Sign up FAST! Login

Given Time, 'Big Data' Promises to Transform Patient Care

data patient

 With limited time and a heavy patient load, physicians face a daunting task trying to identify a patient's multiple needs during a single office visit.


David Bates, M.D.

Now, growing implementation of electronic health records (EHRs) and other health information technology, combined with rapidly evolving clinical analytics techniques, promises to make this task easier. By collecting and analyzing data that present a more comprehensive, detailed medical picture of entire patient populations, physician practices can monitor their patient panels more efficiently -- often without scheduling additional office visits.

"Big data" in health care refers to a wide-ranging combination of clinical, genetic and genomic, outcomes, claims, social and other data that is collected from multiple sources. Bringing diverse sets of data together will allow physicians to use predictive analytics and identify patients who are, for example, most likely to be "high-cost" -- that is, frequent utilizers of costly health care services.

AAFP News recently reached out to David Bates, M.D., to discuss how physicians can use big data to anticipate patient needs. A general internist, Bates is a professor and chief of the Division of General Medicine at Brigham and Women's Hospital in Boston, as well as a professor in the Department of Health Policy and Management at Harvard Medical School. He has written widely on the use of technology in health care, most recently as co-author of an article on big data( published last month in Health Affairs.

Q: How long will it take before big data becomes a standard part of patient care?

A: It will depend on the pace of health care reform, but I expect it will proceed quickly, and the big data approach will take hold in three to four years in routine care. At first it will help with treatment of high-cost patients and to predict readmissions. For healthy patients, it will take make more time because for economic reasons, big data companies are not focused on it.

Q: How can a solo practitioner or a small practice utilize large data sets to improve patient care?

A: A doctor using an EHR will contract with a different company to meet their big data or analytics needs. Claims data are very useful for analytics. Physicians that have access to claims data will be at a great advantage. Some states have established an all-payer claims database.

A limitation with claims is they are often not available until six weeks after the fact, sometimes longer. (In such cases,) it would be really valuable to have more real-time big data that includes clinical records and data from health information exchanges.

If someone in North Dakota is admitted to a hospital, and the hospital belongs to a health information exchange, the way that things should work is that a practice's big data company would get a feed about the patient in real time. The provider would be alerted that the patient is in the hospital and would know to schedule a follow-up visit. (The provider also) might get a suggestion about how soon the patient needs to be seen.

Q: Can aggressive follow-up care be provided without big data?

A: Definitely, but resources are scarce, and the better you manage them the better off you'll be. A lot of people get admitted to the hospital, but there's only a small proportion of them that you should be focusing on. A healthy patient may go in to get his appendix taken out. That person is not so critical to see right after discharge in primary care.

On the other hand, consider a patient with a high potential for readmission (who) is discharged from the hospital on Wednesday. You might want to see (the patient) that Friday to make sure (he or she is) on the right track. The data we have suggest that we don't do that well. Doctors are often not involved at that point. The hospital may call the physician's office to let them know about a patient, but that call might be handled by someone in the practice who doesn't recognize the importance of the situation, (with the end result being that) the patient does not have an appointment for three weeks. That's what happens in a lot of cases.

Q: What are the benefits for chronic care patients?

A: Big data can highlight treatment optimization for complex diseases. For cardiac patients, you might get a suggestion that you should add a new therapy or adjust the patient's medication based on their genetic data or clinical laboratory results.

In one scenario, the practice receives a call from the hospital where one of its patients was just hospitalized, which could let (the practice) know that the patient, based on analytics, is at high risk for another hospitalization so that the practice needs to fit (the patient) in right away for a visit, and perhaps also call (him or her) beforehand. Physician practices will need mechanisms to handle requests like this, which can be a struggle to develop because we are more accustomed to thinking about the patients who are in our offices rather than (all) those who are at risk.

Also, we are starting to realize the high costs associated with mental health. Nearly half of the patients who are projected to be high-cost have multiple chronic diseases or some behavioral health issues. If you don't manage the behavioral health issue, it may be impossible to get them to take their diuretic. Patients also often have social issues such as needing a ride to the clinic, which could mean they miss visits. All these issues can potentially be better addressed using big data profiling to identify a patient's specific needs.

