When Googling for Medical is Not Right

Most of us use google for our medical search, what is wrong in it? But my point here is google is more a computer than a medical one. We all appreciate google search and would definitely continue using it provided it can give fewer and more meaningful results. I will describe my problem with an example but before that I want to talk a bit on how much google is interested in health field and how good computer scientists in general are.

Google Health was started with right intention of storing PHR online but had a premature sad demise. Many a scholarly articles are written on its autopsy report. One such is in HIMSS which says that not everything can be googlized. But most pertinent would be by its creator, while talking to TechCrunch he says ‘it is not social’ in storing health records. He adds, there is no fun in storing health records while patients want encouragement and even pressure of their friends to be healthy. Today it’s contemporary (Health vault of Microsoft) on Facebook page says “take control of your health”.

There might be some occupational hazards or even landmines waiting for engineers in healthcare field. First of all, although modern medicine as practiced worldwide is multidisciplinary there are real, not imaginary but imperceptible barriers for others. For example, although many medical issues are solved by consensus, very few are unanimous. Doctors in charge of a patient make a decision on what is right for him. It might vary if he takes a second opinion. Secondly an error might result in a tragedy for which doctor is held morally responsible.

For computers scientists it is relatively easy to write a code or implement an algorithm. How to ask right questions are openly debated probably on stackoverflow and a consensus or about are reached quickly. There is no such luck in patient care. Endless open corridors confuse not only the patient but also his doctor. Simply put medicine is too complex for a set of algorithms. No human can be defined in any code or a number. So where is the balance between the two?

The solution is from doctors only. As a group doctors understand patients better than engineers can ever understand patients. Patience and endless waiting for results are virtues of medical field which is probably worth learning for young and ambitious who forage into medical field expecting breakthroughs. 

Doctors prepare to treat patients while adapting to the developments in other fields. While electromagnetic resonance quickly evolved into MRI doctor is busy using it for patient care. Similarly there are several efforts to develop large scale standardized terminologies in the field of medicine, anatomy or drugs by doctors whose efforts are awaiting solutions related to computer science. Artificial intelligence solutions namely Natural Language Processing (NLP) and Decision Support Systems (DSS) are necessary for managing the ever expanding complexity of medical practice.

SNOMED CT (Standardized Nomenclature of Medicine Clinical Terms) is just an example of such an effort. Association of American Pathologists (about 30 000) in collaboration with their colleagues at NHS in UK started it in 1965, 50 years ago when some of us are in primary school or many not even born. Today it has over 426 000 clinical terms for which not only a 9 digit unique number is assigned to each but also their relationships are worked out. For all that effort, it is available freely for public. Now, since India is also a member of International Health Terminology Standard Development Organization (IHTSDO) and all of we in India have free access to it.

Ontological principles determine classification of knowledge in terminologies and Computer science as a field has contributed enormously. However, philosophy remains core limiting factor in any such breakthroughs and hence they are in high demand now. New insights in logic are awaited for which ancient Indian logic might give refreshingly fresh breather to existing problems. 

Now I refer to the example I promised earlier. I type eye for google search and I get 1.41 billion results distributed by 10 for each page. Most of us do not complain because after all it fetches all those that have eye as a word. One can only go up to 3rd page before giving up and nobody expects to scan all results. After all how can one ridiculously go through 140 million pages on eye? Something needs to be done about it; I hope all of us agree. Let me say what is expected in medical search for eye assuming none of us want google results as it is now. I am looking for some related terms to eye to pop up in first page only. How can one achieve it?

The solution as promised is from doctors. SNOMED CT has all clinical terms and their relationships. If its algorithm is implemented as NLP then results are what every doctor desires. One such solution is from ai-med which gave 30 results for eye in tags that are google search enabled. It means if I select Proptosis from the result by clicking on it I get google results for it. I find this two layered search much better than routine google search. So why should all of us be content with google search alone?

Unfortunately, internet as democratic as it seems has evolved into asymmetric knowledge society to which google is also part of. Distribution and utilization of knowledge is limited by big organizations with huge resources. Our societies are still grappling to dismantle this divide and slowly bring a change. However, it has not happened completely yet.  In the meantime we doctors need to be practical and accept as solutions come by whether from small or big sources.

Querying for eye on ai-med

By |2015-12-10T21:15:54+00:00December 10th, 2015|Uncategorized|0 Comments

About the Author:

I am a medical doctor with experience in practice and research. After graduating from medical school I was trained in research for about 8 years. Following this I practiced as well as developed medical applications for diagnosis. I developed PAIRS which is hosted at www.ai-med.in

Leave A Comment