Either type of pseudodisease can be viewed as an extreme form of length-time bias. Whenever you think about these biases, assume that screening doesn't genuinely increase lifespan and try to see how the bias might make it seem to do so. In studies subject to length-time bias, you compare outcomes between patients whose disease was diagnosed by screening and patients whose disease was diagnosed by symptoms. To have length-bias, the disease also must have a heterogeneous natural history. Sometimes it's very aggressive, with a short preclinical phase, a short symptomatic phase, and rapidly kills the patient. Sometimes it's indolent with a long pre-clinical phase, a long symptomatic phase, and only kills the patient very slowly. The indolent cases are much more likely to be detected by screening because they stay in the pre-clinical phase much longer, whereas the aggressive cases progress so rapidly through the pre-clinical phase that we don't have time to catch them on screening and instead they present with symptoms. Diagnosis by screening is associated with longer survival, not because screening is effective (we started by assuming that it isn't), but because indolent disease causes both diagnosis by screening and longer survival.
Pseudodisease is the extreme version of indolent disease. Its pre-clinical phase isn't just long, it's forever. The disease would never become symptomatic or kill the patient. Diagnosing pseudodisease can only cause harm, yet if you compare all those diagnosed by screening and all those diagnosed by symptoms, the ones diagnosed by screening will do much better because some of them have pseudodisease.
MAK
