Hwang D, Liu CJ, Pu C, Chou Y, Chou P. Persistence of Topical Glaucoma MedicationA Nationwide Population-Based Cohort Study in Taiwan. JAMA Ophthalmol. 2014;132(12):1446-1452. doi:10.1001/jamaophthalmol.2014.3333
Medication persistence is an important factor for treatment effect in patients with glaucoma. Evaluating risk factors for refill discontinuation might be helpful for improving persistence and preventing blindness in patients with glaucoma.
To estimate the persistence rate with topical glaucoma medication 2 years after diagnosis and evaluate risk factors for nonpersistence among patients in Taiwan with open-angle glaucoma and ocular hypertension.
Design, Setting, and Participants
A retrospective population-based study using claims data from the National Health Insurance Research Database. One million patients were randomly selected from the registered beneficiaries of the National Health Insurance Research Database in 2000. All patients with newly diagnosed open-angle glaucoma and ocular hypertension were included and followed up until December 31, 2008. Patients were included in the analysis only if they had follow-up data for more than 2 years after diagnosis.
Main Outcomes and Measures
Nonpersistence was defined as the patient not refilling any topical glaucoma medication for more than 90 days. Patient characteristics, prescription-related clinical factors, and physician and hospital characteristics were identified and considered in the analysis. The rate of persistence was estimated and risk factors for nonpersistence were investigated using Cox proportional regression models.
A total of 3134 patients were identified and observed in the study. After a 2-year follow-up, 759 patients (24.2%) persisted with their glaucoma medications. Multivariate analysis showed that patients’ living or working areas (P < .001), number of glaucoma medications (P < .001), prescription of pilocarpine hydrochloride (adjusted ratio of persistence = 0.72; 95% CI, 0.59-0.88) or prostaglandin analogs (adjusted ratio of persistence = 2.04; 95% CI, 1.82-2.33), the year in which glaucoma diagnosis was made (adjusted ratios of persistence for patients whose condition was diagnosed after 2004 = 1.18; 95% CI, 1.09-1.27), sex of the main physicians (adjusted ratios of persistence for male ophthalmologists = 0.82; 95% CI, 0.74-0.90), treatment in hospitals (P < .001), and continuity of care index (P < .001) were associated with patients’ persistence with glaucoma medications.
Conclusions and Relevance
The rate of persistence for glaucoma medications is low in Taiwan, although health care costs, including the cost of medication, are mostly covered by the nationwide health insurance system. This study suggests that factors other than cost, such as physician-patient relationship and patient education, may play an important role in the persistence of topical glaucoma medication.
Glaucoma is a progressive optic neuropathy that can lead to irreversible blindness. Intraocular pressure (IOP)–lowering therapy is the only effective strategy recognized to date.1 Most patients require daily topical medications for IOP control. The variation in patient adherence to taking medication may directly influence the effect of treatment and poor adherence may lead to greater disease progression and vision loss.2,3 Most studies have revealed that the medical persistence in patients with glaucoma is relatively low, with the lowest persistence rate reported as 10% at 12 months.4- 11 Previous studies have most often discussed drug-related factors affecting persistence rates among patients with glaucoma, including pharmacological class, dosing frequency, and fixed combination medication.5,6,8,12- 14 Younger age, lower socioeconomic status, and uncertain diagnosis have also been associated with poorer medical persistence.7,8 In addition, the attitude of physicians in clinics, communication and trust between physicians and patients, and even schedule arrangements have been associated with adherence among patients with glaucoma.15- 17 However, few studies have explored the impact of physician demographics, accreditation level of hospitals, and continuity of care on the persistence of glaucoma medication.
The medical environment, such as health policies, the referral system, physicians’ attitudes toward treatment, patients’ perceptions of disease and treatment, and patients’ medical uses, is very different between Eastern and Western countries. Quek and colleagues7 investigated the persistence of glaucoma medical therapy in Singapore and demonstrated that the rate of persistence for glaucoma medication was 22.5% at 1 year, decreasing to 11.5% after 3 years.
The government in Taiwan implemented the National Health Insurance (NHI) program after 1995. This nationwide single-payer program is compulsory, has significantly reduced medical outgoings, and has increased medical use in Taiwan.18 The quality of patient care and professional competency has improved at different levels of medical use because accreditation is completed every 2 to 3 years.19 In the present study, we used the administration database derived from the NHI program to estimate the rate of persistence for glaucoma medical therapy in Taiwan and evaluate risk factors for discontinuation in several dimensions including disease-related factors, patient characteristics, physician-related characteristics, and continuity of care.
The NHI program covers more than 99% of residents and medical resources in Taiwan. All claims, including inpatient and ambulatory reimbursements, details of pharmacological prescriptions, surgical procedures and examinations, and registry data for beneficiaries, physicians, and medical resources, were collected from the National Health Insurance Research Database (NHIRD). The validity, representativeness, and clinical consistency of this database have been reported.20,21 Regarding concerns of privacy, all data that could possibly be used to identify patients or physicians were encrypted before being released to researchers. The institutional review board at the Taipei Veterans General Hospital ruled that approval was not required for this study and participants’ consent was not required owing to the retrospective nature of the study.
We conducted this retrospective cohort study by retrieving data for 1 million participants randomly selected from the registered beneficiaries of the NHI program in 2000. This cohort represented 4% of all residents in Taiwan. The age and sex distributions of this cohort were similar to that of all beneficiaries at that time.22 All NHIRD data were evaluated for participants with claims before December 31, 2008.
We collected data for patients diagnosed as having glaucoma or ocular hypertension (International Classification of Diseases, Ninth Revision, Clinical Modification, 365.X or 364.22) first coded after January 1, 2001 (ie, no previous glaucoma/ocular hypertension diagnosis from January 1, 1996, to the date of diagnosis). A patient was deemed in need of persistent glaucoma medical treatment if more than 3 reimbursing claims with glaucoma diagnoses and associated drug prescriptions were identified. The index date was defined as the date when the patient’s condition was first diagnosed and treated with glaucoma medication.
In Taiwan, topical medication is usually used as the first-line treatment of open-angle glaucoma. Laser trabeculoplasty is used occasionally for medicated patients whose IOPs require further reduction. Of all identified patients with open-angle glaucoma in our database, only 3.2% received laser treatment during the study period. However, without detailed medical records, the necessity of persistent glaucoma medication could not be ascertained in patients with secondary glaucoma or patients who received laser treatment or glaucoma surgery during the study period. Thus, we excluded all patients with secondary glaucoma (International Classification of Diseases, Ninth Revision, Clinical Modification: 364.22, 365.3X, 365.4X, 365.5X, and 365.6X; 6.1% of all patients with glaucoma) or a history of glaucoma laser or incision surgery. Patients with primary angle-closure glaucoma (International Classification of Diseases, Ninth Revision, Clinical Modification: 365.2X; 20.1% of all patients with glaucoma) were excluded from this study because laser iridotomy was regarded as the primary treatment of choice.
The data for all patients were observed and traced for at least 2 years after the index dates. Patients who died or left the insurance program within 2 years after the initial glaucoma diagnosis were excluded from the analysis. Patients with incomplete registry data, such as unknown sex, inconsistent birth date, or incomplete insurance information, were also excluded from this study.
Persistence with medication was defined as the continuous refilling of a topical glaucoma medication with a time interval equal to or less than 90 days based on the pharmacy claim data. Nonpersistence was defined as a gap of more than 90 days for drug refills. To make our results comparable with various previous studies, we also tried to define the prescription gap as more than 60 days or 120 days by calculating the medication possession ratios, which equal the sum of supply days of all glaucoma medications divided by the total days during the follow-up period as a sensitivity analysis.
Patient demographic data included the age at which the glaucoma diagnosis was first coded, sex, and urbanization of the working area. Age was stratified into the following 4 groups: younger than 20 years, 20 to 39 years of age, 40 to 59 years of age, and 60 years or older. Living or working areas were categorized as urban, suburban, or rural based on population density, the availability of medical resources, age distribution, and education levels of residents living in these areas. Diagnosis of dementia was confirmed if the diagnosis code was observed 3 or more times in ambulatory claims or once in the in-patient claims in the database.
Drug-related factors were analyzed based on prescriptions at the last visit of drug refill including the number and class of topical glaucoma medications and whether commercial fixed combination drugs were used. Patients using 1 fixed combination medication, such as Cosopt (dorzolamide/timolol), were categorized as using 2 medications. To confirm the relationship between prescribed medications and a patient’s persistence, we repeated the analysis using the initial prescription and compared the results with what were obtained by using the final prescription as a sensitivity analysis.
The age and sex of the ophthalmologist a particular patient most frequently visited, accreditation levels of medical utilities, and the continuity of care index (COCI) were also analyzed. The age of the ophthalmologist was categorized as younger than 45 years or 45 years or older. The COCI was calculated using the Bice and Boxerman23 equation:Where N is the total number of physician (ie, ophthalmologists) visits, nj is the number of physican visits, j represents a given physician, and m is the number of different physicians the patient visited during the study period. The index is between 0 and 1, with 1 indicating that the patient had perfect continuity of care and 0 indicating that the patient had poor continuity of care.
To determine the association between each factor and drug persistence, a univariate analysis was performed with the χ2 test and a multivariate analysis calculated an adjusted hazard ratio for nonpersistence using Cox proportional regression models. The ratios of persistence and 95% CIs were calculated as the reciprocal of the adjusted hazard ratios (1 divided by the adjusted hazard ratio). The SAS statistical package version 9.2 was used for all statistical calculations. The 2-sided significance level was set at P = .05.
A total of 4100 patients newly diagnosed as having glaucoma were identified from the database. Of these patients, 592 cases (14.4%) died or left the insurance program within 2 years and 3134 of them met the inclusion criteria. Among these patients with glaucoma, 759 (24.2%) were persistent in using their glaucoma medication for at least 2 years after the first prescription. In contrast, 2375 patients discontinued their drug refill for more than 90 days at any point during the follow-up. Most of the nonpersistent cases (91.7%) discontinued their drugs in the first year and 57.7% of the nonpersistent patients restarted their medication during the follow-up.
Table 1 provides the results of the univariate analysis of each variable included in our analysis. After adjusting all factors with the multivariate analysis, several risk factors for persistence were identified in the Cox regression model (Table 2). Patients who lived or worked in an urban area had a higher rate of persistence than those who lived or worked in a suburban or rural area (P < .001). The greater the number of topical medications the patient was using, the higher the rate of persistence; the use of pilocarpine and prostaglandin analogs correlated with the patients’ persistence (the adjusted ratios of persistence were 0.72 and 2.04, respectively).
Patients who were treated primarily by female ophthalmologists or who visited medical centers were less likely to discontinue their medical treatments. In addition, the higher the COCI, the higher the rate of persistence (P < .001). This result indicates that patients who followed up at 1 physician’s clinic tended to have better persistence than those who often changed their doctors.
In regard to the sensitivity analysis, the rate of persistence among patients with glaucoma was 14.9% and 30.8% at 2 years if nonpersistence was defined by a refill interval of more than 60 days and more than 120 days, respectively. The calculated medication possession ratio for glaucoma medications within 2 years was lower than 1 in 1358 patients (43.3%). A second multivariate analysis using these definitions to identify factors for persistence had similar results. The significant drug-associated factors were also unchanged when we repeated the analysis based on the first prescription claims data.
Persistence with medical therapy is one of the important aspects of adherence to treatment among patients with glaucoma. Calculating the medication possession ratios or proportion of days covered could estimate the persistence of medication in chronic disease. A commonly adopted way to explore exogenous causes, which would impede patients’ adherence to treatment, is analyzing the association between risk factors and the interval before patients discontinued their drugs. Various time intervals between drug refilling have been used to define discontinuation of medical treatments in patients with glaucoma. The longer the interval in a refilling gap, the higher the identified persistence rate.9 The refilling gap was most commonly defined as 90 days.4,7,8,10,13
The persistence of glaucoma medication ranges from 19% to 64% in the United States and from 69% to 84% in Europe, with the mean time of therapy ranging from 10.8 to 21.8 months.5,6,8,12- 14,24- 26 In our study, 24.2% of patients with newly diagnosed glaucoma persisted with their medication after 2 years of follow-up and 92% of nonpersistence occurred within the first year of initial diagnosis. Although some studies using a 90-day gap analysis or calculating medication possession ratios have comparable results with ours, a detailed comparison with previous studies has not been performed owing to variations in study methods.5,6,10,13 To our knowledge, only the Quek and colleagues7 study used similar methods to those of our study. They followed up 2781 patients treated with monotherapy in Singapore and reported that the persistence was 22.5% after 1 year of follow-up, which is also comparable with our estimation.
We found that patients with glaucoma who live or work in urban areas have a higher persistence with their topical medications. This finding is consistent with previous research regarding adherence with glaucoma medication.27- 29 People living or working in urban areas may have a higher socioeconomic status, which has been associated with medical adherence. In addition, more medical resources are available in urban areas compared with the other 2 areas, which makes it easier to access medical care and refill drug prescriptions.
Having dementia has been reported to be related to the persistence of medical treatment in many diseases. However, differences in medical persistence between patients with glaucoma with or without dementia were not found in our study. One possible explanation is that most patients with dementia probably live with family members or are institutionalized and are not making adherence decisions themselves.
Previous studies investigating the impact of drug-related factors on the persistence of glaucoma treatment usually focused on patients with newly diagnosed glaucoma receiving monotherapy.5,8,12,14,30 Studies of this kind may not reflect the situation in the real world because patients with glaucoma often switch from one medication to another and many patients need multiple medications to control their IOPs.31 With a retrospective population-based cohort study, it would be difficult, if not impossible, to refine the impact of each class of medication on drug persistence. Instead, this study revealed the rate of drug persistence in patients with glaucoma in a real-world scenario and allowed us to investigate the impact of other factors on persistence in glaucoma treatment after adjusting for drug-related factors.
