Participants were provided the clear medication bottle and electronic cap used to record the date and time when the eye medication was removed for use by the patient.
The percentage of days on which a dose was taken within 4 hours of the mean dosing time for 407 participants who successfully completed the 3-month study visit. The vertical line represents the 75% threshold used to define nonadherence in this study.
eTable. Baseline Demographics of Study Participants
eFigure. Relationship of Participant-Reported to Measured Adherence
Boland MV, Chang DS, Frazier T, Plyler R, Friedman DS. Electronic Monitoring to Assess Adherence With Once-Daily Glaucoma Medications and Risk Factors for NonadherenceThe Automated Dosing Reminder Study. JAMA Ophthalmol. 2014;132(7):838-844. doi:10.1001/jamaophthalmol.2014.856
Adherence with topical glaucoma medications is low in some patients. To identify these patients, we need more information on risk factors for nonadherence.
To assess adherence with once-daily glaucoma medication.
Design, Setting, and Participants
Prospective cohort study involving patients who were recruited from a university-based glaucoma clinic and treated with once-daily prostaglandin eyedrops. Baseline characteristics were recorded, and each patient was provided a medication bottle to hold his or her eyedrop bottle throughout the study. An electronic cap recorded each time the larger bottle holding the eyedrops was opened. Participants were administered standardized tests of depression and mental status and answered questions about their health and their attitudes toward medication adherence. They used the electronic monitors for 3 months, during which their adherence with medications was assessed.
Main Outcomes and Measures
Adherence with daily glaucoma medication and factors associated with poor adherence.
Of the 491 participants enrolled, 407 (82.9%) successfully completed the 3-month adherence assessment. Of this group, 337 (82.8%) took their medication correctly on at least 75% of days and were deemed adherent. Compared with this adherent group, the 70 participants (17.2%) who were nonadherent were slightly younger, were more likely to be of African descent, took medications for a shorter time, had a lower level of educational attainment, and had worse scores on mental status and depression scales (P < .05 for all). Nonadherent participants were less likely to be able to name their glaucoma medications, reported a lower estimate of adherence, and were more likely to admit some missed doses over the past 2 weeks or in general (P < .05 for all). Those who were nonadherent were also less likely to agree that remembering their eyedrops is easy, more likely to strongly agree with the statement that eyedrops can cause problems, and less likely to agree that they followed physicians’ orders.
Conclusions and Relevance
Electronic monitoring of patient adherence documented that a sizable number of patients with glaucoma do not take their medications as prescribed. Factors were identified that may prove useful in targeting those nonadherent patients for interventions.
Topical medications lower intraocular pressure and delay the onset1 or slow progression of glaucoma.2 This benefit can be achieved only if a sequence of events takes place. First, the physician must recognize the need for treatment and prescribe therapy. Second, the patient must decide that taking the medication is important. Third, he or she must obtain the medication. Fourth, the patient must remember to take the eyedrops and, fifth, instill them.
The barriers to adherence in glaucoma are complex, as noted by Tsai et al,3 who reported 71 obstacles that could be categorized as situational or environmental factors, medication regimen, patient factors, and provider factors. Another assessment identified that patients face other important barriers to adherence: lack of education regarding glaucoma, lack of faith in the effectiveness of eyedrops, problems with eyedrop administration, forgetfulness, and the practical issues of medication administration—timing, location, and packaging.4
Multiple lines of evidence demonstrate patients are not taking their glaucoma medications as intended. Analyses of pharmacy claims data from insurance companies and health plans have consistently shown that glaucoma prescriptions are often not filled or refilled.5- 11 Others have found that pharmacy claims data do not always correlate well with other measures of adherence.12
Electronic monitoring of medication dosing is a more proximal measure of adherence.13,14 Most studies using electronic monitoring have demonstrated lower than desired adherence,15- 19 although this has not been universal.20 An evidence-based review of adherence with glaucoma medications found that nonadherence ranged from 5% to 80%.21 Although it is reasonable to assume that some degree of nonadherence with glaucoma medications would lead to worse visual outcomes, this link has not been clearly established.22- 24
Given the likelihood that nonadherence is indeed a risk factor for adverse visual outcomes, a previous study evaluated an intervention designed to improve adherence that was successful.25 While that intervention improved adherence among poorly adherent participants by nearly 20%, it was also labor intensive and applicable to a single brand of medication. The present study uses a simpler approach to remind patients to take their medications by assessing the effect of automated, voice, or text messaging on adherence. We report the characteristics and initial adherence of a cohort recruited to identify participants who might benefit from such an intervention.
The Automated Dosing Reminder Study (ADRS) was reviewed and approved by The Johns Hopkins University School of Medicine Institutional Review Board.
