Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion

Helena Martí; Jon Pérez-Bárcena; Miquel Fiol; Jaume Marrugat; Carmen Navarro; Elena Aldasoro; Adolfo Cabadés; Antonio Segura; Rafael Masiá; Javier Turumbay; Lluís Cirera; José M. Arteagoitia; Carlos A. Tomás; Gema Vega; Joan Sala; Enrique de los Arcos; María J. Tormo; Iraida Hurtado-de-Saracho; Mercedes Francés-Sempere; Roberto Elosua; Eva Alonso; Fernando Arós; Covadonga Audicana; Mikel Basterretxea; Eduardo Castillo; Santiago Esnaola; Miguel A. García Calabuig; Miren Josebe Laresgoiti; Nerea Larrañaga; María José Lasa; Iñaki Lecuona; Nerea Muniozguren; María Cres Tobalina; Jesús M. San Vicente; Emilio Sanz; José Garía; Consuelo Martínez; Josefa Contreras; Fuensanta Aliaga; José M. Alonso; Manuel Belda; Rafael Bañón; José Antonio Castaño; Antonio Castilla; Juan Manuel Chiner; José M. Clemente; Amparo Egea; Francisco Felices; Manuel Fuentes; José Galcerá; Arcadio García; Francisco García; Pedro García; Gurmesindo González; Faustino Herrero; Pedro Jara; José Antonio Melgarejo; José Muñoz; Silvestre Nicolás; Juan Ortega; Miguel de Paco; Emilio Pérez; Pascual Rodríguez; Francisco Javier Rodríguez; Julio Ródenas; Francisco Ruiz-Martínez; José Antonio Ruiz; Fernando de San Eustaquio; Josefa Segura; José Antonio Serrano; Gines Torres; Luis Carlos Torres; Juan Vidal; Manuel Villegas; Amparo Albiñana; Carlos Antón; Joaquín Arguedas; Santiago Borrás; Susana Bosch; José P. Calabuig; Enrique Castellanos; Javier Cebrián; Francisco J. Domingo; Ildefonso Echánove; Enrique Ferrer; Mercedes Francés; Eva Gómez; Julián González; José Gregori; Javier Haba; Antonio López; Vicente López; Immaculada Melchor; Rafael Montero; Julio Montes; Ramón Navarro; Manuel Palanca; Miguel Palencia; Miguel Pérez

Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion

Helena Martí, Jon Pérez-Bárcena, Miquel Fiol, Jaume Marrugat, Carmen Navarro, Elena Aldasoro, Adolfo Cabadés, Antonio Segura, Rafael Masiá, Javier Turumbay, Lluís Cirera, José M. Arteagoitia, Carlos A. Tomás, Gema Vega, Joan Sala, Enrique de los Arcos, María J. Tormo, Iraida Hurtado-de-Saracho, Mercedes Francés-Sempere, Roberto ElosuaEva Alonso, Fernando Arós, Covadonga Audicana, Mikel Basterretxea, Eduardo Castillo, Santiago Esnaola, Miguel A. García Calabuig, Miren Josebe Laresgoiti, Nerea Larrañaga, María José Lasa, Iñaki Lecuona, Nerea Muniozguren, María Cres Tobalina, Jesús M. San Vicente, Emilio Sanz, José Garía, Consuelo Martínez, Josefa Contreras, Fuensanta Aliaga, José M. Alonso, Manuel Belda, Rafael Bañón, José Antonio Castaño, Antonio Castilla, Juan Manuel Chiner, José M. Clemente, Amparo Egea, Francisco Felices, Manuel Fuentes, José Galcerá, Arcadio García, Francisco García, Pedro García, Gurmesindo González, Faustino Herrero, Pedro Jara, José Antonio Melgarejo, José Muñoz, Silvestre Nicolás, Juan Ortega, Miguel de Paco, Emilio Pérez, Pascual Rodríguez, Francisco Javier Rodríguez, Julio Ródenas, Francisco Ruiz-Martínez, José Antonio Ruiz, Fernando de San Eustaquio, Josefa Segura, José Antonio Serrano, Gines Torres, Luis Carlos Torres, Juan Vidal, Manuel Villegas, Amparo Albiñana, Carlos Antón, Joaquín Arguedas, Santiago Borrás, Susana Bosch, José P. Calabuig, Enrique Castellanos, Javier Cebrián, Francisco J. Domingo, Ildefonso Echánove, Enrique Ferrer, Mercedes Francés, Eva Gómez, Julián González, José Gregori, Javier Haba, Antonio López, Vicente López, Immaculada Melchor, Rafael Montero, Julio Montes, Ramón Navarro, Manuel Palanca, Miguel Palencia, Miguel Pérez

