• adverse drug events;
  • levothyroxine;
  • medication safety;
  • monitoring;
  • thyroid


Objectives  Hypothyroid patients managed with excessive or insufficient thyroid replacement therapy are often difficult to clinically recognize. Monitoring may prevent or minimize the consequences of adverse drug events (ADEs). We sought to develop an explicit model of medication monitoring and to evaluate monitoring processes and ADEs in patients taking levothyroxine.

Methods  A retrospective chart review of 400 outpatients receiving levothyroxine therapy between 1 January 2000 and 1 January 2001 at a large North American tertiary care hospital. We measured the proportion of patients satisfying minimum monitoring criteria, experiencing specific monitoring errors and having levothyroxine-related ADEs. Explicit monitoring criteria were derived from the literature and through expert opinion. Adverse drug events were identified using structured implicit reviews.

Results  Overall, only 56% (95% confidence interval [95% CI] 51–62%) of the patients prescribed levothyroxine received the minimal recommended monitoring. Errors were identified at all stages of the monitoring model. Patients who received the recommended monitoring had fewer levothyroxine-related ADEs (1% vs. 6%, P = 0.013) than those who did not. Minority status (white people 2% vs. black people 4% vs. Hispanics 14%, P = 0.023) and primary language (English 3% vs. Non-English 20%, P = 0.002) were the patient characteristics associated with levothyroxine-related ADEs.

Conclusion  Only half of outpatients taking levothyroxine at one tertiary care hospital received the recommended monitoring during one year of follow-up. Levothyroxine-related ADEs were more frequent in patients with lower-quality monitoring and in minorities and non-English speakers.