Racial and Ethnic Disparities in Adherence to Glaucoma Follow-up Visits in a County Hospital Population

Objectives To identify predictors of inconsistent attendance at glaucoma follow-up visits in a county hospital population.

Methods Prospective recruitment from August 1, 2008, through January 31, 2009, of 152 individuals with glaucoma, with 1-to-1 matching of patients (those with inconsistent follow-up) and controls (those with consistent follow-up). Data were collected via oral questionnaire. Survey results were correlated with attendance at follow-up examinations, using the t test, χ² test, and multivariate stepwise logistic regression analysis to calculate the odds ratios (ORs) and 95% confidence intervals.

Results After adjusting for covariates in the logistic regression analysis, factors independently associated with inconsistent follow-up included black race (adjusted OR, 7.16; 95% confidence interval, 1.64-31.24), Latino ethnicity (adjusted OR, 4.77; 1.12-20.29), unfamiliarity with necessary treatment duration (adjusted OR, 3.54; 1.26-9.94), lack of knowledge of the permanency of glaucoma-induced vision loss (adjusted OR, 3.09; 1.18-8.04), and perception that it is not important to attend all follow-up visits (adjusted OR, 3.54; 1.26-9.94).

Conclusions Demographic factors, including race and ethnicity, may directly or indirectly affect adherence to recommended glaucoma follow-up visits. Lack of information regarding irreversible vision loss from glaucoma, need for lifelong treatment, and lack of visual symptoms may be significant barriers to follow-up in this population. Targeted glaucoma education by physicians may improve follow-up, thereby decreasing the morbidity associated with glaucomatous disease.

Glaucoma is the second leading cause of blindness after cataract, accounting for 12.3% of the blindness worldwide, and is the leading cause of irreversible vision loss.¹ Lowering intraocular pressure (IOP) significantly reduces disease progression and development of adverse outcomes, such as vision loss and blindness.^2-4 Several clinical studies^5-7 have demonstrated that the use of IOP-lowering medications decreases the likelihood of measurable disease progression, and it is postulated that patients adherent to therapy will have more favorable outcomes than those who are nonadherent.

Regular patient follow-up is critical for physicians to assess the course of glaucomatous disease and appropriately adjust therapy as needed.^8,9 The American Academy of Ophthalmology suggests at least 2 follow-up visits per year in patients with primary open-angle glaucoma in the Preferred Practice Pattern guidelines.¹⁰ Inconsistent adherence to recommended follow-up hinders the ability of the physician to track disease progression, presumably increasing the likelihood of adverse disease outcomes.¹¹ Kosoko et al¹¹ reported that patients at a glaucoma clinic in inner-city Baltimore, Maryland, with inconsistent follow-up were more likely to have mild disease than those who demonstrated consistent follow-up. Those with inconsistent follow-up also were more likely to express displeasure with long wait times in clinics. Although those with inconsistent follow-up were less likely to have been prescribed glaucoma medications than those with consistent follow-up, it is unclear whether this was due to more severe disease in case than control individuals. A questionnaire-based study¹² at the Aravind Eye Care System in Coimbatore, Tamil Nadu, India, showed that little or no formal education, not using glaucoma medications, and the perception that glaucoma follow-up is less important when patients are adherent to their prescribed glaucoma medication regimen were all independent predictors of inconsistent glaucoma follow-up.

Disparities in the prevalence of glaucoma by race and ethnicity have been well documented in several large population-based studies.^13-17 Previous studies^15,18-20 have found that blacks and Latinos are at greater risk for glaucoma than whites. Black race is a known risk factor for glaucoma, evidenced by increased disease prevalence, earlier disease development, and increased progression to blindness. Latinos also experience an exponentially greater incidence of disease development with age compared with whites.^13,16,17 Several such studies were conducted at academic medical centers serving patient populations predominantly of 1 minority racial/ethnic group who were then compared with whites. No large prevalence surveys, however, have directly compared glaucoma prevalence in blacks and Latinos in 1 city or county population, to our knowledge.

We conducted a case-control study of individuals with definitive or suspected glaucoma to assess differences between those with inconsistent follow-up (ie, cases) and those with consistent follow-up (ie, controls) at San Francisco General Hospital (SFGH), a county hospital serving a patient population that is diverse with regard to race, ethnicity, and language. Data were examined for independent predictors of inconsistent glaucoma follow-up, and patient-reported barriers to such follow-up were assessed.

We recruited 152 individuals diagnosed as having glaucoma or suspected of having glaucoma examined at the SFGH Glaucoma Clinic in San Francisco, California, from August 1, 2008, through January 31, 2009, with 1-to-1 matching of cases (those with inconsistent follow-up) and controls (those with consistent follow-up). This clinic receives referrals from within SFGH and its satellite community centers throughout San Francisco. The clinic is located in an inner-city county hospital that provides comprehensive health care services as the safety net system throughout the San Francisco area for patients covered by Medi-Cal government-sponsored health insurance and those who are homeless, uninsured, and/or immigrants.

Table 1 summarizes criteria for classifying adherence to recommended follow-up patterns at SFGH based on disease severity. Classification according to these guidelines was based on data gathered from medical records. Eligibility criteria included having a medical record documenting the dates of all glaucoma follow-up visits scheduled and attended in the past 12 months, age of 40 years or older, and 1 of the following diagnoses made at SFGH: primary open-angle glaucoma, primary angle-closure glaucoma, exfoliative glaucoma, pigmentary glaucoma, low-tension glaucoma, or ocular hypertension. These selection criteria were designed to exclude patients who would not require ongoing follow-up at the SFGH Glaucoma Clinic, such as those seeking a second opinion or for urgent care. The chief of the SFGH Glaucoma Service evaluated disease severity (ie, mild, moderate, or severe) according to the American Academy of Ophthalmology Preferred Practice Patterns guidelines for primary open-angle glaucoma based on visual field testing, cup-to-disc ratios, and applanation tonometry.¹⁰

All study participants were interviewed in their preferred language (ie, English, Spanish, Mandarin, Cantonese, Vietnamese, or Tagalog) by a trained member of the multilingual research team. Data were collected by oral questionnaire regarding patient demographics, perceived barriers to follow-up, and reasons for nonattendance. Patients self-reported their racial/ethnic classifications. In the case of multiethnic individuals, patients were asked with which classification they most identified. Oral informed consent was obtained from all patients before the interview. Questions were standardized across languages. The questions were based on a questionnaire used at the Aravind Eye Hospital Glaucoma Clinic in Coimbatore, Tamil Nadu, India, and adapted to the present study population.¹² The questionnaire was validated in a pilot study with a randomly selected cohort of 14 patients who met the aforementioned eligibility criteria. Questions were adapted to the SFGH patient population, response coding classifications were calibrated, and protocol feasibility was tested. The participants in the pilot study were not included among the 152 cases and controls who comprised the study group. On the basis of pilot study results, it was determined that 150 patients would be adequate to identify predictors of inconsistent adherence with an odds ratio (OR) of 2 or greater with a power of 80% and an α of .05. Human subjects approval for this study was obtained from the institutional review boards of the following organizations: SFGH; University of California, San Francisco; and the Stanford University School of Medicine. The study followed the tenets of the Treaty of Helsinki.

Statistical analysis was conducted using IBM SPSS Statistics statistical software, version 18.0 (SPSS Inc, Chicago, Illinois). Proportions were compared using the t and χ² tests, and adjusted ORs and 95% confidence intervals (CIs) were calculated using a stepwise multivariate logistic regression model. Variables with P <.20 were initially included in the multivariate regression model with age and sex then successively eliminated based on higher P values. All variables except age and sex had P <.05 in the final regression model.

Characteristics of the patients with inconsistent follow-up (cases) and consistent follow-up (controls) are summarized in Table 2. Of the 186 patients recruited for the study, 14 were involved in the pilot study, an additional 16 were unreachable, and 4 declined to participate, citing time constraints. A total of 152 patients completed the oral questionnaire in a 1:1 matched ratio and were included in the final analysis.