We found that patients who were prescribed more glaucoma drugs had better persistence than those prescribed fewer drugs, with a highly significant association and dose-effect relationship. This finding seems to contradict the common perception that noncompliance with glaucoma medication is higher with a greater dosing frequency.29,32,33 The study design for determining the persistence of drug refilling in a population-based database with a variety of drug combinations is different from that used to evaluate drug compliance in a particular group of patients receiving monotherapy. Patients who need more glaucoma medications are those whose IOP is more difficult to control or those with more advanced visual field defects. These patients may pay more attention to their ocular condition out of fear of going blind. Besides, a better way to address the association between dosing frequency and adherence is by performing prospective clinical studies that evaluate the actual dispensing of glaucoma medication using an electronic monitoring device. We did not find a significant association between a patient’s persistence and the use of fixed combination medications even after adjusting the year of diagnosis, which might be owing to the small number of patients using fixed combination medications in our database.
Patients with glaucoma treated primarily by female ophthalmologists were more persistent with their treatment than those treated primarily by male ophthalmologists. As far as we know, this association has never been reported in patients with glaucoma or any other ocular disease. Nevertheless, previous studies regarding hypertension and hyperlipidemia found that the physician’s sex matters in risk factor control.34- 36 Studies in Asian and other non-Asian countries have found that the health care behaviors, consultation styles, and education efforts are different between male and female physicians.37,38 Perhaps it is one of the reasons why patients treated by female physicians tend to have better medical persistence.
Patients with glaucoma treated primarily at medical utilities with higher accreditation levels had better persistence in our study. The first possible explanation is that patients treated at such hospitals have more advanced disease.39 Second, the quality of health care is an important indicator for judging the accreditation level of a medical utility in Taiwan; thus, the patient-physician relationship, communication, and disease education may be better in hospitals with higher accreditation levels. Third, because copayment is higher in hospitals with higher accreditation levels, patients who seek medical help at these hospitals may share certain characteristics such as a higher socioeconomic status or paying more attention to their own health.
Our results also showed that the COCI is directly related to medical persistence in patients with glaucoma. This finding has also been reported in other systemic diseases.40,41 One explanation is that treatment by the same doctor implies better physician-patient relationship and communication, which may help improve patients’ knowledge about their disease status and the importance of adherence.
Without a medical record review, a major limitation of a database-derived study is the possibility of inaccurate diagnoses and a lack of information regarding disease severity. We only included patients with a glaucoma diagnosis code on at least 3 consecutive visits to increase the validity of the study population. Nevertheless, we may have missed a few patients who stopped seeking medical help after the first or second visit, resulting in selection bias and an overestimation of persistence. In addition, patients who had laser surgery were excluded from our study, which may have limited the applicability of our study results to those whose IOP was easily controlled with medical therapy. We cannot affirmatively distinguish the incident from prevalent cases in our study. Based on our study design, none of the patients had glaucoma diagnosis coding for at least 5 years before they were enrolled in this study. Although we might have enrolled some patients who had discontinued their glaucoma treatment for more than 5 years, the number of such patients was likely to be small and unlikely to bias the study results. Furthermore, our findings cannot be extrapolated to patients with angle-closure glaucoma or secondary glaucoma. Further studies based on detailed clinical records are needed to confirm the findings of this study.
The present population-based study highlights that low persistence of glaucoma medical therapy is still a considerable problem, even under the nationwide health insurance system. The severity of glaucoma, socioeconomic status of patients, and patient-physician relationships were associated with the rate of persistence in patients with glaucoma. Clinicians and policy makers should pay more attention to these issues; strategies, such as public education or medical regulation, may help improve drug persistence in patients with glaucoma.
Corresponding Author: Catherine Jui-Ling Liu, MD, Department of Ophthalmology, Taipei Veterans General Hospital, 201 Shih-Pai Rd, Section 2, Taipei, Taiwan 11217 (firstname.lastname@example.org).
Submitted for Publication: April 22, 2014; final revision received July 4, 2014; accepted July 9, 2014.
Published Online: September 11, 2014. doi:10.1001/jamaophthalmol.2014.3333.
Author Contributions: Dr Hwang had full access to all of 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: All authors.
Acquisition, analysis, or interpretation of data: Hwang, Pu, Y.-J. Chou, P. Chou.
Drafting of the manuscript: Hwang, Pu, Y.-J. Chou, P. Chou.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Hwang, Liu, Pu, Y.-J. Chou, P. Chou.
Obtained funding: Hwang, Pu, Y.-J. Chou, P. Chou.
Administrative, technical, or material support: Hwang, Pu, Y.-J. Chou, P. Chou.
Study supervision: Hwang, Pu, Y.-J. Chou, P. Chou.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: The National Health Research Institutes supplied data for this study.
Role of the Funder/Sponsor: The funder 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; and decision to submit the manuscript for publication.