Abstract
Purpose:
To assess the proportion of newly diagnosed cases of PACG with and without prior diagnosis of anatomical narrow angles (ANA) and identify sociodemographic risk factors for late detection (PACG without prior ANA diagnosis).
Design:
Retrospective cohort study.
Methods:
102,617 patients with PACG were identified from the Optum Clinformatics® Data Mart Database (2007-2019). Newly diagnosed PACG met the following criteria: 1) diagnosis by an ophthalmologist, 2) observable for at least 12 months prior to diagnosis, 3) no history of treatment before diagnosis unless preceded by a diagnosis of ANA. Multivariable logistic regression modeling was performed to identify sociodemographic risk factors for late detection.
Main Outcome Measures:
Proportion of newly diagnosed PACG without prior ANA diagnosis and sociodemographic factors associated with late detection.
Results:
There were 31,044 eligible patients. Over 70% of PACG was detected without prior ANA diagnosis, regardless of age, sex, and race. Odds of late detection was significantly higher (p<0.001) among males (OR=1.32, [1.25-1.40]), Blacks (OR=1.25, [1.15-1.37]), and patients aged 80 years or older (OR=1.28, [1.11-1.47]), or living in Southern (OR=1.30, [1.22-1.40]) or Pacific (OR=1.27, [1.16-1.36]) regions. Findings were similar for cases of PACG with record of gonioscopy and treatment or with a 24-month lookback period.
Conclusion:
Most patients newly diagnosed with PACG in the United States do not have a prior diagnosis of ANA. The elderly, males, and Blacks are at higher risk of late detection. There exists a need for increased disease awareness among providers and more accessible tools to detect patients at risk for PACG.
Precis
More than 70% of PACG in the United States is newly diagnosed without prior diagnosis of anatomical narrow angles. The elderly, males, Blacks, and patients in Southern/Pacific regions are more vulnerable to late ANA detection.
Introduction
Primary angle closure glaucoma (PACG) is a leading cause of permanent vision loss worldwide, currently affecting approximately 20 million people1; 2 Studies estimate that there are approximately a half-million people with PACG in the United States. However, the burden of PACG on patient populations and healthcare systems in the United States is relatively understudied compared to other regions of the world where the disease is endemic. The perceived low impact of PACG in the United States may be partially attributable to population-based epidemiological studies that report low PACG prevalence among Blacks and non-Hispanic Whites.3; 4 However, gonioscopy was performed selectively on participants in these studies, raising the possibility that PACG was misclassified as primary open angle glaucoma (POAG) and PACG prevalence was underestimated.3; 4 Appropriate awareness about early detection of patients at risk for PACG remains crucial as PACG carries a three-fold increased risk for severe bilateral visual impairment compared to POAG1. The disease is also associated with severe ocular morbidity and high rates of unilateral blindness on diagnosis, even in the United States.5-8
Anatomical narrow angles (ANA) can lead to angle closure, characterized by apposition between the iris and trabecular meshwork, which impedes aqueous outflow and cause elevated intraocular pressure (IOP), an important risk factor for glaucomatous optic neuropathy. When detected early, ANA with elevated IOP can be effectively treated with IOP-lowering medications, laser peripheral iridotomy (LPI), and lens extraction surgery, thereby reducing the risk of developing glaucoma.9 However, there is evidence that gonioscopy is under-performed by eyecare providers in the United States despite guidelines by the American Academy of Ophthalmology (AAO) on ANA screening.10; 11 Aside from low utilization of gonioscopy, there is sparse information about other factors that delay the detection of ANA until after glaucomatous damage has occurred.
In this study, we use national healthcare claims data to examine cases of newly diagnosed PACG in the United States with and without a prior diagnosis of ANA to assess the general effectiveness of healthcare systems for detecting at-risk patients with ANA prior to development of PACG. While the United States has the highest total health expenditure per capita of any country, little is known about its benefit for detecting, monitoring, and treating patients with ANA before development of PACG.12 There is sparse evidence that the United States has substantially lower race-specific PACG prevalence compared to other countries. Our aim is to assess the proportion of newly diagnosed PACG with and without prior ANA diagnosis and identify sociodemographic risk factors for late detection (PACG without prior ANA diagnosis).