Department of Paediatrics, Obstetrics and Gynaecology and Preventive Medicine and Public Health

Research output: Contribution to journal › Article › Research › peer-review

6 Citations (Scopus)

Abstract

Introduction and objectives. Analysis of the effect of treatment in observational studies is complex due to differences between treated and nontreated patients. Calculating the probability of receiving treatment conditioned on relevant covariates (propensity score [PS]) has been proposed as a method to control for these differences. We report an application of PS to assess the association between reperfusion treatment and 28-day case fatality in patients with acute myocardial infarction (AMI). Method. We describe the procedure used to calculate PS for receiving reperfusion treatment, and different strategies to analyze the association between PS and case fatality with regression modeling and matching. Data were from a population-based registry of 6307 patients with AMI in Spain during 1997-98. Results. The PS for reperfusion was calculated in 5622 patients. In the multivariate analysis, reperfusion was associated with lower case fatality (OR= 0.59; 95% confidence interval [95% CI]: 0.46-077). When PS was included as a covariate, this association became non-significant (OR = 0.76; 95% CI: 0.57-1.01). In the subgroup of matched patients with a similar PS (n = 3138), treatment was not associated with case fatality (OR = 0.95; 95% CI: 0.72-1.26). When the influence of cases with missing data on PS was controlled for, reperfusion treatment was associated with lower fatality (OR = 0.66; 95% CI: 0.55-0.80). Conclusions. Calculating propensity score is a method that controls for differences between treated and nontreated patients. This score has limitations when matching is incomplete and when data are missing. Results of the present example suggest that reperfusion treatment reduces AMI case fatality.

Original language	English
Pages (from-to)	126-136
Journal	Revista Espanola de Cardiologia
Volume	58
Issue number	2
Publication status	Published - 1 Jan 2005

Keywords

Myocardial infarction
Prognosis
Reperfusion
Thrombolysis

Cite this

Martí, H., Pérez-Bárcena, J., Fiol, M., Marrugat, J., Navarro, C., Aldasoro, E., Cabadés, A., Segura, A., Masiá, R., Turumbay, J., Cirera, L., Arteagoitia, J. M., Tomás, C. A., Vega, G., Sala, J., de los Arcos, E., Tormo, M. J., Hurtado-de-Saracho, I., Francés-Sempere, M., ... Pérez, M. (2005). Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion. Revista Espanola de Cardiologia, 58(2), 126-136.

@article{267e7dd6f1d9406e96d8e58ca1ee15bb,

title = "Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion",

abstract = "Introduction and objectives. Analysis of the effect of treatment in observational studies is complex due to differences between treated and nontreated patients. Calculating the probability of receiving treatment conditioned on relevant covariates (propensity score [PS]) has been proposed as a method to control for these differences. We report an application of PS to assess the association between reperfusion treatment and 28-day case fatality in patients with acute myocardial infarction (AMI). Method. We describe the procedure used to calculate PS for receiving reperfusion treatment, and different strategies to analyze the association between PS and case fatality with regression modeling and matching. Data were from a population-based registry of 6307 patients with AMI in Spain during 1997-98. Results. The PS for reperfusion was calculated in 5622 patients. In the multivariate analysis, reperfusion was associated with lower case fatality (OR= 0.59; 95% confidence interval [95% CI]: 0.46-077). When PS was included as a covariate, this association became non-significant (OR = 0.76; 95% CI: 0.57-1.01). In the subgroup of matched patients with a similar PS (n = 3138), treatment was not associated with case fatality (OR = 0.95; 95% CI: 0.72-1.26). When the influence of cases with missing data on PS was controlled for, reperfusion treatment was associated with lower fatality (OR = 0.66; 95% CI: 0.55-0.80). Conclusions. Calculating propensity score is a method that controls for differences between treated and nontreated patients. This score has limitations when matching is incomplete and when data are missing. Results of the present example suggest that reperfusion treatment reduces AMI case fatality.",

