Diabetic retinopathy (DR) is the leading cause of preventable blindness in working age adults.1 Twenty-nine percent of adult patients with diabetes in the United States have DR on fundus exam, and the incidence of DR is set to increase with growing diabetic burden.2 Projections for DR in 2050 will reach 16 million adults older than 40 years, of whom 3.4 million will have vision-threatening retinopathy.3 DR represents an additional heavy cost burden in the United States, with an estimated cost of $8,021 per patient and additional burden from employee absenteeism.4
Despite the potential detriment to quality of life, many patients with diabetes do not receive appropriate care to prevent visual impairment. A national cross-sectional study of diabetics with macular edema found that only 60% of participants received an eye examination with pupillary dilation in the prior year, and less than half had visited a diabetic nutritionist or dietician.5 Close follow-up and tight glycemic and blood pressure control can reduce the incidence and progression of DR.6 Prompt treatment with laser photocoagulation reduces the risk of visual loss by 50%, and early vitrectomy can improve visual recovery.6 Thus, better understanding of the reasons behind poor follow-up for diabetic eye care will help inform policies, which can help prevent blindness in this vulnerable population.
Multiple studies have examined the factors underlying non-compliance from DR screening and reported low levels of follow-up.7–12 Factors involved in noncompliance for screening of DR and follow-up of diabetic care are multifaceted, including lack of knowledge about diabetes, economic burden, and access to care.10,13 Access to care can also impact diagnosis, with diabetics living in ophthalmologist-scarce counties more likely to have vision-threatening retinopathy.14
A handful of studies have compared follow-up after established care of DR.7,15 The most encompassing study, the Diabetic Retinopathy Barometer Study, surveyed awareness, treatment, and prevention of diabetic eye disease in patients and providers from eight countries.16 The study found limitations in access including lack of education and lack of availability of screening programs. Comparing barriers to care could be useful in informing policies for improving compliance for DR. In this study, we aim to investigate patient-reported barriers to follow-up in a DR clinic in an urban safety-net hospital serving an ethnically diverse population.
Patients and Methods
Prospective Survey in Retina Clinic
Patients treated at Zuckerberg San Francisco General Hospital retina clinic, a safety-net metropolitan county hospital serving an ethnically diverse population, were prospectively enrolled in this study from July 2015 to January 2016. Inclusion criteria for the study were age of 18 years or older, a diagnosis of DR or macular edema, and follow-up of at least 1 year. Institutional review board approval from the Human Research Protection Program at the University of California in San Francisco was obtained for this study, and this study adhered to the tenets of the Declaration of Helsinki. The study was explained to patients who were willing to participate.
Patients were consented, enrolled, and orally interviewed by a trained member of the multilingual provider team in their preferred language (eg, Chinese, English, and Spanish). This survey was adapted from a previously validated glaucoma survey conducted at the same site.19 The survey encompassed 21 questions regarding demographic data (gender, age, race, language), housing, health insurance, employment, education, travel to clinic, and diabetes history and care (Figure A). Available clinical and demographic data were verified through retrospective chart review. Self-reported patient knowledge of DR, perceptions of clinical care, selection of any of 16 barriers to care, and suggestions for improvement were also recorded. Patients were scheduled for follow-up according to American Academy of Ophthalmology preferred-practice guidelines or at physician discretion. Non-compliance was defined as failure to attend 80% or more of scheduled appointments or if they failed to reschedule missed or patient-cancelled appointments within 1 month of the desired follow-up with reminder call — similar to previous studies of compliance.15,18
Survey data was stored in REDCap, a secure HIPAA-compliant database for analysis. Detailed survey methodology and preliminary findings on clinical factors associated with rates of non-compliance are currently in submission.19 Descriptive statistics, such as frequencies and percentages, were used to summarize patient demographics and patient-reported barriers to follow-up. Univariate and multivariate logistic analyses were used to identify barriers independently associated with non-compliance to follow-up. Logistic analysis was also used to identify demographic factors associated with each reported barrier. Chi-squared test was conducted for patient strategies to follow-up and transport questions. Significance was set at P value less than .05. Data were analyzed using Stata SE (StataCorp, College Station, TX).
Out of 232 eligible patients seen in the DR clinic, 209 patients participated in the questionnaire. The mean age was 58.2 years (standard deviation [SD] ± 9.7). Sixty-three percent (132 of 209) of participants were male. The mean hemoglobin A1c was 8.5% (SD ± 0.14). Among the patients, 15.8% had mild nonproliferative DR, 17.7% had moderate nonproliferative DR, 7.7% had severe nonproliferative DR, 55.5% had proliferative DR (PDR), and 3.3% had no retinopathy. The racial makeup of the study was 42% Hispanic/Latino, 32% Asian, 17% White/Caucasian, and 10% African-American/Black.
