Correlation isn't causation. There are literally a billion possible factors that could be responsible besides a direct relationship. Of course, the same applies to any research demonstrating the efficacy of Omega-3s, too.

This is the problem with a lot of nutrition studies: a huge number of completely false associations were and still are being invented because of an inability to control for important factors. For example, the "fiber theory" of health has turned out to be almost completely false: it appeared to be "healthy" because eating fiber meant eating fewer (digested) carbohydrates. The fiber itself did almost nothing for health.

It gets worse when massive chains of correlations are assumed to be causes. A is correlated with B, B is correlated with C, C is correlated with D, therefore A causes D. The whole mess of saturated fats, cholesterol levels, heart attacks, and heart disease is a case of this.