The right time-dependent statistics: this is the moment

The right time-dependent statistics: this is the moment European Journal of Cardio-Thoracic Surgery 0 (2018) 1 LETTER TO THE EDITOR The imbalance in baseline covariates is a common phenomenon in non- The right time-dependent statistics: this is the randomized observational studies and prohibits direct comparison, as acknowledged by the authors. Propensity score matching can then be the moment next best thing if a randomized controlled trial cannot be conducted because of practical and/or ethical reasons [5]. However, we question the quality of the a a,b a Jamie L.R. Romeo , Grigorios Papageorgiou , Johanna J.M. Takkenberg and matching procedure in this study as the authors neglected to explain why a a, M. Mostafa Mokhles * mere 25 of 43 (58%) patients was matched. After matching, important (near) Department of Cardiothoracic Surgery, Erasmus University Medical Center, statistically significant imbalances in baseline covariates still remained, poten- Rotterdam, The Netherlands tially confounding late PR. Most notably, echocardiographic follow-up time Department of Biostatistics, Erasmus University Medical Center, Rotterdam, was included, which is fundamentally erroneous as it is highly likely to correl- The Netherlands ate with severity of PR. We again like to compliment the authors on their efforts to reduce Received 5 March 2018; accepted 6 April 2018 chronic PR using an innovative surgical procedure. However, we do believe that potentially positive results are unnecessarily obscured by a poor statistical Keywords: Tetralogy of Fallot � Pulmonary valve replacement � Pulmonary approach and could benefit significantly from the use of mixed effects models. regurgitation With great interest we read the recent article by Kim et al. [1] published in the Funding European Journal of Cardio-Thoracic Surgery. Kim et al. presented the results of 43 patients who underwent a thoroughly-explained valvular repair technique This work was supported by a NWO Veni grant of the Netherlands during complete correction of tetralogy of Fallot. All patients survived, and Organisation for Scientific Research (NWO 916.160.87 to M.M.M.). pulmonary regurgitation (PR) grades and gradient, assessed immediately after the operation by echocardiography, were satisfactory. After a mean follow-up duration of 131.9 months, only 1 reoperation was performed which was not REFERENCES valve-related. Despite huge improvements in the surgical repair of tetralogy of Fallot, severe regurgitation continued to have detrimental effects on mortality and morbidity, often requiring reintervention. Therefore, we respect every ef- [1] Kim H, Sung SC, Choi KH, Lee HD, Kim G, Ko H, Lee YS. Long-term fort directed at alleviating this well-known burden. results of pulmonary valve annular enlargement with valve repair in tet- However, we found some major pitfalls in the subsequent statistical analyses. ralogy of Fallot. Eur J Cardiothorac Surg 2018; doi:10.1093/ejcts/ezx497. The severity of PR is assessed qualitatively as also recommended by the [2] Lancellotti P, Tribouilloy C, Hagendorff A, Popescu BA, Edvardsen T, guidelines [2]. Except for a rough categorization of severity, the numbers 0–4 Pierard LA et al. Recommendations for the echocardiographic assess- contain no quantitative information about PR and hold little relationship to ment of native valvular regurgitation: an executive summary from the each other. Therefore, the arithmetic mean value of PR which the authors re- European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc port does not convey any meaningful information. Any subsequent statistical Imaging 2013;14:611–44. test designed for continuous measurements is thereafter non-interpretable. [3] Andrinopoulou ER, Rizopoulos D, Jin R, Bogers AJ, Lesaffre E, Takkenberg Furthermore, the authors choose to report and analyse only the last known JJ. An introduction to mixed models and joint modeling: analysis of valve echocardiographic measurements. Hereby, they effectively treated continuous function over time. Ann Thorac Surg 2012;93:1765–72. processes as independent events without any regard for time by reporting [4] Mokhles MM, Rizopoulos D, Andrinopoulou ER, Bekkers JA, Roos- and comparing the incidence of PR with the v test and gradients with the Hesselink JW, Lesaffre E et al. Autograft and pulmonary allograft per- Student’s t-test. Unequal timing of echocardiographic assessment ‘always’ formance in the second post-operative decade after the Ross procedure: needs to be accounted for. This is important because the implicit but funda- insights from the Rotterdam Prospective Cohort Study. Eur Heart J 2012; mental assumption is that valve function ‘develops’ and does not ‘happen’. 33:2213–24. An ideal approach towards repeatedly measured outcomes is the use of lin- [5] Blackstone EH. Comparing apples and oranges. J Thorac Cardiovasc Surg ear (gradient in mmHg) and generalized (ordinal PR) mixed effects modelling, 2002;123:8–15. introduced by Andrinopoulou et al. [3] and demonstrated by Mokhles et al. [4]. Correlation between repeated measurements in the same patient and dis- balances in timing and number of observations can be accounted for by *Corresponding author. Department of Cardiothoracic Surgery, Erasmus including random effects. Patient characteristics (e.g. operative technique) can University Medical Center, ‘s Gravendijkwal 230, 3015 CE Rotterdam, Netherlands. be related to changes in individual temporal evolutions. Furthermore, we Tel: +31-010-7030771; fax: +31-010-7033993; e-mail: m.mokhles@erasmusmc.nl believe that general reservations towards its complexity are unnecessary. (M.M. Mokhles). Nowadays, even well-known statistical software packages (e.g. SPSS) provide elaborate possibilities and documentation without the need for extensive cod- doi:10.1093/ejcts/ezy184 ing skills. The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. Downloaded from https://academic.oup.com/ejcts/advance-article-abstract/doi/10.1093/ejcts/ezy184/4990543 by Ed 'DeepDyve' Gillespie user on 07 June 2018 LETTER TO THE EDITOR http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png European Journal of Cardio-Thoracic Surgery Oxford University Press

