In order to understand the relationship between evidence based practice and informatics, it is important to evaluate what evidence practice is. One definition of evidence based practice is “the conscientious, explicit and judicious use of current best evidence in making decisions” (Ammenwerth, 2015). This definition was created to describe evidence based medicine, where both external evidence and individual expertise help guide a provider to make the best clinical decisions. It is a method that is used to improve the patient care that is administered by the provider. Health informatics also seeks to improve health care, however it does so through the use of technical interventions instead. Evidence based decision making can be applied to health informatics as well, where evidence on efficiency or effectiveness can be used when making decisions regarding implementation of certain health IT systems. This is important because health IT “indirectly affects the patient by influencing clinical processes and clinical decision making in a vivid health care environment” (Ammenwerth, 2015).
Evidence based health informatics begins with the current based evidence. This evidence is scientific evidence that has been derived “from well-designed systematic evaluation studies or systematic reviews” (Ammenwerth, 2015). These studies have to evaluate health IT starting with the development phase through needs assessments, test runs, and simulation studies. Additionally, early use and routine use reviews are needed for evidence based informatics. This means that performance measurements, usability studies, usage pattern analysis, and cost analysis are important to evidence based health informatics as well. The result of evidence based informatics is that many guidelines are developed with existing evidence in mind.
As evidence based guidelines in health informatics is examined, HIE can be used as an example. As we enter the age of standardized use of electronic health records and care becomes more and more specialized, electronically sharing data from multiple points of care is advantageous to many providers. HIE is “the key component of health informatics through which information from various electronic record systems is shared and is potentially transformative for the healthcare system” (Braunstein, 2014, p. 55). Development of HIE began due to the challenges faced when exchanging data between different systems as a result of changing and interoperable interfaces. While care coordination has been considered an essential concept for managing chronic diseases, in practice it is challenging when “every provider has their own EMR, and that EMR can come from any one of the hundreds of companies certified for Meaningful Use” (Braunstein, 2014, p. 57). While these electronic record systems were certified for use by the Office of the National Coordinator, they did not work well together, which is what created these challenges. As the government began to recognize the importance of HIE systems, a greater movement towards interoperability guidelines were placed. Through EHR practice, governing agencies have been able to refine guidelines that benefit interoperability and support the development of HIE. One example of this occurring is this March, where the Center for Medicare & Medicaid Services released new rules that established “policies that break down barriers in the nation’s health system to enable better patient access to their health information, improve interoperability and unleash innovation, while reducing burden on payers and providers” (CMS.gov).
Some existing guidelines for HIE are set by eHealth Initiative, the Health Information Management Systems Society, and the Office of the National Coordinator for Health Information Technology. These guidelines focus on the classification and goals of the HIE, such as the architecture of the systems and function of the systems. The architecture of HIEs include Centralized, Federated, and Hybrid, and these classifications denote where the data is stored and how it is accessed. In the centralized model, the data is stored in a central repository. This differs from the federated HIE, where all of the clinical data is stored at the source. Hybrid HIEs use both storage methods for the data, typically having a patient index that directs to where the data is stored. Functionality is classified through how the data is exchanged and the three classifications are directed exchange, query based exchange, and consumer mediated exchange.
While these classifications adequately cover the data storage and exchange methods that are used for HIEs, improvement still needs to be made since they do not fully cover the scope of HIE. One way to find areas of improvement is by examining existing HIE systems, such as the Indiana Health Information Exchange, and evaluate their systems and outcomes to help refine existing guidelines for future HIE. The Indiana Health Information Exchange is an excellent example for this because it is “the country’s largest HIE” and considered “the premier example of centralized HIE” (Braunstein, 2014, p. 64). This system uses a centralized model to offer a variety of services across the state, such as a portal for lab results and clinical reports, a community health record, a population health management system, a diagnostic-imaging sharing service, and analytics tools for managing care. The IHIE has many value adding services, and these services are one of the ways that this HIE stands out. The guidelines for HIE should be adjusted to include service specific information. For example, if one were to create an HIE that had the services of result delivery and surveillance of population health for COVID-19, guidelines for services that deliver test results to patients and services that aggregate population health data would assist in the development of the HIE.
