Did you ever feel that doctors, nurses, CDI and Quality Specialists need to process too much information (TMI)? This May newsletter is dedicated to those who are overwhelmed on a daily basis with information overload and how we can help them. Computers in this digital age of medicine can now support advanced analytics and machine learning (what most people refer to as “Artificial Intelligence” or “A.I.”). Hiteks accomplishes A.I. through advanced search with a proprietary CDI and quality relevance ranking; we are the only vendor which allows CDI and HIM functions to keep full control of their rule-based knowledge. As this year’s Epic XGM meeting unfolds and we prepare for the annual ACDIS conference, we recognize that point of care physician and nurse advice specific to the patient being treated and prioritized for CDI and Quality review, is going to be key to the functionality of medical records systems. Enjoy the articles and we hope to see you May 21-24 at the San Antonio ACDIS meeting!

Voice Recognition versus Text Analytics (aka Natural Language Processing -NLP) in the age of TMI
There are two Holy Grail technologies for medicine:
1) The first is voice recognition which involves a speech engine to process human language spoken into a microphone. The speech recognition technology then captures the voice and applies phonetics to decipher the accent and intonations which then translates the human voice to written text. So the engine attempts to represent what the voice said in writing. This process involves a class of technology which has improved tremendously over the past 20 years and a small number of vendors currently have majority of market share in health care. Physicians still need to review and finalize the transcribed text, but this review can be done most efficiently in the EHR where the notes are brought into the context of other typed or transcribed notes along with the discrete EHR data (vitals, labs, meds) important to reference to finalize the note. Physicians prefer to receive CDI suggestions right in the EHR in the context of their note review process to speed their documentation of the clinical care as soon as it is provided.

2) The second technology is text analytics or Natural Language Processing which involved analyzing documents or snippets of text which contain information that needs to be prepared for coding and to be understood by computer system for further processing. The NLP technology class of tools has nothing to do with Voice Recognition technology, but vendors who sell both want to increase their sales and profits by attempting to combine them (often ineffectively) into one integrated technology that does both Speech Engine plus Text Analytics.

The reality is that each class of technology is much different and involves different support, engineering and product roadmaps. Furthermore, companies who provide speech engines are already overwhelmed in supporting a complex product with constant security and interface needs required to keep them running. Due to the standardization of workflows now in EHRs, health systems can choose best of breed technologies for each technology class instead of trying to bundle vendor solutions.

In fact, if they choose Hiteks for their best of breed NLP, they get the added advantage of advanced search using Hiteks’ proprietary engine which applies a CDI and Quality relevance ranking to return important advice back to the physicians. Health care providers should be weary of using NLP from vendors who purchased the technology from legacy companies or licensed it from academia. This legacy and academic NLP is not optimized for commercial use nor the multiple use cases for which it is required to function. Providers need the best functioning system for text analytics due to its ability to help in clinical decision-making at the point of care which impacts quality and revenue. NLP can be applied in real-time to facilitate CDI advice, search, coding, and research, or a combination of these. Some vendors like Hiteks includes all of the functionality to support each of these use cases and enable enterprise-level capability for health systems.
Purchasing Decisions around EHR-integrated NLP solutions to deal with TMI
Healthcare purchasing managers are often not domain experts in each of the technologies that they are asked to help purchase. When they conduct RFIs and RFPs to identify new technology, is it important for vendors to provide enough education for purchasers to get a comprehensive understanding of the possibilities of new technology such as natural language processing.

This includes how NLP is different from voice recognition and natural language understanding which may be used to improve speech recognition. To avoid the painstaking task for purchasers to define the questions for NLP and Search, we provide these examples and references to show how real-time NLP and Search impacts physician and nurse advice at the point of care through computer assisted physician documentation. The questions around functionality include accuracy, speed and configurability.

Configurability involves being able to specify different parameters or rules that a computer would use to process information and apply algorithms specific to different diseases or document sections such as pediatrics, orthopedics, dermatology, etc., without having to change the software code which often requires many months of software QA and validation. Hiteks and EHRs like Epic also present Computer Assisted Physician Documentation (CAPD) and Clinical Document Improvement workflow innovation negating the need for back-end CDI, coding and Quality tools. Hiteks’ Insight is the only solution that allows CDI, Quality and HIM Functions to work entirely within their EHR system keep control of their own documentation integrity and coding rules, helping organizations to shorten their revenue cycle time and increase health systems revenues, Case Mix Index and Risk Adjustment Factor scores.

Other coding and voice recognition vendor solutions use slow legacy technology to force the automated query and alerting rules onto an organization with too many false positives and no option to configure them per factors such as patient age, insurance type, CDI Handbooks, Coding Clinics or accumulated experience of the health system. Physician CDI and Quality advice are most efficient when done in real-time at the point of care. A blazing fast system response time of 1 second immediately suggests additional clinical clarifications to add to existing documentation ensuring the highest physician response and compliance.

Once any designated documents like Progress Notes, Radiology Reports, Echo Reports and Pathology Reports are saved in the EHR or brought in through transcription for review, Hiteks’ NLP and Artificial Intelligence engine immediately checks documentation for missed diagnoses in Pediatrics, OBGYN, Orthopedics, Emergency, Dermatology, ICU and other specialties, along with other general conditions.

The system presents the evidence in the same screen as the Notes and Reports, along with any labs, vitals or medications for easy review saving time. This way evidence for improvement is clinically aligned so that the physician and CDI Specialist can do their review concurrently for faster, more accurate ICD-10, CPT and Quality coding and alerting. With over 90% system accuracy, addendums are prevented later and both CDI and Quality Specialists are literally on the same page as the physician ensuring consistency when reviewing Patient Notes and prioritized diagnosis work-lists.

Physician documentation time inside the EHR is reduced by 40% and Hiteks has double the impact on reduction of manual queries than competing vendors, so doctors can spend more time with their patients. Get started with Hiteks at or click on “Get more info” within Epic’s App Orchard for ConcurDI For NoteReader CDI and AdvocateMD Powering NoteReader with advanced workflow, financial and clinical metrics reporting.   VIEW PDF
Personalized Medicine using Digital Tools Integrated with EHR Workflow to address TMI
You have heard that the EHR swaps which took place during the last 15 years are no longer occurring because decision-makers realize that the main vendors that have established market share (Epic, Cerner, MEDITECH) all provide the same basic functionality and the workflows are being optimized in each of the systems along with the addition of analytics to make the data more useful for decision-making.

The difference will be in how well these EHRs interface with analytics providers. The analytics of the actual data collected and stored in the EHRs is the focus of Hiteks and here are some helpful descriptions of how you can leverage and optimize the analytics for personalized medicine.

Because drug and medical device manufacturers are making their products very specific to individual patient types and classification of diseases, it is important for a computer system to allow for analyzing medical record data in real time to incorporate all of these personalized aspects of the drugs and medical devices which requires more data to better tailor their technology to the indications approved by their products. This includes not only structured medication or lab data but also diagnostics through pathology reports, radiology reports, progress notes, encounter notes, etc., to appropriately match the complete criteria that are required for the indications of the drug or medical device directly to the patient. This match best occurs in real time when the physician documenting their care reviews the patient chart for applicability of any innovative medicines or medical devices with the patient at hand.

The key distinguishing features about Hiteks’ approach to data analysis compared to other data providers: all the clinical indicators are analyzed including laboratory results, treatment course, vital signs, and clinical findings. To achieve truly personalized medicine, multiple data points are required by the analytics system to process. These include the following common examples:

1. Wearables EKG Pulse Pulse Ox Glucose 24hr ambulatory BP Sleep monitoring EEG Autonomic nervous system
2. Genetic testing Patient genomics Microbiome genomics Tumor cell genomics
3. Microbiome culture Example #1 of TMI in Hypertension requiring multiple data points accessible and analyzed by the computer to make appropriate suggestions to physicians. Hypertension’s incidence is 29% in American adults over the age of 19, according to CDC:

Current phenotypes for Hypertension determining different treatment protocols include the following list, all requiring knowledge of patient response to various medications and diagnostic tests. The text analytics technology needs to synthesize and reconcile data points from Notes, Meds, Diagnostics, and Vitals to ensure the appropriate diagnosis is presented to the physician. Ranking the results of the analytics through Hiteks' relevance ranking allows for this.

1) Low renin: Treated through Diuretics, CA Channel blockers
2) High renin: Treated through ACE inhibitors, ARB
3) Evolving phenotypes determining different treatment protocols
a. Dipping (normotensive at night)
b. Non-dipping (hypertension throughout 24 hr cycle)
c. White coat hypertension: Abnormal only in physician office, normal in everyday life
d. Sleep apnea induced: Treatment required for sleep apnea
e. Arterial: Treated through Beta blockers, CA Channel blockers
f. Arteriolar: Treated through Nitric oxide supplements
g. Endothelial dysfunction: Treated through Anti-inflammatory agents, lipid-lowering agents, Glycemic-lowering agents
Follow us


Contact us
Copyright ©Hiteks 2018, All rights reserved.