I am trying to see if the use of 2 particular drugs for Type 2 diabetes results in fewer hospitalizations, but once I extract my data, I am using Stata, download and run the .dat file, set the age limit, set the icd10code to “E11”, I have no observations again on the drug used and numbers of nights in the hospital. I need help and guidance on it.
I took a look at your most recent IPUMS MEPS extract, which includes data from the 2018 MEPS in a hierarchical extract format. There are only 2,680 condition records with an ICD10 code of E11. I suspect that after making additional restrictions to your analytical sample, you may not be left with any cases matching the requirements because the sample is simply too small.
I would suggest reviewing our documentation on MEPS record types. There are multiple record types in a hierarchical MEPS extract: event records (E), condition-event link records (J), event-medication-round link records (K), person records (P), and condition records (X). Age is a person-level variable, while ICD10 code is a condition-level variable. Make sure that when you are placing restrictions on the analytical sample, you do not accidentally make mis-matched requirements of variables and record types. For example, you cannot make a restriction on the age value of a condition record, or the ICD10 code of a person record. Condition records do have the associated SERIAL and PERNUM values attached, so you can use those and other variables to make restrictions on the values of variables on multiple record types.