How Intelligent Document Automation Can Improve Clinical Communications

By Chris Larkin

Each time a patient interacts with a medical professional, a record of the visit is created and stored within that medical facility’s recordkeeping system. When the patient is transferred, that record must accompany the patient to inform the next stage of care. The transfer of these records between the electronic health record (EHR) systems used in different care settings can be a cumbersome task. Each provider’s EHR has been customized in terms of what data is needed as well as how it must be formatted in the database to be usable. Furthermore, much of a patient’s story is captured in non-standardized narrative format within discharge summaries, chart notes, visit summaries, consultation reports, and physician orders.

The easiest way to transfer records between incompatible systems is to generate those records in document format. To send those documents, most providers still resort to using an antiquated piece of technology that is found in almost all medical offices: the fax machine.

Fax provides a tried-and-true solution in that documents can be delivered from one location to another without the need to adopt new technology or worry about another office’s system compatibility. Unfortunately, there’s a major downside: The recipient of the fax is left with a paper document that needs to be entered into the system manually.

To make this practice simpler, safer, and more accurate, healthcare offices need to transition to a seamless digital exchange of the patient’s protected health information (PHI) into the EHR, but this progress is hindered by a number of factors, including the lack of full interoperability between systems.

The many drawbacks of the fax machine

Patient clinical encounters are accompanied by a wide range of documents, including prior authorizations, admission and intake forms, records requests, and discharge plans. Fax is the most frequently used method to receive inbound documents from external providers and payers. According to a survey of 400 healthcare professionals commissioned by Concord Technologies in December 2019, the majority of inbound clinical documents are filed to patient charts in the EHR manually. This workflow involves staff scanning documents, uploading them into the EHR, and then rekeying relevant patient information from the scanned images into the system.

This labor-intensive work is tedious and prone to error, and it negatively impacts workflow, staff efficiency, physicians, and patients in several crucial ways:

• Allows patient record errors, including filing or documenting information in the wrong patient file, data entry errors, and improper documentation that can lead to gaps in the patient record
• Creates gaps in communication during transitions of care from one healthcare provider or setting to another
• Results in poorly documented or lost test results
• Creates possible compliance and patient privacy issues
• Offers subpar engagement for both physician and patient
• Generates less-than-optimum patient outcomes

In addition to these inherent challenges, multiple teams and departments are responsible for working with clinical, administrative, and financial documents coming from numerous healthcare organizations. As a result, inbound documents may contain information that is not needed by a specific employee, while some large documents may contain multiple individual documents that must be separated and evaluated by different people. This redundancy can mean that too much information is given to recipients who do not need it, causing further processing delays and potential patient privacy failures as relevant files are separated and redistributed.

Gaps in clinical communication equal gaps in care

Transitions of care are a particularly significant source of error. Information flow is critical within service areas, yet often, necessary information does not follow a patient during a transfer or discharge, especially from surgery to long-term care facilities. In addition, home health is scaling up to become an alternative to lengthy hospital stays as a means of saving costs, decreasing hospital-acquired infections, and improving outcomes. As these options become more commonplace, full integration of data into EHRs will become even more critical.

According to the Agency for Healthcare Research and Quality, communication problems and inadequate information flow are two of the most common root causes for medical errors. In the postacute care (PAC) environment, 83% of facilities report communication problems directly related to receiving inadequate patient information from referring hospitals during the referral process. One in three PACs report they do not receive any documents from a referring hospital, and fewer than one-third of patient discharge summaries ever reach the PAC team.

An inadequate information flow can cause problems that affect the availability of critical knowledge needed for prescribing decisions, timely and reliable communication of test results, and coordination of medical orders at points of interface or transfers of care. These communication problems can occur among members of a healthcare team in a single location, between providers at different locations, between healthcare teams and other nonclinician providers (such as labs or imaging centers), between healthcare providers and patients, or at any other point along the communication chain. The resulting administrative and medical errors raise healthcare costs and may lead to inferior patient outcomes, including patient harm and readmissions.

