A Dose of IT That Doctors Desperately Need

This article first appeared in the September 2008 issue of it magazine. This is part of the Folk Technology series that I write every month.




While search, retrieval and storage technology has advanced tremendously on the enterprise front, doctors feel left behind. Their requirements are unique, and addressing their needs could save lives and improve the success rates in the medical domain. So, are there any adventurous technopreneurs who’ll take up this challenge?

In the past few years, the approach to medicine has changed drastically. You may have observed the tendency of doctors to prescribe numerous clinical and radiological examinations before setting out on a remedial course. For example, doctors increasingly rely upon images like x-rays, MRI (magnetic resonance imaging), ultrasound scans, CT (computed tomography) scans, PET (positron emission tomography) scans, etc, for diagnosis. This is indicative of the move towards evidence-based medicine.

When you visit a doctor, your medical condition is diagnosed based on the symptoms you exhibit at the time and your medical history. The doctor goes on to form one or more hypothesis on your possible medical ailment and proceeds to collect evidence to support or reject each hypothesis. So, for example, if a patient has severe pain in the abdomen accompanied by fever, the doctor may prescribe urine tests, blood tests and an ultrasound to narrow down the options. Once a doctor has pinpointed the specific medical condition, the relevant medicine is prescribed. The doctor may go in for surgical procedures, depending on the diagnosis. In each of the stages -- from hypothesis formation and diagnosis to medicinal or surgical intervention, doctors today are relying more and more on technology to help them. But there is scope for more technology in this space.

Data explosion

Many hospitals in India are moving to the digitised storage of patient data. The patient’s electronic medical records containing doctor reports, diagnosis notes, prescriptions, radiological images and other diagnostic data, are stored in electronic format. This enables easy search and retrieval, which means that every patient’s record is available to any doctor in the country. If you go in with abdominal pain and fever, the doctor can, within seconds, retrieve all similar cases, after suitable anonymisation, from across the country. This knowledge will help the doctor diagnose and treat you better.

Digital medical images, in terms of storage requirements, by far constitute the largest component of patient data. Medical images are viewed and interpreted by radiologists and subsequently by the attending physician. In lung cancer screening, for example, 40 to 400 CT images (depending on the resolution used) of the lung are obtained. It places a huge strain on the radiologist to read all the images, compare with lung pathology atlases, and give an opinion. It is valuable to store these images along with the expert opinion on them so that they may be referred to in the future, to compare cases and prescribe a particular treatment.

The clinical domain is seeing an explosion of data in recent times. Thanks to multimedia and network technologies, dedicated medical data grids are being set up to connect hospitals, doctors and pharmaceutical companies. The data comprises patient records, peer-reviewed scientific journals, clinical studies, disease data, image repositories comprising radiological images, pathological images, laproscopy images, and cardiograms, along with case notes. Patient records with clinical studies and research publications can provide useful clues for diagnosis and treatment. There is a strong realisation that evidence-based medicine requires the availability of comprehensive data on the patient, combined with peer reviewed scientific evidence.

Data for doctors at their
finger tips

There are three types of data that are valuable to doctors:
• Case histories of similar ailments that they or their colleagues may have treated in the past.
• Scientific studies about latest treatment methods.
• Information on drugs and data on drug interactions and adverse reactions.

Typically, doctors examine a patient for about 10 minutes. During this time even if the medical network with the huge amount of data in it is made available to the doctors, they find it hard to extract information that is relevant to the case at hand. So usually, they rely on their medical training and personal experience to guide them. Only in cases where rare conditions are exhibited by the patient, do they take time to search through literature or case histories to determine the exact condition.

The main problem is that efficient technology for searching medical text and images does not exist. A Google-like interface for searching medical data doesn’t work because of the specialised nature of queries.

Doctors wouldn’t query something as general as ‘malaria’, but might like to query on specific points like, ‘the adverse effects observed in using Rifampin with quinine in treating malaria’. Also, in response to their queries, they wouldn’t want to be shown thousands of documents but would prefer the relevant portions that address their particular query. So the system needs to work more like a question answering system rather than like a generic query-based retrieval system that today’s technology provides.

Similarly, if doctors want to know how lung cancer patients responded to a particular drug, they wouldn’t want to screen thousands of lung CT images. They need a system to select the most relevant images and present those to them. Doctors would rather give their valuable time to a patient than spend unnecessary time in searching through a maze of data. Tools for manipulating medical images are still quite rudimentary. While work has been done for the handling of photographic images, sufficient attention has not been paid to images in the medical domain. For accurate diagnoses, subtle differences between images need to be captured and effective methods for doing that need to be invented.

Today we are at a stage where the infrastructure for the storage, retrieval and search of medical data needs to be set up. Technologies that can do this even at a hospital level would be keenly adopted. There is a huge technology gap that can be filled by budding technologists in start-ups, focused on addressing the needs of medical practitioners.

In a recent study conducted by Nobhojit Roy, Neha Madhiwalla and Sanjay Pai on drug promotional practices in Mumbai, published in the Indian Journal of Medical Ethics, it was found that doctors often did not have the entire information about a drug’s interactions and adverse reactions. Their only source of information on drugs are medical representatives who often do not give the complete information and have vested interest in promoting a drug. It is the patient who suffers because of this lack of information.

There is a need among doctors today for tools that allow them to access disease and drug related information quickly. Doctors would value technology that allows them to treat their patients better. In rural areas, access to a medical online grid with a data search and retrieval capability would help multiply the reach and effectiveness of a doctor many times over.


(Part of the monthly Folk Technology Series that I write for i.t. magazine)