The right time-dependent statistics: this is the moment

Free
1 page

Loading next page...
1 Page
 
/lp/ou_press/the-right-time-dependent-statistics-this-is-the-moment-0srdvWo0K0
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.
ISSN
1010-7940
eISSN
1873-734X
D.O.I.
10.1093/ejcts/ezy184
Publisher site
See Article on Publisher Site

Abstract

European Journal of Cardio-Thoracic Surgery 0 (2018) 1 LETTER TO THE EDITOR The imbalance in baseline covariates is a common phenomenon in non- The right time-dependent statistics: this is the randomized observational studies and prohibits direct comparison, as acknowledged by the authors. Propensity score matching can then be the moment next best thing if a randomized controlled trial cannot be conducted because of practical and/or ethical reasons [5]. However, we question the quality of the a a,b a Jamie L.R. Romeo , Grigorios Papageorgiou , Johanna J.M. Takkenberg and matching procedure in this study as the authors neglected to explain why a a, M. Mostafa Mokhles * mere 25 of 43 (58%) patients was matched. After matching, important (near) Department of Cardiothoracic Surgery, Erasmus University Medical Center, statistically significant imbalances in baseline covariates still remained, poten- Rotterdam, The Netherlands tially confounding late PR. Most notably, echocardiographic follow-up time Department of Biostatistics, Erasmus University Medical Center, Rotterdam, was included, which is fundamentally erroneous as it is highly likely to correl- The Netherlands ate with severity of PR. We again like to compliment the authors on their efforts to reduce Received 5 March 2018; accepted 6 April 2018 chronic PR using an innovative surgical procedure. However, we do believe that potentially positive results are unnecessarily obscured by a poor statistical Keywords: Tetralogy of Fallot � Pulmonary valve replacement � Pulmonary approach and could benefit significantly from the use of mixed effects models. regurgitation With great interest we read the recent article by Kim et al. [1] published in the Funding European Journal of Cardio-Thoracic Surgery. Kim et al. presented the results of 43 patients who underwent a thoroughly-explained valvular repair technique This work was supported by a NWO Veni grant of the Netherlands during complete correction of tetralogy of Fallot. All patients survived, and Organisation for Scientific Research (NWO 916.160.87 to M.M.M.). pulmonary regurgitation (PR) grades and gradient, assessed immediately after the operation by echocardiography, were satisfactory. After a mean follow-up duration of 131.9 months, only 1 reoperation was performed which was not REFERENCES valve-related. Despite huge improvements in the surgical repair of tetralogy of Fallot, severe regurgitation continued to have detrimental effects on mortality and morbidity, often requiring reintervention. Therefore, we respect every ef- [1] Kim H, Sung SC, Choi KH, Lee HD, Kim G, Ko H, Lee YS. Long-term fort directed at alleviating this well-known burden. results of pulmonary valve annular enlargement with valve repair in tet- However, we found some major pitfalls in the subsequent statistical analyses. ralogy of Fallot. Eur J Cardiothorac Surg 2018; doi:10.1093/ejcts/ezx497. The severity of PR