The first clue to unravel this mystery is in our media records. Specifically, films of non-agrarian populations through the latter half of the 20th century. By visually inspecting these people, we can observe an inflection point where people on average were, by our current standards, 'thin', and then started appearing slightly heavier (our equivalent of 'healthy'). This inflection point appears to be roughly around the 1970's in the US. It is worth noting that the US was the first country to recognize the growing incidents of obesity and overweight individuals in society. In fact they were often the objects of ridicule for having the fattest population, before obesity and overweightedness started appearing in other western nations. The super obese, as characterized as people who are physically immobile due to large amounts of body fat, started appearing in the 1980's in the US. The 1990's media shows many children who are overweight. An average person on the street as shown from the same media sources is also visibly overweight, although the baggy clothing fashion of the time sometimes makes it difficult to tell. The chins and the guts however are difficult to hide. A pattern emerges when considering this linear progression... Something happened in the 1970's (at least observable) in the US, that spurred this widespread weight gaining tendencies. Let us for now shelve this observation and examine some of the other clues.
The second clue we have is the increased rate of overweighteness and significantly higher occurrences of diabetes among children. As already stated, this pattern appears to arise in the 1990's in the US, possibly slightly before in the 1980's. We can theorize on two broad categories of sources - one due to environmental factors and another on genetic factors. If one was to make a case for the genetic causes of overweightedness and diabetes, the genetic makeup of the parents and the genetic makeup of the children would need to be studied. Possibly a mutation or a set of genetic markers for increased weight gain would be investigated. I would imagine that such studies could show that these children were predisposed to weight gain due to genes expressed in a certain way that allowed this to happen. The conclusion might even argue that such genes are necessary in nature for populations that experience famines. The cause might be due to something that affected the parents and in turn which inadvertently produced the result in the child. Therefore it is a second order effect due to something that took place in the parents in the 1980's and the 1990's. If on the other hand environmental causes were investigated, one could say that there was something directly introduced to the children in the 1980's and the 1990's as a first order effect. Both arguments have merit and in fact both can be true, as first order and second order effects can both contribute to the underlying result. Let us look at the third clue.
Widespread overweightedness, though originally only observable in the Unites States, became commonplace in other western nations in later decades. The phenomena however was not uniform across populations in Europe, likely due to differing habits, genetics, culinary preferences and other varying influences of the underlying cause(s). But if I was start from somewhere and perhaps pick a straw-man example, one notable nation with this problem would be the UK. Lack of access to medical records forces me to consider only visual inspection of film records similar to how it was done for clue #1. However it would be prudent for anyone interested in this topic to correlate medical records and look for patterns in weight, waist size, and other chronic diseases such as diabetes. My observations show that the overweight problem began in the 1990's for the UK. The incidents of beer-bellies and bad diet before this point had already existed in the population. But the overweight pattern was not observable in populations that did not have such habits, until this point in time. Therefore we can surmise that there was an external stimulus that either introduced the same underlying causes affecting the US population, or that caused a different set of effects that had the same result as the US population.
So what can be said when combining these clues? A possible linear narrative is that there was a fundamental cause arising in the 1970's in the US that broadly impacted large segments of society. This caused the next generation of people in the 1980's and 1990's to have weight problems at an earlier age. The effects as observed in the 1970's US population became visible in the UK population in the 1990's presumably due to the same underlying causes. Therefore the question to solving this mystery is in pinpointing the impetus that spurred this chain of events in the 1970's in the US.