Analog to digital: A modern solution to a persistent problem

Healthcare organizations need a solution that will enable secure, efficient information exchange among numerous, highly customized EHRs and financial and administrative systems, each with its own workflow, processes, and preferences for document processing. One solution is intelligent document automation technology that includes cloud-based fax, digitized document workflow, and data extraction using artificial intelligence (AI) capabilities. Replacing traditional analog fax communication with secure, encrypted, HIPAA-compliant delivery of PHI via a cloud-based platform ensures reliable data transmission to and from the EHR and other healthcare applications while eliminating the costs of traditional on-premise fax infrastructure. Digital fax is not only compatible with existing health information technology, but it also allows providers to reduce manual work and increase accuracy in document processing. This leads to both improved clinical care quality and more effectively distributed administrative time and costs.

AI: Automating document processing for better clinical outcomes

The AI-driven data extraction and classification technology, when implemented in conjunction with cloud-based fax, can automatically identify different healthcare documents and extract patient information, which is currently locked in transmitted/faxed documents. This data includes patient name, Social Security number, medical record number, and date of birth, as well as physician name, national provider identifier, and date of service. These critical identifiers are needed for correctly processing inbound referrals, physician orders, clinical notes, and medical records requests. Extracting this information allows for intelligent indexing of incoming documents so they can be linked to, and accessible from, the appropriate patient record.

The benefits of this approach for clinical care are many. For instance, faxed physician orders that are printed out and processed on paper can easily get lost among numerous other documents. Centralized queue inspection in a secure web portal helps ensure that all incoming documents are properly categorized and labeled so they can be handled in order of priority. For example, if a digital document is clearly labeled as a physician order for Chris Larkin, a staff member can process it much easier and faster compared to opening and reading it to understand what it is and who the patient is. When physician orders are processed accurately and in a timely fashion, the patient is better served.

Regarding prior authorizations, the sooner the provider knows if a procedure or therapy is approved, the sooner the treatment can begin. With intelligent document automation technology allowing for faster information processing, the wait time can be reduced. This is especially vital for patients with comorbidities and in serious clinical cases where time is a determining factor in positive clinical outcomes.

Building on the current advances in document processing, AI technology is poised to add value beyond automation and improved communication. For example, AI-enabled document automation software can analyze faxed documents for patterns on a large scale while it rapidly processes a wide sample of patients. For example, it can search documents for keywords, such as “acute” or “allergies,” or organize patients by other criteria such as ZIP code, and act based on findings in the data. Documents contain specific information that should raise or lower their priority. Machine learning technology can securely and privately read the document and assign priority to it based on the content to speed up clinical and administrative processes.

Another area where AI holds great promise to improve outcomes is helping to avoid “false negatives” by scanning a patient’s medical records over time and finding significant combinations of symptoms. For example, if complaints of fatigue, general malaise, and shortness of breath exist in the records of patients with a high risk for congestive heart failure, they can be addressed more quickly, allowing critical early interventions to take place.

The next step in medical communications

The key to realizing the benefits of this new technology is to get it to fully support the daily work of clinicians and administrators, facilitating information exchange for processing prescriptions, medical orders, billing, reporting, analytics, research, and much more. With the use of AI and natural language processing techniques, providers can streamline administrative and clinical processes, and lift the administrative burden on the healthcare system. Accomplishing this goal will mean that manual entry and human error can be reduced significantly, if not eliminated altogether. The information contained in outbound and inbound documents can be prioritized, addressed, processed, and delivered appropriately, speeding clinical and administrative processes.

The potential benefits of using AI technology for document processing are far-reaching, promising the creation of complete and accurate patient records that follow individuals throughout their care continuum for improved outcomes. While the technology will never replace the human element that is so essential in the patient and caregiver interaction, it will alleviate a tiresome administrative burden and give medical personnel more opportunities to embrace sensitivity, creativity, and sincerity that only human beings can offer.

Chris Larkin is chief technology officer of Concord Technologies, a leading provider of cloud-based fax and intelligent document automation software for healthcare providers and enterprises in regulated industries.