A big confounder in studies based on diagnosis numbers is that you can confuse better diagnosis with higher incidence, and earlier diagnosis with longer survival. No room for confusion with Death though, it's very cut and dried.
Suppose that today the average person tells their doctor about a symptom of Example Cancer (which is incurable) six months before it kills them. 1 million people per year in Standard Country die of Example Cancer, with an average of six months between diagnosis and death.
Now, let's imagine I invent a machine, it can scan seemingly healthy people and tell them if they've got Example Cancer on average 12 months earlier. Nothing changed in terms of whether people get Example Cancer, it's still incurable, but now we've improved time between Diagnosis and Death by 200% but even scarier the incidence of Example Cancer, the number of people who know they have it, has also increased by 200%. It's an epidemic!
That machine is pretty unrealistic. A more realistic machine also gives false positives for Example Cancer. Now the number of people living with Example Cancer has increased by 500% but good news, most of those people don't die of it, because they had what medics would call "Sub-clinical incidence" meaning, sure, you had the disease but it didn't actually affect your life so who cares?
A big confounding factor in counting incidence based on cancer deaths would be improved treatment. Whether by medical advances or simple earlier diagnosis having better cure rates.
Suppose that today the average person tells their doctor about a symptom of Example Cancer (which is incurable) six months before it kills them. 1 million people per year in Standard Country die of Example Cancer, with an average of six months between diagnosis and death.
Now, let's imagine I invent a machine, it can scan seemingly healthy people and tell them if they've got Example Cancer on average 12 months earlier. Nothing changed in terms of whether people get Example Cancer, it's still incurable, but now we've improved time between Diagnosis and Death by 200% but even scarier the incidence of Example Cancer, the number of people who know they have it, has also increased by 200%. It's an epidemic!
That machine is pretty unrealistic. A more realistic machine also gives false positives for Example Cancer. Now the number of people living with Example Cancer has increased by 500% but good news, most of those people don't die of it, because they had what medics would call "Sub-clinical incidence" meaning, sure, you had the disease but it didn't actually affect your life so who cares?