keywords = "Myocardial infarction, Prognosis, Reperfusion, Thrombolysis",

author = "Helena Mart{\'i} and Jon P{\'e}rez-B{\'a}rcena and Miquel Fiol and Jaume Marrugat and Carmen Navarro and Elena Aldasoro and Adolfo Cabad{\'e}s and Antonio Segura and Rafael Masi{\'a} and Javier Turumbay and Llu{\'i}s Cirera and Arteagoitia, {Jos{\'e} M.} and Tom{\'a}s, {Carlos A.} and Gema Vega and Joan Sala and {de los Arcos}, Enrique and Tormo, {Mar{\'i}a J.} and Iraida Hurtado-de-Saracho and Mercedes Franc{\'e}s-Sempere and Roberto Elosua and Eva Alonso and Fernando Ar{\'o}s and Covadonga Audicana and Mikel Basterretxea and Eduardo Castillo and Santiago Esnaola and {Garc{\'i}a Calabuig}, {Miguel A.} and Laresgoiti, {Miren Josebe} and Nerea Larra{\~n}aga and Lasa, {Mar{\'i}a Jos{\'e}} and I{\~n}aki Lecuona and Nerea Muniozguren and Tobalina, {Mar{\'i}a Cres} and {San Vicente}, {Jes{\'u}s M.} and Emilio Sanz and Jos{\'e} Gar{\'i}a and Consuelo Mart{\'i}nez and Josefa Contreras and Fuensanta Aliaga and Alonso, {Jos{\'e} M.} and Manuel Belda and Rafael Ba{\~n}{\'o}n and Casta{\~n}o, {Jos{\'e} Antonio} and Antonio Castilla and Chiner, {Juan Manuel} and Clemente, {Jos{\'e} M.} and Amparo Egea and Francisco Felices and Manuel Fuentes and Jos{\'e} Galcer{\'a} and Arcadio Garc{\'i}a and Francisco Garc{\'i}a and Pedro Garc{\'i}a and Gurmesindo Gonz{\'a}lez and Faustino Herrero and Pedro Jara and Melgarejo, {Jos{\'e} Antonio} and Jos{\'e} Mu{\~n}oz and Silvestre Nicol{\'a}s and Juan Ortega and {de Paco}, Miguel and Emilio P{\'e}rez and Pascual Rodr{\'i}guez and Rodr{\'i}guez, {Francisco Javier} and Julio R{\'o}denas and Francisco Ruiz-Mart{\'i}nez and Ruiz, {Jos{\'e} Antonio} and {de San Eustaquio}, Fernando and Josefa Segura and Serrano, {Jos{\'e} Antonio} and Gines Torres and Torres, {Luis Carlos} and Juan Vidal and Manuel Villegas and Amparo Albi{\~n}ana and Carlos Ant{\'o}n and Joaqu{\'i}n Arguedas and Santiago Borr{\'a}s and Susana Bosch and Calabuig, {Jos{\'e} P.} and Enrique Castellanos and Javier Cebri{\'a}n and Domingo, {Francisco J.} and Ildefonso Ech{\'a}nove and Enrique Ferrer and Mercedes Franc{\'e}s and Eva G{\'o}mez and Juli{\'a}n Gonz{\'a}lez and Jos{\'e} Gregori and Javier Haba and Antonio L{\'o}pez and Vicente L{\'o}pez and Immaculada Melchor and Rafael Montero and Julio Montes and Ram{\'o}n Navarro and Manuel Palanca and Miguel Palencia and Miguel P{\'e}rez",

