Bad Actors Getting Your Health Data Is the FBI’s Latest Worry
In February 2015, the health insurer Anthem revealed that criminal hackers had gained access to the company's servers, exposing the personal information of nearly 79 million patients. It's the largest known healthcare breach in history.
FBI agents worry that the vast amounts of healthcare data being generated for precision medicine efforts could leave the U.S. vulnerable to cyber and biological attacks.
That year, the data of millions more would be compromised in one cyberattack after another on American insurers and other healthcare organizations. In fact, for the past several years, the number of reported data breaches has increased each year, from 199 in 2010 to 344 in 2017, according to a September 2018 analysis in the Journal of the American Medical Association.
The FBI's Edward You sees this as a worrying trend. He says hackers aren't just interested in your social security or credit card number. They're increasingly interested in stealing your medical information. Hackers can currently use this information to make fake identities, file fraudulent insurance claims, and order and sell expensive drugs and medical equipment. But beyond that, a new kind of cybersecurity threat is around the corner.
Mr. You and others worry that the vast amounts of healthcare data being generated for precision medicine efforts could leave the U.S. vulnerable to cyber and biological attacks. In the wrong hands, this data could be used to exploit or extort an individual, discriminate against certain groups of people, make targeted bioweapons, or give another country an economic advantage.
Precision medicine, of course, is the idea that medical treatments can be tailored to individuals based on their genetics, environment, lifestyle or other traits. But to do that requires collecting and analyzing huge quantities of health data from diverse populations. One research effort, called All of Us, launched by the U.S. National Institutes of Health last year, aims to collect genomic and other healthcare data from one million participants with the goal of advancing personalized medical care.
Other initiatives are underway by academic institutions and healthcare organizations. Electronic medical records, genetic tests, wearable health trackers, mobile apps, and social media are all sources of valuable healthcare data that a bad actor could potentially use to learn more about an individual or group of people.
"When you aggregate all of that data together, that becomes a very powerful profile of who you are," Mr. You says.
A supervisory special agent in the biological countermeasures unit within the FBI's weapons of mass destruction directorate, it's Mr. You's job to imagine worst-case bioterror scenarios and figure out how to prevent and prepare for them.
That used to mean focusing on threats like anthrax, Ebola, and smallpox—pathogens that could be used to intentionally infect people—"basically the dangerous bugs," as he puts it. In recent years, advances in gene editing and synthetic biology have given rise to fears that rogue, or even well-intentioned, scientists could create a virulent virus that's intentionally, or unintentionally, released outside the lab.
"If a foreign source, especially a criminal one, has your biological information, then they might have some particular insights into what your future medical needs might be and exploit that."
While Mr. You is still tracking those threats, he's been traveling around the country talking to scientists, lawyers, software engineers, cyber security professionals, government officials and CEOs about new security threats—those posed by genetic and other biological data.
Emerging threats
Mr. You says one possible situation he can imagine is the potential for nefarious actors to use an individual's sensitive medical information to extort or blackmail that person.
"If a foreign source, especially a criminal one, has your biological information, then they might have some particular insights into what your future medical needs might be and exploit that," he says. For instance, "what happens if you have a singular medical condition and an outside entity says they have a treatment for your condition?" You could get talked into paying a huge sum of money for a treatment that ends up being bogus.
Or what if hackers got a hold of a politician or high-profile CEO's health records? Say that person had a disease-causing genetic mutation that could affect their ability to carry out their job in the future and hackers threatened to expose that information. These scenarios may seem far-fetched, but Mr. You thinks they're becoming increasingly plausible.
On a wider scale, Kavita Berger, a scientist at Gryphon Scientific, a Washington, D.C.-area life sciences consulting firm, worries that data from different populations could be used to discriminate against certain groups of people, like minorities and immigrants.
For instance, the advocacy group Human Rights Watch in 2017 flagged a concerning trend in China's Xinjiang territory, a region with a history of government repression. Police there had purchased 12 DNA sequencers and were collecting and cataloging DNA samples from people to build a national database.