Q: Besides monitoring specific clinical issues, how does use of big data improve care for high-risk patients?

A: For many physicians, if your patient goes somewhere else for care, you are not going to know about it. If you are connected to a health information exchange and have some relationship with a big data company, you should be able to find (that information) out sooner.

Other tools like accelerometers, which are in most smart phones, can monitor how much a patient is moving, and a (global positioning) system can be used to assess where they've been. If your patient allows their smartphone data to be used, big data techniques can be used to process it. Your practice might receive a message that says, "Miss Smith is not doing well. She's made no phone calls and has barely moved from bed in two days. It might be worth checking out what's going on with her."

We know that when you do not hear from a patient for some time, it is often a sign that something is seriously wrong. Historically, we've focused more on patients who come to see us. Big data approaches can help in population management so that you are thinking about all your patients and can identify risky situations even in those who don't come in.

Q: How are the patient data presented? Are all the data captured in a registry or sorted by condition or patient type?

A: A condition-specific register is useful for most patients. For patients with multiple conditions, big data can help analyze them all and allow physicians to figure out what is likely to happen and what the patient's needs are.

Single-condition registries have limitations. A patient may have diabetes that is under control, but the real issue is asthma or depression. That's what analytics are for: to help you look at patients across multiple dimensions. Physicians are good at many things, but no one is smart enough to put together all the pieces of information about patients because we're so busy. Analytics can help us figure out what is truly important and which issues to target at a particular time.

Q: Who should be monitoring the data inside a practice -- the physician or support staff?

A: You won't need a big data manager. The staff should work with the company that is responsible for collecting and providing the data, which should come to the practice in ways that make it easy to use.

Only a small proportion of the information will be given directly to the doctor. More and more primary care physicians are working in practices that are medical homes or medical home-like, and other people on the team should be able to access the data.

Q: Can big data be useful for treating healthy patients?

A: Absolutely. In the future, it will be very valuable to predict who's likely to have a health issue down the road, and big data are very helpful for predicting risk and helping patients determine what habits they might want to change. The patient might find that for (his or her situation), the things that would make the most difference would be changing (his or her) diet or exercising regularly.

The data might tell us that we should do some other type of screening when a patient comes in for a physical. Let's say a 50-year-old man comes in for a (regular) screening visit. You might get a message from the data company saying that six of his relatives had four different types of cancers, suggesting the possibility of Lynch syndrome. Because of this, you might conduct a genetic test that you might not otherwise have done.

Big data gives you information about patients, and you have to figure out how to manage it within your own setting. You are managing risk in different way.

This too from 

David W. Bates1,*,  Suchi Saria2,  Lucila Ohno-Machado3,  Anand Shah4and  Gabriel Escobar5

Big Data In Health Care: Using Analytics To Identify And Manage High-Risk And High-Cost Patients

AbstractThe US health care system is rapidly adopting electronic health records, which will dramatically increase the quantity of clinical data that are available electronically. Simultaneously, rapid progress has been made in clinical analytics—techniques for analyzing large quantities of data and gleaning new insights from that analysis—which is part of what is known as big data. As a result, there are unprecedented opportunities to use big data to reduce the costs of health care in the United States. We present six use cases—that is, key examples—where some of the clearest opportunities exist to reduce costs through the use of big data: high-cost patients, readmissions, triage, decompensation (when a patient’s condition worsens), adverse events, and treatment optimization for diseases affecting multiple organ systems. We discuss the types of insights that are likely to emerge from clinical analytics, the types of data needed to obtain such insights, and the infrastructure—analytics, algorithms, registries, assessment scores, monitoring devices, and so forth—that organizations will need to perform the necessary analyses and to implement changes that will improve care while reducing costs. Our findings have policy implications for regulatory oversight, ways to address privacy concerns, and the support of research on analytics.

Full text here: 

Stashed in:

To save this post, select a stash from drop-down menu or type in a new one:

You May Also Like: