ohdsi-studies / ticagrelorvsclopidogrel Goto Github PK

I need to validate the outcome defiinition

Imputation death

Some negative controls seem to have potential relationship with the drug or cardiovascular diseases.
So the negative controls should be revised

Adding 28-day blanking period for MI should be considered

Ref

1. Sahlén, Anders, Christoph Varenhorst, Bo Lagerqvist, Henrik Renlund, Elmir Omerovic, David Erlinge, Lars Wallentin, Stefan K. James, and Tomas Jernberg. “Outcomes in Patients Treated with Ticagrelor or Clopidogrel after Acute Myocardial Infarction: Experiences from SWEDEHEART Registry.” European Heart Journal 37, no. 44 (November 21, 2016): 3335–42. https://doi.org/10.1093/eurheartj/ehw284.
   2.Friberg L, Benson L, Rosenqvist M, Lip GY. Assessment of female sex as a risk factor in atrial fibrillation in Sweden: nationwide retrospective cohort study. Br Med J 2012;344:e3522.
   3.Gulliksson M, Wedel H, Koster M, Svardsudd K. Hazard function and secular trends in the risk of recurrent acute myocardial infarction: 30 years of follow-up of more than 775,000 incidents. Circ Cardiovasc Qual Outcomes 2009;2:178 – 185.

All the OHDSI source packages are available in the drat repository , that would be my best bet for installing packages offline. The biggest challenge is the dependencies of those packages that live in CRAN.

There’s a function called insertEnvironmentSnapshotInPackage in the OhdsiRTools package that creates a list of all packages (and version numbers) that need to be installed to run a study, and adds that list to the study package. An example can be found here.

I need to add mortality for another secondary outcome and dyspnea for positive control

I need to add interaction analysis for gender, MI only, PPI use, and etc.

Though I'm not sure I must do, but it would be better to assess drug continuation and the change of the drugs

I fixed what I thought to be the reason of the error but couldn't fix the error.
Do you have any ideas why target and comparator cohorts are empty in my case?
Thank you.

1-on-1 matching using a non-conditioned outcome model might be better, because that usually gives you more power than variable-ratio matching with a conditional model (because the conditioning is very data-hungry).

In the Method Evaluation, we see evidence to support that: http://data.ohdsi.org/MethodEvalViewer/

I get the error below. I wonder if you've tested the package with the package cohortMethod 3.0.2.
Or is it the error only occuring with me?
Thank you!

Error in fitOutcomeModel(population =args$population, cohortMethodData = args$cohortMethodData, : Requested stratified analysis, but no stratumID column found in population. Please use matchonPs or stratifyByPs to create strata.

I need to specify index year for the cohort

I need to add ITT analysis

Add interaction term analysis for maintenance aspirin dosage based on this paper

Should gout be added for safety profile?

Ajit recommended to include positive control synthesis for HR and CI calibration. This helps avoid CIs that don’t span 1 but have calibrated p-value > 0.05.

broad and narrow definition for the outcome should be added for sensitivity analysis.

add concomitant use of unfractionated heparin, low-molecular heparin, bivalirudin, glycoprotein IIb/IIIa inhibitor in baseline characteristics

modify minimum required Time-At-Risk in one year analysis from 1 to 364 days for one-year analysis

• Sometimes, DB doesn't have visit ID linkages between the drug exposure and the PCI.
  This can be relaxed by time-based rule.

The docker image should be updated

Someone is using R3.5.1 and have installed RTools 3.4 and 3.5, but still he's getting the following error message.

WARNING: Rtools is required to build R packages, but no version of Rtools compatible with R 3.5.1 was found. (Only the following incompatible version(s) of Rtools were found:3.4,3.5) Please download and install the appropriate version of Rtools from http://cran.r-project.org/bin/windows/Rtools/.

Redshift does not like st. amount_value != '' because in our data, amount_value is numeric and redshift is fussy about comparing numeric to character ('') Changing to st.amount_value::char != '' fixed the problem.

Should be more careful with missing / zero days_supply and quantity, which is common in some DBs.

[inserting into #cov_6]
https://github.com/chandryou/TicagrelorVsClopidogrel/blob/master/inst/sql/sql_server/CreateSubgroups.sql

As Ajit recommended, I should include more covariates such as observation, devices, and visit

Docker image for reproducibility should be maintained.

The newest version of rJava (0.9-11) was also causing issues with DatabaseConnector. Rolling back to an earlier release helped.

I'm pretty sure this error is thrown because I use oracle instead of ms sql.
The error report says that "COHORT_NUMBER" is an invalid identifier so I wanted to locate this and change it to the some other word, but couldn't find it.

I wonder where this "Contructing target and comparator cohorts" message comes from (from which .R files or .sql files?)

I wonder if you could give me a hint or any ideas to fix it.

Thank you!

errorReport_SCYOU.txt

It would be better if I show the baseline characteristics or add interaction term analysis between Drug-Eluting Stent vs Bare Metal Stent

Current time-at-risk for primary analysis is described below.

CohortMethod::createCreateStudyPopulationArgs(removeSubjectsWithPriorOutcome = FALSE, firstExposureOnly = FALSE, washoutPeriod = 0, removeDuplicateSubjects = 'keep first', minDaysAtRisk = 364, riskWindowStart = 1, addExposureDaysToStart = FALSE, riskWindowEnd = 365, addExposureDaysToEnd = FALSE, censorAtNewRiskWindow = FALSE)

I think this should be revised from cohort generation, (2013.3.1~2017.1.1).

It seems that analysis 13 doesn't work.

I need to summarizing tool for the result

Extraction of full demographics