HIV and ID Observations

The ID group at Mayo Clinic just published a small but important study on the use of remote ID telemedicine consults for hospitals that have no ID services on-site.

The consults were “asynchronous”, meaning that the ID consultants at the main hospital finished them within 24 hours — they didn’t have to respond immediately. Importantly, all the institutions shared the same electronic medical record, so the consultants could review notes, lab results, and other tests.

Comparing 100 cases who had “eConsults” with 300 historic controls, the investigators found that the cases with ID eConsults had a substantially better clinical outcome — a 70% (!!!) lower 30-day mortality.

These results didn’t come from some Magic Dust available only to survivors of the bitter cold Minnesota winters — the primary interventions recommended by the consultants included very straightforward (for an ID doctor) advice, such as antibiotic changes, antibiotic duration change, antibiotic deescalation, additional laboratory testing, and consultation with services other than infectious diseases.

The authors acknowledge the important limitations of the study, primarily the fact that the non-randomized design leaves it open to unmeasured confounders influencing the outcome.

Furthermore, since they utilized historical controls, changes in practice could have led to the improved outcomes, not the eConsults themselves.

Still, limitations notwithstanding, let’s imagine the effect isn’t quite this large — it’s still very impressive. As noted by Dr. Saurab Patel, the upper bound of the 95% confidence interval on the reduction in mortality was 30%! People who study health care quality and outcomes would take that any day of the week, thank you very much.

Furthermore, satisfaction among the clinicians ordering the consults was extremely high. Indeed, we find the same thing at our hospital, where we provide eConsults to our primary care clinicians and specialists for non-urgent questions that arise during outpatient care. I attended a meeting of primary care providers on the use of eConsults, and the enthusiasm for this service was off-the-charts high — they love them!

So based on a study like this and our experience, every clinical ID Division in academic medical centers and every ID private practice must be lining up to provide eConsults, telemedicine, or other forms of remote advice, right?

Well, not quite:

Hey #IDTwitter, do you currently do "eConsults" or provide telemedicine services for patients you have never seen for a face-to-face visit? Can be for other clinicians, or directly to patients, or both. Take the poll, then explain why or why not. Thanks!

— Paul Sax (@PaulSaxMD) January 19, 2020

Only just over half the respondents do these kinds of consults currently, even though arguably the non-procedural and cognitive-based aspects of ID practice fit perfectly into this remote care model.

In the Discussion section of the paper, the authors cite two major reasons why a substantial fraction of ID doctors don’t do these sort of consultations:

Cost and technical challenges are major barriers to the widespread adoption of telemedicine.

Let’s take the second of these first, the technical challenges.

If you don’t have the same electronic medical record as your consulters, providing this kind of formal consultative service would be difficult, if not impossible. What we ID doctors do best is carefully review the relevant history, laboratory, and imaging data, and make recommendations from synthesizing this information. It’s critical that this review is as reliable as possible — especially since confirming details by talking to the patient may not be possible.

(Reminder:  ID doctors take the best histories. Thank you.)

Another challenge is institution-specific credentialing. Even affiliated hospitals in the same healthcare system may have different criteria and systems for credentialing their clinicians. This makes doing a remote consult on inpatients at other facilities far from straightforward — and doing it across state lines might actually be prohibited if the consultant is not licensed in that state.

Now on to the thorniest issue, which is cost. It’s not simply that we’re too busy. If these kinds of clinical services were valued highly enough, we would make time in our schedules to do them, pushing out lower value activities.

In the responses to my poll, the most enthusiastic endorsements of remote consultations come from settings where American-style fee-for-service compensation plays comparably minor role (if that) in a doctor’s salary. Europe, Canada, and here in the U.S., the VA — you get the idea. Says Dr. Ilan Schwartz from Canada:

We cover a catchment area from central Alberta to NW Territories — a geographic area almost 1/3 that of Canada! For some assessments I prefer ability to examine the patient, but for others, telehealth consult is in their best interest. Also sometimes roads are really icy, and if I can spare patients both the inconvenience and the risk of shleping to Edmonton, I will.

No surprise here! What Ilan does in Western Canada makes all kinds of sense, reminiscent of the ECHO research project using telemedicine to treat hepatitis C in remote areas of New Mexico.

But in a differently funded healthcare system — ours — where payment is overwhelmingly based on racking up face-to-face encounters and procedures, the time spent doing eConsults and telemedicine is time not spent in revenue-generating patient care.

And in most practices, eConsults score a big fat zero when it comes to RVUs, that loathsome metric for measuring a U.S. doctor’s “productivity”.

Time for that to change!

Because doing what’s best for patient care should motivate how we spend our clinical time, not racking up RVUs.