Phase I trials, for example, are usually carried out under constant medical supervision and push the dosage as high as they can until 50% of the cohort will no longer tolerate it. That alone is usually thousands or tens of thousands per day per patient.
By the time you get to phase III where you have hundreds to thousands of patients, you're shelling out up to $1k per patient for the incentive and $100-1k per checkup for the length of the trial. A 1k person phase III can easily cost tens of millions in labor just to administer the drug and collect samples once a month for a year. Data provenance and normalization is important so a not insignificant chunk of this time is largely unavoidable paperwork.
>it takes a hundred dollars of drug revenue 17 years from now to motivate someone to invest one dollar today. No wonder drugs are expensive.
That is a 10,000% return on investment required to break even.
http://www.cureffi.org/2019/04/29/financial-modeling-in-rare...
and
http://sitn.hms.harvard.edu/flash/2020/modern-drug-discovery...
are decent reads.
Mostly it's the complexity of the compounds involved, the supply chain, and regulatory costs.