July 2019: What makes the best consultants?
First, a bit of housekeeping. There weren’t and won’t be updates for May or June. I spent those months finishing my ID fellowship at Baylor College of Medicine, completing laboratory experiments needed to turn my last two years of research into a publishable paper, moving across the country with a 1-year old, selling a condo, buying a house (with a garden! and also rabbits that eat the things in my garden!), and starting a faculty position at the University of Nebraska Medical Center. But, now we’re back from hiatus, and I don’t anticipate any further disruptions to the update schedule this year.
Second, there may be some changes to the journals I review if it turns out that the UNMC and the Texas Medical Center libraries have different subscriptions and I’m stuck behind paywalls. I’ll let you know about any such methods updates in that month’s header post.
Finally, a correction from the last update: I erroneously described a recent paper by de Carmago and colleagues (30949690) as a retrospective cohort study; one of my astute former co-fellows pointed out that this was actually a prospective study.
On with the show!
(The following is a brief discussion of the medical literature regarding subspecialist consultation skills. If you’re not interested in that, skip to the bottom of this post for links to this months’ literature reviews)
It’s the first days of August, new Infectious Diseases fellows are starting their training around the country, and so I thought I should do some reading about what makes a quality medical consultant. By this, I specifically mean how to be the person that other clinicians hope will be on the other end of the line when they pick up the phone to call a subspecialist. This is a skill; I even recall it in my fellowship’s curriculum documents, buried in the learning objectives alongside phrases like “medical professionalism” and “business of medicine.” But while my first month of fellowship included great 101-style orientation lectures on antiretroviral therapy, antibiotic spectra, and laboratory microbiology, there was lamentably no How to Be a Fantastic Consultant, and in this I doubt I’m alone.
What would the advice in that talk look like? I’d thought about this a lot by the time I started ID fellowship; more specifically, I’d spent a lot of intern year thinking about the kind of fellow I didn’t want to be. An intern on the general medicine wards spends a lot of time calling for help, and at least half the time it’s at the bidding of your senior resident or attending and you don’t quite grasp the nuance of what they want from the consultant, or don’t understand why the consult needs to be made (or worse, you’re pretty sure it isn’t necessary, but are tasked with making that case anyway). I learned to dread calling certain folks who’d refined pushback to an art – and conversely to treasure the senior residents and fellows who were consistently patient, enthusiastic, kind, interested in teaching, and most of all invested in the wellbeing of my patients the moment they picked up that phone.
Given that experience, the standard I held myself to as fellow was that, when called for a consult, the first thing I’d say as soon as I could work it into the conversation was, “Sure, I’d be happy to see him/her. How can we help?” And while I think that served well enough, this month I thought I ought to search the medical literature for actual data to bring to the conversation.
The earliest reference I could find on quality consultation is a JAMA IM paper from 1983 by two Harvard cardiologists and internist from Stanford, appropriately titled Ten Commandments for Effective Consultations. I love the opening lines:
“Although medical house staff and practicing internists spend much of their time performing consultations, few training programs offer formal instruction in the subtleties of this art. Most physicians learn how to perform a consultation through trial and error, resulting in considerable variability in consultative skills. Thus, some consultants are much sought after by their colleagues, while others have trouble translating their expertise into effective consultations.”
So what are the commandments? (1) Determine the question, (2) Establish urgency, (3) Look for yourself (with notes that “A consultant should not expect to make brilliant diagnostic conclusions based on an assessment of data already in the medical chart” and “Before ascribing ECG changes to lead placement, the consultant should repeat the ECG rather than leave a note recommending that a surgical colleague perform one”), (4) Be as brief as appropriate, (5) Be specific (“[studies have shown that] leaving a long list of suggestions decreased the likelihood than any of them would be followed, including the crucial ones… resist the temptation to suggest tests that are not crucial to the case just validate [your] ‘expert’ status”), (6) Provide contingency plans, (7) Honor thy turf, (8) Teach with tact (“While copies of references are often graciously received, they are not a replacement for discussing the principles of the case with the requesting physicians”), (9) Talk is cheap and effective, and (10) Follow up.
This paper is short and conversational, and while one or two of the remarks haven’t aged well, I think it’s still worth reading for any new fellow or attending consultant.
Moving from expert opinion to data, this study in the Journal of Hospital Medicine surveyed hospitalists across four major medical centers about their current and ideal interactions with medical subspecialty fellows and attendings (disclosure: co-author Nainesh Shah was one of my chief residents). Their 122 survey respondents had a mean 5.6 years of hospitalist experience and reported making an average 3 consults per shift. Eight-six percent of hospitalists reported wanting to receive consultants’ recommendations in-person or by phone, but over 90% reported that in-person delivery of consultants’ recommendations occurred in less than half of cases, and almost 40% reported that they didn’t even receive a phone call at least half of cases. With regards to teaching, not only did 83% of hospitalists want to receive more teaching from their consultants, but 74% hoped that consultants would initiate the teaching interaction if they didn’t. Two-thirds of hospitalists reported experiencing pushback (reluctance or resistant to perform the consultation; e.g. statements that the consult should not have been called, should have been called at another time or to another service, or is not accompanied by sufficient data or a sufficiently clear question) at least “sometimes”; 22% perceived it occurring “most of the time.” The hospitalists reported the main barriers to effective consultation were pushback, followed by lack of in-person interactions and high hospitalist and consultant workloads. Factors facilitating good consultation included working on an interesting clinical case, more free time, and current and prior positive interactions with the consultant. Hospitalists found interacting with consultant attendings superior to work with fellows in all aspects of patient care, but particularly regarding the issues of pushback and professionalism.
