I'm not sure what statistic you are referring to, but for cancer at least the definition of 'cure' is typically a number of years of life after detection (e.g. 5). In jurisdictions with less aggressive testing, this can give a worse measurement of cancer survival rates than the reality, since earlier detection by definition gives a higher chance of surviving to a time X years after detection.

https://www.bmj.com/bmj/section-pdf/187734?path=/bmj/346/789...