Recall your last visit to the doctor’s office. You probably were weighed and had your blood pressure checked. The physician might have monitored a response to a new medication. Perhaps you were asked about your family’s cancer history, or you discussed how a fractured limb is healing.
Even if it was a routine checkup, that doctor’s appointment resulted in numerous data points, which could have been recorded electronically or with paper and pen. Your creation of health data likely didn’t end at the office. Throw on a Fitbit, Microsoft Band, or Apple Watch, and every step you take becomes a note in your fitness record.
Every activity could yield some form of data from every organ; factoring in the complexity of the human anatomy, health care could be one of the largest data sources out there. But at this point, data interpretation has been unable to keep pace with data collection, which occurs in many formats. The data’s out there, but there’s just not a fast, effective way for doctors and patients to utilize it.
“The old approach was get my data organized, get my data normalized, get my data cleansed, load it, and transform it,” said Michael Sandoval, CEO of Atigeo, a Bellevue company that makes a data-analysis platform. “The new approach … is to transform the data, and then load the right pieces of it at run time, at the moment you need it, and only the collections you need.”
Atigeo’s platform, called xPatterns, is Sandoval’s enhanced solution. XPatterns can aggregate and decipher patient data from numerous sources, thus giving the patient, physician, and health care organization a centralized report on the patient’s medical history in real time. Furthermore, machine learning and other analytical tools are employed to reveal patterns and predictions by culling data from sets that would take humans years — perhaps hundreds — to sift through.
Sandoval’s company is one player in the fast-growing medical data field. As more and more facilities update to electronic records, it’s reasonable to expect that a patient’s medical history should follow him or her from doctor to doctor, thus rendering the check-in form unnecessary. From an organizational standpoint, platforms like xPatterns can be used to better understand an individual’s health trends while simultaneously to the larger pool of a patient population.
To further that effort, Ascension Ventures, the investment arm of St. Louis-based Ascension Health, in September participated in an $18.4 million investment round in Atigeo to hone its health-care platform (xPatterns is also used by defense, energy, finance, and cyber security firms). “The goal at the end of the day is to make health care data a commodity so that it can be leveraged by our patients and providers so they can increase the quality of care,” said Gerry Lewis, vice president of information technology strategy for the Catholic health care group.
With xPatterns, Ascension recently created a predictive model for readmissions, which will help the organization hone preventive care. Ascension also is looking at referral patterns and correlations between pain scores and patient satisfaction.
These are early stages in what Sandoval believes will become data-driven health care that leaves doctor and patient only with higher-level decision making. “The data can refine and provide an optimal prescriptive protocol and sets of opportunities to the patient and to the doctor,” Sandoval said. From there, you “take on the wisdom and the experience of the doctor to help guide you. Even if you reduce it from 10,000 down to two choices, (making the) ideal choice … is more human-inference oriented.”
This process holds considerable weight for Sandoval, who lost his first wife to breast cancer.
“You have a structured, procedural approach to breast cancer, even today. Women walk in, and a twenty-something woman is treated exactly like a 75-year-old woman. It’s Russian roulette whether you get the right (treatment),” Sandoval said. “The reality is, with genomics, with epidemiology, clinical data, lab data, and other forms of medical research — if you could take that unstructured data, marry all of them together, and drop my medical record in the middle of that, I can find something that is highly effective for an individual patient, rather than something generalized for breast cancer patients.”
Sandoval’s argument for large-scale data analysis is analogous to one favoring factory automation: Let robots or algorithms handle the mundane, while humans can focus on design, implementation, or other more psychological elements of a process. But machine learning places a great deal of emphasis on an algorithm’s ability to correctly interpret notes and determine which ones are important. The system works pretty well for Netflix and Amazon, but stakes are a bit higher in health care than in television streaming.
There certainly are benefits a system like xPatterns. One, patients can have access to their entire medical records, regardless of provider, format, or longevity. In a 2007 Markle Foundation survey, 91 percent of respondents felt patients should have online access to their medical records but just 1 percent actually did. Wading through records also is time-consuming for doctors already stretched thin; family practitioners can have more than 100 patient interactions a week.
“There’s too much data for any one person or any group of people to ever process and connect the dots,” Sandoval said. He envisions every medical interaction beginning with an xPatterns analysis of a patient’s medical history. New symptoms can be cross-referenced against previous maladies and population data, and doctor and patient alike can avoid redundant analysis or treatment.
Atigeo has secured multiyear contracts worth tens of millions of dollars — Sandoval chose not to name any clients — and the company plans to double its staff to 200 this year. Furthermore, it announced Wednesday the purchase of YaData, a Las Vegas big-data firm with a strong foothold in the defense sector. As big data proliferates multiple industries, Sandoval’s positioning Atigeo to be a simple, adaptable platform that can be used in many capacities. “We’ve exposed the capability for people to build applications very readily, very easily,” Sandoval said. “We want high schoolers to be able to leverage the platform.”
This post has been edited to reflect that Ascension Ventures was an investor, not the sole investor, in the $18.4 million financing round. Gerry Lewis’ title has been corrected.