data at DOC was sharing the binder.
Out-of-date is also the impression I get at most encounters with the Whatcom and Washington healthcare systems. Paper files are largely gone, but digital files are not interchangeable, because each practice was allowed to use its own standards instead of adopting a universal standard that guaranteed compatibility across organizations and systems. Compounding the problem, funding ran out before work was completed on all systems.
It would be good news for patients to learn that U.S. healthcare is indeed moving towards evidence-based medicine, if only today’s evidence didn’t show how slow and cumbersome the move has been. The information technology evolution in healthcare remains in the last century despite a $30 billion federal infusion* in the 21st century.
Although there are other philosophies of medicine, the one most suitable for algorithms, big data and the Law of Large Numbers is evidence-based medicine defined by the “Stedman’s Medical Dictionary for the Health Professions and Nursing” as:
“Good doctors use both individual and clinical expertise and the best available external evidence, and neither alone is enough. Without current best evidence, practice risks becoming rapidly out of date, to the detriment of patients.”
The following Jan. 13, 1996, excerpt from the British Medical Journal article, “Evidence-Based Medicine: What It Is and What It Isn’t,” provides a still-valid 20-year old definition. Evidence-based medicine is:
“The process of applying relevant information derived from peer reviewed medical literature to address a specific clinical problem; the application of simple rules of science and common sense to determine the validity of the information; and the application of information to the clinical problem.”
Since 1996, the Internet, personal computers and wearable devices such as Fitbit have created mountains of data that remain unexploited, data that could provide a clear and swift means of achieving evidence-based medical practice, the current mantra of healthcare reform. Aside from the volume of data and the cost of processing it, societal issues have obstructed accessing the data for evidence-based use: It’s a genuine conundrum. Americans’ obsession with privacy in general and healthcare information privacy in particular creates resistance to sharing health information. Even though the identity of the patient is of no value in processing the data, the public fears data mishandling may reveal an individual’s identity and all of his or her health information.
Safeguarding Medical Information
The elephant in the room, for me, has always been the unfortunate implementation of the Health Insurance Portability and Accountability Act of 1996 (HIPAA). The act purported to provide data privacy and security for safeguarding medical information. Sadly, its passage coincided with primary care’s adoption of the evidenced-based medicine model, the one ultimately at odds with the other. I’ve always thought HIPAA is the great eclipse blocking the rising sun of abundant healthcare data.
Ironically, the HIPAA rule intended to protect patients with existing conditions from being discriminated against by health insurers has been turned into a rule that guarantees that insurers receive your health information to the exclusion of others’ unless named individually by you in a signed release.
HIPAA and self-interested data hoarders such as healthcare insurers and electronic health recordkeeping software corporations have created an essentially closed market, starving evidence-based medicine of the big data necessary to fully implement comprehensive, data-based medical care better able to prevent disruptive healthcare applications. Such situations are discussed by author and healthcare entrepreneur Jonathan Bush in his 2014 book, “Where Does It Hurt? An Entrepreneur’s Guide To Fixing Healthcare.”
At the risk of being accused of being a conspiracy theorist, I find health insurers are among the potential profiteers of huge quantities of valuable healthcare data. Along with health insurers there are the electronic health records systems providers, which acquire massive amounts of healthcare data they covet as their property for future data mining. You have to ask yourself, as many have started to, “Who owns my healthcare data?”
Data Ownership Gold Mine
In Washington state, the ownership of my healthcare data remains undetermined, and there’s still a chance for patients to acquire ownership. Some other states say explicitly who owns it. Within healthcare, there is a contest between Electronic Healthcare Record providers, hospitals and insurers over medical records control. See: “Who Owns Medical Records: 50 State Comparison” at www.healthinfolaw.org/comparative. (New Hampshire is the only state that has ruled healthcare information is the property of the patient.)
So, what’s the rush about fully implementing data-based medicine after 20 years? With the benefit of hindsight, consider the plight of clinicians addressed in the British Medical Journal of January 1996, which is still with us today:
“The difficulties that clinicians face in keeping abreast of all the medical advances reported in primary journals are obvious from a comparison of the time required for reading (for general medicine, enough to examine 19 articles per day, 365 days per year) with the time available (well under an hour a week by British medical consultants, even on self reports.)”
This circumstance can only be worse in 2017:
Because of this disparity between the quantity of information and the physician’s time to access it, there are “… ongoing challenges with evidence-based medicine [in] that some healthcare providers do not follow the evidence. This happens partly because the current balance of evidence for and against treatments shifts constantly, and it is impossible to learn about every change. Even when the evidence is unequivocally against a treatment, it usually takes ten years for other treatments to be adopted. In other cases, significant change can require a generation of physicians to retire or die, and be replaced by physicians who were trained with more recent evidence.” Application of Evidence in Clinical Settings, Wikipedia, April 2017.
In the Public Health Reports journal of March-April 2015, the authors of “Big Data and Public Health: Navigating Privacy Laws to Maximize Potential” provide an example of how big BIG data can be:
‘The University of Pittsburgh is demonstrating the value of secondary uses (of big data) through Project Tycho, which put 88 million disease reports published since 1888 in the Centers for Disease Control and Prevention’s (CDC’s) Morbidity and Mortality Weekly Report, an open-access database.”
So, what does this mean to patients and physicians today? It means the same thing today as it meant 14 years ago as in the Journal of the American Medical Information Association in 2003:
“We believe that decision support delivered using information systems, ideally with the electronic medical record as the platform, will finally provide decision makers with tools making it possible to achieve large gains in performance, narrow gaps between knowledge and practice, and improve safety. These reviews have summarized the evidence that computerized decision support (Ed. algorithms) works, in part, based on evidence domain.” “Ten Commandments for Effective Clinical Decision Support: Making the Practice of Evidence-based Medicine a Reality” by Blackford Middleton, M.D,. M.Sc., M.P.H.
