The uses and abuses of life-table methods in vascular surgery.

This paper details the calculations involved in a life-table analysis and appraises the current use of this technique in vascular research. We suggest that the presentation of patency curves should be standardized and recommend that: curves are presented as step rather than continuous graphs; the number of patients remaining in a study are shown at intervals along patency curves; the length of time for which a curve is claimed to hold is more rigorously defined; an attempt is made to increase, where possible, the frequency of patient review, particularly in the immediate postoperative period; ways of increasing the total sample size (e.g. multicentre studies) are investigated. We also recommend that a set of definitive guidelines on the presentation of graft patency data acceptable to all the major vascular journals be sought.