year = "2005",

month = jan,

day = "1",

language = "English",

volume = "58",

pages = "126--136",

number = "2",

}

Martí, H, Pérez-Bárcena, J, Fiol, M, Marrugat, J, Navarro, C, Aldasoro, E, Cabadés, A, Segura, A, Masiá, R, Turumbay, J, Cirera, L, Arteagoitia, JM, Tomás, CA, Vega, G, Sala, J, de los Arcos, E, Tormo, MJ, Hurtado-de-Saracho, I, Francés-Sempere, M, Elosua, R, Alonso, E, Arós, F, Audicana, C, Basterretxea, M, Castillo, E, Esnaola, S, García Calabuig, MA, Laresgoiti, MJ, Larrañaga, N, Lasa, MJ, Lecuona, I, Muniozguren, N, Tobalina, MC, San Vicente, JM, Sanz, E, Garía, J, Martínez, C, Contreras, J, Aliaga, F, Alonso, JM, Belda, M, Bañón, R, Castaño, JA, Castilla, A, Chiner, JM, Clemente, JM, Egea, A, Felices, F, Fuentes, M, Galcerá, J, García, A, García, F, García, P, González, G, Herrero, F, Jara, P, Melgarejo, JA, Muñoz, J, Nicolás, S, Ortega, J, de Paco, M, Pérez, E, Rodríguez, P, Rodríguez, FJ, Ródenas, J, Ruiz-Martínez, F, Ruiz, JA, de San Eustaquio, F, Segura, J, Serrano, JA, Torres, G, Torres, LC, Vidal, J, Villegas, M, Albiñana, A, Antón, C, Arguedas, J, Borrás, S, Bosch, S, Calabuig, JP, Castellanos, E, Cebrián, J, Domingo, FJ, Echánove, I, Ferrer, E, Francés, M, Gómez, E, González, J, Gregori, J, Haba, J, López, A, López, V, Melchor, I, Montero, R, Montes, J, Navarro, R, Palanca, M, Palencia, M & Pérez, M 2005, 'Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion', Revista Espanola de Cardiologia, vol. 58, no. 2, pp. 126-136.

Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion. / Martí, Helena; Pérez-Bárcena, Jon; Fiol, Miquel et al.

In: Revista Espanola de Cardiologia, Vol. 58, No. 2, 01.01.2005, p. 126-136.

Research output: Contribution to journal › Article › Research › peer-review

TY - JOUR

T1 - Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion

AU - Martí, Helena

AU - Pérez-Bárcena, Jon

AU - Fiol, Miquel

AU - Marrugat, Jaume

AU - Navarro, Carmen

AU - Aldasoro, Elena

AU - Cabadés, Adolfo

AU - Segura, Antonio

AU - Masiá, Rafael

AU - Turumbay, Javier

AU - Cirera, Lluís

AU - Arteagoitia, José M.

AU - Tomás, Carlos A.

AU - Vega, Gema

AU - Sala, Joan

AU - de los Arcos, Enrique

AU - Tormo, María J.

AU - Hurtado-de-Saracho, Iraida

AU - Francés-Sempere, Mercedes

AU - Elosua, Roberto

AU - Alonso, Eva

AU - Arós, Fernando

AU - Audicana, Covadonga

AU - Basterretxea, Mikel

AU - Castillo, Eduardo

AU - Esnaola, Santiago

AU - García Calabuig, Miguel A.

AU - Laresgoiti, Miren Josebe

AU - Larrañaga, Nerea

AU - Lasa, María José

AU - Lecuona, Iñaki

AU - Muniozguren, Nerea

AU - Tobalina, María Cres

AU - San Vicente, Jesús M.

AU - Sanz, Emilio

AU - Garía, José

AU - Martínez, Consuelo

AU - Contreras, Josefa

AU - Aliaga, Fuensanta

AU - Alonso, José M.

AU - Belda, Manuel

AU - Bañón, Rafael

AU - Castaño, José Antonio

AU - Castilla, Antonio

AU - Chiner, Juan Manuel

AU - Clemente, José M.

AU - Egea, Amparo

AU - Felices, Francisco

AU - Fuentes, Manuel

AU - Galcerá, José

AU - García, Arcadio

AU - García, Francisco

AU - García, Pedro

AU - González, Gurmesindo

AU - Herrero, Faustino

AU - Jara, Pedro

AU - Melgarejo, José Antonio

AU - Muñoz, José

AU - Nicolás, Silvestre

AU - Ortega, Juan

AU - de Paco, Miguel

AU - Pérez, Emilio

AU - Rodríguez, Pascual

AU - Rodríguez, Francisco Javier

AU - Ródenas, Julio

AU - Ruiz-Martínez, Francisco

AU - Ruiz, José Antonio

AU - de San Eustaquio, Fernando

AU - Segura, Josefa

AU - Serrano, José Antonio

AU - Torres, Gines

AU - Torres, Luis Carlos

AU - Vidal, Juan

AU - Villegas, Manuel

AU - Albiñana, Amparo

AU - Antón, Carlos

AU - Arguedas, Joaquín

AU - Borrás, Santiago

AU - Bosch, Susana

AU - Calabuig, José P.