Methods
Data
Optum’s Clinformatics® Data Mart (CDM) is derived from a database of administrative health claims data warehouse of commercial and Medicare Advantage health claims. The database includes approximately 17 to 19 million annual covered lives, for a total of over 65 million unique lives over a 13-year period (1/2007 through 12/2019). Available clinical data included first dates of ANA and PACG diagnoses, healthcare provider type, IOP-lowering medications, in-office procedures including gonioscopy, cataract and glaucoma surgeries, and first/last date of enrollment. Available sociodemographic data included sex, race, education level, income, net worth, insurance product, and census division regions. Race and ethnicity are combined in a single variable in the Optum database based on proprietary algorithms relying on the policyholder’s zip code in combination with the individual’s first, middle, and last names (E-Tech 7.3, Ethnic Technologies, Hackensack, NJ), public records, and self-reported surveys. The University of Southern California Institutional Review Board determined that this study was exempt from IRB approval. The study adhered to the tenets of the Declaration of Helsinki and complied with the Health Insurance Portability and Accountability Act.
Study Population Selection and Definitions
Inclusion in the study population required an index diagnosis of PACG based on International Classification of Diseases Ninth Revision (ICD-9) or Tenth Revision (ICD-10) codes (Supplementary Table 1). The index date of diagnosis was defined as the date of the first claim associated with a PACG diagnosis. Inclusion also required continuous enrollment and observability, calculated as the number of months between the first date of enrollment minus the index date of PACG diagnosis, for at least 12 months prior to the index date of diagnosis.
Newly diagnosed PACG cases met the following criteria: (1) PACG diagnosis provided by an ophthalmologist, (2) continuous enrollment and observability during a 12-month lookback period, (3) no history of drops, laser peripheral iridotomy (LPI), or cataract and glaucoma surgery based on Current Procedural Terminology (CPT) codes (Supplementary Table 1) before the index date of diagnosis unless preceded by a diagnosis of ANA. Provider type for criterion 1 was established based on provider type codes in the Optum database. The study was limited to PACG diagnosed by ophthalmologists to increase the fidelity of PACG diagnosis codes. Criterion 2 was implemented to establish cases as newly diagnosed, based on the standard-of-care practice of monitoring established patients with PACG at least once per year. This criterion also minimizes overestimation of newly diagnosed PACG without prior ANA diagnosis by limiting the misclassification of patients with previously diagnosed PACG who are newly enrolled in Optum. This criterion minimizes the mislabeling of patients diagnosed with PACG prior to enrollment as newly diagnosed PACG. Criterion 3 was required to ensure: 1) patients who received conventional ANA treatments without the diagnosis of ANA would not be designated as newly diagnosed PACG, 2) patients diagnosed and treated for ANA before being diagnosed with PACG would not be excluded.
Late detection was defined as newly diagnosed PACG without a prior diagnosis ANA based on ICD-9 or ICD-10 codes (Supplementary Table 1) by any type of healthcare provider. Provider type was relaxed for ANA compared to PACG as ANA is commonly detected by optometrists and limiting diagnoses to only ophthalmologists could overestimate late detection. The proportion of PACG cases without prior ANA diagnosis among all PACG cases was calculated and stratified by age and sex or age and race.
Statistical Analysis
Continuous data were expressed as means and standard deviations. Categorical data were expressed in proportions and percentages. Univariable and multivariable logistic regression analyses were performed to determine odds ratios (OR) for late detection. Multivariable logistic regression analyses controlled for age, sex, race, education, net worth, insurance product, geographic location, length of total observability, and record of gonioscopy on or prior to the index date of PACG diagnosis. Income was excluded from multivariable models due to collinearity with net worth. Adjusted odds ratios were considered significant if they indicated at least a 25% increase (OR ≥ 1.25) or decrease (OR ≤ 0.80) in relative odds compared with reference and were associated with p-value < 0.01. All analyses were conducted using R 4.0.3 or on Unix workstations running SAS 9.4 (SAS Institute; Cary, NC).
Sensitivity Analyses
Two sensitivity analyses were performed using stricter definitions of PACG to validate and assess the dependence of study findings on disease definitions. The first sensitivity analysis (A) was performed on patients with definite PACG, defined as patients with newly diagnosed PACG who also: 1) received gonioscopy before or on the date of PACG diagnosis; 2) received treatment with drops, laser, or surgery within one year of PACG diagnosis. These conditions increase the fidelity of PACG diagnoses as gonioscopy is the clinical standard for detecting ANA and treatment initiated within a relatively short period of time after diagnosis increases the validity of glaucoma diagnoses. The second sensitivity analysis (B) was performed on patients who fit the definition of newly diagnosed PACG with a lookback period of 24 months, instead of 12 months. This condition increases the certainty that cases of PACG were newly diagnosed and that the proportion of PACG without prior ANA diagnosis was not overestimated.
