Today, medical practitioners are analyzing big data using algorithms to improve patient care. Data mining helps identify treatments that are effective for particular conditions, identify patterns related to drug side effects, and a lot of other useful information that can improve the quality of healthcare. However, the challenge that health practitioners face is with regard to data entry. Entering enormous amounts of data can be really tedious for busy physicians. Fortunately, modern technologies are helping in addressing the challenges of big data entry.
Limitations of Online Clinical Decision Support Tools
Patient data include critical clinical data such as symptoms, physical signs, orders, and progress notes, medical test reports, billing details, and more. Digging such data to find useful information can be done with the help of online decision support tools or â€˜Big Data Algorithmsâ€™. There are more than 50 freely available clinical decision support tools including clinical calculators and medical e-tools, scores, scales, tables, and more.
Algorithms include bradycardia treatment algorithm, GI bleed complication risk, hypernatremia treatment algorithm, myocardial infarction probability (Goldman), and much more. Similarly, you have free calculators for alveolar-arterial oxygen gradient (A-a gradient), anion gap, blood alcohol concentration, blood oxygen content, and so on. E-tools are available to calculate Alvarado appendicitis score, Apache II score, bleeding probability after TPA for MI, and coronary disease probability. Physicians can also find tables for several conditions, for instance, endocarditis prophylaxis pertaining to antibiotic regimens, which cardiac conditions require it and which procedures require it.
These algorithms and e-medical tools help identify the physical condition of a patient and even predict the likelihood of death because of a condition. However, they have many limitations:
- A lot of patient data has to be entered for driving a result. For instance, 156 data elements were used to develop an algorithm for diagnosing heart attacks, with the final model using 40 data elements! Entering such massive amounts of data is difficult for physicians managing an overcrowded emergency department.
- Feeding data from the Electronic Health Record (EHR) into the decision support tools is not a feasible option. Often, there is confusion about the type of data to be entered or whether the needed inputs are recorded in the right format and with the right details. Sometimes, inputs are missing.
Computer-assisted Physician Documentation to Harness Big Data
‘Computer-assisted physician documentation (CAPD)’ is one of the latest technologies helping to resolve the issues associated with big data algorithms. It uses real-time natural language understanding (NLU) to convert clinical notes to computer-readable codes. The main advantages of CAPB are:
- Automatically populates decision support algorithms
- Real time decision support by automatically identifying a missing input and immediately prompting the clinician to add it
- Produces more specific ICD-10-compliant clinical documentation
- Has the ability to identify missing data by grabbing it from another source or inferring it from existing data, or prompting the clinician to add it during the normal course of documentation
- Saves time by auto-generation of the doctorâ€™s notes, billing notes, regulatory reporting codes, and a provisional set of orders based on the doctorâ€™s notes.
Another popular technology used in healthcare for addressing data entry issue is â€˜voice recognition softwareâ€™, which allows physicians to enter clinical note by speaking rather than typing. Though, reduced turnaround times and timely documentation are touted as the main advantages of this technology, it can result in erroneous entries. Clinicians still rely on medical transcription services to correct and edit the documents generated by voice recognition systems.