- Email: [email protected]
Reporting on the modes of data collection
Correspondence
Reporting on the modes of data collection
Science Photo Library
To gather more insight into the methods of data collection currently used in medical research, we analysed research studies published in four high-impact general medical journals and three epidemiological journals in 2008–09.1 Two epidemiologists independently assessed the modes of data collection reported in each research paper. Surprisingly, the proportion of inconsistencies between the two reviewers was high, especially for papers published in general medical journals (about 30%). Further examination revealed that these inconsistencies were mainly due to unclear reporting of the methods used, with phrases such as “Information was collected on [list of variables]”, “Race/ethnicity was assessed by the investigator or study coordinator”, and “Sociodemographic, clinical, treatment (…), and laboratory data are collected” without any specification. Did they use questionnaires or interviews, were any measurements taken, or was it all hearsay? The choice of the method of data collection for a particular study depends on several factors, including, but not limited to, the type of study, sensitivity of the topic of interest, and costs of the measurements.2 Valid measurement of exposures, outcomes, and potential confounders is essential in medical research to prevent biased results.3 Since different methods of data collection yield various amounts of measurement error, detailed reporting on the methods used is of great importance to assess the quality of the study by both readers and reviewers or editors. Additionally, an adequate description of the methods of data collection used enables other research groups to replicate the original study. In our view, researchers should improve the description of the modes of data collection used in their studies. 30
In addition, we encourage medical journals to pay more attention to the way in which the methods are reported to improve the possibilities of critical appraisal. We declare that we have no conflicts of interest.
The Lancet has rightly emphasised that authors of papers must make a substantial contribution.3 It is also important that all major contributors are recognised, either as an author or with an acknowledgment.
*Marleen M H J van Gelder, Reini W Bretveld, Nel Roeleveld
I have received payment for lectures or speakers’ bureaux from Dako, Ipsogen, Hammersmith Hospital, and Netherlands Cancer Institute.
[email protected] Department of Epidemiology, Biostatistics and HTA, Radboud University Nijmegen Medical Centre, 6500 HB Nijmegen, Netherlands 1
2
3
Van Gelder MMHJ, Bretveld RW, Roeleveld N. Web-based questionnaires: the future in epidemiology? Am J Epidemiol 2010; 172: 1292–98. Armstrong BK, White E, Saracci R. Principles of exposure measurement in epidemiology. New York: Oxford University Press, 1992. Rothman KJ, Greenland S, Lash TL. Validity in epidemiologic studies. In: Rothman KJ, Greenland S, Lash TL. Modern epidemiology, 3rd edn. Philadelphia: Lippincott Williams & Wilkins, 2008: 137–46.
Andrew H S Lee [email protected] Department of Histopathology, Nottingham University Hospitals, City Hospital Campus, Nottingham NG5 1PB, UK 1 2
3
Kang S. Anecdotes in medicine—15 years of Lancet case reports. Lancet 2010; 376: 1448–49. Wilson R, Liston J, Cooke J, et al. Clinical guidelines for breast cancer screening assessment. Sheffield: NHS Cancer Screening Programmes, 2005. The Lancet. The role and responsibilities of coauthors. Lancet 2008; 372: 778.
All important contributions to papers should be recognised In her Comment (Oct 30, p 1448),1 Seema Kang discusses Case Reports in The Lancet according to specialty interest and mentions that “more than one specialty might be involved in managing a case”. She comments on the absence of contributions from general practice, but makes no comment on the contribution of other specialities such as radiology and pathology. Many patients require multidisciplinary investigation or management. For example, in the diagnosis of breast disease, the triple approach of clinical examination, radiology, and pathology is recommended.2 Over the past 2 years (November, 2008, to October, 2010) histopathology had an important role in the diagnosis in 36 Case Reports in The Lancet. A histopathologist was an author in 20. In the other 16 there was no apparent recognition of histopathology in terms of authorship or acknowledgment, although it had an essential role in ten and was contributory in six.
Department of Error Tatem AJ, Smith DL, Gething PW, Kabaria CW, Snow RW, Hay SI. Ranking of elimination feasibility between malaria-endemic countries. Lancet 2010; 376: 1579–91—In this Series paper (Nov 6), the fourth sentence under the heading “Estimation of intensity of endemic P falciparum transmission” should have read: “R0 is a measure of maximum potential transmission, and if R0 is 1 or greater, then endemic malaria transmission can be sustained”. Additionally, the title for figure 1 should have read “Categorical map of Plasmodium falciparum reproductive number, PfRc, indicating the extent to which transmission needs to be reduced for elimination”. These corrections have been made to the online version as of Dec 31, 2010. The InFACT Global H1N1 Collaboration. InFACT: a global critical care research response to H1N1. Lancet 2010; 375: 11–13—In this Comment (Jan 2), the name of the third member of the InFACT Global H1N1 Collaboration was misspelled. The correct spelling is “Neill K J Adhikari”. This correction has been made to the online version as of Dec 31, 2010.