Results
A total of 102,617 unique patients with the diagnosis of PACG appeared from 2007 up to 2019 (Figure 1). There were 31,044 patients with newly diagnosed PACG in the primary analysis after patients were excluded based on provider type (38,657), a 12-month lookback period (29,195), and treatment history (3,721). There were 13,952 unique patients with definite PACG in sensitivity analysis A after patients were excluded based on absence of gonioscopy (11,525) and treatment following diagnosis (5,567). There were 20,363 unique patients with newly diagnosed PACG in sensitivity analysis B after patients were excluded based on a 24-month lookback period (40,687) and treatment history (2,910).
Figure 1.
Attrition diagram of patients with diagnosis of PACG.
The proportion of PACG without prior ANA diagnosis among all PACG was 75.2% overall across all ages, sexes, and races. The proportion of PACG without prior ANA diagnosis exceeded 70% (range 72.0% to 91.5%) for males, females, and overall when stratified by age (Table 1). The proportion was higher among males (78.2%) compared to females (73.4%). The proportion approximated or exceeded 70% (range 70.1% to 90.7%) for all races when stratified by age (Table 2). The proportion was highest among Blacks (79.3%) followed by non-Hispanic Whites (75.0%), Asians (73.3%), and Hispanics (72.1%).
Table 1.
Proportion of PACG without prior ANA diagnosis stratified by sex and age.
| Age | Male | Female | Overall | ||||||
|---|---|---|---|---|---|---|---|---|---|
| N | n | % | N | n | % | N | n | % | |
| <40 | 317 | 290 | 91.48 | 394 | 334 | 84.77 | 711 | 624 | 87.76 |
| 40-49 | 573 | 467 | 81.50 | 1181 | 854 | 72.31 | 1754 | 1321 | 75.31 |
| 50-59 | 1567 | 1257 | 80.22 | 3018 | 2187 | 72.47 | 4586 | 3445 | 75.12 |
| 60-69 | 3053 | 2375 | 77.79 | 5484 | 3948 | 71.99 | 8540 | 6325 | 74.06 |
| 70-79 | 4036 | 3048 | 75.52 | 6373 | 4586 | 71.96 | 10409 | 7634 | 73.34 |
| 80+ | 1885 | 1504 | 79.79 | 3158 | 2477 | 78.44 | 5044 | 3982 | 78.95 |
| Overall | 11431 | 8941 | 78.22 | 19608 | 14386 | 73.37 | 31044 | 23331 | 75.15 |
N = newly-diagnosed PACG; n = late-detected PACG
Table 2.
Proportion of PACG without prior ANA diagnosis stratified by race and age.
| Age | Non-Hispanic White | Asian | Black | Hispanic | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| N | n | % | N | n | % | N | n | % | N | n | % | |
| <40 | 391 | 346 | 88.49 | 70 | 58 | 82.86 | 80 | 71 | 88.75 | 97 | 88 | 90.72 |
| 40-49 | 1060 | 778 | 73.40 | 167 | 125 | 74.85 | 160 | 132 | 82.50 | 241 | 180 | 74.69 |
| 50-59 | 2751 | 2034 | 73.94 | 317 | 235 | 74.13 | 550 | 449 | 81.64 | 623 | 446 | 71.59 |
| 60-69 | 4929 | 3649 | 74.03 | 680 | 485 | 71.32 | 1104 | 871 | 78.89 | 1043 | 731 | 70.09 |
| 70-79 | 5913 | 4332 | 73.26 | 835 | 607 | 72.69 | 1393 | 1081 | 77.60 | 1387 | 976 | 70.37 |
| 80+ | 3181 | 2535 | 79.69 | 305 | 231 | 75.74 | 643 | 512 | 79.63 | 561 | 427 | 76.11 |
| Overall | 18225 | 13674 | 75.03 | 2374 | 1741 | 73.34 | 3930 | 3116 | 79.29 | 3952 | 2848 | 72.06 |
N = newly-diagnosed PACG; n = late-detected PACG
On multivariable analysis (Table 3), the following parameters were significantly associated with higher odds of late detection (p ≤ 0.001): patients under the age of 40 (OR = 2.08, 95% CI [1.62, 2.70]) or over the age of 80 (OR = 1.28, 95% CI [1.11, 1.47]) compared to patients aged 40 to 59, males (OR = 1.32, 95% CI [1.25, 1.40]) compared to females, Blacks (OR = 1.25, 95% CI [1.15, 1.37]) compared to non-Hispanic Whites, patients with an exclusive provider organization (EPO) (OR = 1.38, 95% CI [1.22, 1.56]) compared to a health maintenance organization (HMO) insurance product, patients living in Southern (OR = 1.30, 95% CI [1.22, 1.40]) or Pacific regions (OR = 1.27, 95% CI [1.16, 1.36]) compared to the Northeast region, and patients without a prior record of gonioscopy (OR = 2.50, 95% CI [2.38, 2.70]).
