Letter responses to Dobson et al published in Lancet

Letter responses to Dobson et al published in Lancet Announcement Date: September 1, 2015

Series of responses to Dobson et al  meta-analyses of individual patient published in the Lancet including concerns over the methods.

Oseltamivir for influenza. Lancet 2015; 386: 113336.(9999):11345. Doi: 10.1016/S01406736(15)002020.

Oseltamivir for influenza. Peter Doshi, Carl Heneghan, Tom Jefferson.

Meta-analyses of individual patient data hold the potential of offering new insights into old data, but the Article by Joanna Dobson and colleagues fails to do this, presenting no new data on the potential treatment effect of oseltamivir in reducing the risk of influenza complications or admittance to hospital. Instead, the authors provide only new interpretations of already public data.2 We are concerned, however, that Dobson and colleagues’ interpretation of their results1 seems to be driven by a methodology not described in the methods section of their Article. Dobson and colleagues list nine data sources: four journal articles, three conference abstracts, secure web-access to participant-level data, and data clarifications from the drug’s manufacturer. These sources do not contain enough detail to support the authors’ comment that “specific diagnostic tests were not necessary” for respiratory complications. For example, one conference abstract that is cited (for a trial of more than 1400 participants) is just a paragraph long. (Read More)

Oseltamivir for influenza. Mark Jones, Rokuro Hama, Chris Del Mar.

In their Article (May 2, p 1729)1 on oseltamivir treatment for influenza, Joanna Dobson and colleagues reported both fewer admittances to hospital for those treated with oseltamivir in the subgroup of patients classified as infected and fewer lower respiratory tract complications in patients treated with antibiotics as a surrogate for severity than in placebo recipients. However, our Cochrane review2 found insufficient evidence of a difference between groups for complications classified as serious or leading to patient withdrawal, or for rate of admittance to hospital. (Read More)

Oseltamivir for influenza. Kubo et al.

In their article, Joanna Dobson and colleagues presented the value of oseltamivir for influenza treatment. But does this end the debate about the efficacy of oseltamivir?

The authors selected a secondary outcome of lower respiratory tract infection more than 48 h after randomisation requiring antibiotics and thus included 143 cases of bronchitis, 30 cases of pneumonia, and five cases of lower respiratory tract infection. However, in general, most cases of bronchitis do not need treatment with antibiotics. It is commendable that the authors are the first to present results from pooled individual patient data without significant heterogeneity, but they seem to be limited and non-specific, as Dobson and colleagues themselves noted in their Article. (Read More)

Oseltamivir for influenza. Radecki.

Joanna Dobson and colleagues1 present an individual patient meta-analysis assessing the efficacy of oseltamivir for the treatment of influenza infection. The authors report that the data for their analysis were acquired from Roche Pharmaceuticals via a secure web portal. Interestingly, the data presented differ substantially from the data reported by the Cochrane Collaboration in their review of clinical study reports.2 Five trials had populations common to both analyses for comparison, and differences in admittance to hospital reported are presented in the table. (Read More)

Oseltamivir for influenza. Arnold S Monto, Joanna Dobson, Stuart Pocock, Richard J Whitley.
We thank the correspondents for their interest in our study.1 We did our meta-analysis of oseltamivir treatment studies in view of the controversy that has arisen about oseltamivir’s rational use. To us, an analysis with the most robust design, namely an individual patient data analysis, would help clarify issues of importance not only to clinicians but also to public health decision makers. (Read More)