The medical treatment of glaucoma is limited by patient adherence to medications. By the end of a year, approximately half of patients are not refilling their medications as directed, and physicians are not able to tell those who are from those who are not.^1,2,3

In recognition of this, the travoprost dosing aid is a device that has been designed to aid drop administration, has integrated visual and auditory alarms to remind patients to use their medication, and tracks drop usage for the treating physician.^4,5,6

We performed a study to try to gauge patient adherence to travoprost while using the dosing aid, and to try to assess the influence that patients' perceptions of being monitored might have on adherence.⁷

We were surprised to find that some of our best patients, who were specifically enrolled in this study because of their perceived good adherence, used only 78% of all prescribed drug doses overall. We also found that patients liked the dosing aid, but did not seem particularly influenced by or concerned that their detailed drop usage history was available to the clinician.

Although in the past, studies were performed in subjects who were unaware that they were being monitored,^8,9 current Institutional Review Boards insist that subjects be fully informed of study design. We therefore told all of our subjects that everyone would receive a dosing aid, but that some of these would not be functional. We then proceeded to recruit 45 subjects who were given a dosing aid and told to use travoprost in both eyes, each evening at a predetermined time.

Designing the study

Twenty subjects were given fully functional dosing aids (group 1), 20 were given dosing aids that had no alarms but still recorded drop usage (group 2), and five were given dosing aids that did not record dosage or display alarms. We hoped that group 2 would believe that their devices were not monitoring their drop usage, allowing us to compare their adherence to that of group 1, and gain some insight into the influence that monitoring might have on drop usage.

Adherence was defined as: 1) any use of drops on a given day, or 2) the use of drops within the predetermined two-hour time window.

Reported adherence rates differed to reality

Of the 45 patients [Age (SD): 67.6 (12.1); 47% male], all but one completed the questionnaire (one patient died). One device malfunctioned in group 1. Two devices were never returned by group 2 patients, despite persistent attempts to obtain these, and three devices malfunctioned.

Table 1: Association of group membership and questionnaire response.

When using definition 1 (any drop usage on a given day), the overall adherence was 78% (95% CI: 71-85). Adherence was nearly identical in both groups: 78% adherence in group 1 (95% CI: 70-88) and 76% adherence in group 2 (95% CI: 65-89). One of the most striking results was that, although most patients missed at least one day of treatment, few reported it in the questionnaire (Tables 1 & 2).

Table 2: Intervals of non-adherence by definition 1.

When using definition 2 (drop usage within the subject chosen two-hour window), adherence was 42% (95% CI: 34-54) and, although adherence was higher by definition in group 1, this failed to reach statistical significance: 51 % adherence in group 1 (95% CI 40-65) vs. 31% adherence in group 2 (95% CI: 20-48) (p>0.05). The relative risk of adherence using definition 2 was 1.63 (95% CI 1.00-2.68) (p=0.052): this suggests a statistically non-significant greater adherence in the first group with fully functional devices. It may be that the lack of alarms exerted a statistically non-significant influence on the timeliness of the dosing in group 2, but that our study lacked the power to show this.

Subjects in group 2 were more likely to admit to not administering eye drops but much less likely to admit that the dosing aid affected how much they used their drops (see Table 1, questions 1 and 4), which is consistent with the lack of alarms compared with group 1.

Most subjects in both groups 1 and 2 thought they were being monitored, undermining the study's goal of investigating the relationship between perception of monitoring and adherence. The increased adherence in patients who believed that they were being monitored was not statistically significant.

For each subject who did not comply on at least one day, the mean number of consecutive days of non-adherence and the maximum number of consecutive days of non-adherence were calculated. Under definition 1, 16 (of a possible 19) patients in group 1 and 12 (of a possible 15) patients in group 2 had at least one day of non-adherence (Table 2). Using definition 2,

Table 3: Intervals of non-adherence by definition 2.

19 patients in group 1 and 15 patients in group 2 had at least one day of non-adherence (Table 3).