Abstract

REJAS, J.; RUIZ, M.  and  PARDO, A.. Standard error of measurement: An alternative to minimally important difference to assess changes in patient-reported-health-outcomes?. An. Med. Interna (Madrid) [online]. 2007, vol.24, n.9, pp.415-420. ISSN 0212-7199.

Objetive: To assess the concordance level between the standard Error of Measurement (EEM) and the minimal important difference (MID) criteria when evaluating the magnitude of change in self-perceived health status due to a clinical intervention. Method: Two scales of patient self-reported health are used [a measure of benign prostatic hyperplasia impact (BPH-PIM) and a life satisfaction scale (LISAT-8)] in two independent studies carried out in order to adapt each scale to Iberian Spanish population. These studies included 129 and 537 males with benign prostatic hyperplasia (BPH) and erectile dysfunction (ED), respectively. After computing EEM and MID values, patients where classified in three groups attending to each criterion: improvement, without change, and worsening. Cohen's kappa agreement index and Kendall's tau-b lineal correlation coefficient where used to assess the level of concordance between criteria. Results: The concordance level between EEM and MID criteria was found to be high for both scales: kappa agreement index attained values of 0.81 and 0.76, while tau-b correlation coefficient attained values of 0.87 and 0.79. Conclusion: The agreement level found allows to state that the EEM criterion is an adequate alternative to the MID criterion when assessing the magnitude of change produced in patient reported health.

Keywords : Minimal important difference; Standard error of measurement; Patient self-reported output; Clinical relevance; Magnitude of change.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )