@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli To clarify, "outpatient PCI" has nothing to do with whether the patient stayed in the hospital overnight or not. Since ~2013, an outpatient PCI is one for which the patient wasnot expected to be in the hospital for 2 midnights (i.e., virtually all elective PCI procedures).(1/x)
@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli Prior to 2013, the rules were a bit more nebulous but a large proportion of PCIs for SIHD/UA would have been coded as outpatient status as well-- at least starting in 2008, when the RAC audits to identify improper use of inpatient status for PCI went into effect. (2/x)
@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli Given these definitions, I suspect that the NIS is missing at least half of all SIHD/UA PCIs over the last few years of the study and a somewhat smaller fraction in the years prior to 2013. (3/x)
@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli As such, it is possible that the vast majority of the trend toward decreasing annual PCI volumes is explained not by a true decrease in procedural volume but a shift toward outpatient coding. (4/x)
@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli In addition to causing marked distortions in the procedural volumes, these same trends would be expected to result in distortion of the clinical and demographic characteristics of the PCI population as well as in the risk-adjusted outcomes for the SIHD/UA PCI group as well. (5/x)
@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli And unfortunately, there is no straightforward way to get the correct volumes, patient characteristics, or risk-adjusted outcomes on a population basis at the present time. NCDR is probably the closest one can get. (6x)
@ATSandhu @venkmurthy @AnkurKalraMD @rwyeh @annalsthorsurg @drjohnm @SajjadRazaMD @JAMANetworkOpen @adnanalkhouli Which is why @rwyeh provided his PSA upstream-- you can't study a procedure with a meaningful outpatient presence using an inpatient database. (7/7).