Table 3.
Univariable and multivariable analysis of risk factors for PACG without prior ANA diagnosis.
| Prior ANA Diagnosis |
No Prior ANA Diagnosis |
Univariable Analysis | Multivariable Analysis | |||
|---|---|---|---|---|---|---|
| N (%) | N (%) | OR (CI) | P value | OR (CI) | P value | |
| Age [years] | ||||||
| <40 | 87 (12.2) | 624 (87.8) | 2.35 (1.84-3.03) | <0.001 | 2.08 (1.62-2.70) | <0.001 |
| 40-59 | 433 (24.7) | 1321 (75.3) | REF | REF | ||
| 50-59 | 1141 (24.9) | 3445 (75.1) | 0.99 (0.87-1.12) | 0.873 | 0.99 (0.87-1.13) | 0.916 |
| 60-69 | 2215 (25.9) | 6325 (74.1) | 0.94 (0.83-1.05) | 0.275 | 0.89 (0.79-1.01) | 0.077 |
| 70-79 | 2775 (26.7) | 7634 (73.3) | 0.90 (0.80-1.01) | 0.083 | 0.90 (0.79-1.02) | 0.105 |
| 80+ | 1062 (21.1) | 3982 (78.9) | 1.23 (1.08-1.40) | 0.002 | 1.28 (1.11-1.47) | 0.001 |
| Sex | ||||||
| Female | 5222 (26.6) | 14386 (73.4) | REF | REF | ||
| Male | 2490 (21.8) | 8941 (78.2) | 1.30 (1.23-1.38) | <0.001 | 1.32 (1.25-1.40) | <0.001 |
| Race | ||||||
| Non-Hispanic White | 4551 (25.0) | 13674 (75.0) | REF | REF | ||
| Asian | 633 (26.7) | 1741 (73.3) | 0.92 (0.83-1.01) | 0.074 | 0.96 (0.87-1.07) | 0.477 |
| Black | 814 (20.7) | 3116 (79.3) | 1.27 (1.17-1.39) | <0.001 | 1.25 (1.15-1.37) | <0.001 |
| Hispanic | 1107 (28.0) | 2845 (72.0) | 0.86 (0.79-0.92) | <0.001 | 0.85 (0.78-0.92) | <0.001 |
| Unknown | 608 (23.7) | 1955 (76.3) | 1.07 (0.97-1.18) | 0.171 | 0.99 (0.85-1.17) | 0.938 |
| Gonioscopy | ||||||
| Yes | 6043 (31.0) | 13476 (69.0) | REF | REF | ||
| No | 1670 (14.5) | 9855 (85.5) | 2.63 (2.50-2.78) | <0.001 | 2.50 (2.38-2.70) | <0.001 |
| Education | ||||||
| High school or less | 2003 (23.5) | 6529 (76.5) | REF | REF | ||
| Some college | 3804 (25.0) | 11410 (75.0) | 0.92 (0.86-0.98) | 0.009 | 0.97 (0.91-1.04) | 0.418 |
| College graduate | 1493 (27.2) | 3990 (72.8) | 0.82 (0.76-0.89) | <0.001 | 0.93 (0.85-1.02) | 0.121 |
| Unknown | 413 (22.8) | 1402 (77.2) | 1.04 (0.92-1.18) | 0.509 | 1.02 (0.82-1.26) | 0.861 |
| Income | ||||||
| <$50k | 2140 (23.5) | 6985 (76.5) | REF | - | ||
| $50-$100K | 2372 (25.5) | 6925 (74.5) | 0.89 (0.84-0.96) | 0.001 | - | |
| >$100K | 1991 (26.7) | 5475 (73.3) | 0.84 (0.79-0.90) | <0.001 | - | |
| Unknown | 1210 (23.5) | 3946 (76.5) | 1.00 (0.92-1.08) | 0.983 | - | |
| Worth | ||||||
| <$149k | 2266 (23.4) | 7438 (76.6) | REF | REF | ||
| $150-499K | 2002 (24.9) | 6047 (75.1) | 0.92 (0.86-0.99) | 0.018 | 0.95 (0.88-1.02) | 0.173 |
| 500K+ | 2289 (27.3) | 6100 (72.7) | 0.81 (0.76-0.87) | <0.001 | 0.90 (0.83-0.98) | 0.012 |
| Unknown | 1156 (23.6) | 3746 (76.4) | 0.99 (0.91-1.07) | 0.756 | 0.94 (0.85-1.04) | 0.203 |
| Insurance | ||||||
| HMO | 3862 (25.5) | 11268 (74.5) | REF | REF | ||
