In December last year, CVS offered to buy Amgen Insurance Company for US $ 69 billion. In January this year, another three giants – Amazon, JPMorgan Chase and Berkshire Hathaway – said they would set up a joint venture to reduce the medical costs of a total of about 1 million employees and improve the results. In March, Sinosure said it would invest more than $ 50 billion to acquire Express Scripts, a medical and welfare administration.
Why are medical transactions so busy these days? At first glance, you may think that it is in pursuit of magnitude. This is often called ” scale” in management. But in fact there is a powerful catalyst, an infinite and infinitely small thing seems to be able to answer all the questions. The answer is data.
More specifically, it is your data: personal biological information, medical history, fluctuating health, and where you have been, consumption habits, sleep status, diet and excretion, etc. The data produced in daily life, laboratory test results, medical images, gene files, liquid biopsies, electrocardiograms … these are only a small part, but they already cover a large amount of data. In addition to medical claims, clinical experiments, prescriptions and academic research, the data generated every day is about 750 trillion bytes, accounting for about 30% of the world’s data.
Massive data has existed for a long time. Thanks to new technologies, advanced measuring instruments, ubiquitous connectivity and cloud technologies, and artificial intelligence, enterprises can finally make use of these data. ” Collecting data is only one aspect.” Eric Topper, director of Scripps Research Institute, said. ” But more important is analysis. Just three to five years ago, the data could only be idle and can now be analyzed and interpreted. This is the biggest change in the medical field. ”
Therefore, acquiring, analyzing and utilizing data has become a new gold rush. A group of technology giants and a group of hot start-up companies are actively testing water.
Verily, a biotech company under Alphabet, Google’s parent company, is tracking the biometric information provided by 10,000 volunteers and trying to establish a ” benchmark” for human health ( rumors say it is also interested in the health insurance industry ). Apple has just announced a new function on the iPhone that can immediately interface with several major medical systems and upload medical data. In addition, Apple is also conducting research in the heart field with Stanford University to test whether wearable devices can detect serious heart diseases.
According to the data of the medical insurance and Medicaid service center, reasonable utilization of the data can ultimately improve the health level of patients and reduce the medical cost, which is expected to increase by 5.3% in 2018 alone. Of course, this is a long-term ideal. But at the very least, it can drive potential related businesses.
David Flanders, managing director of BDO, pointed out that Facebook and Google, which enjoy huge amounts of data, make money through advertising, and he estimated the relevant businesses to be worth $ 200 billion. ” The medical industry is 15 times larger,” he said. ” Health care spending reached 3 trillion US dollars. In theory, if you go the right way, you can build 15 Facebook and 15 Google. This is why competition is fierce. ”
This is why so many traditional medical companies, from hospitals to insurance companies to welfare management agencies to medical and instrument manufacturers, are eager to restructure and innovate, knowing that the volume of related industries can reach one fifth of the total economic volume. Industry restructuring not only means redrawing the map of the medical industry at the enterprise level, but also affects everyone.
Paul Doherty, Accenture’s chief technology and innovation officer, is optimistic. He estimates that due to the existence of ” information asymmetry”, patients with their own biological data will benefit from it and enjoy more dividends.
In order to better understand how to achieve a balance of power in the future and to further explore the current situation, Fortune magazine interviewed more than 30 senior executives involved in various segments of the medical industry, as well as entrepreneurs, doctors, patients and other experts. The following are all kinds of statements made by the big data revolution in promoting the pharmaceutical industry.
Data Pills: A New Paradigm for Patient Diagnosis
Around jacoby’s 8th birthday, Lin Desai Amos began to pay attention to his son’s abnormality. Jacoby is very lively. He often plays hockey and lacrosse, but suddenly he becomes lazy and always wants to go to the bathroom. After the doctor measured jacoby’s blood sugar, he informed his family to send the child to the emergency department. The children began to lose consciousness on the road.
Amos later learned that the blood sugar level in jacoby was as high as 735 mg / dl, while the health range was 70 – 140 mg / dl. Fortunately, jacoby’s illness did not develop into diabetic ketoacidosis, or DKA for short. DKA is a potentially fatal complication. As blood sugar level continues to rise, blood becomes acidic, leading to organ failure.
