How Medical Databases Lead to Better Individual Treatment

A patient of mine asked me whether the cancer treatment she was receiving was the best one for her. I explained that cancer treatments are usually selected based on data collected from real patients with similar cancers. “Wait,” she told me. “I don’t want my personal data to be included in any database.”

I explained that in the past, doctors made medical decisions for patients based on their own experience of what worked and what didn’t. The problem was that one doctor’s experience may be different from others. Over the years, society has come to realize that medical decisions based on more evidence give better outcomes.

Some evidence comes from clinical research studies published in medical journals. However, a large amount comes from patient data collected across the country. For example, we know about specific cancers from registries that collect data from hospitals, clinical laboratories, doctors’ offices, and governments. This data allows us to understand how rates of cancers change over time, as well as differences between locations, genders, and races and the effects of different treatments.

In the early 1970s, the National Cancer Institute began collecting data for the SEER (Surveillance, epidemiology, and end results) database. At first, it was cumbersome. Data was collected from patients’ paper records and eventually entered manually into a database. Then, a couple of decades ago, along came the electronic medical record — the EMR — with an early boost from the US Veterans Administration. Even though SEER, covers less than 50% of the population today, it is still the best-coordinated cancer database in the world.

President Obama’s American Recovery and Reinvestment Act of 2009 really pushed EMRs forward by giving doctors incentives and hospital mandates to use electronic records. This Act also made clear that it was legal and in the public interest to collect patient data and aggregate it without a person’s permission so long as there was no way to identify the person.

I have used this data, and I know there is no way to identify a specific person. But there are still problems. EMR software programs don’t always communicate with each other, making it difficult to collect data across geographic areas.

Another problem is that our understanding of cancer has changed. For example, not long ago, we grouped all colon cancers into one bucket. Today, we recognize dozens of variations, often by genetic differences. Some types of colon cancer might respond well to one type of treatment while other types do not. So, registries now need to be specific about listing variations. We also don’t do well at capturing alternative treatments offered outside of traditional medical establishments. And we could do better collecting data from rural populations and from people with varied cultural, ethnic, or income differences.

Society works best when we collaborate. Large, anonymous patient databases serve us all by identifying which treatments work and which do not. That information can only come from examining data from large populations of people who suffer from similar ailments.

– Dr. Michael Wilkes with a Second Opinion