SciELO - Scientific Electronic Library Online

 
vol.33 issue1Case Report of Rescue Therapy for Bile Duct Perforation using Fully Covered Self-Expanding Metal Stent after Failure of Conservative ManagementCarta al Editor author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Journal

Article

Indicators

Related links

  • On index processCited by Google
  • Have no similar articlesSimilars in SciELO
  • On index processSimilars in Google

Share


Revista colombiana de Gastroenterología

Print version ISSN 0120-9957

Rev Col Gastroenterol vol.33 no.1 Bogotá Jan./Mar. 2018

 

Letter to the Editor

Methodological review: association study in gastroenterological surgery

Daniel Carbajal Zamora1 

Sandra López Jiménez1 

Víctor Soto Cáceres2 

Víctor Torres Anaya3 

1 Student of human medicine in the Faculty of Human Medicine at Universidad San Martín de Porres in Chiclayo, Peru

2 Medical epidemiologist at Lambayeque Regional Hospital in Chiclayo, Peru

3 Medical epidemiologist at the Regional Health Management of Lambayeque in Chiclayo, Peru


Dear editor:

In an article published recently in your review, Vargas et al. conducted an investigation into factors associated with the conversion of laparoscopic cholecystectomy (LC) into an open procedure. 1 The research contributed important information about factors such as age and leukocytosis that contribute to failure of procedures such as cholecystectomies to that they become open surgeries. This research was conducted with the intention of contributing to reports on factors that contribute to taking measures during performance of elective surgery such as laparoscopic cholecystectomies.

However, the article’s, materials and methods section mentioned a descriptive study and in the statistical analysis performed it was indicated that bivariate relative risk (RR) tests were applied. These tests aimed to determine risk factors for conversion of laparoscopic surgery to open surgery. These details of the investigation deserve to be discussed.

First, when a descriptive study is mentioned, the use of bivariate RR should not be included since the RR is defined as a ratio between two probabilities or two risks so that the concept of risk is equivalent to the concept of epidemiological incidence which is feasible only for prospective studies (which differ from the article in question). Its use is restricted to randomized controlled trials and cohort studies 2.

In this way, it would connote an analysis that measures the prevalence ratio (PR) for this type of studies. In a cross-sectional study, the sample is selected without a priori knowledge of the condition of each subject regarding the exposure and the event of interest. For this reason, it does not measure causality. Therefore, once the sample is selected, the study proceeds to determine whether the condition of each subject is exposed or not exposed to the presence or absence of the event of interest, measuring both variables simultaneously. 2,3 Cross-sectional studies do not allow calculation of incidence rates since they lack a sense of prospective observation, therefore it is methodologically incorrect to try to calculate RR as a measure of effect. 2

Second, the article also said, “... patients with ages over 50 years have a 55% higher risk of having LC converted into open surgery (odds ratio [OR] = 0.55).” (1) In addition, it has been observed that the aforementioned research has a random convenience sample of 59 clinical histories. However, the article indicates that the use of odds ratios (OR) is compatible for case and control studies which use a methodology that is different from one presented in the article, as well as previously indicating that there was an analysis of bivariate RR. 2

Another detail to highlight relates to interpretation of results by the authors of this study. Despite the fact that it is a descriptive study, they evaluate risk instead of only determining an association among factors without determining causality of those factors. This is why there are several sections within the results that emphasize, “... ages have 55% higher risk ... (odds ratio [OR] = 0.55) ...” and “... leukocytosis has a 40% higher risk of being converted (OR = 0.40)…”, but instead of signifying risk factors, these should be protective factors with respect to their research. The authors interpreted that those who are over 50 years old are 45% less likely to have a conversion from CL to open surgery and that those who have leukocytosis are 60% less likely to have a conversion from a laparoscopy to open surgery.

For interpretation of OR corresponding to division when presenting two interventions with equivalent risks, the odds should be the same and the OR should be equal to one. Therefore, the farther away the OR is from one, the greater or smaller its magnitude will be, and the magnitude of the effect will be greater than its value. 4 OR has been used in the medical literature due to its usefulness for estimating the relationship between two binary variables and because of the fact that it allows evaluation of the effects that other variables have in that relation by means of logistic regression methods. 5

For the above, an inadequate interpretation could lead the reader to an erroneous view. Since there is no clear support to establish adequate methodology, it is recommended that these statistical and methodological considerations be taken into account.

Referencias

1. Vargas RL, Agudelo SM, Lizcano CR, et al. Factores asociados con la conversión de la colecistectomía laparoscópica a colecistectomía abierta. Rev Col Gastroenterol. 2017;32(1):20-3. https://doi.org/10.22516/25007440.125. [ Links ]

2. Cerda J, Vera C, Rada G. Odds ratio: aspectos teóricos y prácticos. Rev Med Chile. 2013;141(10):1329-35. DOI: 10.4067/S0034-98872013001000014. [ Links ]

3. Hernández B, Velasco-Mondragón HE. Encuestas transversales. Salud Pública Mex. 2000;42(5):447-55. [ Links ]

4. Rivera MS, Larrondo FJG, Ortega JPR. Evaluación de los resultados en un artículo sobre tratamiento. Rev Med Chile. 2005;133(5):593-6. [ Links ]

5. Abraira V. Medidas del efecto de un tratamiento (II): odds ratio y número necesario para tratar. SEMERGEN. 2001;27(8):418-20. DOI: 10.1016/S1138-3593(01)74000-2. [ Links ]

Received: November 22, 2017; Accepted: November 23, 2017

Creative Commons License Este es un artículo publicado en acceso abierto bajo una licencia Creative Commons