On June 9, 2017 at 10:25:39 AM, Marsolo, Keith (keith.mars...@cchmc.org<mailto:keith.mars...@cchmc.org>) wrote:
Hi Russ. Thanks for reaching out. You’re correct in that CDM v3.1 lacks a dose field (and dose units). As I mentioned in a few recent calls, we’re going to address that in the next version of the data model. Quantity is really number of units (pills, tablets, etc.). I think we can work on improving the documentation. There are two options in the short term: 1. Add DOSE and DOSE_UNIT to your tables. It won’t break any of our existing query tools, and you’ll already have most of the ETL worked out once their added officially. 2. Use the preferred RxNorm CUI term types (listed in Implementation Guidance), which encode Ingredient + strength (dose) + dose form. This would give you all the information you need (and is one of the reasons why those fields aren’t present). This 2nd option is the preferred one. I agree that the more specific TTYs (BPCK and GPCK certainly, and SBD and SCD in most cases) get us to a manufactured dose. I’d call them equivalent to NDCs for most research purposes. More to help with discussion, option 2 helps you estimate duration of exposure, but gives limited information on how many units a patient takes at a time. That is to say, if I know the drug is, say, ranitidine 150 mg caps, and the patient got 60 caps with 3 refills, I can infer from general knowledge H2A usage and of how payers operate that the dose is probably 1 cap bid, but there’s nothing in the data that tells me that; I can only make that call because I assume the payer won’t reimburse for >30 days at a time and the prescriber wrote to that benchmark. If I’m wrong and the mail-order plan fills 60 days, I’m sunk. And it’s even more fun if the BPCK is, say, Teva amoxicillin 40 mg/ml solution, 250 ml and the unit is “bottle”. So for now option 1 is the choice you need if you want to study dose. The OMOP effective_drug_dose goes a step further, in that it allows normalization, so that for instance we can report pediatric doses as 80 mg/kg/dose, rather than just 1200 mg. I’m not sure PCORnet would contemplate that step, since it’s not focussed on pediatrics, but the nice thing about an approach like effective_drug_dose is that it’s transparent: if you don’t want to normalize, it accepts absolute dose, and if you do, it accepts normalized dose. But all nicety comes with its downside: query writers need to account for both if they’re selecting on dose level. (So do analysts, but they need to do that anyhow; what the normalized dose does is not require the dataset to also include weight or BSA values.) On Jun 9, 2017, at 9:25 AM, Russ Waitman <rwait...@kumc.edu<mailto:rwait...@kumc.edu>> wrote: Hi Keith and DRNOC, I see an issue perhaps with this new field or I am missing something. The Prescribing table is managing information at a level above NDC so we can represent inpatient and outpatient medication orders. From a research perspective, one would want to know the dose and units of a drug a patient receives. That was what I thought we’d be getting with a unit field to match up with the quantity field. Instead, it appears the guidance is to populate this with things like “Pill”. That really doesn’t to me offer any research value to understand how much of a drug a patient received. Furthermore, when you’re dealing with things at the prescribing/order level, you have no idea what the final dispense formulation would be. That’s going to be up the dispensing pharmacy in many cases or the hospital pharmacy compounding. So, we in general will have things out of epic at the level of gentamicin 80 MG q12h etc. I would have thought we’d see 80 in the quantity and MG in the unit. But this seems otherwise and of less value. Is there a thread and documentation where this was discussed? If I look at other modelling work in this space like omop http://www.ohdsi.org/web/wiki/doku.php?id=documentation:cdm:drug_exposure<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.ohdsi.org_web_wiki_doku.php-3Fid-3Ddocumentation-3Acdm-3Adrug-5Fexposure&d=DwMFAg&c=P0c35rBvlN7D8BNx7kSJTg&r=ZV2YJfZerKSuowURj_x5M4Js9LgfYQJD64exlqG9Igo&m=8Xu_7ef_aXuHCmecd2O25RjEejoXqRNEDEe6Ei_GnqA&s=j612nRM75J5h1A5FqyMuEhE0e_FRk3nCj0ubKZAb5LY&e=> They have a nice field for effective_drug_dose No float Numerical value of Drug dose for this Drug Exposure record. With a corresponding dose_unit_concept_ id No integer A foreign key to a predefined concept in the Standardized Vocabularies reflecting the unit the effective_drug_dose value is expressed. We seem to lack that or have inconsistency as I currently read this new 3.1 field. Sincerely, Russ Waitman, PhD Director of Medical Informatics Associate Vice Chancellor for Enterprise Analytics Professor, Department of Internal Medicine University of Kansas Medical Center, Kansas City, Kansas 913-945-7087 (office) rwait...@kumc.edu<mailto:rwait...@kumc.edu> http://www.kumc.edu/ea-mi/<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.kumc.edu_ea-2Dmi_&d=DwMFAg&c=P0c35rBvlN7D8BNx7kSJTg&r=ZV2YJfZerKSuowURj_x5M4Js9LgfYQJD64exlqG9Igo&m=8Xu_7ef_aXuHCmecd2O25RjEejoXqRNEDEe6Ei_GnqA&s=rcizV5gO02DMZ57wf8g33vmgGWVMG7bi-RW63QD-tls&e=> http://informatics.kumc.edu<https://urldefense.proofpoint.com/v2/url?u=http-3A__informatics.kumc.edu_&d=DwMFAg&c=P0c35rBvlN7D8BNx7kSJTg&r=ZV2YJfZerKSuowURj_x5M4Js9LgfYQJD64exlqG9Igo&m=8Xu_7ef_aXuHCmecd2O25RjEejoXqRNEDEe6Ei_GnqA&s=H7tuZYjHvubnfpruOz00QFcK-HDEeb17ZrNIbWy1IeM&e=> http://www.gpcnetwork.org<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.gpcnetwork.org_&d=DwMFAg&c=P0c35rBvlN7D8BNx7kSJTg&r=ZV2YJfZerKSuowURj_x5M4Js9LgfYQJD64exlqG9Igo&m=8Xu_7ef_aXuHCmecd2O25RjEejoXqRNEDEe6Ei_GnqA&s=7BnZCtVEu3WG1Zysent9UZ5CDMrF3Gi77m8W_FojCmI&e=> – a PCORnet collaborative www.pcornet.org<https://urldefense.proofpoint.com/v2/url?u=http-3A__www.pcornet.org&d=DwMFAg&c=P0c35rBvlN7D8BNx7kSJTg&r=ZV2YJfZerKSuowURj_x5M4Js9LgfYQJD64exlqG9Igo&m=8Xu_7ef_aXuHCmecd2O25RjEejoXqRNEDEe6Ei_GnqA&s=qSKVNwF0GiDdd_d4-UaaTDZm9JRBfsxW3NOd-ugv1Fs&e=> PRESCRIBING Table Specification Field Name RDBMS Data Type SAS Data Type Predefined Value Sets and Descriptive Text for Categorical Fields Definition / Comments Data Element Provenance RX_QUANTITY_UNIT RDBMS Text(2) SAS Char(2) PI=Pill TA=Tablet VI=Vial LI=Liquid SO=Solution SU=Suspension OI=Ointment CR=Cream PO=Powder PA=Patch IN=Inhaler KI=Kit DE=Device NI=No information UN=Unknown OT=Other The unit associated with the quantity prescribed. New field added in v3.1. PCORnet, based on NDC attributes — Regards, Charlie
_______________________________________________ Gpc-dev mailing list Gpc-dev@listserv.kumc.edu http://listserv.kumc.edu/mailman/listinfo/gpc-dev
