Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Not a cohort study
- Andrew J Hall (27 February 2006)
-
Immunological risk factors for meningococcal disease
- Prasanta Padhan (27 February 2006)
-
Retrospective assessment of behavioral factors
- Vilhjalmur Rafnsson (12 March 2006)
-
Religion as a marker for behaviour?
- Fiona Reynolds (17 March 2006)
-
Response from authors
- Robert R Booy, Pietro Coen, James Stuart (6 April 2006)
|
|
|||
|
Andrew J Hall, Professor of epidemiology London School of Hygiene and Tropical Medicine WC1E 7HT
Send response to journal:
|
A cohort study can be distinguished from a case control study by the method of recruitment of the participants. In a cohort study it is on the basis of an exposure and the participants are then followed up for the outcome. In a case control study it is on the basis of the outcome and the participants are then investigated for exposures. In this study participants were recruited because they had meningococcal disease - the outcome - it can not therefore be a cohort study but is a case-control study. Competing interests: None declared |
|||
|
|
|||
|
Prasanta Padhan, MD,Senior Resident Department of Internal Medicine,JIPMER,Pondicherry,India,605006.
Send response to journal:
|
Dear Editor, It has long been known that deficiency of proteins in the classical, alternative, or terminal complement pathways predispose to invasive meningococcal disease. The relative risk has been estimated at several thousands; however, deficiency of complement proteins is extremely rare in the general population.But certain factors in the patient history or course of meningococcal disease may help identify individuals with complement deficiency.Infection with an unusual serogroup organism such as W135 and recurrent disease is strongly suggestive of complement deficiency,particularly terminal complement component deficiency.Complement deficiencies show an autosomal recessive inheritance pattern, with the exception of properdin deficiency, which is X linked. Individuals without spleens (asplenic) and those taking immunosuppressive drugs such as prednisone or ciclosporin are also more prone for meningococcal disease. Competing interests: None declared |
|||
|
|
|||
|
Vilhjalmur Rafnsson, Professor of Preventive Medicine University of Iceland, Neshaga 16, 107 Reykjavik, Iceland
Send response to journal:
|
It has already been pointed out that the study is a case-control study and not a cohort study (1). It should be added that the two concepts are used interchangeably in the text. The cases in the study are incidence cases with meningococcal disease originating from a large background population or a study base. However, the controls were selected by general practitioners from their lists of patients, but not matched randomly from the background population. This procedure made it uncertain whether the controls were representative of the background population regarding prevalence of risk and protective factors. Instead of addressing this as shortcoming the authors maintained that the recruitment of controls was population based. Information on significant risk and protective factors in the study was obtained from the participants or their heads of household, after the diagnoses of the cases. This method has the inherent weakness of retrospective study involving life-threatening disease. The authors explained the protective effect of attendance at a religious event as due to other associated lifestyle factors. However, it is not clear from the study what was meant by other lifestyle factors or whether those attending religious events belonged to the group who did not practice multiple intimate kissing contacts. In the study multiple intimate kissing contacts were significant risk factor for meningococcal disease. 1. Hall AJ. Not a cohort study. Rapid Responses for Tully et al., 332 (7539) 445-450. Competing interests: None declared |
|||
|
|
|||
|
Fiona Reynolds, Public Health SpT Greater Manchester Health Protection Unit, Peel House, Albert Street, Eccles, Manchester, M30 0NJ
Send response to journal:
|
This was an interesting article which raises a couple of points for consideration. The study was carried out between January 1999 and June 2000. The meningococcal C vaccine was introduced in November 1999. The first to be vaccinated were adolescents, and the entire programme was completed by the end of 2000(1) . This issue is not mentioned in the study and would possibly have been a confounding factor. There were also issues arising in the collection of the data, which may have led to bias – not just of the recall variety. Data was collected from participants and parents around behavioural factors, such as kissing and smoking. Although recall bias may have been adjusted for by using a ‘short recall period’, this cannot eliminate the impact of ‘social acceptability bias’. Will teenagers admit to behaviours that are frowned upon by their community, their peers or their parents? The use of religious ceremony attendance as a marker for social behaviour also raises a couple of questions: firstly, was the same protective value of attendance found across all religious denominations?; and secondly, can religious ceremony attendance really be an accurate indicator for what young people do with the rest of their time? Finally, as already stated by two contributors, this study was a case -control study rather than a cohort study, as claimed in the title (2,3). 1. Ramsey M, Andrews, NJ, Trotter CL, Kazcmarski, EB and Miller E. (2003) Herd immunity from meningococcal serogroup C conjugate vaccination in England: database analysis, BMJ 2003;326:365-366 2. Hall AJ. Not a cohort study. Rapid Responses for Tully et al., 332 (7539) 445-450. 3. Raffnsson V, Retrospective assessment of behavioural factors. Rapid Responses for Tully et al., 332 (7539) 445-450. Competing interests: None declared |
|||
|
|
|||
|
Robert R Booy, Co-Director NCIRS, Westmead, .Australia NSW 2145, Pietro Coen, James Stuart
Send response to journal:
|
We are grateful to several contributors who support our position that this is a case control study. The wording “matched cohort study” was inserted during the editorial process at the request of an expert reviewer. We designed, analysed and interpreted the study as a matched case control study. Rafnsson is concerned that controls were not population based. Although not a random sample of the population that gave rise to the cases, controls were selected from the general population and were individually matched. Selection bias introduced by matching would have been removed in the analysis through conditional logistic regression. The “protective” effect of attending religious events was independent of multiple intimate kissing. In other words the protective effect remained after adjusting for other risk factors. We discussed possible explanations for this finding. We did not have enough data to conduct sub- analyses on different religious groups. In response to Reynolds’ question about vaccination, a protective effect of vaccination was observed as stated and was included in the multivariable analysis. We reduced recall bias by minimising the time between interview and illness for cases and asking about the immediate 2 weeks preceding interview for controls. We controlled for the difference in interview time for the matched pairs by including a seasonal variable in the multivariable analysis. Social acceptability bias is certainly possible but it might be surprising if controls were less likely to admit to multiple kissing than teenagers who had been ill. This would be necessary to explain the results through such a bias. Competing interests: None declared |
|||