Artificial Intelligence (AI) is a means of storing and rapidly accessing massive amounts of data with the ability to access and utilize the appropriate data for decision-making. Just as the human brain makes decisions based on its collective stored data, the more (quality) data that is available, the better will be the decisions. Since the AI technology has the capacity to store far more data than the normal human brain, AI specialists and experts are working to program this technology to sort, analyze and make better evidence-based decisions in the medical field.
For example, when used in medicine AI has given physicians the ability to diagnose illnesses such as cancer at earlier stages when it may be more treatable. By storing millions of normal versus abnormal radiographic images in its memory, an AI machine can detect an early abnormality in tissue much earlier than would be seen by the human eye. Based on a patient’s personal demographics such as age, genetic makeup, gender, etc., the AI system can then offer a treatment plan, subject to the review and approval of the physician.
AI is Enhancing EMR Software and Other Healthcare Technology
- While the advent of electronic medical records (EMR) has improved recordkeeping and communication between providers, providing a treasure trove of essential data, the downside is the enormous amount of time spent almost daily by providers keeping up with entering information into patient charts. The data demanded by payers comes with a huge cost of provider time and energy, which many physicians believe should be spent instead on delivering care for patients. EMR has all too often had the unintended results of reducing patient “face-time” as well as valuable time spent conferring with colleagues.
- With the help of AI, machine learning can relieve providers of hours of data drudgery with programs such as TalkEHR, which uses voice-recognition technology to provide automatic charting. Among the benefits of this technology is the reduction of a backed-up inbox, with its redundancy of note-making and alerts – much of which doubtless contributes to an increase in physician burnout. AI can help providers prioritize and remove some tasks from providers’ inboxes, such as prescription refills or results notifications.
- Advanced text analytics as well as more sophisticate voice-recognition software will be able to analyze the basic unstructured data in medical records to aid providers and other clinicians in accurately extracting events and other factors. Some natural-language processing tools that are already on the market are able to not only recognize certain words but distinguish their meanings and relationships to each other. EMR analytics have successfully produced risk scoring and stratification for identifying database relationships.
- AI technologies are now able to create dictionaries and ontologies, customized for the specific user(s) while able to identify complexities of entity relationships between patient records and the subsequent bills. These technologies are especially useful to recovery auditors and can include the following: drugs and their dosages, ICD codes, CPT codes, medical and laboratory tests.
- As a result, a clinician is now provided with a timeline from the medical records, documenting various health care encounters, providing that these lead to the charges within the medical bills. Information contained can then be cross-referenced from the records for certain line items in the bill, resulting in better patient care while reducing underlying costs.
Will AI Replace Doctors?
Not everyone is convinced that AI is always benign or that it won’t inevitably result in replacing some healthcare workers through the automation of their jobs. As will be the case with most doctors and other healthcare professionals, most researchers believe that AI will augment rather than supplant coders’ abilities to code and bill with more precision and accuracy, ensuring better reimbursement rates.
However, it is likely that automation will inevitably displace some less-experienced or skilled workers, particularly in the back office, as automation takes over many of their more routine duties, according to an Infosys report. Establishing ethical standards, then, is necessary prior to deploying AI.
AI’s Cautions and Concerns
Already, medical coders are using computer-assisted coding (CAC) to create an abridged version of patient records rather than wade through the entire collection of data. While this certainly speeds things up, mistakes can still happen: due to time constraints to process as many charts as quickly as possible, accuracy is put at risk.
When creating predictive models, developers need to beware of using bad data, such as what might be extrapolated from the EMR, resulting in worse models. Human intervention and oversight is therefore critical at every stage of development and usage of AI in a medical setting. Ideally, physician and machine will form a partnership, with AI doing what it should do best: provide accurate predictions – with the physician explaining, interpreting and taking action.
How Does a Medical Billing Service Fit into the AI Picture?
Programs such as TalkEHR use patient medical and personal information, including age, gender, test results, examination and prescription history and diagnoses as well as the provider’s previous practice to determine the most appropriate procedure code for a particular patient visit. However, as any experienced coder will probably attest, coding simple patient charts are not the best use of their time when they could better focus on tasks beyond the ability of most machines. While AI, at least for now, promises to save time and aggravation for providers, it still falls to billing staff to verify accuracy and compliance to avoid interruptions in the revenue cycle.
Related Article: 3 Effective Tips to Enhance Patient Engagement with Medical Practice
This is where a good medical billing service, such as M-Scribe, really shines in the ability to quantify certain factors and use AI and other technologies to identify and extract data from documents for coding and billing. The need for experienced coders and billing analysts isn’t going away anytime soon, as there must be human oversight and monitoring of EMR’s AI technology.