People who work with data tend to think in terms which are very structured and linear. They prefer to have B to follow A and C to follow B not just sometimes, but pretty much all of the time. But, healthcare data simply doesn’t work that way. Healthcare data can render linear analysis useless as it is both diverse and complex. There are several characteristics of healthcare data which make it both unique and often difficult to work with. Let’s discuss just a few of those characteristics.
Multiple Places
Healthcare data tends to be found in a variety of different places. The data tends to come from all over an organisation, from EMRs or human resources software to varying departments such as oncology, radiology or pharmacy. Compiling all of this information into one single central system can make it both more accessible and manageable. Programs such as SQL Training London can help you gain the required knowledge to sort very large databases and create a central system.
Healthcare data can also occur in a range of different formats. For example, healthcare data can be in text format, paper, digital, numeric, multimedia, or even videos. Departments such as radiology use images, old medical records tend to be on paper, and EMRs today can hold hundreds and hundreds of rows of both textual and numerical data. All of this can make healthcare data very difficult to deal with, especially when the data exists in multiple systems in different formats.
Structured and Unstructured Data
For years, the documenting of medical and clinical findings on paper has trained the healthcare industry to capture data in the most convenient way for the care provider. However, the problem with this method is that it has little thought for how the data can eventually be analysed, resulting in the data capture being anything but consistent.
Although EMRs are used to standardising the process of capturing data, many care providers can be reluctant to adopt a one-size-fits-all approach when it comes to documentation. Because of this, unstructured capturing of data is often allowed as a way of appeasing frustrated EMR users and avoiding creating hindrances in the process of care delivery. As a result, a large amount of the data captured in this matter is difficult to analyse and aggregate in a manner which is consistent. However as EMR products improve with users becoming more accustomed to standard workflows and care providers increasingly get more used to entering data in structured fields, data for analytics is expected to improve.
Inconsistent or Variable Definitions
Often, healthcare data can be made more difficult to analyse as it contains inconsistent or variable definitions. For example, one group of clinicians may define a group of asthmatic patients in a different manner than another group of clinicians or healthcare practitioners. If you ask two clinicians which criteria are necessary to identify a certain condition, you may well get three or four different answers. Oftentimes in the healthcare industry, there is a level of consensus about a particular treatment or definition of certain cohorts.
Along with this, experts are constantly discovering new knowledge even when there is consensus. As healthcare professionals and researchers learn more about the workings of the human body, an understanding of what is important begins to change along with ideas about what to measure, how to measure it, when to measure it and the specific goals which should be targeted. Although there are best practices defined in the industry, there is always discussion and change regarding the way those things are defined. This means that working with healthcare data often means that you’re trying to create order out of chaos and hit an unpredictable, moving target.
Complexity
Since claims data has been around for many years, it has been scrubbed and standardised. But, this type of data can be incomplete, with clinical data from sources such as EMRs giving a more all-round and complete picture of a patient’s story.
Whilst developing a range of standard processes which improve quality is one of the main goals in the healthcare industry, the number of data variables that are involved often make achieving this goal a huge challenge. When you work with healthcare data, you are not working with a finite number of identical parts in order to create identical outcomes, rather you will be looking at a series of complex, individual systems which make up the human body. Managing the data which relates to each of these systems and turning it into something useful across a population will require a far more complex set of tools than data management in other industries.
What is the Future for Healthcare Data?
Although healthcare data analysts are being provided with more and more useful tools to help them with certain aspects of working with the data such as Excel-Easy, healthcare data is definitely not set to get simpler in the future. If anything, the list of things which make dealing with healthcare data challenging will continue to grow. Healthcare is an industry which faces a set of unique challenges, and as a result, unique data challenges arise.
Because healthcare data is so complex, it’s important to recognise that many traditional methods of dealing with data that may work well for other industries may not work so well in healthcare. When it comes to an approach to healthcare data, one which can handle the unstructured sources, structured and unstructured data, inconsistency, variability, and complexity with constant changes is needed. The solution for the unpredictable change and complexity within the healthcare industry is an agile and flexible approach which is tuned and adapted for healthcare. Rather than thinking in direct, linear terms, those who work with healthcare data should have the ability to change directions fluidly when the environment does.
Understanding the core issues with healthcare data is important in order to learn how to best manage it. Along with that, having an understanding of the fact that some of these issues are never likely to change enables healthcare data analysts and managers to come up with solutions to make working within these limitations easier.