It’s not just hospitalists: Sara et al interviewed 27 residents and 16 fellows at Stanford’s Children’s hospital about what made the ‘perfect consult’, and while both residents and fellows felt that multiple communication modalities (e.g. in-person, by phone, by text, by note in the chart) could be appropriate depending on the context, they agreed that in-person communication of consultant recommendations reduced misunderstandings, improved patient safety, and led to better teaching. They also agreed that fostering interpersonal relationships improved the quality of consultation, and that fellow pushback (often triggered by a resident consulting with unclear question or limited data – which again, I would say is as much an issue with the fellow as the resident) not only negatively affected fellow teaching during that consultation but poisoned future resident-fellow interactions. Similarly, a 2015 study in which researchers interviewed residents and fellows across two major medical centers found consultant pushback widely cited as a negative initial interaction that harmed or foreclosed on future teaching opportunities. Specifically, quotes from the interviewed interns suggest that pushback is not an effective teaching tool and that this approach erodes collegiality and trust between house staff:
"As an intern calling, it’s always like a game. . .How many questions can you answer from a consult
service? Eighty per cent is pretty good. . . but if you can only answer half your questions, you
just feel really dumb. You feel like that was a really bad experience talking to them and you’re
kind of worried that they might not see the patient.”
“They’ll come off as. . . being jerks. . . That was the initial consult and then on further consults . . .
you find out they’re a nice person. . . you kind of tell them: “Why were you giving me such a hard
time before?” They’re like: “I was just pushing you to get you to learn.”
I want to note here that aside from pushback not being an effective teaching tool for the shamed referring clinician, it’s poor clinical care. Delays in inpatient consultation prolong length of stay (see this study) and hence prolong patient’s exposure to the hazards of the hospital. Moreover, infections are dynamic illnesses: as far as I’m concerned the ID doc who cultivates such a reputation for lousy demeanor that primary teams put off calling them until they can’t is in part responsible for their patients’ poorer outcomes.
The theme of consultant communication extends not only to the referring clinician but also to the patient, their family, and the other clinicians involved in the case. Yes, consultants take a subsidiary role to the primary team, and that means it generally behooves us ‘stay in our lane’ when it comes to updating patients and families on the therapeutic plan. On the other hand, when Roche et al surveyed ICU nurses and patient families about the quality of subspecialist consultations in the ICU, half of nurses and two thirds of families reported no direct communication with the consultants at all, and nearly half of the families knew so little about the consultation they felt incapable of evaluating it. Is it optimal care to ask a non-specialist (the primary team) to deliver your recommendations via a game of telephone with the nurses and patient families, rather than letting them discuss your recommendations with you, the expert, in person?
What about a consultant’s notes? Boulware et al invited physicians at four Minnesota teaching hospitals to complete a survey on inpatient consultations, of whom 323 responded (54% attendings with 5+ years of experience, 17% with <5 years of experience, 30% house staff). Sixty-nine percent of respondents preferred bulleted or numbered recommendations rather than paragraphs, and more preferred separate rather than integrated assessments and plans (48% vs 37%). When discussing consultant’s recommendations, 71% agreed that it was very important recommendations be simple and concise and 64% highly valued the rationale behind the consultant’s decisions, while only 7% highly valued references to the medical literature. When consultants recommended a new medication, 80% agreed on the importance of using generic names, 80% on the importance of specifying the dose and duration of the drug, and 76% on recommendations for alternative therapies where appropriate. When active consultation was no longer needed, everyone thought a sign-off note was needed and 74% agreed the note should include recommendations for followup specialty care. In a review titled The Role of The Medical Consultant, Cohn gives more explicit recommendations (with additional supporting references from the 80s that I haven’t cited here), including: (1) Respond to consults within 24 hours, (2) Limit recommendations to 5 or fewer and identify the crucial (versus routine) recommendations, (3) Make specific, relevant recommendations using definitive language, (4) Specific the dose, route, frequency, and duration of recommended medications, and (5) Frequently follow up via progress notes and direct verbal communication with the primary team.
Finally, it’s critical to define the expectations implicit in the consultation, which often vary between institutions and between consulting services. For example, a 2007 JAMA IM study interviewing 153 surgeons and 170 nonsurgeons found that nonsurgeons more often wanted the consultants to limit their input to a specific question (69% vs 41%) and to refrain from writing orders without obtaining permission from the primary team (59% vs 37%), whereas surgeons were more likely to desire a comanagement relationship (59% vs 24%) and to want daily progress notes from the consultant regardless of the patient’s acuity of illness (78% vs 67%).