Systematic and Patient Barriers to Follow-Up
We first analyzed self-reported barriers to follow-up across all patients in our study cohort. The most common self-reported barrier to follow-up cited by patients was long waiting times in clinic (46.4% [97 of 209]), followed by other medical/physical conditions (35.9% [75 of 209]), forgot to come to clinic (28.2% [59 of 209]), and inability to leave work (9.1% [19 of 209]) (Figure 1).
Self-reported patient barriers to follow-up. Barriers not included had a less-than-5% response rate (ie, “having the office reschedule my appointments,” “lost wages,” “other serious personal issues,” etc.)
We next determined which self-reported factors were significantly associated with poor compliance to follow-up as defined by our pre-specified criteria. In univariate models, barriers significantly associated with non-compliance included “other medical or physical condition” (odds ratio: [OR] 1.99; 95% confidence interval [CI], 1.12–3.53) and “forgot to come” (OR: 3.82; 95% CI, 2.00–7.29). In multivariate analysis, both of these barriers were independent risk factors for non-compliance after adjusting for age, gender, ethnicity, insurance type, and severity of diabetic retinopathy (Table 1).
Odds Associated With Non-Compliance for Each Reported Barrier
We also wished to determine which self-reported barriers for follow-up were associated with severe vision-threatening retinopathy by assessing the association of PDR with self-reported barriers to follow-up. PDR was associated with reporting the barrier of “other medical or physical condition” in both univariate (OR: 4.60; 95% CI, 1.79–11.8) and multivariate analysis (OR: 4.35; 95% CI, 1.53–12.4). Having private insurance was associated with reporting “forgot to come” as a barrier to follow-up in univariate analysis (OR: 10.9; 95% CI, 1.12–105.7). Of note, only five patients with private insurance were included in our study. No other patient characteristics (ie, age, gender, language, insurance type) were associated with the remaining barriers in logistic analyses.
Ethnic and Racial Disparities Within Barriers to Care
We next wished to compare whether there was a significant association between race and self-reported barriers to follow-up. When comparing cited barriers by race, the proportion of Black/African-Americans who cited “forgot to come” (52% [11 of 21]) was significantly higher than the proportion of White/Caucasians (14% [five of 35]; χ2 = 9.33; P < .01) (Figure 2). There were no differences in barrier selection by ethnicity (as defined by primary language).
Top barriers by race. NS = not significant, * = proportion of respondents significantly different from proportion of White / Caucasian respondents.
Patient Feedback and Transportation
We included a section on the survey to ask for patient feedback on ways to improve appointment follow-up. Patient suggestions for facilitating follow-up were to provide reminders (49.3% [103 of 209]) and provide a mobile van pick-up (43.5% [91 of 209]), with limited need for education outreach (19% [40 of 209]) (Figure 3). Patients cited ophthalmologists and clinic staff as their primary source of knowledge for DR (93.8% [196 of 203]), followed by other sources (13.9% [29 of 203]), media (6.2% [13 of 203]), pamphlets / posters (5.3% [11 of 203]), and family / friends (4.3% [9 of 203]).
Patient suggestions for, “What would help you attend regular exams?”
In addition, we surmised that access to care may be affected by modes of transportation to the clinic. Patients spent on average 29 minutes (SD ± 19 minutes) traveling from home to clinic. Convenience of transport was the only transportation-specific factor that differed significantly between patients who requested a mobile van pick-up and those who did not. Patients requesting mobile van transport were more likely to state transport was “not convenient” or “very inconvenient” (χ2 = 7.43; P < .01).
A key component of prevention of DR progression is close follow-up, glycemic control, and appropriate treatment. Although early identification can prevent irreversible retinal damage, barriers to care can cause patients to miss crucial components of DR treatment. Within a diverse urban clinic population, our study found that patient barriers such as other medical and physical conditions and forgetting the appointment can significantly impact compliance of DR follow-up. Patients with severe DR were more likely to cite other medical/physical conditions as a barrier, which may reflect additional comorbidities associated with advanced diabetes.
Our findings were reflective of similar barriers previously reported in the literature. Barriers to DR care identified in the Diabetic Retinopathy Barometry Study (DRBS) included high cost of care, long waiting times, travel difficulties, and lack of available health services.16 Large-scale studies in Singapore have identified lack of knowledge, economic burden, and higher perceived barriers as associated with severity of DR.20 Factors influencing timely fundus examinations in southern Chinese patients with PDR included lack of knowledge, worries about insulin, and economic burden.13 We hope our study will help clinicians understand and correct the drivers of non-compliance and lead to greater likelihood of receiving eye care.
