HIV and ID Observations

In a welcome update, CDC just presented revised data on the number of Lyme Disease cases annually in the United States. Here are the key facts:

Cases reported by clinicians: 30,000

Estimated number of cases using additional information from insurance claims, laboratories, and patient reports: 300,000

Go ahead, check my math — that’s a whopping 10:1 ratio of actual to reported patients with Lyme, highlighting that this reporting is something we clinicians just don’t do very well.

And it’s no surprise to me at all.

Here’s why:

• Clinicians are busy doing other stuff, most importantly taking care of their patients.
• Lyme is one of many reportable diseases — take a look at this list!
• It’s annoying. If you asked 100 doctors and nurses what they like least about practicing medicine, 74.39% would say “the paperwork.”
• The forms can’t be completed quickly, as they request detailed information about dates of exposure and onset of illness, presence of diverse associated symptoms, and evaluations for ancillary diseases. Plus, they’re paper forms — a particular problem in an increasingly web-based and electronic medical record world. As for the specific dates of possible exposure and onset of symptoms? “I just lie,” one PCP told me.
• “Soft calls” won’t get reported at all. Think about the febrile, achy patient in the summer who’s just spent a weekend hiking in the woods, saying ticks were everywhere, then gets treated (appropriately) for suspected Lyme. That may be the billing code the clinician uses, but often as not, no Lyme test is sent.
• No one’s paying for these reports. Forgive the mercenary tone to the comment, but as these folks will tell you, humans are very predictable creatures — give them some incentive to do something, and they’ll often do it, even if the task is unpleasant.

I don’t have an immediate solution to this problem of Lyme underreporting, but it’s important we try to solve it — to quote Paul Mead from CDC, “Lyme disease is a tremendous public health problem in the United States,” and I completely agree.

So let’s look at a disease with extremely accurate surveillance, and that’s HIV.

HIV reporting is tied to receipt of government funds for HIV treatment and prevention. That’s a huge incentive to get the numbers right. If anything, some sites are known to overreport their cases, which is why name-based reporting is such an important accuracy check on the data.

Lest you think that HIV is different because, well, it’s HIV, believe me there are plenty of places where HIV reporting is woefully inaccurate. Back in the mid-1990s, I joined a group from the Harvard School of Public Health to participate in a scientific symposium on HIV in India. There was a reported explosion of HIV cases in certain regions of the country, and their doctors and researchers set up the conference to address this growing problem.

The first presentation was from a public health official representing the government. He showed data demonstrating that the number of cases in India was still relatively small and confined to isolated regions and “risk groups” — mostly commercial sex workers and injection drug users.

The next presentation was from an active ID clinician at one of the largest hospitals in Mumbai. He showed slide after slide demonstrating the ravages of advanced AIDS in diverse populations (emphatically not all commercial sex workers and IDUs) and then finished with a few data slides on cases at his hospital. Turns out that he had seen as many cases at his one site as had been officially reported to the government from the entire country.

When asked how to reconcile his volume of experience with the previous lecture on the national epidemiology, he responded with two other questions.

So why should we report our cases? What good does it do us or our patients?

Once we can answer those questions for Lyme — as we most certainly can with HIV — then the underreporting problem will be solved.