AU - Castellanos, Enrique

AU - Cebrián, Javier

AU - Domingo, Francisco J.

AU - Echánove, Ildefonso

AU - Ferrer, Enrique

AU - Francés, Mercedes

AU - Gómez, Eva

AU - González, Julián

AU - Gregori, José

AU - Haba, Javier

AU - López, Antonio

AU - López, Vicente

AU - Melchor, Immaculada

AU - Montero, Rafael

AU - Montes, Julio

AU - Navarro, Ramón

AU - Palanca, Manuel

AU - Palencia, Miguel

AU - Pérez, Miguel

PY - 2005/1/1

Y1 - 2005/1/1

N2 - Introduction and objectives. Analysis of the effect of treatment in observational studies is complex due to differences between treated and nontreated patients. Calculating the probability of receiving treatment conditioned on relevant covariates (propensity score [PS]) has been proposed as a method to control for these differences. We report an application of PS to assess the association between reperfusion treatment and 28-day case fatality in patients with acute myocardial infarction (AMI). Method. We describe the procedure used to calculate PS for receiving reperfusion treatment, and different strategies to analyze the association between PS and case fatality with regression modeling and matching. Data were from a population-based registry of 6307 patients with AMI in Spain during 1997-98. Results. The PS for reperfusion was calculated in 5622 patients. In the multivariate analysis, reperfusion was associated with lower case fatality (OR= 0.59; 95% confidence interval [95% CI]: 0.46-077). When PS was included as a covariate, this association became non-significant (OR = 0.76; 95% CI: 0.57-1.01). In the subgroup of matched patients with a similar PS (n = 3138), treatment was not associated with case fatality (OR = 0.95; 95% CI: 0.72-1.26). When the influence of cases with missing data on PS was controlled for, reperfusion treatment was associated with lower fatality (OR = 0.66; 95% CI: 0.55-0.80). Conclusions. Calculating propensity score is a method that controls for differences between treated and nontreated patients. This score has limitations when matching is incomplete and when data are missing. Results of the present example suggest that reperfusion treatment reduces AMI case fatality.

AB - Introduction and objectives. Analysis of the effect of treatment in observational studies is complex due to differences between treated and nontreated patients. Calculating the probability of receiving treatment conditioned on relevant covariates (propensity score [PS]) has been proposed as a method to control for these differences. We report an application of PS to assess the association between reperfusion treatment and 28-day case fatality in patients with acute myocardial infarction (AMI). Method. We describe the procedure used to calculate PS for receiving reperfusion treatment, and different strategies to analyze the association between PS and case fatality with regression modeling and matching. Data were from a population-based registry of 6307 patients with AMI in Spain during 1997-98. Results. The PS for reperfusion was calculated in 5622 patients. In the multivariate analysis, reperfusion was associated with lower case fatality (OR= 0.59; 95% confidence interval [95% CI]: 0.46-077). When PS was included as a covariate, this association became non-significant (OR = 0.76; 95% CI: 0.57-1.01). In the subgroup of matched patients with a similar PS (n = 3138), treatment was not associated with case fatality (OR = 0.95; 95% CI: 0.72-1.26). When the influence of cases with missing data on PS was controlled for, reperfusion treatment was associated with lower fatality (OR = 0.66; 95% CI: 0.55-0.80). Conclusions. Calculating propensity score is a method that controls for differences between treated and nontreated patients. This score has limitations when matching is incomplete and when data are missing. Results of the present example suggest that reperfusion treatment reduces AMI case fatality.

KW - Myocardial infarction

KW - Prognosis

KW - Reperfusion

KW - Thrombolysis

M3 - Article

VL - 58

SP - 126

EP - 136

IS - 2

ER -

Analysis with the propensity score of the association between likelihood of treatment and event of interest in observational studies. An example with myocardial reperfusion

Abstract

Keywords

Fingerprint

Cite this