| PPO | 938 (23.0) | 3137 (77.0) | 1.15 (1.06-1.24) | 0.001 | 1.12 (1.02-1.22) | 0.013 |
| EPO | 402 (21.0) | 1516 (79.0) | 1.29 (1.15-1.45) | <0.001 | 1.38 (1.22-1.56) | <0.001 |
| Other | 2511 (25.3) | 7410 (74.7) | 1.01 (0.95-1.07) | 0.702 | 1.00 (0.93-1.07) | 0.949 |
| Location | ||||||
| Northeast | 2078 (28.3) | 5275 (71.7) | REF | REF | ||
| South | 3108 (22.8) | 10539 (77.2) | 1.34 (1.25-1.42) | <0.001 | 1.30 (1.22-1.40) | <0.001 |
| Midwest | 1348 (24.8) | 4087 (75.2) | 1.19 (1.10-1.29) | <0.001 | 1.13 (1.03-1.23) | 0.007 |
| Mountain | 441 (28.3) | 1116 (71.7) | 1.00 (0.88-1.13) | 0.960 | 0.99 (0.87-1.13) | 0.887 |
| Pacific | 731 (24.2) | 2292 (75.8) | 1.24 (1.12-1.36) | <0.001 | 1.27 (1.16-1.36) | <0.001 |
| Total observability | ||||||
| Mean (SD) [years] | 3.9 (2.5) | 3.2 (2.1) | 0.88 (0.87-0.89) | <0.001 | 0.88 (0.87-0.89) | <0.001 |
Statistically significant p-values and odds ratios are bolded; OR = odds ratio; CI = Confidence Interval; EPO = exclusive provider organization; HMO = health maintenance organization; PPO = preferred provider organization.
The results from sensitivity analysis A (definite PACG) closely resembled results from the primary analysis. The proportion of definite PACG without prior ANA diagnosis was 71.8% overall (Supplementary Table 2). On multivariable logistic regression analysis, the only difference from the primary analysis was that patients with EPO insurance had reduced odds (OR = 1.20, 95% CI [1.01, 1.42]) of late detection (Supplementary Table 3).
The results from sensitivity analysis B (newly diagnosed PACG with 24-month lookback period) closely resembled results from the primary analysis. The proportion of PACG without prior ANA diagnosis was 72.2% overall (Supplementary Table 4). On multivariable logistic regression analysis, the only difference from the primary analysis was that patients over 80 years of age had reduced odds (OR = 1.16, 95% CI [0.98, 1.36]) of late detection (Supplementary Table 5).
Discussion
In this study, we used national healthcare claims data to assess the proportion of newly diagnosed cases of PACG by ophthalmologists in the United States with and without a prior diagnosis of ANA. Overall, the proportion of PACG without prior diagnosis of ANA was high, exceeding 70% of all PACG cases regardless of age, sex, and race. Odds of late detection were higher among males, Blacks, and patients aged 80 years or older, without record of gonioscopy, or living in Southern or Pacific regions. We believe these results provide evidence supporting the need for increased disease awareness among eyecare providers and more accessible tools to detect patients at high risk for PACG.