With an average age of 57 years in our study and consideration of the comorbidities associated with DR, mobility can be a hindering factor in clinic attendance. Patients with more severe DR were more likely to cite physical condition as an obstacle to care, reflecting a particularly vulnerable population in whom follow-up is even more important. These patients likely have advanced diabetes and PDR due to poor glycemic control, and thus medical limitations. In suggestions from patients, we found that clinic reminders and transport to clinic were the most common suggestions for facilitating follow-up. Thompson et al., also found that patients asked for reminders and assistance with parking/transportation.15 Services such as mobile van pick-up can assist particularly patients who otherwise might miss appointments due to physical limitations and transportation difficulties. The SFGH clinic currently offers taxi cab vouchers on an as-needed basis at the discretion of physician or staff.
In this study, 28% of patients cited forgetting as a reason for lack of follow-up. Although the burden of remembering appointment dates is often placed on the patient and family, the clinic plays a crucial role in ensuring patients are reminded. During this study, the SFGH clinic provided reminders of appointments by mail. Implementing changes such as phone call reminders of appointments could reduce no-shows from forgotten appointments.
On the other hand, patients rated the need for additional retinopathy education as minimal in our study. A national Canadian study found that a discussion of diabetes with health professionals increased eye screening.21 Although we did not survey ophthalmologists in our in study, Harnett et al. have demonstrated a gap in patient- and physician-perceived barriers to care of DR.9 Patients were most likely to cite finances as a barrier, whereas ophthalmologists most commonly cited need for patient education. Similarly, the DRBS noted lack of patient knowledge of DR as the most common barrier to optimal eye health per providers.16 This discrepancy may result from patients lacking an understanding of the consequences of going without treatment or follow-up. In our study, 94% of patients stated their source of DR knowledge came from ophthalmologist and clinic staff. Facilitating a discussion on the importance of continued care to preserving vision may stress to patients the risks of forgetting an appointment.
Often-dismissed systemic barriers such as long waiting times can have a disproportionate impact on access to eye care, with a linear decline in patient satisfaction regardless of cost of care.22 Although we do not track waiting times in clinic, anecdotal conversation with clinic staff and consultation with electronic medical records schedules found that wait times of 30 minutes to an hour to see a provider were common within our clinic. Thompson et al. similarly surveyed glaucoma, macular degeneration, and DR in a suburban, highly educated population and found long waiting times as the most common barrier, followed by rescheduling difficulties.15 External financial pressures, such as overbooking to meet clinic quotas or preempt cancellations, may be driving extended wait times. Improvements can be derived from wider distribution of appointments, rearrangement of new and follow-up slots, and dilation-free examination.23
We did not observe an association in language / ethnicity and follow-up or specific barriers, although one geographically proximate study with a large Asian population found low follow-up in this group.15 However, we did note that African-American patients were most likely to cite forgetting to come as a barrier. Possible confounders may include household income or background, which were not measured in this study. High rates of non-compliance have been reported in the literature, with one-half of patients in a majority African-American study population in Alabama who did not follow-up with eye care despite recommendations from DR screening.8 In that study, older age and knowing one's glycated hemoglobin were associated with adherence; whereas agreeing to assistance in making a follow-up appointment was associated with nonadherence.
A potential weakness of this study is that we did not track non-compliance and its impact on disease progression and clinical outcome. The vulnerabilities we identified in this study may not translate to clinical retinopathy progression and vision loss. Further, our results are not generalizable to the entire diabetic population as we examined a safety-net hospital population with high Hispanic population and geographic limitation. Finally, other factors impacting compliance such as weather and financial constraints were not examined.
Implementation of further action is likely necessary to raise participation rates, starting with appointment reminders, streamlined scheduling and assistance with transport. Prior studies on enhancing DR follow-up have attempted patient contracts and phone call follow-up with mixed success.7,24 Multiple reminders have been found to increase DR exam adherence in diabetics.25 A pilot study that incorporated visual feedback of patient's own retina images was found to improve Hgb A1c levels 3 months later.26 In our study, patients with PDR most at risk of vision loss struggle with comorbidities that limit their ability to attend clinics. We recommend when scheduling a patient's follow-up to ensure patients have access to appropriate transportation. Second, to raise compliance rates, implementing an appointment reminder system for adequate follow-up is necessary. This should also accompany a conscious education of patients on consequences of DR progression and the importance of follow-up. Finally, our findings also suggest a heavy burden on health care providers to ensure efficient clinic times and examine their patient flow to reduce no-show rates from long waiting times.
Continuity and follow-up of DR can be impeded by both systemic and patient-centric barriers. Long waiting times, comorbidities, and transportation needs impact follow-up of particularly vulnerable patients with retinopathy in an urban clinic. Debilitation and vision loss from DR remain a pressing concern, and adequate measures to address follow-up will be needed as demand for retina services continues to grow. Incorporating appointment reminders and mobile van pick-up into DR clinic may address these vulnerabilities for certain diverse urban populations.