Falls count and counting falls: Making sense of data about falls

Abstract

It is often challenging to make sense of research reports on falls. The choice of statistical method depends on whether the outcome is binary (faller: yes/no), a rate (falls per person-time in view), ordinal (number of falls per person) or time to fall (first). The most useful methods for analysing falls are those that estimate parameters as they provide an estimated value for risk associated with different levels of a factor or intervention. Less useful are statistics that simply provide a yes/no answer as to whether the factor or intervention affects risk (hypothesis testing). As falls are negative events, when parameters such as odds ratios (OR), incidence rate ratios (IRR), hazard ratios (HR), proportional odds ratios (POR) or cumulative odds ratios (COR) are greater than 1.0, they indicate that the factor is associated with a higher risk of falls; when <1.0, the factor or the intervention is associated with a lower risk of falls. All of these statistical parameters can be used to identify risk factors for falls or to evaluate effective interventions.