Skip to main content

Archived Comments for: Dutch healthcare reform: did it result in better patient experiences in hospitals? a comparison of the consumer quality index over time

Back to article

  1. The conclusion that competition caused improved satisfaction is flawed

    Robert Grant, Kingston University & St George's, University of London

    1 June 2012

    The analyses carried out by Ikkersheim and Koolman are thorough and entirely appropriate, but the conclusion of this paper that competition and publication of satisfaction statistics caused improvements in patient satisfaction is far from justified. The conclusions are subject to four major flaws:

    1) post hoc ergo propter hoc - in particular, the latest measurement (2009) has significantly younger participants

    2) the improvement is tiny and probably meaningless: the time effect is about 1% of the scale, the publishing of stats effect slightly less and the "competition" effect is 0.1%. Because the numbers of patients involved is large, everything is statistically significant

    3) The definition of the CQI scale changed during the period being analysed. Table 4 reveals that this reversed some of the results, which suggest great caution is required when interpreting the results

    4) Competition is defined based on the division of 'market share' among geographically nearby hospitals. This is therefore an inadvertent proxy for urban / rural location rather than competition. It seems likely that networks linking healthcare providers to the public will be more complex than simple geographical distance, particularly in a small and highly urbanised country such as the Netherlands.

    These are typical of the difficulties in analysing natural experiments, which is why statisticians never tire of reminding colleagues that "correlation is not causation".

    Competing interests

    No competing interests

Advertisement