The high proportion of PACG cases without prior ANA diagnosis (> 70%) across all age, sex, and racial groups suggests that most cases of ANA progressing to PACG remain undetected until after glaucomatous damage has occurred. The importance of this finding is highlighted by the severe ocular morbidity associated with PACG and high rates of unilateral blindness on diagnosis.5-8 The problem posed by late detection is compounded by the fact there are effective treatments to alleviate ANA and reduce risk of disease progression once it is detected.9; 13 One possible explanation for the high proportion of PACG without prior ANA diagnosis is that most high-risk cases of ANA are detected and treated prior to development of elevated IOP and glaucomatous optic neuropathy in the United States. However, there is a paucity of robust epidemiological data on PACG to support this assumption, and the proportion of PACG to POAG among Chinese Americans is similar to or higher than other ethnically Chinese populations.14-17 In addition, it is well-established that the majority of glaucoma in the United States, including PACG, remains undetected, despite screening efforts by eyecare providers and financial expenditures by healthcare systems.10; 18
We observed racial differences in the proportion of PACG without prior ANA diagnosis; specifically, odds of late detection was significantly higher among Blacks compared to other races before and after adjusting for other sociodemographic factors. One interpretation of these findings is that some racial populations are more likely to be evaluated for ANA than others. It is widely accepted that the prevalence and burden of PACG is greatest among Asians.19 Conversely, data from the Salisbury Eye Evaluation Study reported low PACG prevalence of less than 0.1% among Blacks and Whites over 73 years of age, although gonioscopy in this study was performed selectively at the discretion of the examiners.4 Studies of PACG prevalence based on healthcare claims data in the United States support racial differences in PACG prevalence, although these differences may be smaller than previously thought.2; 3; 20-23 Therefore, perceived risk of ANA and PACG by eyecare providers may contribute to the racial bias observed among cases of PACG without prior ANA diagnosis. An alternative explanation is there are racial anatomical differences that influence the likelihood eyecare providers will perform gonisocopy.24 25 Finally, the observed differences may be related to lower utilization of eyecare services among some racial minority groups26. While it is important to reflect on racial differences in the odds of late detection, the high proportion across all races suggests that objective clinical methods that do not rely on perceived disease risk by eyecare providers are needed to improve detection of high-risk ANA prior to development of PACG.
The use of gonioscopy before the date of PACG diagnosis was strongly associated with lower odds of late detection. This is unsurprising given that gonioscopy remains the clinical standard for detecting ANA and distinguishing PACG from POAG. AAO Preferred Practice Patterns recommend the use of bilateral gonioscopy to evaluate angle anatomy in all patients suspected of ANA. However, the CPT code for gonioscopy was only associated with 62.9% of total cases of PACG used in our original analysis. This led us to perform a sensitivity analysis of definite PACG cases, confirmed by record of gonioscopy and treatment (eye drops, LPI, or surgery) initiated after PACG diagnosis, which yielded results similar to the primary analysis. Our finding on the low rates of gonioscopy among newly diagnosed cases of PACG is consistent with previous findings by Coleman et al., who reported gonioscopy screening rates as low as 46% prior to glaucoma surgery performed in the United States, despite AAO recommendations.10 It is important to reiterate that gonioscopy is an effective method for detecting ANA but remains under-performed and/or under-billed despite recommendations by the AAO.
Our findings on regional differences in odds of late detection resemble previously reported findings that both incidence and prevalence of POAG and PACG are highest in the New England and Mid-Atlantic regions, even after controlling for access to care and number of ophthalmologists per capita.27 We observed similar regional differences in our study, with lower odds of late detection in the Northeast (combined New England and Mid-Atlantic) compared to Southern and Pacific regions, even after adjusting for gonioscopy use. We speculate that these regional differences are related to increased awareness about ANA, possibly due to increased exposure to patients with the disease in training or practice. Interestingly, odds of late detection was higher in the Pacific region, despite the higher residential density of Asians in some parts of the region.
