How does that make it a garbage study? Just because it doesn't divide up "vegetables" into tinier classes doesn't mean it didn't valuably disprove a perfectly well-formed and plausible hypothesis about vegetables as a group.
It's the "vegetables as a group" part. The "vegetable group" is ill-defined because some vegetables (such as potatoes and corn/maize) have nutritional profiles more akin to grains, with higher glycemic index and lower nutrient density. I'd be very surprised if eating lots of broccoli and cabbage doesn't lower cancer rates; lots of potatoes, not so much.