Participants in the ADRS were recruited from among patients being seen in the Glaucoma Center of Excellence at the Wilmer Eye Institute. To be included, a patient had to be at least 18 years old, be using a prostaglandin agent once per day (with or without other topical medications), and have no expectation of surgery. He or she also had to agree to have his or her medication use monitored and be willing to receive automated dosing reminders. Patients were not compensated for their participation and provided written informed consent.
A study coordinator used the medical record and a patient interview to record glaucoma diagnosis, duration of medical treatment of glaucoma, highest level of educational attainment, race, glaucoma medications, systemic medications, family history of glaucoma, baseline intraocular pressure (IOP; before medication, if known), most recent IOP, cup-disc ratios, target IOP, and ocular surgical history.
The coordinator also administered a self-assessment of general health, the Center for Epidemiologic Studies Depression Scale (CES-D10),26 and the Mini-Mental State Examination (MMSE).27 The CES-D10 is a 10-question instrument used to screen for depression in which a 5-point scale is used to assess how frequently the patient “felt depressed” or “felt tearful,” among others. Higher scores indicate a higher probability of depression. Similarly, the MMSE assesses patient orientation, memory, and cognitive function, with lower scores assigned for worse performance. Participants also responded to the following questions regarding medication adherence and their knowledge of glaucoma:
Do you personally know anyone who is blind?
Do you personally know anyone who is blind from glaucoma?
What are the names of your glaucoma medications? (Did the participant correctly name his or her medications?)
How many times per day do you take your Lumigan/Travatan/Xalatan?
How many drops do you use in each eye at each time you use it?
Over the past month, what percentage of your drops do you think you took correctly?
How many years have you been using drops for glaucoma?
In the past two weeks, I used my glaucoma drops (every day, all but one or two, more than half, about half, less than half, or almost never).
Some days I don't take all of my glaucoma medicines (strongly agree to strongly disagree).
When I don't take my glaucoma medicine, the main reason(s) is/are.
I'm the sort of person who follows doctors' orders exactly (strongly agree to strongly disagree).
Using glaucoma medications every single day may cause long-term side effects and problems (strongly agree to strongly disagree).
I have an easy time remembering to take my Lumigan/Travatan/Xalatan once per day (strongly agree to strongly disagree).
I don't like the idea of using glaucoma drops (strongly agree to strongly disagree).
Quiz Ref IDTo monitor medication use, we gave participants a 100-mL clear plastic bottle in which they were instructed to put their prostaglandin eyedrop bottle (Figure 1). Use of a second bottle is necessary since no system is available to monitor more than a single brand of eyedrop. Participants were also provided an electronic bottle cap (MEMS Cap; AARDEX Group Ltd) to record the use of their eyedrops. The MEMS Cap is an electronic device that records the date and time whenever it is removed from the medication bottle. The event logs are then extracted from the device during study visits via a device connected to a computer. Participants were instructed to keep their prostaglandin medication in the clear bottle and to retain the MEMS Cap on that bottle except when they were actually taking the medication once per day, just as they did before the study. They were also called 1 week after enrollment to assess whether and how they were using the provided medication bottle and MEMS Cap. Adherence with any other eyedrops being taken was not monitored.
After 3 months, the patients either returned the cap via US mail or brought the cap with them to a clinic visit. In either case, a coordinator downloaded the dosing data from the MEMS Cap. Quiz Ref IDAdherence was then assessed in 3 ways: as the percentage of (1) days on which a dose was taken within 4 hours of the mean dosing time for that patient, (2) the total number of prescribed doses taken, and (3) days on which the correct number of doses was taken. If any of these numbers was below 75% or if the second value was above 125%, the patient was determined to be “nonadherent” (only 2 participants were labeled nonadherent by the latter 2 criteria alone). The latter 2 criteria also allowed us to identify participants taking too much medication. These values were calculated on those doses taken within a window starting 2 weeks after the baseline visit and ending 2 weeks before the 3-month follow-up to minimize any positive effect of a clinic visit on adherence. The above definition of nonadherence is empirical and based on previous work in this area.19
At the 3-month point in the study, participants were asked about their experience with the dosing monitor and the use of their glaucoma medication:
I find the device easy to use when I put in my drops (strongly agree to strongly disagree).
I feel that using the monitor is causing me to take my drops less than I would if I were not using the dosing monitor (strongly agree to strongly disagree).
I have used my drops without using the device during the study (strongly agree to strongly disagree).
In the past three months, I used my glaucoma drops (every day, all but one or two, more than half, about half, less than half, or almost never).
Some days, I forget to take one of my doses of glaucoma medications (strongly agree to strongly disagree).
