Skip to content

What you have told us.

May 30, 2012

The discussions in this blog have been creative and informative.  To give you a brief summary, we have paraphrased or quoted what you have told us.  Please continue discussions on these topics and we will be posting additional critical questions. 

 

What you have told us so far about…

…duplicative studies:

Past/Future sign

  • Can duplication be a specific review criterion during NIH peer review?
  • Studies should replicate (not duplicate) by confirming results and improving research quality in the replication.
  • Replication studies should be more carefully formulated in advance.
  • Journals should more carefully address issues of duplication and design.
  • Reviewers of grant applications should appraise projects by direct measures of scientific contribution.  

 

…false positive results:

  • False positive single study results may be common but in the broader questions, false positives are fewer.
  • Patience is required to confirm results to make policy recommendations.
  • Use better statistical tools to establish how well replication leads to findings that can be generalized.
  • Register aims in a public system when study is started to distinguish a priori hypotheses from data mining.
  • Improve initial design with larger sample sizes to improve statistical power.

 

…cost-efficient study designs:

Need New Ideas cubes

  • We don’t need alternatives to cohort designs but smarter, cost-efficient use of them.
  • Maybe research should be more inclusive of both mental and physical causes of disease.
  • To identify targets for early intervention, including primordial, studies should begin at middle-age or younger.
  • To encourage innovative research, a separate funding mechanism should be established for discovery epidemiology.
  • Cheaper studies are not always cost effective since they reduce the range of research questions that can be asked.
  • Study applications to NIH could have a formal systematic literature review of the research question to demonstrate the need for the proposed research.
  • Instead of thinking about a less expensive study design, we might think about how technology will affect the study designs we have, and potentially reduce costs.
  • Crowd sourcing designs should be explored by epidemiologists and biases carefully evaluated.

 

…electronic medical records (EMRs) for research:

laptop-and-stethoscope

  • EMRs will be the cardiovascular epidemiology of the future.
  • Experiences from countries with nearly universal use of a common EMR system will provide evidence on the value for EMR research.
  • Uses could include identifying rarer cases for case-control studies, taking advantages of the large numbers various exposure-outcome relationships, and obtaining prevalence for conditions that are almost always require hospitalization. 
  • EMRs should be very useful for studying quality of care and finding unexpected associations requiring large samples.
  • Potential advantages include efficiency and reduced cost, longitudinal follow-up, linking of EMR data to biorepositories.
  • Research within an EMR will require robust standardization and will be hindered by fragmentation of care.
  • Some exposures are poorly measured including physical activity, behavioral characteristics, and biological or imaging markers that are obtained only for clinical purposes.
  • When universal, EMRs can provide data on the whole population and will reconnect epidemiology, public health and clinical research.
  •  A hybrid model with both cohort studies and linked EMR will take best advantage of each design.
  • Further advances to other electronic data collection systems (home, cell phone, etc) could provide significant advances in important data needed for research.
  • Many, if not most, electronic diagnoses require validation.
  • There could be great potential for EMRs to be used as larger surveillance systems for CVD. 
  • Long term thinking on CV epidemiology may lead to less use of traditional models of cohort studies and more use of national EMR systems. 
  • While a broad generalization, most questions that can be answered in traditional cohort studies have already been answered.
  • Using EMR data for epidemiology is possible, but with realism and careful selection of appropriate studies and outcomes.
  •  At a minimum, we might say that an EMR should always precede more expensive study designs and then identify the factors that might require moving to the cohort study or randomized trial.

 

…separate data collection and analysis funding:

analyzing-graphs-and-charts

  • Separate funding works well for studies with more straight-forward data system, but not in complex longer-term studies requiring cross-talk between data collectors and analyzers. 
  • Investigators who design and conduct the study should have a larger role in analytic plan and writing up the data.
  • If cohort studies were open to all, with confidentiality controls, they would be more widely used. 
  • Why not have NHLBI be a generator of open-source datasets on cardiovascular conditions?
One Comment leave one →
  1. July 4, 2012 12:38 am

    How can you tell a false negitave result during the studies? Is there any signs which stipulate that the results gathered by a study may not be accurate and is there anyway to determine this might be the case whilst the study is being carried out?
    It would be benificial to all kinds of medical trials/studies with live patients if this we were able to detect a patients results may be altered or incorrect during the process!

Leave a Comment

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: