Title: An International Survey of Patients with Thalassemia Major and Their Views About Sustaining Life-Long Desferrioxamine Use Authors: Alexandra Ward (alexward@caroresearch.com) J Jaime Caro (jcaro@caroresearch.com) Traci Craig Green (tcraig@po-box.mcgill.ca) Krista Huybrechts (khuybrechts@caroresearch.com) Alejandro Arana (alejandro.arana@pharma.novartis.com) Suzanne Wait (suzanne.wait@bms.com) Androulla Eleftheriou (thalassaemia@cytanet.com.cy) Version: 1 Date: 19 Mar 2002 This manuscript addresses the important problem of non-adherence to life-long medication regimens and provides insight into patients' reasons for non-adherence. The sample size is robust (n=1888) and comprised of responses from 10 countries. The manuscript demonstrates that non-adherence is an international problem, and that patients beliefs and attitudes about medication are the most important factors in non-adherence. I recommend acceptance of the manuscript, pending inclusion of the following information: Compulsory revisions: 1) More information on the procedures used to translate the questionnaire into Greek, Italian, Arabic, Farsi and Chinese is needed. For example, was back-translation used to validate the initial translation process? Also, were native speakers used in each country or was the translation process done in one location? 2) The proportion of patients who missed at least one dose of desferrioxamine in the preceding month was the primary outcome measure. Why did the authors use this particular cut-off for non-adherence? Was there a clinical rationale for defining non-adherence as missing one dose or was this decision based on statistical considerations? 3) The authors report that age was highly related to missing at least one dose, and that non-adherence was significantly different by country. (Figure 2, chi-square=159.8). The age distributions of the sample in each country were different so it is important to determine if differences in non-adherence between countries are the result of differences in the age distributions. Logistic regression was used to determine if country and age were independent predictors of non-adherence but the results are not presented. I would recommend adding a table describing the logistic regression results. Discretionary revisions: 1) Sixty-five percent of patients returned the questionnaire - Are there differences between those who responded and those who did not? Age was highly associated with non-adherence, thus it would be useful to know if non-responders were older or younger than those who responded to the questionnaire. Since the sample was selected from patient organization membership lists, information on the age and
gender of non-responders might be available.