We’re talking here again about algorithms, (as first discussed in Part 1, May 2017 Whatcom Watch) the heart of evidence-based medicine.
Scientific algorithms yield good results, because they obey their rules and yield more consistent results. A popular algorithm reference service used by physicians states:
“… (an algorithm) facilitates diagnostic and therapeutic decision making for a wide range of common and often complex problems faced in outpatient and inpatient medicine. Comprehensive algorithmic decision trees guide you through more than 250 disorders organized by sign, symptom, problem, or laboratory abnormality. The brief text accompanying each algorithm explains the key steps of the decision-making process, giving you the clear, clinical guidelines you need to successfully manage even your toughest cases.”
“Decision Making in Medicine: An Algorithmic Approach,” (Clinical Decision Making Series).
The outlook for evidence-based medicine and big data may not be as bleak as my findings portend. According to the March 31, 2017, Tableau Software website:
“… 2016 was a landmark year for big data with more organizations storing, processing, and extracting value from data of all forms and sizes. In 2017, systems that support large volumes of both structured and unstructured data will continue to rise. The market will demand platforms that help data custodians govern and secure big data while empowering end users to analyze that data. These systems will mature to operate well inside of enterprise IT systems and standards.”
Algorithms and other information technologies are means by which medical knowledge can be widely spread and put into the hands of non-doctors, such as nurses, paramedics and first-responders. Despite algorithms, doctoring will still remain as much art as technology, and a doctor must get involved at some point.
Algorithms Deceptively Simple
Algorithms create decision trees that can be constructed around medical problems that lead to a limited number of diagnostic possibilities, as illustrated in the following flowchart from Medscape.com. One advantage of algorithms is that they enable people who are not doctors, such as physician’s assistants, practical nurses and nurse practitioners to manage a disease or condition.
In particular, algorithms unify diagnostic and therapeutic reasoning and enable colleagues to follow each other’s management, thereby minimizing inefficient, ineffective or even harmful variations in cure and care.
In its broadest form, evidence-based medicine is the application of the scientific method to healthcare decision-making. Physicians are the presumed conduits of this method to patients, and while it works pretty well some of the time, physicians are only human and prone to undisciplined thinking and faulty judgment. Throw the threat of malpractice litigation into the physician/patient dynamic and physicians change their procedures in multiple ways: ordering more tests, referring more cases to specialists, applying conventional treatments even when they are unlikely to help. These actions protect the physicians more than they benefit the patient, creating potential conflicts of interest.
Judgment and Decision Making
“Doctors, like all the rest of us humans, are vulnerable to cause-and-effect thinking and too often succumb to decisions from the fast thinking experience part of their brain instead of summoning the slow thinking remembering part of their brain where their medical education and inherent wisdom reside. In plain English — doctors are just human.” This notion of how we think is the result of research of Nobel laureate Daniel Kahneman, an Israeli-American psychologist notable for his work on the psychology of judgment and decision-making and author of “Thinking Fast and Slow.” In his book, Kahneman features (some might say picks on) medical doctors as examples of the consequences of thinking fast and slow — of using the experiencing mind versus the remembering mind — in making critical decisions.
Other sources paint a similar picture of the faults and foibles of American healthcare and its need to implement evidence-based healthcare and its companion facilitator, big data:
“[A] major cause of physicians and other healthcare providers treating patients in ways unsupported by the evidence is that these healthcare providers are subject to the same cognitive biases as all other humans. They may reject the evidence because they have a vivid memory of a rare but shocking outcome, such as a patient dying after refusing treatment. They may over treat to “do something” or to address a patient’s emotional needs. They may worry about malpractice charges based on a discrepancy between what the patient expects and what the evidence recommends. They may also over treat or provide ineffective treatments because the treatment feels biologically plausible.” Application of Evidence in Clinical Setting, Wikipedia, April 2017
Of course, algorithms and computerization of healthcare create fear in the minds of many patients, but if you look deeply into what’s going on in healthcare today, as I have attempted to do, you would see that computerization, IT, big data, algorithms and evidence-based medicine are the only way to go. Everything about healthcare has gotten too big and too complicated to manage without these 21st century tools.
Problems Plague Medical Computers
Opponents of computerization are not without their justifications, as I have witnessed first hand. I can’t count all the doctors who can’t use their computers, even to take consultation notes. For consultation notes some doctors employ scribes who sit it on the consultation to take notes. Other doctors must call in a young receptionist to find and display a critical MRI image before consulting with a patient. Because the hematologist who administered my late wife’s chemotherapy was unable to display her monthly MRI brain images, we had to see a second physician, a friendly radiologist to learn exactly what the most recent images meant.
One expert explains that computers create problems because they are brought into a fragmented information technology environment without proper training, and where interfaces between equipment and systems are poor or non-existent.
Many fear bringing big data, science and computer technology together in healthcare will make healthcare unacceptably impersonal. I maintain that healthcare is already highly impersonal, and that adopting technology will allow physicians to customize healthcare and refocus on the art of healing.
Patients need to decide how much science and how much art they want in their healthcare. As a patient myself, when it comes to diagnosis and treatment, I want 100 percent science. But in the application of this science, I want the arts of compassion, empathy and good bedside manner that make scientific medicine a humane experience.
*”The Digital Doctor: Hope, Hype and Harm at the Dawn of Medicine’s Computer Age,” Robert M. Wachter, McGraw-Hill Education Books, 2015.
Robert A. Duke is author of “Waking Up Dying: Caregiving When There Is No Tomorrow,” he lives in Bellingham. His email: firstname.lastname@example.org