Older age and female sex are well-known risk factors for ANA and PACG.20 The increased risk conferred by these two factors is related to progressive narrowing of the anterior chamber angle associated with aging and biometric differences between male and female eyes.24 We observed higher odds of late detection in the 80 years and older group when compared to the 40 to 59 years age groups, despite the known increased risk of PACG in older age groups. This highlights the importance of routine eyecare in aging populations and prompt removal of cataracts when clinically indicated. Additionally, males had higher odds of late detection compared to females, which likely reflects lower frequency of eye exams by males compared to females, resulting in missed screening and diagnosis opportunities.28
Our study has several limitations. First, we relied on broad definitions of ANA and PACG utilized by practicing ophthalmologists, which may not be consistent with specific definitions of primary angle closure disease (PACD) developed for epidemiological studies.29 While a study of the current scope and size is difficult if not impossible with more traditional forms of data, we acknowledge that healthcare claims data comprise a wide range of real-world physician diagnosis and coding patterns. This led us to perform two sensitivity analyses to validate our study definitions and findings. In both analyses, results were consistent with the primary analysis, which supports that our findings are robust to variations in provider- and study-level disease definitions. Second, we were unable to adjust for ophthalmologist-per-capita density when assessing regional data, which may have helped explain regional differences in the proportion of PACG without prior ANA diagnosis. Finally, we imposed a 12-month minimum lookback period on all patients to establish PACG cases as newly diagnosed. As previously mentioned, this minimizes bias related to previously diagnosed PACG being misclassified as newly diagnosed; however, this may also introduce bias as patients who were not receiving care prior to enrollment in Optum could be excluded.
Our study findings suggest that the majority of ANA progressing to PACG in the United States remains undetected until glaucomatous damage has occurred, and that the elderly, males, Blacks, and patients living in certain geographic regions are more vulnerable to late detection. These findings have important implications for eyecare providers and healthcare systems as PACG prevalence is expected to increase worldwide due to aging of the general population.2 Therefore, there is an urgent need to develop more accessible tools for detecting patients at risk for PACG. However, as most patients with angle closure on gonioscopy are at low risk of developing PACG, there is also a need to develop more precise methods to identify high-risk individuals13; 30. These methods could help eyecare providers and healthcare systems reallocate healthcare resources away from screening and surveillance of low-risk cases of ANA toward care and treatment of high-risk cases to reduce PACG prevalence and associated vision loss.
Supplementary Material
Supplementary Table 1. Diagnosis, procedure, and treatment codes.
Supplementary Table 2. Sensitivity analysis A: Proportion of definite PACG without prior ANA diagnosis.
Supplementary Table 3. Sensitivity analysis A: Risk factors for definite PACG without prior ANA diagnosis.
Supplementary Table 4. Sensitivity analysis B: Proportion of PACG without prior ANA diagnosis (24- month lookback period).
Supplementary Table 5. Sensitivity analysis B: Risk factors for PACG without prior ANA diagnosis (24- month lookback period).
Acknowledgements
This work was supported by grants K23 EY029763 from the National Eye Institute, National Institute of Health, Bethesda, Maryland; a Young Clinician Scientist Research Award from the American Glaucoma Society; and an unrestricted grant to the Department of Ophthalmology from Research to Prevent Blindness, New York, NY.
Footnotes
Conflict of Interest: No conflicting relationship exists for any author
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Quigley HA, Broman AT. 2006. The number of people with glaucoma worldwide in 2010 and 2020. Br J Ophthalmol. 90(3):262–267. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Tham YC, Li X, Wong TY, Quigley HA, Aung T, Cheng CY. 2014. Global prevalence of glaucoma and projections of glaucoma burden through 2040: A systematic review and meta-analysis. Ophthalmology. 121(11):2081–2090. [DOI] [PubMed] [Google Scholar]
