Tag Archives: microsoft healthvault

A sustainable business model for health information exchange

Personal Health Records (PHRs) have received a lot of attention lately with the release of Google Health and Microsoft HealthVault. PHRs allow individuals to securely organize their own health information and to control who has access. PHRs allow users to import data from many large repositories, including insurers, pharmacies, and laboratories, but mostly do not connect to care providers, excepting a few integrated delivery networks such as Kaiser and Beth Israel in Boston.

RHIOs are less well known but have been around much longer. RHIOs exist at the community level to facilitate the exchange of health information between care providers (e.g. doctor notes), though some also aggregate data from insurers, pharmacies, and labs. The primary audience here is care providers, rather than consumers. In theory, a network of RHIOs could form together to create a National Health Information Network (NHIN), but this is unlikely to occur without significant federal funding. More than 140 RHIOs exist today, but only a handful are considered sustainable (look at the Regenstrief Institute and MA-SHARE). Interoperability between EHR systems is complex and expensive, and though it will improve quality of care, there will be a negative return on investment until a significant proportion of providers is sharing data in a standardized way. Read The State Of Regional Health Information Organizations: Current Activities And Financing by Julia Adler-Milstein, et al for a better understanding of the current RHIO landscape.

Given the similar goals of PHRs and RHIOs, I have speculated that these ideas may eventually end up merging. John Halamka, CIO of Harvard Medical School, commented on Google Health’s new sharing and auditing features in his Cool Technology of the Week update:

I’ve already invited my primary care doctor, my family, and a few of my clinical systems colleagues who built the BIDMC-Google interface. Thus, in one morning I’ve become my own regional health information organization, sharing medical records across multiple organizations with perfect privacy controls.

However, Dr. John Glaser, CIO of Partners Healthcare, estimates that only 1 out of 10 American’s with access to an electronic health record will access it, meaning that consumers are unlikely to drive the healthcare industry’s to adoption of EHR and HIE (see Glaser’s interview with Information Weekly).

Though PHRs do not have a huge chunk of the health consumer market, 6 million users (10% of the estimated 60 million American’s with electronic health records) is large enough to give Google, Microsoft, Dossia and Revolution Health some leverage with data furnishers. PHRs also capture a monetizable audience (consumers) whether through advertising or sales of health applications. RHIOs, on the other hand, make participating providers pay to use their service, but are not large enough to offer providers significant incentives for involvement. Most rely on government funding to stay afloat. I am of the opinion that unless government steps in to fund a largescale network, PHRs have settled on a more sustainable business model and could more easily extend their data aggregation efforts into doctors notes and other administrative data handled by RHIOs. In this case, both providers and consumers would be the primary audience of the platform. By using the PHR for both consumer and clinical needs, the record becomes immensely more useful, assuming that audit trails, quality control, and security measures are in place.

health, crowds, and data mining

Google released Google Flu Trends yesterday, which analyzes search terms for indicators of flu activity. With the onset of flu season, people start searching for keywords such as “flu vaccine” which Google detects and charts. The example below reveals that we are just a couple weeks away from a time of year that has experienced a large outbreak:

Google Flu Trends Sample

The true genius behind this system is that Google is not directly involved in data collection. Data is collected passively as searches are submitted by users. Incredibly, Google Flu Trends reliably performs flu surveillance up to 2 weeks faster than the CDC (US Center for Disease Control)! For details on Google’s tracking method, check out their blog post Tracking Flu Trends.

In a similar fashion, David Bates of Harvard Medical School is creating an epidemic surveillance system that analyzes electronic health records of several Boston-area medical centers every night. When an outbreak is in the works, not all the sick people go to one hospital. 2 might show up at one hospital and 3 at another. The next day several more go. By the time authorities are aware of an outbreak, it is weeks too late. Performing surveillance on data from several hospitals simultaneously greatly expands quantity of information available and can potentially prevent outbreaks from occurring.

Data mining in health that transcends a single unit (like a hospital) has only just begun. Personal health record systems like Google Health and Microsoft HealthVault optionally aggregate health data from a variety of sources (e.g. hospitals, clinics, insurers, pharmacies). Determining health trends is one of Google’s primary goals with this system:

Google will use aggregate data to publish trend statistics and associations. (http://www.google.com/intl/en-US/health/privacy.html)

Once again, while Google and Microsoft are both investing heavily in platform development and partner recruitment, the data is entered, imported, and managed by the consumer. For an interesting post on the positive and negative ramifications of Google Health, check out Tree of Knowledge.