The Amos family lived in a suburb of Denver, but the doctor’s response after the terrible incident was extremely casual and unclear. Her family took a crash course in type 1 diabetes to learn how hyperglycemia and hypoglycemia can be life – threatening. They learned that to calculate jacoby’s sugar intake, they had to check their blood sugar several times a day, prick their fingers with a diabetes test strip, and then write it down in the diary.
Stress, training, insulin, various foods … Coupled with the interaction of various influencing factors, the blood sugar level is notoriously difficult to control. Keeping the blood sugar level in jacoby within a healthy range is exhausting and terrible. Amos calculated carefully everywhere, but the results were not good. jacoby’s blood sugar was like riding a roller coaster, sometimes soaring to extremely high levels ( he would feel tired ) and sometimes falling to extremely low levels ( he would feel dizzy ).
But the life and death of jacoby depends on these figures. Amos hopes to know the blood sugar level all the time. In the weeks after jacoby’s diagnosis, she tried her best to test her son’s blood sugar, which can be measured more than 20 times a day, far exceeding the number of times covered by insurance.
Millions of American families are hit by similar diseases every year. However, starting from 2015, the emerging smart phone technology has partially eased the concern. Dexcom, a California company, wirelessly connects a continuous blood glucose monitor ( which has been in existence for more than 10 years ) with a smart phone ( or smart watch ). Users can monitor, draw and share blood glucose data. The latest data can be obtained every five minutes. They are on call all day long and can give an alarm when blood glucose reaches a dangerous level.
Some experts believe that the online equipment will tell patients too much information at any time, but Amos said half jokingly about how he ” tracks” his son’s blood sugar level, saying that the equipment can be called a lifeguard. She said that jacoby can now read the third grade of primary school normally, so long as you pay attention to the blood sugar data on Apple Watch, you don’t have to run many trips to the school hospital one day. ( Amos is also watching on the iPhone at any time. ) And now jacoby can sleep all night without getting up every few hours to prick his finger to measure blood sugar.
The equipment provides more than peace of mind. The collected data can really help Amos and jacoby understand the diabetes situation and how to control it. They can find out which foods cause blood sugar to soar and when insulin injections can better control complex carbohydrates such as pizza. Yes, his diabetes has not been cured. But at least it can be predicted now that there are few accidents. Obviously, this technology has been fully integrated into their daily life. ” It’s no different from car seat belts and bicycle helmets.” Amos said.
This is just one of the ways that smart phones and networking devices change the relationship between patients and health data. It can also help patients improve their health in the process.
Digital diabetes prevention and treatment platforms such as Virta and Omada Health also connect communities and health coaches, who can remotely monitor weight, blood sugar, diet and medication. Now there are Proteus Digital Health edible sensors, which are worth filming in ” Black Mirror”. This sensor can help patients ( as well as doctors and family members ) to check whether to take medicine at any time.
Michael Andres, Chief Innovation and Digital Officer of Froedtert Hospital and Wisconsin Medical Network School, said that every tablet starts the application when it contacts gastric acid, and the sensor displays the whole process like a game. According to Froedtert test, 98.6% of hepatitis C patients took drugs on time with sensors during taking expensive drugs such as Gilead Harvoni. Taking medicine on time not only helps cure diseases, but also saves a lot of money, because taking medicine for one more month will cost tens of thousands of dollars more. This is why Froedtert ( not the patient ) pays for the sensor.
The popular Apple Watch or Android wearable health devices have become popular, reminding people of various health issues, from sleep apnea to hypertension, and even severe arrhythmia. People are paying more attention to their genomes, hoping to predict the risks of certain diseases ( such as cancer and Alzheimer’s disease ), some of which are temporarily impossible to realize and some of which are incomplete ( and controversial ).
Consumers began to use the very cheap and convenient gene detection kits of companies such as Colour Genomics and 23andMe, and the detection results will indicate a higher risk of suffering from certain diseases. Supporters believe that understanding risks can sometimes help people take corresponding preventive measures to reduce risks.
All of a sudden, family-like genetic testing became a necessity: on the recent Black Friday, 23andMe’s standard DNA testing sales ranked among the top five in Amazon, almost catching up with Amazon’s own smart speaker Echo Dot and multifunctional pressure cooker Instant Pot.
The selling point of all innovative technologies is simple: consumer – led.
