Collecting healthcare data is difficult for a variety of reasons. Where the industry is right now, healthcare data is “owned” by a variety of people, and in a variety of ways. The necessity to make changes to the industry as a whole requires changes that aren’t simple and won’t be met with open arms by all those involved with inputting of data.
In electronic medical records and other forms of healthcare databases, data is recorded both numerically and textually. However, much of the data has a long history of being written out. Converting this data, which is inaccessible to most healthcare professionals due to being located a physical site, requires the time, effort and dedication to transfer into digital form.
To understand better these two different forms of data, numerical data is defined as anything containing a simple number, such as date of birth, age, height, weight, temperature, blood pressure and amount of medicine being administered. Textual data is data in all other forms, including all written information, MRI images, x-rays, etc. This sort of data is much more difficult to delineate into specific areas or categories because it isn’t always an absolute, like numbers are.
This is where we introduce systems like the International Classification of Diseases (ICD), which assigns numbers and codes to different textual data in the healthcare industry. For example, the current ICD-10 code for the Ebola virus is A98.4.
This coding system will certainly help to define healthcare data more precisely, and allow it to all be readable by any system. This is the type of uniformity that is crucial to produce a readable and analytical system. However, making sure that all healthcare professionals are on the same page can be a different story altogether, and adds to the obstacles that make collection of data difficult.
In addition to the two major forms of data (which are hopefully more easily defined by the ICD), data can also come in the form of doctor’s or nurse’s opinion, which cannot always be categorized into the coding system. There certainly are standards in healthcare practices, but there is also variation in the ways doctors might diagnose something or note specifications of a patient. One hopes for consistency across the board, but that can’t happen when from practice to practice, from doctor to doctor, or even from patient to patient there are personalities and uniquenesses that make this impossible.
On top of the obstacle of data collection and varietals, there is also the legality in collecting healthcare data. Different providers “own” their own data. The patient has gone to that doctor and agreed that that doctor/practice/hospital can have their data, but not anyone else. Confidentiality agreements are signed, and the data cannot be shared unless the patient agrees to it.
Undergoing this data transformation is a huge undertaking for everyone in the healthcare field. It means learning new systems to track this data, learning new codes for doing so, and learning how to use and measure this data. While it is certainly a part of the solution, it means doctors, nurses and other healthcare experts will need to learn what is virtually an entire new language to record their data. This will be a challenge, but one that has the potential to benefit both patient and professional.
Though the current system of data collection for healthcare information is not ready for the diversity in how facts and opinions are recorded, technology specifically focused on this classification of data is moving forward and may one day allow for the variety of textual data and still produce a plethora of usable analytics. Until then, finding standardization within the system is necessary and will turn out very useful information that can be beneficial to everyone involved.