7
The most fundamental mission of Wolfram|Alpha is to be the internet’s hub for all things computable. With this in mind, our medical data team has been combing through peer-reviewed journals, population-based surveys, and credible online health calculators to bring you the most complete, up-to-date, and easy-to-use library of medical calculations available anywhere on the web. This endeavor has been ongoing since the launch of Wolfram|Alpha more than a year ago, and can be demonstrated through queries such as “heart disease risk”, “male age 27, 175 lbs”, or “basal metabolic rate”.
Over the past couple of months, we have worked to implement over 20 new equations. For example, hematocrit levels outside the normal range are indicative of any number of health concerns ranging from dehydration to kidney disease. In circumstances where estimates of hematocrit are in need and only certain parameters are known, Wolfram|Alpha can be used to fill in the gaps and assess whether the estimated value falls within the normal range, given a number of personal attributes such as weight, height, sex, or age:
Calcium in the blood is also a very important indicator of various health conditions, including complications of various types of wounds, hyperparathyroidism, and even osteosclerosis. Given total calcium and serum protein levels, Wolfram|Alpha can estimate the blood concentration of unbound ionized versus protein-bound serum calcium: More »
9
A new medical diagnosis or drug treatment can often leave us with more questions than answers. A few weeks ago we introduced a disease dataset within Wolfram|Alpha that can be helpful for those wondering how their condition and treatment plans compare to those of other patients. Most notably, this dataset includes the fraction of patients within the United States that have been diagnosed with a medical condition in a given year. For each condition, Wolfram|Alpha has various levels of information, including commonly reported symptoms, co-occurring diseases, and lab tests used for diagnosis. Beyond this, Wolfram|Alpha also has carefully curated data on drug treatments. For example:
The data displayed from these inputs gives classes of drugs prescribed or administered to patients during health care provider visits. Wolfram|Alpha ranks the drug classes by the number of patients to whom they were administered. For example, “hypertension drug treatment”, initially shows us that, of all the patients diagnosed with hypertension, 25% were prescribed angiotensin converting enzyme inhibitors, 22% HMG-CoA reductase inhibitors, 21% cardioselective beta blockers, 19% antihypertensive combinations, and 16% calcium channel blocking agents. (That’s over 100% total because some patients are prescribed more than one medication.)
Looking above the ranked drug table we can see that there are a handful of useful options. Click “Show drugs”, and the table opens up and displays a ranked table of brand-name drugs prescribed within each class. From this table, you can see interesting differences in drug-prescribing patterns between the sexes. For example, the angiotensin converting enzyme inhibitor Lisinopril is more commonly prescribed to male hypertension patients than females, but looking further down the list, we can see that female patients are more commonly prescribed Enalapril than are males.
Wolfram|Alpha can also can also provide generic options for prescription drug treatments. More »
7
How many people are diagnosed with diabetes in a given year? Is hypertension more common in men than in women? What drugs are most commonly prescribed for anemia?
In order to address questions like these and many more, Wolfram|Alpha has now assimilated data from two different surveys conducted by the CDC: the national ambulatory medical care survey (NAMCS) and its hospital-focused counterpart, the national hospital ambulatory medical care survey (NHAMCS). Together, these surveys provide information on common reasons why people visit the doctor’s office, drug treatments that are highly correlated with a particular disease, and which diseases are most commonly diagnosed within specific races, ethnicities, and genders.
Through Wolfram|Alpha, you can investigate data on thousands of diseases and medical conditions, such as these:
Instead of looking at all the information at once, you can also try more targeted inputs, such as “fraction of US population affected by lung cancer”:
From this output, we can see that approximately .21% of all U.S. patients are diagnosed with lung cancer each year. More »
4
We recently added data on health indicators for more than 200 countries and territories. We now have World Health Organization data on health care workers, immunizations, water and sanitation, preventive care, tobacco use, weight, and more.
Data is also now available on specific types of health care personnel, such as physicians, nurses, and dentists, and Wolfram|Alpha can also compute per capita figures for each type of health professional. Check out the figures on midwives in South Africa or dentists in Iceland—or for a particularly interesting view, try asking about doctors per capita in all countries.
Other intriguing indicators include figures on hospital beds, drinking water and sanitation, tobacco use, weight and obesity, and reproduction and contraception.
Some data, such as for infant immunizations (including DTP, MCV, hepatitis B, and Hib), spans several years—which allows you to see dramatic increases in immunizations in many developing countries, as well as surprising declines in some first-world nations. More »
180
When we launched Wolfram|Alpha in May 2009, it already contained trillions of pieces of information—the result of nearly five years of sustained data-gathering, on top of more than two decades of formula and algorithm development in Mathematica. Since then, we’ve successfully released a new build of Wolfram|Alpha’s codebase each week, incorporating not only hundreds of minor behind-the-scenes enhancements and bug fixes, but also a steady stream of major new features and datasets.
We’ve highlighted some of these new additions in this blog, but many more have entered the system with little fanfare. As we near the end of 2009, we wanted to look back at seven months of new Wolfram|Alpha features and functionality.
3
One of Wolfram|Alpha’s primary sources for medical test data is the National Health and Nutritional Examination Survey (NHANES), an annual survey of thousands of people, from throughout the United States, conducted by the National Center for Health Statistics (NCHS). Wolfram|Alpha’s presentation of this data is unique in that it doesn’t just report reference ranges, but allows you to see where your own measurements and test numbers fall within the survey’s distribution of results. (Wolfram|Alpha does not give advice, medical or otherwise.)
At the most basic level, an input of “cholesterol test” returns the survey’s distribution of total cholesterol values:
1
A trip to the doctor’s office can sometimes leave patients with more questions than answers, specifically if their doctor has requested they undergo medical tests. Wolfram|Alpha is a helpful reference for understanding what the tests measure and how to interpret the results. Wolfram|Alpha allows you to query information on a specific medical test or a panel of tests, compare tests and results for patients with specific characteristics, compute your estimated risk for heart disease, and find the diagnosis corresponding to an ICD-9 code. Wolfram|Alpha can take into account specific patient characteristics like gender, age, smoker, non-smoker, pregnant, diabetic, obese, and underweight. Wolfram|Alpha can give you a snapshot of available data that might help you understand how your results compare to others’. (Wolfram|Alpha does not give any advice, medical or otherwise.)
First we will demonstrate how you can use Wolfram|Alpha to learn more about a specific type of test your doctor has ordered. By entering the name of the test into Wolfram|Alpha, such as “CBC”, we can learn what the test measures. In this case, the test measures the number of cells commonly found in a blood sample, such as red blood cells and platelets.
