It is Genetic susceptibility and the way reacts to envtl Triggers.

Like they say, where there is a smoke, there is a fire.
You need to look within your genome, where there is a fuse blown out in genes/ then you can tell which child would have problems w meds or food/nurture.

BTW, I found something that may interest you, as you always quote asthma example.
Here it is~
Not All Patients Who Take Asthma Medications Have Asthma

LONDON (Reuters Health) Nov 23 - The prescription of asthma medications does not predict accurately whether a patient has asthma, according to a report in the November British Journal of General Practice.

Prescription databases are often used to assess quality of care, the authors explain, but the degree to which asthma drugs identify asthma patients has not been known.

Dr. Wolfgang Himmel and colleagues from the University of Gottingen in Germany investigated the correlation between certain drugs and the diagnosis of asthma, using prescription data from a retrospective chart review of eight general practices and one community respiratory medicine practice in northern Germany.

Asthma drugs were prescribed to 632 patients during the study period, only 309 of whom were diagnosed as having asthma or ****asthma-like symptoms, the authors report. One third of the patients were diagnosed with chronic obstructive pulmonary disease (COPD), and 34 patients had no diagnosis documented in the context of the asthma drug prescription.

The only highly specific drug class for asthma was cromoglycate plus betamimetics in fixed combinations, the researchers note. All but two patients who received a drug in this class had asthma. Rather than identify asthmatics, the investigators say, the use of systemic or inhaled corticosteroids identified patients with COPD.

"These results may reflect (a) shortcomings in the quality of asthma treatment, (b) shortcomings in the doctors' diagnostic reliability, or (c) shortcomings in the quality of the suggested prescribing indicator," the researchers suggest.

"As long as identification of patients is not, or is only to a limited degree, made possible through their prescriptions," they advise, "diagnosis-linked prescribing data should be used whenever possible, whereas prescribing data without diagnosis linkage should be interpreted with caution."

Br J Gen Pract 2001;51:879-883>>