The follow-up patterns of the 152 study patients varied across ethnic groups (Table 2). The number (percentage) of black, Latino, Asian, and white patients classified as cases was 21 (27.6%), 31 (40.8%), 18 (23.7%), and 6 (7.9%), respectively. The proportion of cases within each race/ethnicity was 61.8% black, 54.4% Latino, 42.9% Asian, and 31.6% white.

Table 3 shows the univariate analysis for the potential predictors of inconsistent follow-up. A stepwise multivariate logistic regression model identified independent predictors of inconsistent follow-up after adjusting for covariates (ie, age, sex, disease severity, employment status, marital status, and health insurance coverage status). These predictors included black race, Latino ethnicity, having little or no recollection of receiving glaucoma-related education from clinic staff, receiving information regarding glaucoma from family members, being unaware of the risk of irreversible blindness due to glaucoma and the need for lifelong treatment, and perceiving follow-up visits not to be important if patients were regularly using medications and did not notice any visual symptoms (Table 4).

Black race and Latino ethnicity were strong predictors of inconsistent follow-up even after adjustment for age and sex (OR, 7.16 [95% CI, 1.64-31.24]; P = .009; and 4.77 [1.12-20.29]; P = .04, respectively). A breakdown of cases and controls by race/ethnicity and educational background (used as a proxy for socioeconomic status) is given in Table 5. Latino patients having a low level of education (ie, no formal education through completion of primary school) and black patients with a medium level of education (ie, any secondary school through completion of a secondary school degree or a general educational development certificate) were more likely to have inconsistent follow-up patterns compared with their white counterparts with similar educational backgrounds (P = .04 for each). Although not statistically significant, a trend was observed of black patients with high educational level (ie, having any undergraduate university or community college coursework or beyond) and Asian/Pacific Islander patients of low educational background demonstrating consistent follow-up patterns (P = .07 and P = .08, respectively).

Factors not found to be predictive of follow-up patterns included employment status, marital status, health insurance coverage status, self-reported ability to pay for medications, history of laser treatment for glaucoma, and self-reported inconvenience of transportation to clinic (Table 3 and Table 4).

Disease severity at the time of diagnosis and rate of disease progression are well-established risk factors for vision loss in patients with glaucoma.^3,4 Recently, much attention has been given to nonadherence to prescribed glaucoma medication regimens as an important predictor of adverse outcomes.^11,21 Remarkably little attention, however, has been given to inconsistent adherence to recommended follow-up visits as a predictive factor for vision loss. The physician treating glaucoma has many treatment options, including medications, laser trabeculoplasty, and incisional glaucoma surgery, all of which can lower IOP. Insufficient IOP lowering with 1 approach because of lack of efficacy of or inconsistent adherence to glaucoma therapy generally results in advancement to the next step in this treatment algorithm. Such assessment and advancement of therapy can only take place, however, if patients are seen on a timely basis for follow-up. Although high-quality confirmatory evidence is lacking, one can hypothesize that inconsistent follow-up is a risk factor for inconsistent outcomes regarding any chronic degenerative disease for which multiple potentially effective therapeutic options exist.

This study demonstrates that race/ethnicity and educational background are important factors that affect the follow-up patterns of patients with glaucoma in an ethnically diverse, resource-limited county hospital population. Black and Latino patients were found, on average, to show less consistent follow-up relative to whites. The proportion of black patients with inconsistent follow-up (61.8%) was approximately double that observed among white patients (31.6%), resulting in a calculated OR of 7.16 (P = .009) even after adjustment for covariates (ie, age, sex, disease severity, employment status, marital status, and health insurance coverage status). Latino patients had a calculated OR of 4.77 (P = .04) compared with their white counterparts after similar covariate adjustment. These results are consistent with the trends of lower follow-up and screening rates observed in black and Latino patients with breast and cervical cancer compared with their white counterparts.^22,23

Patients' understanding of glaucoma disease mechanisms greatly influences their adherence to recommended follow-up visits. In particular, those unaware of the chronic and insidious nature of glaucoma and the need for permanent treatment despite regular use of prescribed medications and lack of symptoms were more likely to have inconsistent follow-up in this study. Uncertainty regarding these issues was highly correlated with patient perceptions regarding insufficient glaucoma-related education from clinic staff. This finding dovetails with the observation in the breast cancer literature that misconceptions regarding breast lumps and appropriate follow-up care were more common among women of ethnic minority backgrounds, those with lower socioeconomic status, and those with less access to health care.²⁴