When I don’t take my glaucoma medicine, the main reason(s) is/are.
Baseline characteristics of those participants who successfully completed the initial 3-month adherence assessment were compared with those who were unavailable for or lost to follow-up. Comparisons were also made on variables related to adherence and glaucoma knowledge between adherent and nonadherent participants at 3 months. Categorical variables and proportions were compared using the Fisher exact test, and the distributions of continuous variables were estimated using the Wilcoxon rank sum test. Responses to Likert-type questions were compared using the Mann-Whitney test. Statistical analyses were performed using R (version 3.0.2; www.R-project.org).
Of the 491 participants, 407 (82.9%) completed the 3-month adherence assessment. The 84 participants who were unavailable for or lost to follow-up contained a lower proportion of males, were less likely to be of European descent, had a lower level of educational attainment, had a shorter mean duration of glaucoma medical therapy, and had a lower mean score on the MMSE (P < .05 for all; eTable in the Supplement). There were no statistically significant differences between the 2 groups with regard to age, glaucoma diagnosis, cup-disc ratio, or mean score on the CES-D10 or on their self-assessed health status (P > .05 for all; eTable in the Supplement). Of the 84 participants who were unavailable for or lost to follow-up, 43 (51.2%) could not be contacted, 15 (17.9%) disliked using the dosing monitor, 3 (3.6%) found the study inconvenient, 3 (3.6%) developed a medical problem, 1 (1.2%) found it too expensive, 1 (1.2%) transferred out of the practice, and 18 (21.4%) provided other reasons.
Of the 407 participants who successfully completed the 3-month visit, 337 (82.8%) were adherent using the definition above. Quiz Ref IDCompared with the adherent group, the nonadherent group was younger, with a mean difference of 3.3 years between groups; was more likely to be of African descent (58.6% vs 28.8%); took medications for a shorter time (7.3 vs 10.2 years); had a lower level of educational attainment; had a lower score on the MMSE (27.7 vs 28.4); and had a higher (worse) score on the CES-D10 (6.7 vs 4.8, P < .05 for all; Table 1). There were no statistically significant differences regarding sex, glaucoma diagnosis, previous surgery, cup-disc ratio, IOP, or self-assessment of health.
The mean measured adherence of the adherent group was 95.0%. The mean adherence of the 70 participants (17.2%) in the nonadherent group was 50.0% (Table 1). The distribution of adherence rates for all participants at the 3-month visit is shown in Figure 2, and the difference between patient-reported and measured adherence in the adherent and nonadherent groups is shown in eFigure in the Supplement. The Pearson correlation coefficient between the patient-reported adherence provided at baseline and the adherence measured at 3 months was 0.5.
Compared with the adherent group, nonadherent participants were less likely to be able to name their glaucoma medications (61.4% vs 82.2%). They also reported, on average, a lower estimate of adherence (87.0% vs 97.0%) and were more likely to admit some missed doses over the past 2 weeks or sometimes forgetting eyedrops in general (P < .05 for all). Those who were nonadherent were also less likely to agree that remembering their eyedrops is easy and also less likely to strongly agree that they followed physicians’ orders. There were no statistically significant differences in the number of doses per day or eyedrops per dose, no difference in their agreement with the statement “I don't like the idea of using glaucoma drops,” and no difference in agreement that eyedrops cause problems (P > .05 for all; Table 2).
At the 3-month visit, there was no difference between the adherent and nonadherent groups in terms of their agreement with the statement “I find the device easy to use when I put in my drops.” The nonadherent group was more likely to agree that they took fewer eyedrops because of the device and that they took eyedrops without using the clear medication bottle and electronic monitor cap. As with the same questions asked at baseline, the nonadherent group was more likely to report some missed doses during the previous 3 months and to agree that they sometimes forget eyedrops (Table 3).
Electronic monitoring of patient adherence once again demonstrates that a sizable number of patients with glaucoma do not take their once-daily glaucoma medications as prescribed. Failure to take glaucoma medications results in higher IOP and therefore almost certainly increases the likelihood of disease worsening. A recent retrospective analysis found that patients with poorer adherence had more severe visual field loss.24 Interventions to improve adherence are therefore an important goal of glaucoma research. The initial adherence-monitoring phase of the ADRS is an important contribution in this area in that it provided us with a cohort of nonadherent participants who could be randomized to an intervention designed to improve adherence. It also demonstrated risk factors that could be used to develop a quantitative tool for identifying patients likely to be nonadherent.