Some large insurance companies have even found that involving patients in data can help improve results and control costs. This is also a bet made by Mark Bertau Lini, chief executive officer of Antai Insurance. Bertau Lini is trying to reach a cooperation with CVS. He believes that if enterprises can clearly inform consumers of the benefits they enjoy, consumers may even actively share data. ” We have made various rules to protect data,” Bertau Lini said. ” But if we tell the customer from another angle,’ If we can have information about you, it will be more convenient to serve,’ he or she will be willing to provide data. This is also why social media plays a big role. ”
With personalized data in hand, the company can work out a health plan together with the patients. Bertau Lini said: ” We will tell the patients that if we work out the plan together, we can exempt the co-payment amount and we don’t need to authorize any more, because the plan is worked out together.”
Command Center: Better Decision Making Through Data
Rodney M was picking out a birthday present for his wife when the pain hit. The 52 – year – old chief executive of a public relations company clutched his chest and struggled to sit on the chair, losing consciousness above his waist and feeling numb. He felt as if he had been hit by a truck.
Rodney had just been treated for cancer two years ago and suffered from a new serious disease – aortic dissection caused by tearing of a key artery that carries blood to the body.
He was lucky enough that an ambulance arrived in time to take him to Howard County General Hospital in Columbia City, Maryland. One in five patients with arterial dissection died before they were sent to the hospital. This was not finished. Rodney was flown to Johns Hopkins Hospital in Baltimore within a few minutes and survived after seven hours of complicated surgery.
At that time, he did not know that it was actually the artificial intelligence-driven data ” command center” of Hopkins Hospital that saved his life. ” 41 minutes. It only took 41 minutes from the alarm of the monitoring system to the plane taking off from the hospital. ” Chief Executive Jim Schulen said he was in charge of the expanding emergency department of the medical center.
The command center extracts information from more than a dozen data streams in real time, including medical records, emergency dispatch service updates, experimental test results, and how many hospital beds there are at a given time. Then, through manually improved algorithms, the system can instantly complete patient classification and arrange according to needs, such as arranging the surgical team in advance in Rodney’s case.
For hospitals, the economic benefits of data management are indisputable. ” Since the Hopkins Hospital Command Center was launched in 2016, the capacity to treat patients with complex cancers has increased by about 60%, the waiting time in the emergency room ( waiting for hospital beds ) has decreased by more than 25%, and the waiting time in the operating room has decreased by 60%,” said Jeff Terry, the designer of Hopkins Command Center and head of the general electric medical program.
Schulen of Hopkins Hospital said that after the application of the technology, the hospital would have 15 to 16 additional beds instead of actual investment. This year General Electric plans to announce the establishment of 10 new command centers covering 30 different hospitals. Terry said that in the next five years, the return on investment in medical centers could reach nearly 4: 1.
Electronic Medical Record ( EHR ) is also the latest breakthrough in recent big data. It may not be so eye – catching, but it is also innovative.
Electronic medical record is a rare technology that everyone hates in medical circle. Doctors complain that they waste too much time and cannot work smoothly with other medical record systems, and most of them cannot read. Since 2015, at least 593 academic papers and a rap video have complained about the technology, causing doctors to run out of energy. If big data can really revolutionize the industry, the opportunity may lie here. Electronic medical records should be transformed from time-consuming work into a viable research tool.
In Lakeland Health, a non-profit community health system in southwestern Michigan, big data did play a role. In 2012, Lakeland started to use electronic medical record system, but it basically consists of paper and pen. The nurse first recorded the vital signs of each patient in the chart, returned to the computer and manually re-entered the hospital’s electronic system. The transcription process took 15 to 20 minutes, and errors often occurred. In mid – 2016, the hospital will start a new process to upload the data automatically through the wristbands of patients, or the nurses will use the bedside of handheld devices to input data.
Arthur barac, Lakeland’s Chief Nursing Information Officer, said there was an obvious change from the beginning: nurses spent less time inputting data and more time caring for patients. But the more radical and surprising change is the decline of the ” blue alert”: the patient’s heart and breathing stops. The blue alert has dropped 56% since the new system developed by Philips, the Anglo – Dutch health giant, was launched in June 2016. Why? Part of the reason is that the alarm system is embedded with artificial intelligence technology, which can not only detect minor changes in vital signs, but also score risks according to the condition of the patient, thus facilitating nurses to give priority to patients with more critical conditions.
” It is important to have an honest understanding of the current situation.” Philip’s Chief Medical Officer for Strategy and Innovation Roy Smythe said. Many new digital and data tools focus not on nursing but on how to make nursing more efficient, intelligent and accurate.