is assessed qualitatively as also recommended by the [2] Lancellotti P, Tribouilloy C, Hagendorff A, Popescu BA, Edvardsen T, guidelines [2]. Except for a rough categorization of severity, the numbers 0–4 Pierard LA et al. Recommendations for the echocardiographic assess- contain no quantitative information about PR and hold little relationship to ment of native valvular regurgitation: an executive summary from the each other. Therefore, the arithmetic mean value of PR which the authors re- European Association of Cardiovascular Imaging. Eur Heart J Cardiovasc port does not convey any meaningful information. Any subsequent statistical Imaging 2013;14:611–44. test designed for continuous measurements is thereafter non-interpretable. [3] Andrinopoulou ER, Rizopoulos D, Jin R, Bogers AJ, Lesaffre E, Takkenberg Furthermore, the authors choose to report and analyse only the last known JJ. An introduction to mixed models and joint modeling: analysis of valve echocardiographic measurements. Hereby, they effectively treated continuous function over time. Ann Thorac Surg 2012;93:1765–72. processes as independent events without any regard for time by reporting [4] Mokhles MM, Rizopoulos D, Andrinopoulou ER, Bekkers JA, Roos- and comparing the incidence of PR with the v test and gradients with the Hesselink JW, Lesaffre E et al. Autograft and pulmonary allograft per- Student’s t-test. Unequal timing of echocardiographic assessment ‘always’ formance in the second post-operative decade after the Ross procedure: needs to be accounted for. This is important because the implicit but funda- insights from the Rotterdam Prospective Cohort Study. Eur Heart J 2012; mental assumption is that valve function ‘develops’ and does not ‘happen’. 33:2213–24. An ideal approach towards repeatedly measured outcomes is the use of lin- [5] Blackstone EH. Comparing apples and oranges. J Thorac Cardiovasc Surg ear (gradient in mmHg) and generalized (ordinal PR) mixed effects modelling, 2002;123:8–15. introduced by Andrinopoulou et al. [3] and demonstrated by Mokhles et al. [4]. Correlation between repeated measurements in the same patient and dis- balances in timing and number of observations can be accounted for by *Corresponding author. Department of Cardiothoracic Surgery, Erasmus including random effects. Patient characteristics (e.g. operative technique) can University Medical Center, ‘s Gravendijkwal 230, 3015 CE Rotterdam, Netherlands. be related to changes in individual temporal evolutions. Furthermore, we Tel: +31-010-7030771; fax: +31-010-7033993; e-mail: m.mokhles@erasmusmc.nl believe that general reservations towards its complexity are unnecessary. (M.M. Mokhles). Nowadays, even well-known statistical software packages (e.g. SPSS) provide elaborate possibilities and documentation without the need for extensive cod- doi:10.1093/ejcts/ezy184 ing skills. The Author(s) 2018. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved. Downloaded from https://academic.oup.com/ejcts/advance-article-abstract/doi/10.1093/ejcts/ezy184/4990543 by Ed 'DeepDyve' Gillespie user on 07 June 2018 LETTER TO THE EDITOR

Journal

European Journal of Cardio-Thoracic SurgeryOxford University Press

Published: Apr 30, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off