CT screening for lung cancer for homeless people: the new JAMA IM paper

Common sense is dead, and careers are made from useless research.

Recently, I read a randomized trial that is everything wrong with medical research. It takes people who were ever homeless and randomizes them to an initiative to encourage CT screening for lung cancer or usual care. The result: if you get a push, you were more likely to get 1 scan. Let’s unpack why I think poorly of this research effort, how it speaks to misplaced priorities and lack of common sense in medicine, and worse, how it signifies putting one’s career ahead of doing right by patients.

First, the authors take 200 people who were homeless. 1 in 3 are currently homeless, and only 40% have stable housing.

Imagine this population and ask yourself: What is the most important thing they need?

An apartment, a job, basic blood pressure control, work up for existing ailments, perhaps, more simply, a shower, a haircut, clean clothes, a hot meal. Now consider 1 in 4 has a drug or alcohol use disorder. 85% are current smokers.

Wow, perhaps they need some addiction medicine support, perhaps mental health care more broadly; Again, stable housing, smoking cessation, blood pressure control, food, etc.

But these researchers think a priority for these people is…. a CT scan for lung cancer. They randomize these people to a patient navigator to get it done, or usual care and look how many get 1 scan. Yay! The authors celebrate.

What are the problems?

1. Lung cancer screening DOES NOT SAVE lives. NLST has null all cause mortality in long term follow up (Pinsky JTO), and NELSON failed to even improve all cancer mortality. I summarize these data in a past post (one of our most popular) and figure below

"I just want to work on my car"

The authors are pushing a screening test that lacks data even in motivated participants with high rates of follow up.

2. Even if lung cancer screening works that is contingent on the rates of follow up scans performed in NLST and NELSON, and follow up work up— biopsy, lung resection, SBRT and the like. Screening can’t work if people don’t follow up. Homeless people are much less likely to follow up than the motivated people in the initial trials.

3. Homeless people have bigger priorities than lung cancer screening. Their risk of all cause death is massively higher. This study is akin to randomizing homeless people to being advised to eat a bar of dark chocolate and then measuring the percentage who ate a bar of dark chocolate— and claiming that homeless people are now better thanks to the flavonoids.

3. The primary endpoint is poor (getting 1 scan is meaningless)

4. Authors wast time with a qualitative analysis

What I really wish were that someone— just one author, perhaps a reviewer or even the journal editor could have been honest and said, perhaps via email,

Listen team, I hate to say this, but this is a bad idea. Lung cancer CT screening is itself debated. Some argue that we shouldn’t offer it to anyone, and I am sympathetic to those arguments. But even if you aren’t: pushing it on homeless people is the last thing they need and might not even do them any good. I know that that American Cancer Society is looking to fund bullshit, feel good ideas, but we have to have some dignity. We can’t just do pointless research even if we get grants for it and if we get another item on our CV. We have to respect ourselves, and we have to respect patients. Homeless people need a lot of things. CT screening for lung cancer is not one of them. Let’s stop before we publish this and look foolish in front of the last few people in medicine with common sense.

But no one sent that email.

And here we are.