Smace agrees with many experts interviewed by Fortune magazine, even those who are extremely interested in digital health. They all warned not to overuse new medical technologies.
” We are over – committed, and we haven’t really achieved much.” Brennan Spigel, a doctor and director of health service research at Sidas Sinai Medical Center, said. ” I think I am a technophile who doubts technology. However, too many people in Silicon Valley’s ” echo chamber” have never contacted patients and do not understand how difficult digital medicine is. ” Spigel is also a professor of medicine and public health at UCLA. He cited his own failure cases as evidence, including West Das Sinai’s failure to connect patients with electronic medical records through wearable devices such as Fitbit, Apple Watch and Withings in 2015. ” We failed to give patients appropriate information and did not sincerely invite them to participate in the project.” Potential participants have little reason to participate in the connection and project because the project has no clear value proposition. ” Digital health is not computer science or engineering science, but social science and behavioral science.”
Eric Topper of the Scripps Institute is also a famous cardiologist, and he also gave a similar warning. ” There are many promises, but most of them have not been fulfilled.” He said the reason was that there were various systematic obstacles. One of the challenges is the rigid and ” long – standing” medical institutions in the United States. Half of American doctors are over 50 years old, and they resist change, ” unless they get more compensation.”
Confronting ” Failure Problem”
According to Amgen, Big Data has already upended California’s biomedical research and development process and has greatly affected new drugs at the research and development stage. It started in 2011 when Sean Harper, then head of research and development, traveled to Iceland. His aim is to solve the ” failure problem” faced by the company, which is also a problem in the industry. In short, 90% of the new drugs developed cannot enter the market.
Developing new drugs is very expensive and inefficient. Enterprises often invest billions of dollars and spend many years trying to verify possible scientific hypotheses. Drug scientists are eager for sudden good luck in chemical experiments. In fact, they do not fully understand the biological complexity they are striving for, nor do they understand why some drugs have obvious effects on mice but no effect on humans.
For Harper, Iceland seems to be able to provide some unique medical data. The data are provided by the Icelandic government, including genetic sequencing information of 160,000 citizens, as well as medical treatment and genealogy. The data are stored and analyzed by deCode, a human genetic equipment company headquartered in Reykjavik. The company was established in 1996 and has been struggling to operate.
Despite the solvency problems, deCode has achieved fruitful results in the field of gene exploration. The large amount of data it has can be used to mine the population with genetic variation and to link the variation with the clinical results of diseases such as cancer and schizophrenia. With the improvement of computer processing capacity and the sharp drop in sequencing costs, Harper found that the company was an undervalued asset for drug research and development, and in 2012 Amgen took it into its pocket with US $ 415 million.
The acquisition completely changed the research and development process of Amgen. Before the acquisition, only 15% of Anjin’s candidates were verified for specific genetic targets. After the acquisition was completed, Amgen began to evaluate all candidate drugs using deCode’s database. The audit found some obviously ineffective drugs. Evidence shows that 5% of the candidates have no therapeutic effect. The management immediately shut down related projects ( including a highly anticipated coronary artery drug, which is about to enter the human test phase ) and gave priority to drugs with clear gene targets. After being confirmed by deCode gene database, Amgen has also passed more than 10 drugs.
Harper said that three-quarters of Amgen’s research and development products now refer to genetic data, most of which come from deCode database. Now the acquisition cost has already been recovered.
Although gene verification of drug targets cannot guarantee final success, scientists still need to find out how to safely and effectively administer drugs to meet a large number of biological challenges, which is indeed a good start. Harper said, ” If the rate of return can be increased by 50%, the change will be great.”
Regent, a biotech company, also wanted to cooperate with deCode. It was about the same time as Amgen acquired deCode and its strategy was similar to Amgen. In the end, Regent did not seek acquisition, but set up its own research center in the form of Regent Genetics Center ( RGC ), spending four years sequencing a large number of explicit groups ( genome of protein coding part ) as much as possible with medical records, and accelerating drug research and development.
Jeff Reid, head of bioinformatics for regenerative metagenomics, said that many people used gene technology when discussing drug research and development, ” but did not design the sample flow in advance.” In short: they have no data. Reed later learned that Regent cooperated with Geisinger Company and joined Regent ( see Basic Care ). Geisinger, a Pennsylvania – based health management system, plans to collect samples from 100,000 patients with complete medical records and sequence them. ” They hope that after obtaining the data, they can effectively improve the care of patients.” Aris ballas, head of the Regenerative Metagenetics Center, said.
