People who are used to reading scientific research papers will have undoubtedly read Conclusion sections of papers that are completely unsupported by the Results section of the same paper.
It has gotten so bad that some leading researchers recommend that Conclusion sections be ignored or not even written. They advocate letting the reader draw their own conclusions. The problem with this solution is that some readers may want to rely the opinion of the study author, probably an expert (of sorts) in the field.
Others suggest that there is nothing to do. These sort of conclusions are the products of biases (conscientious or not) that we all still have.
High school students are taught in science classes, to think about alternative conclusions before even conducting an experiment. But everywhere else, even at top universities, it seems, that drawing conclusions in hindsight is acceptable practice.
We say No! Our paper writing software Banjo nudges users to write their conclusions, all different possible ones, while creating their protocol. The alternative conclusions (what you would have concluded had the results turned out different) gets peer reviewed with the paper.
Better science is better healthcare. Go Science!