The identification of modifiable risk factors for inconsistent follow-up may allow assessment of targeted solutions, such as increasing physician-initiated patient education in an effort to eliminate the knowledge gap and to decrease inconsistent adherence to follow-up. Given our results, targeted interventions aimed at patients with multiple predictors of inconsistent follow-up, such as minority race/ethnicity, low educational background, and those with limited knowledge of glaucoma, have the greatest potential of increasing rates of glaucoma disease follow-up. Additional larger studies in the future can help refine recommendations regarding groups to target for improvement in follow-up.

Several prior studies^11,12 have reported that mild glaucomatous disease and lack of glaucoma symptoms are predictors of inconsistent adherence to follow-up. In contrast, individuals with severe disease and those with ocular pain and/or vision changes at the time of their initial glaucoma diagnosis were more likely to show inconsistent adherence to recommended follow-up in this study. Although causality is difficult to assess in such a study, a possible explanation for these findings is that inconsistent adherence to appointments in the past led to greater disease severity and also predicted future follow-up behavior. This hypothesis is supported by reports from the Institute of Medicine^25-27 and the US Department of Health and Human Services²⁸ documenting that racial/ethnic minorities and patients of lower socioeconomic status are more likely to have limited access to care, to have frequently missed opportunities for preventive care, and to report insufficient patient-physician communication. These trends are believed to contribute to the high rates of late-stage breast, endometrial, cervical, and colorectal cancer diagnoses in those patients.^25-28 Whether similar patterns are evident in patients with glaucoma, thus allowing for greater disease progression and adverse symptoms at the initial office visit, is an area for future investigation.

This study suggests that blacks and Latinos, 2 racial/ethnic groups previously identified as being at high risk for glaucomatous disease, may benefit from specialized education and counseling to improve follow-up. Targeted physician-initiated educational interventions might primarily focus on 3 variables: the risk of irreversible blindness from glaucoma, the long-term nature of treatment, and the importance of attending follow-up visits despite the ongoing use of medications and lack of visual symptoms. The effect of such interventions should be assessed, with the primary outcome being visual preservation. Glaucoma education by physicians may be important in correcting the pervasive lack of awareness regarding the insidious nature of disease progression, which is predictive of inconsistent follow-up. Our work suggests that advocating for patient-centered education is important when caring for a multilingual patient population in a resource-limited setting. With health system reform in full swing, pressure is growing to deliver high-quality, cost-effective care. An understanding of the population-specific pattern of each practice is likely a necessary prerequisite to providing high-yield interventions in an effort to prevent avoidable blindness.

One limitation of this study is the small number of patients classified as white cases. Despite the aim of the study to interview more patients in this reference category, limited numbers of white patients with inconsistent follow-up were available. Such a small sample size in 1 group is clearly suboptimal for many reasons, including the resultant wide CIs, which limit the race/ethnicity analysis. Future studies will aim for larger sample sizes to address this issue.

Generalizing the current findings to other practices must be done with caution given site-specific variability in demographics, patient counseling, and availability of care. However, this study demonstrates the adaptability of our research method, originally applied at Aravind Eye Hospital in India, to other geographic locations. We believe this survey will be a useful tool in determining predictors of and barriers associated with inconsistent patient adherence to glaucoma follow-up visits in a variety of settings. The application of these findings in examining the influence of educational interventions on adherence to follow-up should be the focus of future studies.

Correspondence: Shan C. Lin, MD, University of California, San Francisco, 10 Koret Way, Room K-325, San Francisco, CA 94143-0730 (LinS@vision.ucsf.edu).

Submitted for Publication: April 20, 2010; final revision received October 5, 2010; accepted October 11, 2010.

Additional Contributions: We thank Rita Popat, PhD, at Stanford University; Bennie Jeng, MD, and Dandan Wang, MD, at the University of California, San Francisco; and the San Francisco General Hospital staff for their invaluable help with this project.