Thread by @Argenscore

📌1. I am excited and proud to be able to present to you: Just published our study:

RISK ESTIMATION IN HEART SURGERY IN THE "REAL WORD": ARGENSCORE ADJUSTED TO THE CENTER.
@mmamas1973 @dr_benoy_n_shah @VictorDayan1 @mirvatalasnag @DrMoritzWvB @pomyers

bit.ly/2T3mw0l

2. You may get a message on your computer saying: some files may be harmful to your computer. This is because you are going directly to open a free-form .pdf file. This file comes from a fully trusted source. Please open the file & don't miss this excellent article.

3. Congratulations @CONAREC_ORG! The only way to know our realities is through multicenter and prospective records. One situation is the RCT and/or trials and another very different situation is the registries ("Real World"). This impacts on the best treatment of our patients.

4. Analysis: 2548 patients from 44 centers of the prospective & multicenter registry in cardiac surgery @CONAREC_ORG XVI. In each center the mean observed mortality was evaluated and the mean estimated mortality of each center was calculated applying both versions of ArgenSCORE.

5. Heterogeneous mortality (1.3%-25%). 28 (63.6%) centers: mortality > than 7%. 34 (77.3%) centers: mortality > 6%.
ArgenSCORE was developed in 1999 in a population w/a mortality of 8%: ArgenSCORE I
ArgenSCORE was recalibrated in 2007 (population w/ 4% mortality): ArgenSCORE II

6. Hypotheses: After recalibrating the model, for a similar absolute additive value, ArgenSCORE I estimates a higher risk & ArgenSCORE II estimates a lower risk. Arg I would better estimate risk in centers w/> mortality. Arg II would better estimate risk in centers w/ < mortality

7. In centers w/mortality rate <6%, where the ArgenSCORE II must be used, the score absolute additive value in a high-risk patient (>8%) is >37 points (this score is the sum of the patient's risk factors). In centers with mortality rate >6%, where the ArgenSCORE I must be used...

8... the score absolute additive value in the same patient is >25.8 points. In other words, for the same clinical risk, the predicted mortality changes according to the center where the procedure will be performed (depending on which version of ArgenSCORE
the center should use).

9. Conversely, in centers with mortality <6%, the score absolute additive value in an intermediate risk patient (4-8%) is >29, while in centers with mortality >6%, this value is only >16.5. 8.

10. In a low-risk patient (<4) the score absolute additive value is ≤29 in centers with mortality rates <6%, whereas in centers with mortality >6%, where the ArgenSCORE I must be used, it is only ≤ 16.5.

11. Interestingly, the different variables that constitute the ArgenSCORE, which are expressed as a sum of additive values to calculate an estimated risk, have different weights in centers with lower mortality than in those with higher mortality (different estimated risks).

12. In centers with mortality lower than 6% it is recommendable to use the recalibrated ArgenSCORE II- and in centers with mortality higher than 6% the original ArgenSCORE I.

📌13. We propose an Algorithm to be used for risk estimation according to the reality of the center (our "real world"):

1. Start prospective incorporation of all your surgical data into a database.

2. Calculate mean OM of your center.

14.
3. Calculate mean EM of your center in a prospective fashion. Use the model or models you trust most.

4. Evaluate the OM/EM ratio of your center.

5. Identify the risk score with the value closest to 1. Use the Z test to calculate significant differences in the OM/EM ratio.

15.
6. Use the risk score with non-significant differences.

7. If all the scores produce significant differences, identify the one closest to 1 and apply the following formula, described by Jin and et al. and others:

EM by ArgenSCORE/CF = EM by recalibrated ArgenSCORE

16.
For example:
Your center has OM/EM ratio 8% / 4% = 2

Your patient has an EM by ArgenSCORE: 3.5%.

Formula: EM by ArgenSCORE/CF = EM by recalibrated ArgenSCORE CF (calibration factor): OM/EM ratio: 4% / 8% = 0.5

3.5 / 0.5 = 7%.

📌17.

Your patient's estimated risk by recalibrated ArgenSCORE is 7%.

Obviously, it is better to recalibrate the score in the entire database of the center and thus obtain a score recalibrated or adjusted to the reality of the center. @mirvatalasnag @mmamas1973 @DrMoritzWvB

@rahatheart1 @EduQuintanaCVS @dompagano
@gbiondizoccai @OuzounianMD @DFCapodanno
@PhilippLurz @vonBardelebenRS @sabeedak1
@CarreroCeleste @drdargaray @PCRonline @djekhar_hadjer @FernandoBiguria @Pooh_Velagapudi @rachkataria @MeliAntoniolli
@Baratta_Sergio @ESC_Journals @pash22

@JCuencaCastillo @rafasadaba @juanjopaganini6 @JAHerediaDelga1 @RMSSAA @daviftenorio @GiuseppeGalati_ @tamaranihici @j_alvarezgarcia @cardiopole @crfheart @NimeshDesaiMD @NunziaBorrell12 @m_taramasso @gbiondizoccai
@DLBHATTMD @drandrewsharp @DrMarthaGulati @MarcDweck @falcardio

18. The Heart Team and the attending physician should
know the actual results of cardiovascular surgery and
interventional cardiology procedures in their centers
to make the best therapeutic decisions in the real
world and thus improve the risk-benefit equation for
the patient.

@dompagano @GlobalCardiac @carlaagatiello1
@PLooserMD @RayburnTom @chadialraies
@shisadoctor @CACIarg @RevEspCardiol
@SOLACI3 @krychtiukmd @CardiologiahoyC
@ClaudiaAlmont17 @hemodinamiasac @SACemergencias @SOLACI3 @cardio_hmc
@PabloAvanzas @Nkn63 @DomingoPascualF
@xgmoll

@mi_cardiologo @varrarte @SECTCV @EstebanDL
@CousiratLiendro @GlobalCardiac @EACTS @PeroneFrancesco @saia_francesco @IHamamsy
@JacoboSilvaGuis @jisaezdeibarra @igeorge1975 @DrIribarne @BavariaMd @PhilGenereuxMD
@sbrugaletta @josejgdnews @jrr242
@ahmedmakhdoum @JACCJournals