"The concern is that this particular province has a huge population of the Muslim minority in China," Ms. Berger says. "Now they have a really huge database of genetic sequences. You have to ask, why does a police station need 12 next-generation sequencers?"
Also alarming is the potential that large amounts of data from different groups of people could lead to customized bioweapons if that data ends up in the wrong hands.
Eleonore Pauwels, a research fellow on emerging cybertechnologies at United Nations University's Centre for Policy Research, says new insights gained from genomic and other data will give scientists a better understanding of how diseases occur and why certain people are more susceptible to certain diseases.
"As you get more and more knowledge about the genomic picture and how the microbiome and the immune system of different populations function, you could get a much deeper understanding about how you could target different populations for treatment but also how you could eventually target them with different forms of bioagents," Ms. Pauwels says.
Economic competitiveness
Another reason hackers might want to gain access to large genomic and other healthcare datasets is to give their country a leg up economically. Many large cyber-attacks on U.S. healthcare organizations have been tied to Chinese hacking groups.
"This is a biological space race and we just haven't woken up to the fact that we're in this race."
"It's becoming clear that China is increasingly interested in getting access to massive data sets that come from different countries," Ms. Pauwels says.
A year after U.S. President Barack Obama conceived of the Precision Medicine Initiative in 2015—later renamed All of Us—China followed suit, announcing the launch of a 15-year, $9 billion precision health effort aimed at turning China into a global leader in genomics.
Chinese genomics companies, too, are expanding their reach outside of Asia. One company, WuXi NextCODE, which has offices in Shanghai, Reykjavik, and Cambridge, Massachusetts, has built an extensive library of genomes from the U.S., China and Iceland, and is now setting its sights on Ireland.
Another Chinese company, BGI, has partnered with Children's Hospital of Philadelphia and Sinai Health System in Toronto, and also formed a collaboration with the Smithsonian Institute to sequence all species on the planet. BGI has built its own advanced genomic sequencing machines to compete with U.S.-based Illumina.
Mr. You says having access to all this data could lead to major breakthroughs in healthcare, such as new blockbuster drugs. "Whoever has the largest, most diverse dataset is truly going to win the day and come up with something very profitable," he says.
Some direct-to-consumer genetic testing companies with offices in the U.S., like Dante Labs, also use BGI to process customers' DNA.
Experts worry that China could race ahead the U.S. in precision medicine because of Chinese laws governing data sharing. Currently, China prohibits the exportation of genetic data without explicit permission from the government. Mr. You says this creates an asymmetry in data sharing between the U.S. and China.
"This is a biological space race and we just haven't woken up to the fact that we're in this race," he said in January at an American Society for Microbiology conference in Washington, D.C. "We don't have access to their data. There is absolutely no reciprocity."
Protecting your data
While Mr. You has been stressing the importance of data security to anyone who will listen, the National Academies of Sciences, Engineering, and Medicine, which makes scientific and policy recommendations on issues of national importance, has commissioned a study on "safeguarding the bioeconomy."
In the meantime, Ms. Berger says organizations that deal with people's health data should assess their security risks and identify potential vulnerabilities in their systems.
As for what individuals can do to protect themselves, she urges people to think about the different ways they're sharing healthcare data—such as via mobile health apps and wearables.
"Ask yourself, what's the benefit of sharing this? What are the potential consequences of sharing this?" she says.
Mr. You also cautions people to think twice before taking consumer DNA tests. They may seem harmless, he says, but at the end of the day, most people don't know where their genetic information is going. "If your genetic sequence is taken, once it's gone, it's gone. There's nothing you can do about it."
Ethan Lindenberger, the Ohio teenager who sought out vaccinations after he was denied them as a child, recently testified before Congress about why his parents became anti-vaxxers. The trouble, he believes, stems from the pervasiveness of misinformation online.
There is evidence that 'educating' people with facts about the benefits of vaccination may not be effective.
"For my mother, her love and affection and care as a parent was used to push an agenda to create a false distress," he told the Senate Committee. His mother read posts on social media saying vaccines are dangerous, and that was enough to persuade her against them.