Do we have literature about consultation quality specific to Infectious Diseases? Yes! Lo et al prospectively studied 465 ID consultations across two major academic medical centers to identify factors associated with the primary team’s adherence to consult recommendations. They found that the overall adherence to our recommendations was high (80%), and that recommendations were more likely to be followed (1) when the recommendations were therapeutic versus diagnostic, (2) when the recommendations related to a specific clinical question, (3) when the recommendations were identified as crucial by the ID team, and (4) when the consultant’s note was legible and organized. The accompanying editorial in CID, by an experienced private practice physician, asserted that adherence to ID recommendations in the private world is governed by similar rules, and emphasized the importance of collegiality and mutual trust and respect between the ID and consulting physicians as the critical foundation for effective consultation.
So, to conclude: what does the medical literature have to say about how to be a fantastic ID consultant?
- Don’t give pushback! Well-intentioned or not, pushback impedes learning, negatively impacts referring physicians’ satisfaction with both the specific encounter and the consultant personally, and may delay and/or compromise patient care. Work to cultivate, not strain, your relationships with primary teams.
- Clarify your role as the consultant. This may not vary much between individual referring physicians within the same department, but it is likely to vary between services and institutions. The internists may not take kindly to you writing any orders in the chart; Oncology may be fine with you ordering diagnostics, but want to informed before any change in antimicrobial therapy; Orthopedic Surgery may be annoyed to discover that you recommended antibiotics without starting them.
- Show up and talk. As one of my ID mentors likes to say, “No one consults a ghost.” The literature tells us that verbal communication is essential on both sides of the consultation interaction: it provides an opportunity to ask and answer questions, facilitating both teaching and high-quality patient care. Moreover, data indicates that other stakeholders, like nurses and patient families, are often ignored by consultants, and that these members of the care team are more likely to value the consultation when brought into the discussion.
- Write clear notes with helpful recommendations. Provide a well-organized plan with brief, specific, unambiguous recommendations (and not too many of them) in bullet form, indicate which recommendations are most crucial, and make sure your antimicrobial recommendations include the generic drug name, dose, route, frequency, and duration of therapy. When you write your sign-off note, make sure your plan isn’t just “we’re not seeing him anymore” and includes things like antibiotic stop dates, the recommended time frame for ID followup, and developments that should trigger reconsultation (e.g. results of pending microbiology data).
- Teach. Teaching is a fundamental role of the ID fellow. Moreover, fielding (and having to look up the answers to) tough, complex questions from your peers is an indispensable means of building the “deep knowledge” that will eventually separate your expertise as an ID specialist from those of your referring colleagues. Studies show that house staff and attending physicians alike want to learn from their consultants, and specifically that they value having the consultants explain the rationales behind their decision-making rather than have reading material emailed to them or cited in the chart.
And now, on to the medical literature:
Antimicrobial agents research this month included the activity of imipenem-relebactam vs KPC and OXA-48 type CREs, AUC/MIC targets for daptomycin in enterococcus, cefazolin in MSSA, and ethambutol in tuberculosis, and the efficacy of pip-tazo versus cefepime or ertapenem for Enterobacter pneumonia.
ID diagnostics research included use of CRP to safely curb antibiotic use in COPD, blood culture bottle incubation of periprosthetic tissue for PJI, transcutaneous versus through-wound biopsies for osteomyelitis in diabetic foot infection, and the clinical import of EPEC and campylobacter on the GI pathogen stool PCR panel.
General ID research this month covered the hazards of excess antibiotic therapy in CAP and HAP, MDR GNR colonization after receipt of unnecessary anti-anaerobic antibiotics, the unexpected benefits of preemptive malaria treatment in African children, improved clinical cure with prolonged treatment for streptococcal PJI, and the epidemiology of CSF leakage in patients with bacterial meningitis.
HIV and STD research in July included genotype-guided reduction of regimens for treatment-experienced patients who have achieved viral suppression, vaginal microbiotia associated with susceptibility to trichomonas infection, the lack of benefit of “CNS penetrating” antiretrovirals in preventing HIV-associated neurocognitive disease, and the lack of benefit of stopping efavirenz in people with HIV who have developed neurocognitive symptoms.
Onc and transplant ID research this month included predictors of MDR Pseudomonas in patients with neutropenia and bacteremia, presentation and outcomes of influenza in immunocompromised hosts, and the discouraging results of a trial of oral antibiotic decolonization for patients with intestinal carriage of ESBL-producing bacteremia.
Antibiotic stewardship and infection control literature this month included the role of vaccination in curbing antibiotic use, documented penicillin allergy as a predictor of suboptimal antibiotic use and poor patient outcomes, 15 is as good as 30 seconds of hand-rubbing with an alcohol-based hand sanitizer, reduction of VRE carriage after decreased carbapenem use, and the benefits of creating a “PPE-free zone” in the rooms of patients on contact isolation.
As always, feel free to share your thoughts by commenting on any of these blog posts, or reach out to me directly by email at contact.idjournalclub@gmail.com or via Twitter at @Cortes_Penfield.