Pictorial outcome measures for the hospital care of older patients - a suggested toolkit

Objective: to propose three pictorial methods of presenting hospital outcom e data, suitable for use in older patients entering medical specialties (in cluding rehabilitation). Patients: 224 patients (mean age 80.6 years, 56% female, 75% emergencies) a dmitted to a department of medicine for the elderly. Presentational techniques: the methods we propose for the presentation of o utcome data are (i) place of discharge, using a two-dimensional diagram; (i i) 'survival' analyses, but using discharge from hospital rather than death as the endpoint; and (iii) 'phase diagrams', a novel method of charting th e progress of a cohort of patients. To illustrate these methods, the relati onship between admission case-mix (with patients put into tertiles on the b asis of their Barthel index score) and outcome is shown graphically. Result: each of the three techniques has different relative strengths, but their pictorial nature allows for rapid interpretation of data, showing, fo r example, the marked influence of case-mix. Separate analyses of subgroups of patients (such as those who die in hospital and those who survive) are also readily attainable by the three methods. Conclusions: the three methods of presenting outcome should be of benefit i n comparing the performance of different units, particularly when case-mix is taken into account. The pictorial methods are complementary both to more conventional patient-based methods (mean duration of stay, median duration of stay, percentile duration of stay, regression analyses etc) and to mode lling techniques using 'census' data from large numbers of patients.