An Introduction to Electronic Health Records

A question that I get asked a lot is “What is an Electronic Health Record (EHR)?” This is a great question, and I hope to answer it in this blog post. This will be UK focussed as that is where I am based and have experience with but the principles are the same in other countries, even if the systems are slightly different. This is a very high level overview and I will go into more detail in future posts. The aim is to provide a basic understanding of what an EHR is and how it is used in research and clinical practice.

What is an Electronic Health Record?

An electronic health record (EHR) is a digital version of a patient’s notes. It contains information about a patient’s medical history, diagnoses, medications, test results, vaccines, and treatments. EHRs are used by healthcare professionals to provide care to patients. So when you visit your doctor, they will both be looking at your EHR to see what has happened in the past and writing down new information from your visit. The EHR is a living document that is updated every time you have an interaction with the healthcare system, and importantly its purpose is for routine clinical care, not research.

How do EHRs differ from paper records?

A paper record is a physical document that you are free to write in and add to as you see fit. An EHR is a digital document that is stored on a computer system, and whilst you can write free text in it, you are strongly encouraged to use forms and structured data to make it easier to search and analyse. All the big EHR providers make use of Clinical Coding Systems like SNOMED-CT and ICD-10 to help with this. SNOMED-CT stands for Systematized Nomenclature of Medicine - Clinical Terms and it is a detailed tree of medical terms that can be used to describe a patient’s symptoms, diagnoses, and treatments. This is primarily what is used by Primary Care (GP) systems in the UK. ICD-10 stands for International Classification of Diseases and it is used to code diseases and other health problems. This is primarily used by Secondary Care (Hospital) systems in the UK.

What this means for your doctor is that they are prompted when they start typing or when filling out a form to pick from a list of terms that are standardised. So, if you go to your GP with a cough, they will type in “cough” and a list of possible terms will come up. They will then pick the one that best describes your cough. This is then stored in your EHR, with the date and time it was recorded, and who recorded it. Things get a little complicated when you have multiple terms that could describe the same thing as it is really up to the individual clinician to choose the term they think is most appropriate. For example, if we just take the term “cough”, and put this into the online SNOMED-CT browser, we can see that 1073 terms come up in the search, and even if we pick up Cough (finding), there are still 37 different children terms underneath this.

What this means is that your simple trip to your doctor can produce a number of records in your EHR - all within one visit. Let’s run a scenario to see how this might work:

“You go to your GP with a cough and fever. Your GP takes your pulse, blood pressure, oxygen saturations and temperature and records these values. They ask you about your cough and fever and record what you say, and then they examine you and record what they find. They then decide that you have a chest infection and prescribe you antibiotics, making sure to record that you have no known allergies to any antibiotics, and asking you to return if you are not feeling better in a week”

Typically, what are called “events” such as diagnoses and symptoms are recorded in the EHR in the “Event Table”, your examination and vital signs are recorded in “Observations Table” and your medications are recorded in the “Medications Table”. These tables are then linked together by a unique identifier for you (in the UK, your NHS number). Your GP is not aware that these tables exist, as they just see the EHR as one big document, but the computer system knows where to put the data so that it can be easily retrieved later. So our visit to the GP would produce set of records that look like this:

Event Table:

Observations Table:

Medications Table:

So you can see that our short visit to the GP has produced 12 rows of data in the EHR, as well as some free-text that sits around this data.

A similar process happens in hospitals, but at least in the UK, ICD-10 and OPCS (which is a classification system for surgical procedures) are used instead of SNOMED-CT. These systems are a lot less verbose than SNOMED-CT, but they are still arranged in a tree structure. Confusingly OPCS and ICD10 have a very similar format (they are both alphanumeric codes) but they are not interchangeable.

How are EHRs used in research?

Now we have seen how EHRs are structured, we can start to think about how they are used in research. The first thing to note is that EHRs are not collected primarily for research purposes, but they can still tell us a lot about health trends, causes, treatments, and outcomes. In this blog, I am not going to go into the details of how consent and data sharing works in the UK, but it is important to note that researchers can normally only access EHR data if they have the appropriate permissions and have gone through the correct ethical approvals.

Researchers can normally only access the structured data in the EHR, and not the free text. This is because the free text is often impossible to anonymise effectively with computer code which is what you would need to do because the data is so massive. To put it into context, I used to work with a dataset of roughly 50,000 people, and the whole dataset was around 70 million rows of data. This is a lot of data to anonymise if you had to check all the free text. As a result, free text is normally not used in research, and researchers have to rely on the structured data to answer their questions.

I plan to write more on the actual steps that researchers take to carry out research using EHR data in future posts. I think this is worthy of a post in itself, as it is a complex process that involves a lot of different steps, and caveats.

This is a fantastic paper by Zhang et al that goes into a lot more detail about how EHRs are used in research, and I would recommend reading this if you are interested in this topic.

Interoperability

One final thing to note is that EHRs are not always interoperable. What I mean by this is that different systems do not always talk to each other. We have a variety of different EHR providers in the UK, and it is not always a simple task to get one set of patient data from one system to another (for example, if you move house and register with a new GP). One big advantage that we do have is that the coding systems are standardised across the NHS, so we are at least speaking the same language when it comes to recording data. Primary Care systems all use SNOMED-CT, and Secondary Care systems all use ICD-10 and OPCS. There is a lot of work going on to try and improve the interoperability of EHRs in the UK, but it is a complex issue and one that will take time to solve, not least because of the sheer volume of data that is generated, and the patchwork of EHR providers that we have.

Conclusion

I hope this post has given you a basic understanding of what an EHR is and how it is used in both clinical practice and research. I plan to write more on this topic in the future, so please let me know if you have any questions or if there is anything you would like me to cover in more detail.