I'll keep the self-promotion to a minimum, but I have been spending lots of time reading on the science consensus on how children actually become fluent readers as part of my upcoming product Pictures Are For Babies (
https://picturesareforbabies.com), a literacy program that uses a deliberate practice engine I created to teach literacy from A-B-C to post-secondary level.
Phonics is all the rage, and I was planning to make it central to my pedagogy, but it turns out the answer is a bit more complicated, especially if you want to work with children with reading difficulties.
Phonics is part of the answer, but it's only the first step. Introducing children to the explicit mapping of graphemes to phonemes (letter to sounds) teaches decoding, but skilled reading is not decoding.
Actual reading is developed through a process called orthographic mapping. The result of this process is storing the grapheme to phoneme mappings in long-term memory for immediate retrieval. The words stored in this way form a sight vocabulary that spans tens of thousands of words in fluent readers.
When taught only phonics, kids run the risk of plateauing in later grades. It's not evident at first because the material they are given is simple and deals with concrete subjects (e.g. "Mike got a bunny for his birthday"). Later material uses many more words that don't follow phonics "rules" and deal with abstract material. Under these circumstances, decoding is too slow and effortful and leaves little remaining capacity to deal with harder tasks like comprehension.
The main cause of issues in developing this sight vocabulary is phonological deficits, not IQ, motivation, intelligence, visual processing, or attention like one might imagine. Kids with these deficits have trouble understanding that words are made up of smaller sound units and cannot work with them. Because of that, they cannot store the mapping efficiently and their vocabulary and fluency is limited.
Thankfully, the best interventions that fix these deficits are not too complicated and can correct the issues with as little as a dozen of hours of correct instruction. The main drawback is that finding and targeting those deficits is time-consuming for the instructors, but my program deals with that through the practice engine, which automates all that work.
The bad news is that most teachers are not aware of this and are simply being moved to phonics, which will not work for all children unless those phonemics deficits are identified and remediated. Worse news is that most commercial products that claim to be evidence-based or backed by the "science of reading" still use phonics and make no mention of orthographic mapping, the actual process that produces fluent readers. Again, phonics instruction is part of the answer, but nowhere near the entire story.
You can look at my pedagogy document for more info. Although it's meant to be about my product, it still contains a primer of the actual research on how full literacy (not just reading, but writing as well) is developed: https://picturesareforbabies.com/home/pedagogy/