His story is an example of how widespread and harmful the current discourse on vaccinations is—and more importantly—how traditional strategies to convince people about the merits of vaccination have largely failed.
As responsible members of society, all of us have implicitly signed on to what ethicists call the "Social Contract" -- we agree to abide by certain moral and political rules of behavior. This is what our societal values, norms, and often governments are based upon. However, with the unprecedented rise of social media, alternative facts, and fake news, it is evident that our understanding—and application—of the social contract must also evolve.
Nowhere is this breakdown of societal norms more visible than in the failure to contain the spread of vaccine-preventable diseases like measles. What started off as unexplained episodes in New York City last October, mostly in communities that are under-vaccinated, has exploded into a national epidemic: 880 cases of measles across 24 states in 2019, according to the CDC (as of May 17, 2019). In fact, the Unites States is only eight months away from losing its "measles free" status, joining Venezuela as the second country out of North and South America with that status.
The U.S. is not the only country facing this growing problem. Such constant and perilous reemergence of measles and other vaccine-preventable diseases in various parts of the world raises doubts about the efficacy of current vaccination policies. In addition to the loss of valuable life, these outbreaks lead to loss of millions of dollars in unnecessary expenditure of scarce healthcare resources. While we may be living through an age of information, we are also navigating an era whose hallmark is a massive onslaught on truth.
There is ample evidence on how these outbreaks start: low-vaccination rates. At the same time, there is evidence that 'educating' people with facts about the benefits of vaccination may not be effective. Indeed, human reasoning has a limit, and facts alone rarely change a person's opinion. In a fascinating report by researchers from the University of Pennsylvania, a small experiment revealed how "behavioral nudges" could inform policy decisions around vaccination.
In the reported experiment, the vaccination rate for employees of a company increased by 1.5 percent when they were prompted to name the date when they planned to get their flu shot. In the same experiment, when employees were prompted to name both a date and a time for their planned flu shot, vaccination rate increased by 4 percent.
A randomized trial revealed the subtle power of "announcements" – direct, brief, assertive statements by physicians that assumed parents were ready to vaccinate their children.
This experiment is a part of an emerging field of behavioral economics—a scientific undertaking that uses insights from psychology to understand human decision-making. The field was born from a humbling realization that humans probably do not possess an unlimited capacity for processing information. Work in this field could inform how we can formulate vaccination policy that is effective, conserves healthcare resources, and is applicable to current societal norms.
Take, for instance, the case of Human Papilloma Virus (HPV) that can cause several types of cancers in both men and women. Research into the quality of physician communication has repeatedly revealed how lukewarm recommendations for HPV vaccination by primary care physicians likely contributes to under-immunization of eligible adolescents and can cause confusion for parents.
A randomized trial revealed the subtle power of "announcements" – direct, brief, assertive statements by physicians that assumed parents were ready to vaccinate their children. These announcements increased vaccination rates by 5.4 percent. Lengthy, open-ended dialogues demonstrated no benefit in vaccination rates. It seems that uncertainty from the physician translates to unwillingness from a parent.
Choice architecture is another compelling concept. The premise is simple: We hardly make any of our decisions in vacuum; the environment in which these decisions are made has an influence. If health systems were designed with these insights in mind, people would be more likely to make better choices—without being forced.
This theory, proposed by Richard Thaler, who won the 2017 Nobel Prize in Economics, was put to the test by physicians at the University of Pennsylvania. In their study, flu vaccination rates at primary care practices increased by 9.5 percent all because the staff implemented "active choice intervention" in their electronic health records—a prompt that nudged doctors and nurses to ask patients if they'd gotten the vaccine yet. This study illustrated how an intervention as simple as a reminder can save lives.
To be sure, some bioethicists do worry about implementing these policies. Are behavioral nudges akin to increased scrutiny or a burden for the disadvantaged? For example, would incentives to quit smoking unfairly target the poor, who are more likely to receive criticism for bad choices?
The measles outbreak is a sober reminder of how devastating it can be when the social contract breaks down.
