May 28th, 2024

More on ID Doctors and Primary Care for People with HIV

People waiting to see their new primary care doctor.

A recently published study suggested that “non-ID doctors do better” when it comes to providing primary care to people with HIV.

At least that was the attention-grabbing subject line of an email summary distributed by a local primary care doctor, Dr. Geoffrey Modest. He periodically sends around detailed descriptions of studies he finds interesting, then posts them on his blog. Sign up, it’s great value!

But does this recent study really say that? And does it help answer the “who should do HIV primary care” debate? Let’s take a quick look at the study design and results.

Under the auspices of the CDC’s Medical Monitoring Project, 6323 adults receiving HIV care from 2019-2021 were included. The investigators selected a random group to undergo a phone or face-to-face interview. Patients from 16 states and Puerto Rico participated.

The clinical endpoints of interest as summarized in the abstract:

… retention in care; antiretroviral therapy prescription; antiretroviral therapy adherence; viral suppression; gonorrhea, chlamydia, and syphilis testing; satisfaction with HIV care; and HIV provider trust.

Compared to the non-ID providers, the ID providers’ patients were less likely to be “retained in care” (83.1% vs 89.7%), and less likely to be screened for STIs (40.1% vs 49.5%), both comparisons statistically significant. Outcomes improved when ID doctors worked with a PA or NP.

By contrast, no significant differences were observed with any of the other metrics. This included by far the most important marker of HIV outcomes in 2024, which is viral suppression.

A study like this, of course, has limitations, most notably the non-randomized design, meaning those patients who receive primary care from ID doctors differed in many ways from those who went to non-ID specialists. While the investigators adjusted for demographic and other differences, residual confounding is always a possibility. I can certainly attest to the fact that at least here in Boston, many HIV patients who have ID doctors as their primary providers do so because they have no previous regular provider of any sort prior to their diagnosis.

And they can’t get a primary care doctor, an issue I’ll come to in a bit.

In addition, the “retention in care” metric is kind of weird, specifically:

… retention in care, defined as ≥2 elements of HIV care (including documented or self-reported outpatient encounters with an HIV care provider, HIV-related laboratory test results, ART prescriptions), at least 90 days apart.

For the record, many of my most stable patients long ago graduated to annual visits and blood test monitoring — admittedly in violation of guidelines that advocate for twice-yearly visits. Before you tsk-tsk at this “infrequent” care, how strong is the evidence that a person with HIV with virologic suppression for decades must have their labs checked twice yearly? I suspect these rock-solid stable patients wouldn’t qualify as being “retained in care” by the study’s definition, but they’re definitely not lost to follow-up.

In other words, my conclusion from this paper is that based on the metrics chosen, the quality of care for most people with HIV is pretty darn similar, regardless of their specialty, and that those who worked as a team — with an NP or PA — did a touch better with STI screening.

The results don’t surprise me much — HIV care in stable patients is, well, usually very stable. Hooray! It does not require a specialist’s expertise and raises the question once again why, in this phenomenally successful current ART era, we need ID doctors to do HIV primary care.

Indeed, my last take on this issue argued that it was time for us to move to a model more like the one we have for oncology or rheumatology patients, with ID specialists still playing a major role when HIV and its complications are active, and patients returning to primary care once stable. A periodic visit with an HIV expert can support primary care clinicians for issues related to ART switches and new treatment options, managing side effects and drug interactions, or clinical applications of advances in the field.

The clinical endpoints of the study also don’t get to the core reason why many of us ID doctors would be thrilled if our patients also had a generalist primary care doctor — especially our older patients with multiple medical comorbidities. Instead of these HIV-related endpoints, what if the study chose non-HIV-related measures of the quality of care, including control of hypertension, diabetes, cardiovascular risk, guidelines-recommended cancer screening, and musculoskeletal pain?

In other words, these and other aging-related endpoints — none ID-related — would be far more relevant to the issue of primary care quality done by an ID specialist than the outcomes the investigators reported.

But before we ask our patients to switch their primary care to generalists, we once again must point out that this will be easier said than done. Primary care providers don’t grow on trees — or, if they do, these trees are pretty rare flora around these parts. A very important person recently asked me, on behalf of another big-time academic who just moved to Boston, for the names of good primary care doctors for him, his wife, and his two adult children.


If she’d asked me for a thoracic surgeon, an ophthalmologist, an oncologist who specializes in breast cancer (to cite three recent successful examples of referrals), no problem. But a primary care doctor? The shortage is so bad that the giant healthcare system I work in, the largest in New England, stopped accepting new patients late last year. The wait even for established patients to get into see their primary care doctors for a non-urgent problem can be 6 months or longer.

You think these beleaguered PCPs could add people with HIV to their panel?

So we’ll continue to do primary care for now because there’s a desperate shortage of generalists. In some patients, we’ll be managing the only active medical problem they have, either because HIV is not yet stable, or they are otherwise healthy and their problem list is very short, with just one item:  HIV.

In the older patients with stable HIV and multiple medical problems, we’ll do our best, with generous use of our subspecialty colleagues as needed — because some primary care is vastly better than none.

9 Responses to “More on ID Doctors and Primary Care for People with HIV”

  1. Qiang Xia says:

    This is not an easy study and there might be some issues with covariate adjustment. Study site is a confounder and is missing from their models. Suppose there are no differences in retention in care by provider type, but by study site. At Site #1, 200 patients are cared by ID physicians and 90% of them are retained in care; 800 patients are cared by non-ID physicians and 90% of them are retained in care. At Site #2, 800 patients are cared by ID physicians and 60% of them are retained in care; 200 patients are cared by non-ID physicians and 60% of them are retained in care. Without controlling for Study site, we would find retention in care among patients cared by ID physicians to be 66%, lower than that for by non-ID physicians.

