Each year, Medical Economics surveys physician readers to find out what irks them most. Topping the latest list: insurance paperwork, followed closely by electronic health records (EHRs).

The reason is the same for both. Insurers and EHRs get between doctors and their patients.

When it comes to medicine’s computer problem, the obstacle is literal. Doctors sit behind a screen, focused on the EHR and not the patient. After hundreds of clicks and little eye contact, the visit ends with both parties feeling dissatisfied.

Hoping to bridge this digital divide, a host of companies — from health-tech startups to multinational conglomerates — have promised to ease the way doctors enter, organize and retrieve medical data.

But to make sense of these high-tech solutions, it’s important to understand the origins of healthcare’s vexing IT problem.

EHR burnout: Who started the fire?

Let’s hearken for a moment to a simpler time in American healthcare. Before the Health Maintenance Organization (HMO) Act of 1973, it was illegal to profit off of health insurance. As such, the documentation process was fairly straightforward: The physician scribbled a few notes in the patient’s chart, and an office staffer used that info to fill out a single-sided billing form, which was mailed to the payer and promptly reimbursed. Equally straightforward was the purpose of health insurance at the time: cover the patient, pay the doctor and protect both parties.

All that began to change in the late 1980s and early ‘90s. The birth of for-profit insurance companies turned health coverage into big businesses. CEOs, hoping to please shareholders by slashing costs, introduced strict prior-authorization requirements, demanded extensive documentation and turned primary care physicians into gatekeepers. They thought that by de-greasing the wheels and slowing down the claims process, doctors and hospitals would think twice before recommending complex and costly services.

Healthcare providers bristled, accusing insurers of putting profits above patient care. In just a few years’ time, the payer-provider relationship had gone from cooperative to confrontational.

Money: the root of all EHRs

As doctors began drowning in paperwork and coding requirements, the tech-boom of the 1990s offered a life raft: The electronic health record. Marketed as a billing and claims tool rather than a patient-care portal, the EHR vendor’s pitch went something like this: You, the doctor, are leaving thousands of dollars on the table by failing to document your full range of medical services. Our system helps you maximize your income.

But the promise of more money wasn’t enough to convince most physicians to put a computer in the exam room. By 2008, less than 10% of providers had installed even a basic EHR system. The following year, Congress stepped in, ratifying the HITECH Act to “promote the adoption and meaningful use of health information technology.”

The government’s sales pitch was markedly different than the EHR vendor’s spiel: You, the doctor, can’t provide excellent care without an electronic medical record. We’ll pay you $44,000 to install an EHR in your office (but penalize those who use it only for billing purposes).

It’s estimated that half the EHRs in medical offices today were purchased because of these financial incentives, which turned out to be a Faustian bargain. Billing and documentation requirements that had once taken doctors only a few minutes now consumed half the day, literally. An observational study in the Annals of Family Medicine found that family medicine doctors spend 6 of 11 hours each workday staring into the abyss of their EHRs.

Listen up: scribes vs. virtual assistants

On my monthly podcast, health care leader Zubin Damania (ZDoggMD to his fans) lamented that doctors today are, “stuck behind a clunky EHR that looks like it was built in the ‘90s or worse. It looks like it has a DOS prompt.”

He’s right. Six in 10 physicians think EHRs need a complete overhaul. Until that happens, a handful of solutions have emerged to get doctors out from behind their computers. Whether these options eliminate or exacerbate to the doctor’s frustration will depend largely on three factors: product quality, cost, and user experience.

Solution No. 1 is the medical scribe.

Usually an aspiring MD fresh out of undergrad, the scribe types notes into the EHR while sitting in the exam room or listening through headphones. This approach has its pros and cons. On the pro side, the scribe can quickly retrieve notes or laboratory results at the physician’s request. And by listening to the doctor-patient interaction, the scribe can judge the most important parts of the conversation and enter them into the EHR. Con: scribes don’t come cheap. Year-round costs range from $30,000 to $50,000. For an independent primary care doctor earning ~$200,000 a year, scribes are cost-prohibitive. Although the job can be outsourced to India for less, security and quality concerns have limited adoption so far.

Solution No. 2 is the “computerized virtual assistant.”

Companies like Nuance Health, makers of the Dragon dictation software, and M*Modal, a 3M company, now offer products with a host of voice-driven features to help doctors. These virtual assistants perform two of three tasks very well:

1. Voice recognition. Anyone with a smartphone is familiar with this function, which quickly translates spoken words into text, saving physicians time otherwise spent typing. The transcripts are mostly accurate, but not perfect, meaning doctors must review and correct each medical-chart entry. For greater accuracy in areas like medication prescribing, most systems feature audio verification, repeating the doctor’s words and asking if the information is correct. This, of course, consumes valuable time.

2. Task automation. Anyone who uses Siri or Alexa can appreciate these action-oriented applications. In clinical situations, the app performs simple commands by connecting spoken words with a specific and previously entered data set, like “bring up the last office visit” or “order Lovastatin 20mg once a day at bedtime.” In terms of time saved, it’s like asking Alexa to “play the theme from Game of Thrones” rather than opening your Spotify app, searching for the song yourself and clicking play. Not much time saved per command, but the minutes add up over hundreds of commands in a day.

Most important, the “smart assistant” allows physicians to keep their eyes on the patient and off the computer screen. For some, the monthly or per-use fee for this service is worth it. Others still prefer using their computer’s drop-down menus, dot-phrases and macros.

3. Artificial intelligence. Some manufacturers of virtual assistants market their systems as the latest in AI technology, implying the system can “listen,” “learn” and create a medical record similar to what a scribe can produce. They can’t, at least not yet. The application of judgment in this context is an exclusively human ability. After all, it’s one thing for Alexa to search for a specific song. It’s much harder (and currently impossible) for Alexa to listen in on your dinner conversation and select the appropriate “mood music.” In a clinical setting, the misuse of judgment creates great risk and liability, which no doctor or technology company is yet willing to assume.

The health-tech Holy Grail

The ideal solution would be a low-cost “virtual assistant” that (a) listens to the patient’s problem and the doctor’s advice, (b) extracts relevant information, (c) converts that data into medical text, and (d) creates a complete medical record with 100% accuracy; no editing required.

Unfortunately, creating such an intelligent device will take a decade or more. A secondary solution involves EHR companies opening their Application Processing Interfaces (APIs) to third-party developers. By allowing one piece of EHR software to interact with another, developers could create more user-friendly apps for doctors. So far, the vendors have refused, fearing the loss of proprietary data and, therefore, the loss of profits.

I hope tech companies keep hacking away at healthcare’s digital divide. Getting physicians out from behind their computers would strengthen the doctor-patient relationship, provide much-needed relief from EHR-related burnout, and greatly improve both the physician’s performance and satisfaction. Nothing irks today’s doctors more than administrative complexity and cumbersome EHRs. Unless the government is willing to step in and mandate change, scribes and virtual assistants are the best we can do.

Robert Pearl is a physician and CEO, Permanente Medical Groups. He is the author of Mistreated: Why We Think We’re Getting Good Health Care–And Why We’re Usually Wrong and can be reached on Twitter @RobertPearlMD. This article originally appeared in Forbes.

Image credit: Shutterstock.com