While this is a valid concern, behavioral economics offers one of the only ethical solutions to increasing vaccination rates by addressing the most critical—and often legal—challenge to universal vaccinations: mandates. Choice architecture and other interventions encourage and inform a choice, allowing an individual to retain his or her right to refuse unwanted treatment. This distinction is especially important, as evidence suggests that people who refuse vaccinations often do so as a result of cognitive biases – systematic errors in thinking resulting from emotional attachment or a lack of information.
For instance, people are prone to "confirmation bias," or a tendency to selectively believe in information that confirms their preexisting theories, rather than the available evidence. At the same time, people do not like mandates. In such situations, choice architecture provides a useful option: people are nudged to make the right choice via the design of health delivery systems, without needing policies that rely on force.
The measles outbreak is a sober reminder of how devastating it can be when the social contract breaks down and people fall prey to misinformation. But all is not lost. As we fight a larger societal battle against alternative facts, we now have another option in the trenches to subtly encourage people to make better choices.
Using insights from research in decision-making, we can all contribute meaningfully in controversial conversations with family, friends, neighbors, colleagues, and our representatives — and push for policies that protect those we care about. A little more than a hundred years ago, thousands of lives were routinely lost to preventive illnesses. We've come too far to let ignorance destroy us now.
New Tech Can Predict Breast Cancer Years in Advance
Every two minutes, a woman is diagnosed with breast cancer. The question is, can those at high risk be identified early enough to survive?
New AI software has predicted risk equally well in both white and black women for the first time.
The current standard practice in medicine is not exactly precise. It relies on age, family history of cancer, and breast density, among other factors, to determine risk. But these factors do not always tell the whole story, leaving many women to slip through the cracks. In addition, a racial gap persists in breast cancer treatment and survival. African-American women are 42 percent more likely to die from the disease despite relatively equal rates of diagnosis.
But now those grim statistics could be changing. A team of researchers from MIT's Computer Science and Artificial Intelligence Laboratory have developed a deep learning model that can more accurately predict a patient's breast cancer risk compared to established clinical guidelines – and it has predicted risk equally well in both white and black women for the first time.
The Lowdown
Study results published in Radiology described how the AI software read mammogram images from more than 60,000 patients at Massachusetts General Hospital to identify subtle differences in breast tissue that pointed to potential risk factors, even in their earliest stages. The team accessed the patients' actual diagnoses and determined that the AI model was able to correctly place 31 percent of all cancer patients in the highest-risk category of developing breast cancer within five years of the examination, compared to just 18 percent for existing models.
"Each image has hundreds of thousands of pixels identifying something that may not necessarily be detected by the human eye," said MIT professor Regina Barzilay, one of the study's lead authors. "We all have limited visual capacities so it seems some machines trained on hundreds of thousands of images with a known outcome can capture correlations the human eye might not notice."
Barzilay, a breast cancer survivor herself, had abnormal tissue patterns on mammograms in 2012 and 2013, but wasn't diagnosed until after a 2014 image reading, illustrating the limitations of human processing alone.
MIT professor Regina Barzilay, a lead author on the new study and a breast cancer survivor herself.
(Courtesy MIT)
Next up: The MIT team is looking at training the model to detect other cancers and health risks. Barzilay recalls how a cardiologist told her during a conference that women with heart diseases had a different pattern of calcification on their mammograms, demonstrating how already existing images can be used to extract other pieces of information about a person's health status.
Integration of the AI model in standard care could help doctors better tailor screening and prevention programs based on actual instead of perceived risk. Patients who might register as higher risk by current guidelines could be identified as lower risk, helping resolve conflicting opinions about how early and how often women should receive mammograms.
Open Questions: While the results were promising, it's unknown how well the model will work on a larger scale, as the study looked at data from just one institution and used mammograms supplied by just one hospital. Some risk factor information was also unavailable for certain patients during the study, leaving researchers unable to fully compare the AI model's performance to that of the traditional standard.
One incentive to wider implementation and study, however, is the bonus that no new hardware is required to use the AI model. With other institutions now showing interest, this software could lead to earlier routine detection and treatment of breast cancer — resulting in more lives saved.