By Bruce Schatz and Richard Berlin

Every day, millions of persons search on Google, shop on Amazon, share on FaceBook. Internet services show the way to measure everyday health for all Americans, by analyzing the trends of specialty populations within geographical regions. This leads to a healthcare infrastructure that will support viable healthcare, acceptable quality at acceptable cost.

Google Flu Trends offers a service assessing risk of catching the flu, using internet queries from their search engine. Google philanthropy worked with the Centers for Disease Control and Prevention (CDC) to adapt this technology from their commercial trends service for determining product popularity from search records. The Google system itself automatically collects information from millions of widely distributed sources, before indexing this to support rapid search capability at their custom supercomputer data center. Amazon offers services assessing the interest in buying an item from their internet store, by analyzing sales popularity within geographic region, customized using overlapping records of which persons bought which items. These services run on their cloud cluster supercomputer, on which they also rent time to other companies. Adapting such analysis to healthcare will match populations with similar properties in similar regions, so that similar trends can be similarly addressed.

In his book on the American Health-care Crisis, former Senator Tom Daschle, proposes establishing the Federal Health Board that will resemble the current Federal Reserve Board for the banking industry. The Federal Reserve accomplishes its mission by gathering vast data sets that measure the nation’s money supply, through regional networks to local banks from individual persons. They employ experts who produce analysis reports of money trends, from detailed surveys across many sources, including periodic data from manufacturers and consumers. The Fed Health will have to similarly gather vast data sets that measure the nation’s health quality, through regional networks to local clinics from individual persons. They will also employ experts who analyze health trends from detailed surveys across many sources, including electronic medical records from health plans and personal health records from individual patients, along with archives of clinical trials and medical literature.

Analysis of these data sets will enable government regulation to address extraordinarily high administrative costs, provide health information services for the uninsured, and track the innumerable healthcare processes that contribute to harmful medical errors. Modern information technology can meet these problems of healthcare, to support evidence-based medicine. Health data is more diffuse and more complex than economic data, so the
Fed Health being established in the early 21st century needs to leverage the latest technology in a way not possible with the Fed that was established in the early 20th century. Gathering the necessary data sets was not technically feasible fifteen years ago during the last wave of national discussions on health reform. The national infrastructure measuring population health can now leverage new technologies in the private sector, particularly those for targeted marketing of internet services.

This national healthcare infrastructure will measure the daily health of all Americans and extrapolate this to the health of populations to provide the data necessary to manage the nation’s health. It will ensure universal access for universal healthcare, by reaching all populations with appropriate interfaces. Seniors can be reached with phone calls to their homes, students with text messages on their cell phones, baby boomers with web forms via their personal computers. Even the uninsured and the under-served can participate with inexpensive devices on the ubiquitous network.

Large-scale health data sets from millions of persons is necessary to properly evaluate the quality of medical care and outcomes. Current methods do not measure up to daily usage across all of America, as shown by numerous examples where a treatment was widely used before science showed that it might cause more harm than good. Estrogen was taken by millions of post-menopausal women over many years; cold and cough medicine was taken by millions of sick children; arthroscopic knee surgery was performed on millions of arthritic seniors and younger athletes. If national data sets had existed, these treatments would have been used only by those populations who would properly benefit from these treatments.

Indexing the nation’s health will use 21st century technologies of internet services to solve 21st century problems of chronic illness. Different groups of people in different populations respond to different treatments in different ways. The interactions with individuals will generate measurements of populations, as the input for massive data mining on large-scale supercomputers. These information technologies will gather the data necessary to support the healthcare infrastructure for viable healthcare. Regulatory standards will be set to encourage treatments with positive outcomes for particular populations and discourage treatments with negative outcomes, thus improving quality while reducing cost. The Federal Health Board can then support the best available quality of life for all Americans, using best practice standards based on analysis of adequate data. Measuring health is like tracking flu; it is time to use the strongest private technologies for the greatest public good.

Bruce Schatz is Head of the Department of Medical Information Science at the University of Illinois at Urbana-Champaign and the Research Scientist for Information Systems at the National Center for Supercomputing Applications, including the period when NCSA developed the Mosaic browser that catalyzed the Web.

Richard Berlin is a general surgeon at a local hospital and former Medical Director of the regional
HMO.  They are writing a book on healthcare infrastructure for the 21st century.