The more scientific literature I read, the more it would seem that the typical hatchery process doesn't significantly contribute to the health of wild populations. In fact, it is likely that they may even add further pressure on wild stocks.
To understand this, the distinction in "fitness"must be made between wild and hatchery fish. Wild fish arise from the natural spawning of two fish, obviously. The healthiest, most viable spawners get the best spawning grounds and the most desirable mates, the combination of good genes and good redds leads to high survival amongst the offspring of this favortable pairing. In contrast, weaker fish get less desirable mates and spawn in marginal spawning grounds. The result is a lower survival rate amongst this less-successful group of fish. This "survival of the fittest" mechanism drives the continuation of a strong population of salmonids.
Hatchery fish on the other hand, arise from a random pairing of a male and a female (assuming hatcheries have enough brood to spawn 1 female/1 male). Through random chance, a desirable female might be paired with the gametes of an undesirable male. The result of this pairing reduces the fitness of desirable female's progeny. In nature, the undesirable male would not have had much of a chance to reproduce and his set of undesirable genes would might be eliminated from the gene pool, thus increasing the overall health of the genepool. Through the hatchery process, the elimination of less-desirable fish at spawning does not occur and healthy fish cannot select healthy mates through random pairing of spawners.
When the juveniles hatch, the close quarters of hatcheries may cause an increase in cortisol levels which may cause juvenile hatchery fish to become more voracious feeders (Garner et al, 2010). The size at release is often larger amongst hatchery fish, and they are potentially more aggressive, driving their wild counterparts out of the optimal feeding areas. The increase in total juvenile numbers may evoke whats called a type III functional response where predators are attracted by the density of the prey (juvenile steelhead), increasing the overall predation on the population. Furthermore, hatchery fish that return to rivers are less successful in spawning, however if they do successfully manage to mate, their less-desirable genes are passed on and reduce the genetic fitness of the population. Furthermore, they take up spawning area and may dig up the redds of wild fish if they should spawn later than a wild pair.
I think hatcheries have their place in several instances:
a) to support a fish an artificial fish population in an otherwise unhealthy rivers such as the Capilano where the Cleveland Dam has blocked access to spawning grounds and stopped cut off the supply of gravel to the lower reaches. Some salmon and steelhead still successfully spawn in-river, but there is little spawning habitat available due to the unfortunate placement of the Cleveland Dam.
b) provide a productive harvest fishery. People like eating fish, those people pay taxes too, so enhancing some rivers to provide this opportunity is fair and provides opportunity for businesses such as tackle shops, guiding, river-side accommodation etc.
I believe the best (only?) way to restore a wild population of fish is through habitat enhancement. This could take the form of installing fish-friendly culverts on coho tributaries, to ensuring the presence of proper gravel in spawning beds, to protecting riparian zones, marshes, estuaries, making sure there's structure in a stream, even stream fertilization.