The EHR and Big Data

In a recent study by MIT staff there is a result that claims to have the ability to gain significant insight from EHR data.

They state:

The results are quite interesting: This is one of the first analyses of large data you get from using electronic health records, and it just became available. This is a big amount of data we got from General Electric. What we tried to look at is, when you go to see the doctor, you’ve got a certain [medical] history, and you’re perhaps looking at a [medical] problem. When you look at that problem, is there any predictive power in the history that comes before? We looked at that from a pure computer science point of view — and it turns out there is predictive power....At the level of the individual, this allows you to compare the medical history to other people, and give additional information to the doctor. Doctors can get additional input from this analysis of the medical history. Of course this is what doctors already do — they look at the past in order to understand what might be the problem. But it’s a mathematical way that guides you, gives you more [than] than you might get by going through [one patient’s medical history].

 Now there are several concerns here.

1. The EHR has become a cut and paste system. EHR records have become extensive restatements by rote of what may be patient complaints or conditions. Physicians all too often just sit and "take" a patient history of a chronic disease such as Type 2 Diabetes by cutting and pasting what they did the last several visits. Perhaps the HbA1c data is changed but otherwise the rest if the same.

2. EHRs for the most part do no provide or encourage a key element in medical analysis; namely examining changes. Patient health all too often is determined by something that has changed. Did the patient gain or lose weight, gain or lose endurance, gain or lose sleep. Simple issues, some of which are quantitative and many subjective.

3. EHRs are distributed and dissonant. Namely a patient may be seeing several physicians. Take the Type 2 Diabetic patient. The ophthalmologist for retinal deterioration, while seeking a dermatologist for follow ups on a melanoma excision. Does the ophthalmologist look for ocular melanomas as well?

4. Images and imaging are frequently left out. This includes both classic imaging as well as pathology images. It may also include genetic results.

Thus the data may be very noisy, may be inaccurate, may contain the wrong results, and thus may be misleading. Moreover the data may become a baseline for policy, such as the PSA test. And the data like the PSA may be all wrong. One should remember the recent SEER data disaster!