The risk factors identified that could be used to develop a risk calculator for nonadherence include patient age, race, educational attainment, duration of glaucoma treatment, MMSE score, and CES-D10 score. These results will also be useful in conjunction with our results from the Travatan Dosing Aid (TDA) study since one of the two can serve as an independent data set for validating a risk calculator. Production of a nonadherence risk prediction tool is important because only a small proportion of patients with glaucoma are not using their medications as prescribed, and any intervention should target that group.
The proportion of nonadherent participants in this study (17.2%) was significantly lower than that found in a previous study (44%).19 One potential difference between these studies is the dosing monitor used to assess adherence. If the MEMS Cap monitor somehow more consistently captures eyedrop use than the TDA used in the previous study, then higher adherence rates would be expected. However, both electronic monitors have been validated regarding accuracy, so this possibility is not likely.28,29 It is also unlikely that there is a differential effect of the MEMS Cap on adherence compared with the TDA since both devices make it obvious to the patients that their use of medication is being monitored.
Another difference between our studies is that participants were given free medication to use with the TDA but were kept on their own medications in this study. The lure of free medication may have resulted in more nonadherent patients participating in the previous study. The patient populations for the 2 studies may also have differed since, while both study populations were recruited primarily from the same practices of the same physicians at the Wilmer Eye Institute, the TDA study also included participants from another institution. Another interesting possibility for the increase in medication adherence is that there was a concerted effort by the Wilmer Glaucoma Service to communicate more deliberately with patients about adherence in the interval between the 2 studies. This effort, based on established principles of patient-physician communication, was the subject of previous work.30,31 This change was not systematic, and there was no evaluation of its effect on adherence, however.
Nearly 20% of participants did not complete the 3-month follow-up visit. This rate is smaller than that of the TDA study, in which 30% of enrolled participants were lost to follow-up,19 but larger than the dropout rate reported by Robin et al20 (3%). Those who did not complete the initial phase of the ADRS were more likely to be female, have a slightly lower (worse) MMSE score, be of African descent, and receive therapy for a shorter time. Given the latter 2 features have been associated with decreased adherence in previous studies,23 our overall estimate of the adherence rate may be too high. On the other hand, the TDA study had a higher dropout rate and found a lower rate of adherence19 than we report for the ADRS, while the study by Robin et al20 had a very low dropout rate and very high adherence. Those who dropped out of our study as well as those who were poorly adherent and refused to participate in the first place are inherently difficult to study but are important to consider in the broader effort to address adherence in glaucoma, however.
Participants in the nonadherent group reported taking fewer eyedrops due to the monitor and were more likely to report taking eyedrops without using the monitor. Both of these behaviors would result in artificially low estimates of adherence for this group. This issue will be resolved in the analysis of the intervention since one would expect that these artificially low estimates of adherence would not improve with an intervention.
Quiz Ref IDAs in previous studies, we found poor correlation between self-reported adherence and adherence measured objectively.19,32 Some information can still be gleaned from the adherence-related questions, however, since the nonadherent group was statistically more likely to admit to missing drops at both the baseline and 3-month points. At baseline, they reported a lower overall adherence rate for the past month, were less likely to agree that they follow physicians’ orders, were more likely to agree that eyedrops cause problems, and were less likely to agree that remembering eyedrops is “easy.” At the 3-month visit, they were also more likely to agree that they forget eyedrops on some days. Given that most patients are taking their eyedrops as prescribed, identifying patients at risk of nonadherence is a critical step. The results from the patient questions and demographic factors may therefore be useful in creating risk calculators that could find those patients most in need of intervention.33
Submitted for Publication: July 26, 2013; final revision received January 26, 2014; accepted January 28, 2014.
Corresponding Author: Michael V. Boland, MD, PhD, Glaucoma Center of Excellence, Wilmer Eye Institute, The Johns Hopkins University School of Medicine, 600 N Wolfe St, Wilmer 131, Baltimore, MD 21287 (firstname.lastname@example.org).
Published Online: May 15, 2014. doi:10.1001/jamaophthalmol.2014.856.
Author Contributions: Dr Boland had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Boland, Frazier, Plyler, Friedman.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Boland, Frazier, Friedman.
Critical revision of the manuscript for important intellectual content: Boland, Chang, Plyler, Friedman.
Statistical analysis: Boland.
Obtained funding: Boland, Frazier.
Administrative, technical, or material support: Boland, Chang, Plyler.
Study supervision: Boland, Plyler.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was supported by a grant from the Microsoft Be Well Fund.
Role of the Sponsors: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; or decision to submit the manuscript for publication.
Disclaimer: Dr Boland is the JAMA Ophthalmology Web Editor, but was not involved in the review process or the acceptance of the manuscript.
Acknowledgments: Rhonda Miller, Glaucoma Center of Excellence at the Wilmer Eye Institute, helped with the data collection and management portions of the study.