Today, Regent has more than 60 partners, including the British Biological Bank, which has recruited 500,000 participants. Ballas and Reid said that it is of vital importance that the scale and diversity of the data they hold are continuously developing and that their internal research data capabilities are also continuously improving. So far, 50 target biological projects have been launched.
Hidden Numbers: Unused Medical Records
Amy abernethy, an oncologist and former Duke University professor, believes that messy health information is meaningless unless two key criteria are met: quality and background. ” Those who do not know the core of medical practice do not know how chaotic it is.” Abernethy, who was chief medical officer of Flatiron Health four years ago, said the startup was invested by Google Ventures ( GV ).
Take the cancer case history that Flatiron specializes in as an example. Many important information in oncology electronic medical records ( actually about half ) may be in doctors’ notes, but notes cannot form specific data fields. All kinds of observation results cannot be sorted into tables by category.
” In the past, electronic medical records were just collection tools and fees. Doctors could only keep their jobs by writing them according to the rules.” Jeffrey Barton, a doctor and chief executive of the Tennessee Cancer Center, explained. Tennessee Cancer Center is a community-based medical institution that treats most cancer patients in the state and is one of hundreds of community cancer centers that use Flatiron system.
Ironically, Flatiron’s real selling point is human beings. Encountered with such data, human beings can often find details that computer systems may miss. Abernethy said that the real challenge is not to collect data, but to ” clean up”. ” It’s really hard to do without understanding the relevant background.”
Flatiron now has data on 20% of active cancer patients in the United States. ” The data structure is very clear.” Daniel Otti, chief executive officer of Roche Pharmaceuticals, said that Roche bought Flatiron in February this year for US $ 1.9 billion. ” Flatiron is unique in being able to collate real data that meet regulatory requirements.” Oti told Fortune. He said Flatiron’s data was very complete, ” theoretically, it can replace one of the’ control groups’ of clinical trials set up by Roche for cancer immunotherapeutic drug Tecentriq”.
In theory, Flatiron’s data system may have a wider impact on clinical trial recruitment. For decades, patient recruitment has been one of the greatest challenges in cancer drug research. For example, if the recruitment goes smoothly, it is easier to match suitable drugs according to the patient’s situation.
IBM Watson has achieved initial results in this field. Mayo Clinic reported in March this year that after using IBM’s advanced cognitive computing system, the number of participants in breast cancer clinical trials increased by 80%. ” Watson helps us to match patients with potential clinical trials faster and more accurately, which was difficult for oncologists in the past.” Christopher Ross, chief information officer of Mayo Clinic, said in an interview with trade publication MobiHealthNews.
Even organizations that have seldom had access to patients before, such as pharmacy welfare management structures, may contribute to improving human health and reducing costs due to a large amount of data.
Take Express Scripts, Missouri’s pharmacy welfare management structure, for example, which was just acquired by Sinosure in March this year. Express Scripts manages 1.4 billion prescriptions for 100 million Americans every year. People know when they don’t take medicine on time. The annual cost of taking medicine not in accordance with the doctor’s advice ranges from US $ 100 billion to US $ 300 billion. The difference in figures is due to different methods of calculation. The reason for the cost is that the patient does not follow the doctor’s advice and has complications, which lead to follow-up treatment.
Express Scripts chief data officer Tom Henry said that 300 factors have been found that may cause patients to give up taking prescriptions. Various factors include basic demographic data ( income level and postal code ) to behavioral data ( the degree of amnesia and procrastination tendency of patients are judged according to questionnaires after patients do not take prescription drugs ) to less intuitive factors, such as the sex of the prescribing person and the patient ( male patients treated by female doctors are more likely to disobey doctor’s orders ).
The company said that the accuracy rate of the algorithm reached 94%, and the algorithm could be used to score patients’ risks and take different reminding services. Henry said that the way to take the medicine was ” relatively mild and not forced.” In summary, Express Scripts said that non-compliance with medical orders decreased by 37%, saving customers US $ 180 million.
Like revolutions in other fields, many people poured in without thinking clearly about the next step or being ready to bear the corresponding consequences. There are also many problems in the social experiments of medical big data, from patient privacy to moral dilemma when warning people that some risks are unavoidable. For many digital medical supporters, big data seems to be a ” magic bullet” waiting hard. The question is where the bullet is going.