BACKGROUND: There is still considerable disagreement on several aspects of monitoring asthma with symptom score and peak expiratory flow (PEF). OBJECTIVE: To investigate patient adherence to a retrospective diary-card monitoring method in patients with poorly controlled persistent asthma, in a clinical management setting; to develop improved methods for fast manual entry data into a computer; and to generate real-time informative graphs of the data. METHODS: In 115 consecutive adult patients we applied a diary-card monitoring method in which the patient records symptom score and PEF. We analyzed the diary cards of 84 patients. We used SigmaPlot software to graph the data, and linear regression to analyze the relationship between days of expected diary-card completion and days of actual correct diary-card completion (completed entries). RE-SULTS: Linear regression gave an overall correlation coefficient (r 2 ) of 0.65. Surprisingly, the r 2 values in the patients with mild, moderate, and severe asthma were 0.24, 0.44, and 0.99, respectively, revealing a striking correlation between adherence and severity. Moreover, when we arbitrarily set 75% as the minimum acceptable rate of days of completed diary-card entries, 68% of the patients were in the over-75% category. Remarkably, 100% of the patients with severe asthma were above the 75% cut-off. The graphing method we tested proved user-friendly, flexible, and quick, allowing computerized processing of 30 days of data sets in 5 min, and generation of high-quality selfexplanatory graphs that facilitate rapid management decision making via visual pattern recognition. CONCLUSIONS: In a clinical setting, retrospective monitoring of patients with moderate and severe persistent asthma by symptom score and PEF is feasible, and patient adherence appears to be good, particularly in patients with severe asthma. We recommend a lower priority on retrospective monitoring in patients with mild persistent asthma. Monitoring should be carried out according to a definite follow-up protocol. Improving the quality and standardization of the monitoring graph is a priority.