there are a few points that makes me go ehh

1. This is a paper done in China. All Chinese Authors. Chinese professors are pressured to publish, the quality of the researches are sometimes questionable.
2. They did not perform primary research. They took genetic data and questionnaire data from UK database. They can't really control the quality of the questionnaires.
3. This study is based on data on Europeans. Even though the scientists are in China. They couldn't gather data (but mostly because there is no genome data for Chinese population)
4. They excluded people in the data who have no diseases. They also excluded people without full genomic data. Usually people who submitted full genomic data have similar behavioral traits, so this makes the data flawed.
5. I'm not familiar with Mendelian randomization, it's not something that we use in molecular biology. I'm guessing their logic is a. we found specific gene sequences correlated to consumption of red meat, or processed meat. b. we found no specific correlation between the presence of these gene variants to the development of three diseases in people with these diseases.

(girlfriend is a molecular biologist)