Scholarlink

Background <p>The BIO|GUARD-MI study was an unblinded trial that investigated the clinical effect of continuous arrhythmia monitoring with an implantable cardiac monitor in patients after myocardial infarction. Unexpectedly, a pre-specified interim analysis identified a higher incidence of non-cardiovascular adverse events in the treatment group. This was assumed due to a performance bias and ultimately led to premature study termination. By study protocol, participants were expected to visit the study site only when invited after detected arrhythmias. Here, we provide a secondary analysis of participant initiated contacts to the investigational site, which may be the basis of the presumed bias, and employ it to identify a subgroup of participants that may be responsible for the bias.</p> Methods <p>We compare the numbers of participants with self-initiated contacts between the study groups in the total cohort and in 24 subgroups, 14 of them based on the country and 10 on clinical or demographic characteristics.</p> Results <p>In the total cohort, similar percentages of participants of the treatment and control groups visited the investigational site on their own initiative (13.3% vs. 9.7%; <i>P</i> = 0.12). However, the subgroup analysis revealed a statistically significant difference in participants from Denmark (24.4% vs. 2.3%) after correction for multiple testing (<i>P</i> = 0.0004). Other variables support the assumption that this affected the reporting of adverse events.</p> Conclusions <p>In the BIO|GUARD-MI trial, a performance bias is evident in participants from Denmark. In the Danish health system, the general practitioner acts as a gate keeper for secondary and tertiary care, which treatment group participants could bypass. This led to more clinical events being reported in them. When conducting an unblinded diagnostic trial, continuous assessment of characteristics that indicate unexpected participant behaviour is important. Such differences may be driven by characteristics of national health systems.</p> Trial registration <p><a href="https://www.clinicaltrials.gov/ct2/show/NCT02341534">https://www.clinicaltrials.gov/ct2/show/NCT02341534</a>, First registered at ClinicalTrials.gov on January 14, 2015.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Pitfalls in the reporting of clinical events in unblinded device trials: analysing performance bias in the BIO|GUARD-MI trial

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月13日

查看全文
  • Christian Jøns,
  • Peter Søgaard,
  • Tom Smilde,
  • Miloš Táborský,
  • Jozsef Faluközy,
  • Marcus Wiemer,
  • Per Dahl Christensen,
  • Andrejs Erglis,
  • Muchtiar Khan,
  • Julio Marti Almor,
  • Susana del Prado Diaz,
  • Ivan Rybar,
  • Laura Vitali Serdoz,
  • Jürgen Schrader,
  • Steffen Behrens

摘要

Background

The BIO|GUARD-MI study was an unblinded trial that investigated the clinical effect of continuous arrhythmia monitoring with an implantable cardiac monitor in patients after myocardial infarction. Unexpectedly, a pre-specified interim analysis identified a higher incidence of non-cardiovascular adverse events in the treatment group. This was assumed due to a performance bias and ultimately led to premature study termination. By study protocol, participants were expected to visit the study site only when invited after detected arrhythmias. Here, we provide a secondary analysis of participant initiated contacts to the investigational site, which may be the basis of the presumed bias, and employ it to identify a subgroup of participants that may be responsible for the bias.

Methods

We compare the numbers of participants with self-initiated contacts between the study groups in the total cohort and in 24 subgroups, 14 of them based on the country and 10 on clinical or demographic characteristics.

Results

In the total cohort, similar percentages of participants of the treatment and control groups visited the investigational site on their own initiative (13.3% vs. 9.7%; P = 0.12). However, the subgroup analysis revealed a statistically significant difference in participants from Denmark (24.4% vs. 2.3%) after correction for multiple testing (P = 0.0004). Other variables support the assumption that this affected the reporting of adverse events.

Conclusions

In the BIO|GUARD-MI trial, a performance bias is evident in participants from Denmark. In the Danish health system, the general practitioner acts as a gate keeper for secondary and tertiary care, which treatment group participants could bypass. This led to more clinical events being reported in them. When conducting an unblinded diagnostic trial, continuous assessment of characteristics that indicate unexpected participant behaviour is important. Such differences may be driven by characteristics of national health systems.

Trial registration

https://www.clinicaltrials.gov/ct2/show/NCT02341534, First registered at ClinicalTrials.gov on January 14, 2015.