www.thelancet.com Vol 377 January 1, 2011
Reporting on the modes of data collection
Science Photo Library
To gather more insight into the methods of data collection currently used in medical research, we analysed research studies published in four high-impact general medical journals and three epidemiological journals in 2008–09.1 Two epidemiologists independently assessed the modes of data collection reported in each research paper. Surprisingly, the proportion of inconsistencies between the two reviewers was high, especially for papers published in general medical journals (about 30%). Further examination revealed that these inconsistencies were mainly due to unclear reporting of the methods used, with phrases such as “Information was collected on [list of variables]”, “Race/ethnicity was assessed by the investigator or study coordinator”, and “Sociodemographic, clinical, treatment (…), and laboratory data are collected” without any specification. Did they use questionnaires or interviews, were any measurements taken, or was it all hearsay? The choice of the method of data collection for a particular study depends on several factors, including, but not limited to, the type of study, sensitivity of the topic of interest, and costs of the measurements.2 Valid measurement of exposures, outcomes, and potential confounders is essential in medical research to prevent biased results.3 Since different methods of data collection yield various amounts of measurement error, detailed reporting on the methods used is of great importance to assess the quality of the study by both readers and reviewers or editors. Additionally, an adequate description of the methods of data collection used enables other research groups to replicate the original study. In our view, researchers should improve the description of the modes of data collection used in their studies. 30
In addition, we encourage medical journals to pay more attention to the way in which the methods are reported to improve the possibilities of critical appraisal. We declare that we have no conflicts of interest.
The Lancet has rightly emphasised that authors of papers must make a substantial contribution.3 It is also important that all major contributors are recognised, either as an author or with an acknowledgment.
*Marleen M H J van Gelder, Reini W Bretveld, Nel Roeleveld
I have received payment for lectures or speakers’ bureaux from Dako, Ipsogen, Hammersmith Hospital, and Netherlands Cancer Institute.
[email protected] Department of Epidemiology, Biostatistics and HTA, Radboud University Nijmegen Medical Centre, 6500 HB Nijmegen, Netherlands 1
2
3
Van Gelder MMHJ, Bretveld RW, Roeleveld N. Web-based questionnaires: the future in epidemiology? Am J Epidemiol 2010; 172: 1292–98. Armstrong BK, White E, Saracci R. Principles of exposure measurement in epidemiology. New York: Oxford University Press, 1992. Rothman KJ, Greenland S, Lash TL. Validity in epidemiologic studies. In: Rothman KJ, Greenland S, Lash TL. Modern epidemiology, 3rd edn. Philadelphia: Lippincott Williams & Wilkins, 2008: 137–46.
Andrew H S Lee [email protected] Department of Histopathology, Nottingham University Hospitals, City Hospital Campus, Nottingham NG5 1PB, UK 1 2
3
Kang S. Anecdotes in medicine—15 years of Lancet case reports. Lancet 2010; 376: 1448–49. Wilson R, Liston J, Cooke J, et al. Clinical guidelines for breast cancer screening assessment. Sheffield: NHS Cancer Screening Programmes, 2005. The Lancet. The role and responsibilities of coauthors. Lancet 2008; 372: 778.
All important contributions to papers should be recognised In her Comment (Oct 30, p 1448),1 Seema Kang discusses Case Reports in The Lancet according to specialty interest and mentions that “more than one specialty might be involved in managing a case”. She comments on the absence of contributions from general practice, but makes no comment on the contribution of other specialities such as radiology and pathology. Many patients require multidisciplinary investigation or management. For example, in the diagnosis of breast disease, the triple approach of clinical examination, radiology, and pathology is recommended.2 Over the past 2 years (November, 2008, to October, 2010) histopathology had an important role in the diagnosis in 36 Case Reports in The Lancet. A histopathologist was an author in 20. In the other 16 there was no apparent recognition of histopathology in terms of authorship or acknowledgment, although it had an essential role in ten and was contributory in six.
Department of Error Tatem AJ, Smith DL, Gething PW, Kabaria CW, Snow RW, Hay SI. Ranking of elimination feasibility between malaria-endemic countries. Lancet 2010; 376: 1579–91—In this Series paper (Nov 6), the fourth sentence under the heading “Estimation of intensity of endemic P falciparum transmission” should have read: “R0 is a measure of maximum potential transmission, and if R0 is 1 or greater, then endemic malaria transmission can be sustained”. Additionally, the title for figure 1 should have read “Categorical map of Plasmodium falciparum reproductive number, PfRc, indicating the extent to which transmission needs to be reduced for elimination”. These corrections have been made to the online version as of Dec 31, 2010. The InFACT Global H1N1 Collaboration. InFACT: a global critical care research response to H1N1. Lancet 2010; 375: 11–13—In this Comment (Jan 2), the name of the third member of the InFACT Global H1N1 Collaboration was misspelled. The correct spelling is “Neill K J Adhikari”. This correction has been made to the online version as of Dec 31, 2010.
www.thelancet.com Vol 377 January 1, 2011