• Defending the survival-rate

    Faithful readers of the blog have written me about my earlier post on Paul Meier, asking me why I’m now being so nice to survival rates when I’ve been so hard on them in the past. Excellent question.

    What they’re referring to is that fact that I keep on screaming about how mortality rates are more important than survival rates when talking about how well we prevent death at a population level. I still stand by that. But it’s important to understand that I, as a health services researcher, function mostly at the population level. That’s my gig.

    Nearly every other doctor, however, has a different job. They’re functioning at the individual level. And when you’re dealing with an individual patient, then survival rates are way more important than mortality rates.

    When a patient is sitting in a room with their physician, and she’s just learned she has cancer, the only thing she cares about is “what can I expect”? She doesn’t care about the population, or the mortality rate. She wants to know, what’s going to happen to me?

    And that’s where survival rates are necessary. Doctors want to be able to say to patients with confidence that you have a certain chance of living five years, or ten years, or more. We want to know that, here in the US, if you’re this age, and have this cancer, and it’s this stage, that this is the probability you’ll live 10 years. That’s what patients want to know, and it’s totally reasonable, and it’s what survival rates can tell us.

    What they can’t tell us is if that time is based on earlier diagnosis or better care (or both). That’s what I care about as a health services researcher. That’s what you should care about when we discuss the quality of the US health care system. But when you’re a patient, or when I am, survival rates are paramount. And they should be.

    • How long people who get X survive would seem to be a truer reflection of care and early diagnosis than mortality rates which may be influenced by demographic, lifestyle, and environmental factors that influence the prevalence of X..

      Are survival rates population level data in the sense that they are based on everyone who gets a particular condition or are they based on nonrandom samples of that population? Is there data about survival rates for various conditions for other countries?