Improve Outcomes at Your Organization with Healthcare Informatics
Institutions that understand how to collect, store, and share data while adhering to ethical and legal guidelines will have an advantage across many facets of healthcare delivery.
Collecting and sharing data allows providers and researchers to study diseases, especially rare subsets of patients. Everything from developing hospital algorithms to detect sepsis or clinical deterioration to running a clinical trial on a new medication requires rich curated data.
The amount of healthcare data collected continues to increase rapidly. Data from hospitals and clinics – along with information from clinical trials related to the testing of new drugs – comprise a large proportion of this information. In addition, there are increasing amounts of data being collected from people outside of the formal medical encounter – data from pharmacies, wearables, sensors, and patient-reported outcomes, to name just a few.
These data are extremely valuable both as a rich source of information to drive healthcare decisions but also as a monetizable resource for hospitals, pharmaceutical corporations, and insurance companies.
Given this rapid growth in the amount and kinds of data being collected, it has never been more important to understand how data are collected, stored, aggregated, transmitted, analyzed, used, and sometimes destroyed. This “data lifecycle” is critical to any healthcare organization, and being able to develop policies and procedures for how data are leveraged is critical.
The entire lifecycle of data is important. How data are obtained remains one of the most vexing problems for healthcare institutions. The electronic health record (EHR) is where most of the rich data are stored, yet there remains no standardized and efficient way to pull these data back out in an analyzable format. Free text remains the state-of-the-art method for cataloging information about patients, and this requires complex natural language processing for downstream extraction of meaning.
Increasingly, data are found outside of the formal EHR – genomic sequencing data, patient-reported outcomes, wearables data – and there must be a standardized way of including these data in any healthcare analytics project.
How we collect, use, and share data requires thoughtful consideration, beyond the technical components.
There are many legal, ethical, and regulatory issues that are important, including adhering to HIPAA requirements (in the US), obtaining patient consent (or a waiver) when collecting data for research, and ethical considerations about data use (and re-use). Any organization that will be successful in leveraging data must pay attention to these issues.
The analysis of data requires high-quality input information, which is why the process of collecting and aggregating data is so critical. Analysis could result in the generation of new algorithms that can be employed inside and outside the hospital – for instance, an algorithm to detect worsening sepsis or a way to predict which patients will be readmitted to the hospital shortly after discharge.
Understanding the different analytical methods as the assumptions and requirements for each will aid in understanding how these algorithms are developed and deployed.