Editor's note: When specific software products are mentioned in this article, they should be considered examples, not endorsements. FPM requested pricing information for products for which it was not publicly available online. If no information appears in this article, it is because the vendor either didn't respond or declined to provide specifics. Several vendors said their prices vary based on how many clinicians a practice has, what EHR system the practice uses, what features are included, and other factors.

Administrative burden is one of the top sources of career dissatisfaction among doctors, with 58% reporting that bureaucratic tasks are their top cause of burnout.¹ Delegating more tasks to the health care team can help,² but finding qualified staff members was hard even before the pandemic and labor shortages are only projected to get worse in the near future.³

Fortunately, non-human help is on the way. Software programs that use artificial intelligence (AI), machine learning, and other forms of digital automation are poised to make several administrative tasks easier for physicians, giving them more time to focus on patients. Some of these programs can help decrease the time required to review the chart and document encounters. Others can ease prior authorization or calculate patient risk scores.

The following are five tasks that tech may soon help physicians complete more quickly and easily.

1. PATIENT RISK SCORE CALCULATION

As more payers move to value-based care, the need for physicians to have an accurate picture of each patient's health risks is growing.⁴ Comprehensive risk stratification includes both objective data and the physician's subjective assessment,⁵ but technology should be able to help with the objective side, especially now that most patients' health records are digitized.

Tech-savvy EHR users may already be able to mine the data in their patient records and develop algorithms that calculate patient risk with less clinician oversight.⁶ But new tools are being developed to do that for physicians who aren't expert EHR users.

Several programs on the market or coming to market can help find diagnoses buried in patient records and convert them into hierarchical condition category (HCC) codes, resulting in higher risk scores. Examples include RCx Rules, Navina, and ForeSee Medical. Several of these programs can be integrated with common EHR platforms.

Navina users showed a 23% increase in diagnoses found and a 38% increase in risk-adjustment scores in an initial evaluation by the American Academy of Family Physicians (AAFP) Innovation Lab.⁷ That program is now entering testing by a larger group of family physicians.

2. VISIT DOCUMENTATION

Advances in voice recognition software and AI have allowed tech companies to develop digital assistants that can handle a wide variety of tasks (think Amazon's Alexa or Apple's Siri). That same technology is now being leveraged to help with one of the leading causes of physician burnout: time-consuming visit documentation.⁸

Traditional voice recognition programs focused on dictation. Physicians had to place the cursor where they wanted the text before they could speak what they wanted typed. Those programs also required physicians to train the software to their voice (i.e., accent). Modern voice recognition software, leveraging AI, does not require up-front training, and a new class of product includes additional AI that can understand a user's intent. Instead of just typing what the physician says word-for-word, it listens for commands, such as “change Heart and Lungs to…” The AI understands the physician is referring to the section of the physical exam for the heart and lungs, and it is able to replace the current note text with whatever the physician says next. This frees physicians from navigating around the EHR with a mouse and keyboard. Instead, they tell the digital assistant what text they want and where.⁹

AI assistant programs include Suki ($149 per user per month for AAFP members, $199 for nonmembers), Robin, Notable, and Nuance Dragon Ambient Experience (a representative said prices vary based on too many factors to provide an estimate, but previously subscriptions were estimated at roughly one-third the cost of a human scribe's salary¹⁰). These programs act as sort of virtual scribes, continually learning how to better anticipate the needs of physician users and their patients. Vendors have generally tried to market these products as more affordable than human scribes, without any of the added concerns about hiring, training, or retention.

In an AAFP Innovation Lab trial of Suki that included 132 primary care clinicians, 60% opted to continue using the program as paying members after their 30-day free trial ran out. Those adopters reported a 72% reduction in their median documentation time per note, saving an average of 3.3 hours per clinician per week.¹¹ They also reported increased satisfaction with their EHRs, workload, and overall practice.

3. PRE-VISIT PLANNING

Another type of AI assistant can help practices prepare for patient visits. Reviewing charts before visits can take a significant amount of time, especially for patients who are medically complex. Clinicians or staff have to navigate the EHR to get to the data, even if they know where it is, and then hunt through notes and reports for data buried in unknown places.

An AI assistant can take over these tasks, reading through the entire patient chart automatically, including scanned and faxed documents, and creating a concise, problem-oriented summary of the patient's history with all of the data for a specific diagnosis grouped together. For example, for a patient with diabetes, the program would create a short list of the last A1C value and date, microalbuminuria test result (if available), ophthalmology note, etc.

The AI assistant can also provide a summary of care gaps for the patient and appropriate billing codes. The result is a document that distills years of medical records into something the clinician can open and review with a single click.

Navina, the program mentioned earlier for risk calculation, also has this pre-visit planning functionality. Navina users in the AAFP Innovation Lab's initial pilot saw a 70% reduction in physicians' visit prep time.¹²

4. PRIOR AUTHORIZATION

In terms of administrative tasks, one of the most common sources of physician frustration is prior authorization. Practices have to complete an average of 41 prior authorizations per physician per week, according to a survey by the American Medical Association.¹³ Two out of five physicians surveyed said they had hired staff to work exclusively on prior authorizations, and 88% described their prior authorization burden as “high” or “extremely high.”

It's not just a costly administrative hassle, either. More than 90% of doctors surveyed said prior authorization requirements delayed patient care at least somewhat regularly. These delays have serious consequences, with 18% of physicians reporting that prior authorizations had caused life-threatening adverse events for their patients.

Emerging technologies can help ease the prior authorization burden and prevent delays in care. Companies such as PriorAuthNow (now Rhyme), Verata Health (now a division of Olive), Experian Health, Infinx, Referral MD, and EviCore have developed software that uses artificial intelligence to learn various payers' prior authorization rules and automatically reach into medical records to pull supporting documentation for prior authorization requests. Most of these programs integrate with common EHRs and are available to demo now.

Developing and maintaining effective automated solutions for prior authorization is challenging because payers frequently change their plans and rules. But it's possible. An analysis of Verata Health's program by the KLAS Research health IT firm found that 65% of its users (whether physicians or staff) were able to get more prior authorizations done in less time with fewer denials.¹⁴

5. DPC MEMBERSHIP MANAGEMENT

In the last decade more family physicians who are frustrated with the administrative burdens imposed by insurance companies have turned to direct primary care (DPC), a membership-based payment model in which patients directly pay their physicians a fee (usually monthly) for primary care.¹⁵ There are resources available to help ease this transition,¹⁶ but DPC comes with its own administrative tasks, such as managing these direct contracts and up-front payments. New tech tools can help with those tasks as well. The AAFP Innovation Lab followed 10 physicians who transitioned from traditional fee-for-service payments to DPC using Hint Health's program, HintOS (packages start at $50 a month; the recommended package is $300 a month), to handle things like membership management, billing, invoicing, and payments. When HintOS users were asked how likely they would be to recommend the program to other physicians making the switch to DPC, the median score was 9.8 on a scale of 1–10.¹⁷

Other potential programs for handling DPC membership management include the Atlas.md EHR ($300 a month per clinician, with nonclinician staff users free), PrognoCIS EHR (starts at $250 a month), and HealthBiller (starts at $49 a month).

READY FOR ADOPTION

The administrative challenges to practicing family medicine are real and significant. But there is hope that a new set of technologies can address some of these challenges head-on. Adopting these technologies could help optimize the physician experience.