December 2, 2015

statins news stories and continuation of statins

The effect of statin-related news stories on people continuing with statin prescriptions and health outcomes in a Danish population is the subject of a paper published in the European Heart Journal. The authors report an association between discontinuing statin use early and heart attack and risk of death from cardiovascular disease. Roundup comments accompanied this analysis.

Title, Date of Publication & Journal

‘Negative statin-related news stories decrease statin persistence and increase myocardial infarction and cardiovascular mortality: a nationwide prospective cohort study’ by Sune Fallgaard Nielsen and Børge Grønne Nordestgaard

European Heart Journal

Wednesday 2 December 2015

Study’s main claims – and are they supported by the data

The paper suggests that there is an association between negative press coverage of statins and patients discontinuing their use – the study cannot prove that negative press coverage causes patients to stop taking statins, because this is an observational cohort study, but that is one possible explanation.

The paper similarly suggests there is an association between negative press coverage of statins, discontinuation of statins and increased numbers of heart attacks and deaths.  Again this kind of study cannot prove causation, but there is biological plausibility underlying an association between stopping statins and a heart attack or death (rather than between a news article and death!).

The paper finds a similar sized association of positive media stories with maintenance of statin use. As there are many more of these positive stories, the overall impact of the all media stories taken together is not one of patients discontinuing use. Perhaps the headline could be “Positive statin-related news stories may help patients maintain their statin regime leading to less chance of myocardial infarction and cardiovascular mortality: a nationwide prospective cohort study”.

Strengths/Limitations

Limitations

This study can only identify associations rather than causative relationships.

Secondly, obviously people with the same statin start date have the same or very similar (depending on location) values for their three important predictors (amounts of bad, neutral or good media coverage). So any temporal change in the event rate (stopping statin within 6 months of first prescription) will potentially confound results. The authors correct for this using a simple linear regression on the date of starting statin. But given the length of observation (see figure 2) and the bunching up of media stories into the final period, such a simple adjustment may not be sufficient.

Strengths

This is an interesting hypothesis and is worthy of this type of investigation. It underlines the responsibility that the media must take if using scare stories to increase readership.

Any specific expertise relevant to studied paper (beyond statistical)?

None

Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry (PSI) and experienced statisticians in academia and research. A list of contributors, including affiliations, is available here.