Search
  • theopiletski

Healing healthcare processes with Intelligent Automation

Automation should be seen as a driver for competitiveness and new opportunities.

The healthcare provider marketplace is an ongoing sequence of consolidations, acquisitions, and mergers. The healthcare systems (think of networks of many thousands of hospitals and facilities) that are joining forces must plan for and manage the process of consolidating different clinical documentation systems. Electronic conversion of electronic health record (EHR) data from a legacy system to a target system and legacy data archiving are usually high priority considerations during the planned merging of systems. While the traditional digital conversion of clinical data is often more helpful vs the manual one, it has limitations, including the technical inability to convert certain kinds of data from a legacy system to a new EHR. Also, conversion efforts are often costly while depending on somewhat manual hence error prone labor, and many healthcare organizations have to balance clinical needs with budgetary constraints. Since I spend most of my time helping businesses improve efficiency and customer experience using artificial intelligence (AI) and intelligent automation (IA) in particular, I thought it is a perfect time to tackle the healthcare industry use-case to make it more cost-efficient, improve customer and employee experience/attrition and significantly improve profitability.



Problem: The challenge of Manual and/or traditional digital EHR Data Migration and Abstraction

Patient chart abstraction is the process of collecting important information from a patient’s medical record and transcribing that information into discrete fields or locations within the new EHR. Chart abstraction today is a manual data entry effort where organizationally-defined, clinically relevant data elements that are not being electronically converted, are collected from the legacy system and manually entered into the new target system. While there are some benefits to manually abstracting clinical information, it appears that the time has come to reuse the learnings from the efforts spent on automation of many other businesses using Intelligent Automation lessons learned and apply them towards EHR process, focusing on the activities found across the following areas:

Make discrete patient data readily available in the electronic chart, which allows providers and staff to care for the patient without needing to reference a paper chart or a legacy EHR.Allow triggering of decision support alerts related to the information entered during abstraction.Allow for faster decommissioning of the legacy EHR since there is less need to reference the legacy system.Mitigate risk to patient safety.



Make discrete patient data readily available in the electronic chart, which allows providers and staff to care for the patient without needing to reference a paper chart or a legacy EHR.

Approach: Utilize OCR technology and machine learning to extract, classify and match data from paper charts before moving them into new systems. Use Robotics (RPA) to automate the integration of data from multiple systems into a single electronic chart system or to attach and store images of poorly handwritten documents that cannot be processed by out-of-the-box OCR solution with a required degree of accuracy and confidence. 

Allow triggering of decision support alerts related to the information entered during abstraction.

Approach: Following data gathering using RPA, OCR, Web scrapping, and other tools, use machine learning based cognitive automation to evaluate the quality of the data, build decision algorithms specific to EMR/Clinical Data business process, to identify gaps and to route exceptions to be processed a by human into workforce orchestration tool.

Allow for faster decommissioning of the legacy EHR since there is less need to reference the legacy system.

Approach: Decommissioning of the legacy EHR today involves a long process of planning, preparation, effective governance and thousands of hours of manual work. Pre-populate a new EHR with legacy data using software robots, while making consecutive API (or robotics) calls to third parties like CMS.gov or AHIMA.org for external data about the particular record.

Mitigate risk to patient safety.

RPA + Personalization + Chatbots. Personalization and CRM software will help to prioritize records based on their importance, priority, and lifecycle to instantly identify, secure and accurately automate patient data migration 24/7 using RPA. And in case of inquiries, Chatbots will support simpler exchanges so that customers can get private, secure, faster, more satisfying customer service while receiving updates on their record status.

Keys to a Successful Data Abstraction and Migration Process using IA and AI

As healthcare executives begin planning for a new implementation or a merger of EHR systems, AI/IA should be added into their toolbox as well as in the planning and budgeting efforts at the project’s inception. Migration and Abstraction planning efforts should always take into consideration the following areas to leverage this technology successfully:

Start the planning efforts early in the process. This will help determine scope and budgeting requirements early in the project’s lifecycle.Decompose your EMR project as a step-by-step business processRun a proof of concept using multiple technologiesIdentify and empower key stakeholders to participate in the decision-making related to abstraction efforts. This should include both practice managers and clinicians - they are the ones to become your SMEs to support your automation efforts.Determine the scope of abstraction. What data will be abstracted if it was done manually, by whom and when? Are there any special abstraction needs for sub-specialties? Which patients will have their charts abstracted? it will help you to have a high-quality data for your machine learning automation.Define oversight/management/human-in-the-loop—who will manage the abstractors? How will they be trained? Will there be a quality review to ensure abstraction is accurate? Because automation level and accuracy is never at 100% and will evolve over time, make sure that your SMEs can uplift it to 100% manually if that's what you or your compliance team is looking for.



0 views

©2019 by AIST Tech