| OHDSI Home | Forums | Wiki | Github |
Hi there,
I see the webpage on drug dose estimation
drug_dose.knit (ohdsi.github.io)
Is it possible to have a chat with the person/team who suggests this information? We want to put effort into extending this information.
Any knowledge that can be shared with this regard is highly appreciated.
@anthonysena @Christian_Reich
Thank you and best wishes,
Theresa
was there any form of publication we can refer to with regards to dose estimation in OMOP CDM other than 2-poster-Medication-dosage-and-exposure-duration.png (11520Ă—8640) (ohdsi.org)
Slide 1 (ohdsi.org)
Not yet. There is that idea about calculating daily doses and finding out all sorts of interesting insights, but nobody has done anything. Want to go for it?
part of my job description I guess 
Bring it on. 
We would need to create some generic script for that. Most ETLs are done in SQL.
We will create formulas to estimate daily doses using drug_strength tables, that is the plan. What can you tell me about “box size”? according to the vocabulary it is this: “The number of units of Clinical Branded Drug or Quantified Clinical or Branded Drug contained in a box as dispensed to the patient”, however, we see (concept id 586584, ceftolozane 1170 MG / tazobactam 500 MG Injection Box of 10), box_size is filled with 10, however, we doubt that this was the amount of units given to the patient, we see this problem mainly with injection boxes, another example (concept id 21053896, Ceftazidime 2000 MG Injection Box of 10 by A A H), box size filled with 10. Any ideas? @Christian_Reich
The box size is the standard packaging size the drug is marketed in. So, it should contain the units. These are pills and tablets etc. for solid drugs, and flasks or bottles for liquids. The standardization also creates standard prescriptions, by which the box usually is sized to contain one, two or three months worth of prescription. But that depends on the country. In some countries, drugs are not packaged at all (like in the US), the pharmacists has to manually count the quantity provided in the prescription into these yellow plastic flasks you can see in every other Hollywood movie.
Bottom line: This is something the ETLer of the source data should tell you. But we should also improve our documentation.
Thank you @Christian_Reich, maybe I should have clarified that the example I gave are not self-administered drugs, but injected by health care personnel (in the hospital e.g.). We have this information from the “drug_strength” table, i.e. the vocabulary, so it should not be source specific?
Ah! That’s different. I didn’t realize you are talking injectables. The vocabulary contains the product the way it is referred to by the pharmacy distribution centers. It contains 10 flasks, each with 1g ceftolozane and 0.5g tazobactam. If they are administered in the hospital usually each administration is a record in itself. So, if you have 2 in quantity that means 2 of those flasks were injected that day, which is in line with how you dose this antibiotic battle tank. Same thing for the 2g ceftazidime injection.
You may want to check if they continue the dosing the next day and the day after, to be sure.
So what you are saying is to disregard box_size in this situation and only focus on quantity? @Christian_Reich I repeat my example from above “however, we see (concept id 586584, ceftolozane 1170 MG / tazobactam 500 MG Injection Box of 10), box_size is filled with 10, however, we doubt that this was the amount of units given to the patient, we see this problem mainly with injection boxes, another example (concept id 21053896, Ceftazidime 2000 MG Injection Box of 10 by A A H), box size filled with 10”
Exactly. The hospital pharmacy buys the boxes with 10 flasks in it each. The 10 is part of the product information. But they use 2 or 3 on a patient only. Not 10. That would be an overdose anyway.
Can you check the subsequent days?
We are not using the drug exposure to come up with a logic for daily dose but only the drug_strength table so far (of course quantity will be an aspect of the daily dose formula). From a clinical point of view, I would also suggest to look at when the next administration took place to help work out the daily dose but we tried to find daily dose formulas based upon the vocabulary.
