Taking patient mortalities into account when assessing change in NYHA classification from baseline

Patient response to treatment is often monitored by comparing metrics of performance immediately before treatment and at follow up. One such metric in patients with heart disease is the New York Heart Association (NYHA) classification, which is graded on a I to IV scale. In order for such an analysis to be unbiased, any missing NYHA classifications at follow up must be 'missing completely at random'. The study aim was to determine how the exclusion of expired patient data may bias the analysis of changes in NYHA class, with attention focused on the effects of ignoring patient mortalities, not only for the sake of simplicity but also that this might inflict the largest degree of bias in favor of the treatment.