Early studies reported high success rates; in one study in 1925, 60% of patients became seizure-free, and another 35% of patients had a 50% reduction in seizure frequency. These studies generally examined a cohort of patients recently treated by the physician (a retrospective study) and selected patients who had successfully maintained the dietary restrictions. However, these studies are difficult to compare to modern trials. One reason is that these older trials suffered from selection bias, as they excluded patients who were unable to start or maintain the diet and thereby selected from patients who would generate better results. In an attempt to control for this bias, modern study design prefers a prospective cohort (the patients in the study are chosen before therapy begins) in which the results are presented for all patients regardless of whether they started or completed the treatment (known as intent-to-treat analysis).
A computer program such as KetoCalculator may be used to help generate recipes. The meals often have four components: heavy whipping cream, a protein-rich food (typically meat), a fruit or vegetable and a fat such as butter, vegetable oil, or mayonnaise. Only low-carbohydrate fruits and vegetables are allowed, which excludes bananas, potatoes, peas, and corn. Suitable fruits are divided into two groups based on the amount of carbohydrate they contain, and vegetables are similarly divided into two groups. Foods within each of these four groups may be freely substituted to allow for variation without needing to recalculate portion sizes. For example, cooked broccoli, Brussels sprouts, cauliflower, and green beans are all equivalent. Fresh, canned, or frozen foods are equivalent, but raw and cooked vegetables differ, and processed foods are an additional complication. Parents are required to be precise when measuring food quantities on an electronic scale accurate to 1 g. The child must eat the whole meal and cannot have extra portions; any snacks must be incorporated into the meal plan. A small amount of MCT oil may be used to help with constipation or to increase ketosis.
Ketogenic "eggnog" is used during induction and is a drink with the required ketogenic ratio. For example, a 4:1 ratio eggnog would contain 60 g of 36% heavy whipping cream, 25 g pasteurised raw egg, saccharin and vanilla flavour. This contains 245 kcal (1,025 kJ), 4 g protein, 2 g carbohydrate and 24 g fat (24:6 = 4:1). The eggnog may also be cooked to make a custard, or frozen to make ice cream.