Practice based evidence is different from EBP because it is scientific evidence that is developed in a real world setting and it is an important research method. In one study where the PBE method was used to explore traumatic brain injury inpatient rehabilitation practices, the researchers were able to look at “information on the types, intensity, and duration of key activities used in interdisciplinary rehabilitation using a separate taxonomy for each discipline” (Horn et al., 2015). This is important because they were able to gain more detailed information on the processes used in addition to the outcomes. In health informatics, this is an important method of research because understanding the processes of the systems that are being developed is just as important as the outcomes that are produced. Understanding the processes will help with efficiency and effectiveness with the development of future projects. Practice based evidence can also be applied to researching existing projects because we can study the “practice” or the systems themselves.
Data mining is the extraction of information from data and it “implies in-depth searching to find additional information that may have previously gone unnoticed during a more routine analyses of the data” (Penny & Smith, 2012). This process is an increasingly popular method for research as computer systems have become more capable of processing large amounts of data. In the focus area of health information exchange, HIEs can be used as tools to aid in the data mining process. For example, the data from a statewide HIE such as Indiana Health Information Exchange was aggregated, we could use data mining to research questions like what populations in the state are hit hardest with COVID-19, and if wealthier counties have fared better than poorer counties during the pandemic. Since the aggregated data from this hypothetical HIE should include anonymized location data and COVID-19 test results, it would be a good data source for data mining.
One issue with data mining and HIE is ensuring that patient data is protected and remains private. However, there are measures that can be taken in order to ensure patient information is protected. One measure is to use secure multiparty computation, which encrypts data, as a result “ensuring that no party learns anything about another’s data values” (Clifton, Kantarcioglu, & Vaidya, 2002). Another measure is to obscure data, which is done by making private data available, but with enough noise added that exact values (or approximations sufficient to allow misuse) cannot be determined” (Clifton et al., 2002). A common way of doing this is aggregation, where there is not enough information to get an individual’s data with only the information from the community level. Each method has its benefits and downfalls. Secure multiparty computation allows for more specific data sets and high levels of privacy but is very costly and difficult to achieve, while obscuring data is much easier to do but you are left with less specific population level data. However, when used in the case of public health research and population level data is the goal, then patient privacy from data mining HIE is achievable.
Clinical decision support systems are important because they can help increase quality of care while decreasing the errors. This is done by delivering “information tools to the point of care” (Hills, Lober, & Painter, 2008). In HIE, this is achieved by ensuring “records follow the patient and clinicians have access to critical health care information when treatment decisions are being made” (Hills et al., 2008). Health information exchange makes it easier for patient data to be accessed by providers at multiple points of care, which helps them make decisions based on the patient’s medical history. For example, when prescribing a medication, access to a patient’s full medication would help the provider avoid errors.
References
Ammenwerth, E. (2015). Evidence-based Health Informatics: How Do We Know What We Know? Methods of Information in Medicine, 54(04), 298-307. doi:10.3414/me14-01-0119
Braunstein, M. L. (2014). Contemporary health informatics. Chicago, IL: AHIMA Press.
Clifton, C., Kantarcioglu, M., & Vaidya, J. (2002, November). Defining privacy for data mining. In National science foundation workshop on next generation data mining (Vol. 1, No. 26, p. 1).
Fact sheet Interoperability and Patient Access Fact Sheet. (2020, March 9). Retrieved October 01, 2020, from https://www.cms.gov/newsroom/fact-sheets/interoperability-and-patient-access-fact-sheet
Hills, R., Lober, W., & Painter, I. (2008). Biosurveillance, Case Reporting, and Decision Support: Public Health Interactions with a Health Information Exchange. In Biosurveillance and Biosecurity (Vol. 5354, pp. 10–21). Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-540-89746-0_2
Horn, S. D., Corrigan, J. D., Bogner, J., Hammond, F. M., Seel, R. T., Smout, R. J., … & Whiteneck, G. G. (2015). Traumatic Brain Injury–Practice Based Evidence study: design and patients, centers, treatments, and outcomes. Archives of physical medicine and rehabilitation, 96(8), S178-S196.
Penny, K. I., & Smith, G. D. (2012). The use of data-mining to identify indicators of health-related quality of life in patients with irritable bowel syndrome. Journal of clinical nursing, 21(19-20), 2761–2771. https://doi.org/10.1111/j.1365-2702.2011.03897.x