    ID physicians
    Numerator: 200*90% + 800*60% = 180 + 480 = 660
    Denominator: 1000
    Retention in care: 66%

    Non-ID physicians
    Numerator: 800*90% + 200*60% = 720 + 120 = 840
    Denominator: 1000
    Retention in care: 84%

    They included a number of covariates in their models from a forward stepwise method, which may not be the best approach. These covariates are more likely to be mediators than confounders, but it is not important to distinguish them here because they need to be controlled for anyway.

  2. Christine Robb says:

    If the recognition and reimbursement for Family Medicine improved in Boston, more of the young medical students would follow this track. When I worked in Massachusetts, there was no department of Family Medicine at your institution, and the specialist snobbery was rampant, as well as the continuing undervaluation of real primary care. I was happy to provide primary care for HIV positive patients with consultation from ID specialists, and think I probably did a better job of addressing their other social and medical concerns because of my broad training. Although I valued the excellent ID consultants, I cannot conceive that their specialty training covered pediatrics, gynecology, orthopedics, and all that training in Family Medicine does. What is needed is a reversal of years of medical education culture which has venerated “specialists” over good family physicians.

  3. Dr Rogerio Luz Coelho (Family Medicine in Brazil) says:

    Nice appraisal of a very obvious study. Here in Brazil we have broken the patents of ARV for almost 30 years. Since mid 2000´s HIV care has been migrated to Primary Care with varying degrees of succcess (mainly because Primary Care in Brazil is a heteregenous affair, when done by the specialist family doctor we have way better outcomes than when done by the non-specialist generalist).

    I see this study more as a testament of the need of a better health care system organization than anything else.

    National health systems should be primary care driven, this poses problemas for every health system on earth since primary care doctors are in short supply everywhere .Seems that market forces alone are not enough to drive salaries and benefits high enough for medical students to choose primary care / family medicine.

    We have the same studies already done in Hypertensive patients (vs Cardiologists) , Diabetes (vs Endocrinologists) , pre natal care (vs Obstetricians) and primary care alwas has better endpoints in “overall care” than the specialist. Why this study did not look into those endpoints baffles me also.

    My final take of this study? Seems it is past the time we consider HIV care as another chronic condition that primary care has the ability / responsability to manage, with the more problematic patients needing a specialist view … HURRAY !!! HIV became chronic care !!!

  4. Loretta S says:

    “Instead of these HIV-related endpoints, what if the study chose non-HIV-related measures of the quality of care, including control of hypertension, diabetes, cardiovascular risk, guidelines-recommended cancer screening, and musculoskeletal pain?” THIS x 100!

  5. Debra K says:

    Long-time primary care doctor here. My perspective is that given the crisis in primary care access that many regions have, it’s not going to be possible to suddenly transfer all people with HIV to new PCPs. Some other solutions are needed, and having ID doctors chip in could be one of them.

    One place I worked had primary care doctors mixed in with ID doctors in the HIV clinic. It was Ryan White supported, and worked great. I don’t know if this would be possible in most settings without this funding.

  6. MW says:

    My perspective, as a community based ID physician is likely a bit different than many other responders here.

    I don’t find this study results surprising in the least. These for the most part were stable patients. The end points examined would be hard to show that a specialist’s knowledge base would be superior. If the health care practitioner has the appropriate, basic knowledge, there is nothing magical about caring for stable HIV patients to expect a difference between a specialist and primary care.
    If the endpoints were different- like inappropriate use of antibiotics for asymptomatic bacteriuria, or for a viral URI; or management of complex genotypes in patients without virologic suppression, now that is where you may see a superiority from an ID background.

    The specialty of ID, in the community at least, is on life support. Reimbursement is down and not improving sufficiently. Training positions go unfilled chronically. Despite rising antimicrobial resistance and predictions of rising infection rates from climate change/ aging populations, ID is failing to attract enough clinicians.

    As ID physicians we need some of these bread and butter, stable patients to survive financially.
    One of my mentors once said to me many years ago when we saw a fairly straightforward case that “They can’t all be FUOs. And you wouldn’t want that.” If a complex, complicated ID case takes an extended time to assess and manage, we need to offset that with some straightforward work.

    We may not provide superior care in most situations of stable patients, but we still will for some. And for the others, that is part of the requirements to keep paying the bills that allow us to see the complex cases, where we are severely underpaid, on a per time basis.

    • Paul Sax says:

      All excellent points. It’s exactly what I meant when I wrote:

      The results don’t surprise me much — HIV care in stable patients is, well, usually very stable.


  7. Juan E. Losa says:

    In Spain we already have a “consensus document on the shared care of patients with HIV infection between primary and hospital care”

  8. Marvin S Rabinowitz, MD says:

    We are all BC internists and are trained to care for IM pts, including outpts. My HIV pts tend to have other med problems so they get seen more than 1-2x /year.
    It is not that hard to keep up to date with IM and turf out gyn and ortho problems. Given the nature of our pt populations, opioid use disorder can be a big part of our pt population. Of course, this applies to us rural id physicians, not urban areas with high inpt consults.

Leave a Reply

Your email address will not be published. Required fields are marked *

Note: This is a moderated forum. By clicking on the "Submit Comment" button below, you agree to abide by the NEJM Journal Watch Terms of Use.

Our physician bloggers cannot respond to requests for personal medical advice, and recommend patients discuss health issues with their individual physicians.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

HIV Information: Author Paul Sax, M.D.

Paul E. Sax, MD

Contributing Editor

NEJM Journal Watch
Infectious Diseases

Biography | Disclosures | Summaries

Learn more about HIV and ID Observations.