- 3.Tielsch JM, Sommer A, Katz J, Royall RM, Quigley HA, Javitt J. 1991. Racial variations in the prevalence of primary open-angle glaucoma. The baltimore eye survey. JAMA. 266(3):369–374. [PubMed] [Google Scholar]
- 4.Friedman DS, Jampel HD, Muñoz B, West SK. 2006. The prevalence of open-angle glaucoma among blacks and whites 73 years and older: The salisbury eye evaluation glaucoma study. Arch Ophthalmol. 124(11):1625–1630. [DOI] [PubMed] [Google Scholar]
- 5.Erie JC, Hodge DO, Gray DT. 1997. The incidence of primary angle-closure glaucoma in olmsted county, minnesota. Arch Ophthalmol. 115(2):177–181. [DOI] [PubMed] [Google Scholar]
- 6.Vijaya L, George R, Arvind H, Baskaran M, Ve Ramesh S, Raju P, Kumaramanickavel G, McCarty C. 2008. Prevalence of primary angle-closure disease in an urban south indian population and comparison with a rural population. The chennai glaucoma study. Ophthalmology. 115(4):655–660.e651. [DOI] [PubMed] [Google Scholar]
- 7.Sakata K, Sakata LM, Sakata VM, Santini C, Hopker LM, Bernardes R, Yabumoto C, Moreira AT. 2007. Prevalence of glaucoma in a south brazilian population: Projeto glaucoma. Invest Ophthalmol Vis Sci. 48(11):4974–4979. [DOI] [PubMed] [Google Scholar]
- 8.Foster PJ, Oen FT, Machin D, Ng TP, Devereux JG, Johnson GJ, Khaw PT, Seah SK. 2000. The prevalence of glaucoma in chinese residents of singapore: A cross-sectional population survey of the tanjong pagar district. Arch Ophthalmol. 118(8):1105–1111. [DOI] [PubMed] [Google Scholar]
- 9.Azuara-Blanco A, Burr J, Ramsay C, Cooper D, Foster PJ, Friedman DS, Scotland G, Javanbakht M, Cochrane C, Norrie J et al. 2016. Effectiveness of early lens extraction for the treatment of primary angle-closure glaucoma (eagle): A randomised controlled trial. Lancet. 388(10052):1389–1397. [DOI] [PubMed] [Google Scholar]
- 10.Coleman AL, Yu F, Evans SJ. 2006. Use of gonioscopy in medicare beneficiaries before glaucoma surgery. J Glaucoma. 15(6):486–493. [DOI] [PubMed] [Google Scholar]
- 11.Hertzog LH, Albrecht KG, LaBree L, Lee PP. 1996. Glaucoma care and conformance with preferred practice patterns. Examination of the private, community-based ophthalmologist. Ophthalmology. 103(7):1009–1013. [DOI] [PubMed] [Google Scholar]
- 12.OECD. 2018. Health spending.
- 13.He M, Jiang Y, Huang S, Chang DS, Munoz B, Aung T, Foster PJ, Friedman DS. 2019. Laser peripheral iridotomy for the prevention of angle closure: A single-centre, randomised controlled trial. Lancet. 393(10181):1609–1618. [DOI] [PubMed] [Google Scholar]
- 14.Seider MI, Pekmezci M, Han Y, Sandhu S, Kwok SY, Lee RY, Lin SC. 2009. High prevalence of narrow angles among chinese-american glaucoma and glaucoma suspect patients. J Glaucoma. 18(8):578–581. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Baskaran M, Foo RC, Cheng CY, Narayanaswamy AK, Zheng YF, Wu R, Saw SM, Foster PJ, Wong TY, Aung T. 2015. The prevalence and types of glaucoma in an urban chinese population: The singapore chinese eye study. JAMA Ophthalmol. 133(8):874–880. [DOI] [PubMed] [Google Scholar]
- 16.Wang YX, Xu L, Yang H, Jonas JB. 2010. Prevalence of glaucoma in north china: The beijing eye study. Am J Ophthalmol. 150(6):917–924. [DOI] [PubMed] [Google Scholar]
- 17.He M, Foster PJ, Ge J, Huang W, Zheng Y, Friedman DS, Lee PS, Khaw PT. 2006. Prevalence and clinical characteristics of glaucoma in adult chinese: A population-based study in liwan district, guangzhou. Invest Ophthalmol Vis Sci. 47(7):2782–2788. [DOI] [PubMed] [Google Scholar]
- 18.Varma R, Ying-Lai M, Francis BA, Nguyen BB, Deneen J, Wilson MR, Azen SP, Group LALES. 2004. Prevalence of open-angle glaucoma and ocular hypertension in latinos: The los angeles latino eye study. Ophthalmology. 111(8):1439–1448. [DOI] [PubMed] [Google Scholar]
- 19.Cheng JW, Zong Y, Zeng YY, Wei RL. 2014. The prevalence of primary angle closure glaucoma in adult asians: A systematic review and meta-analysis. PLoS One. 9(7):e103222. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Seah SK, Foster PJ, Chew PT, Jap A, Oen F, Fam HB, Lim AS. 1997. Incidence of acute primary angle-closure glaucoma in singapore. An island-wide survey. Arch Ophthalmol. 115(11):1436–1440. [DOI] [PubMed] [Google Scholar]
- 21.Stein JD, Kim DS, Niziol LM, Talwar N, Nan B, Musch DC, Richards JE. 2011. Differences in rates of glaucoma among asian americans and other racial groups, and among various asian ethnic groups. Ophthalmology. 118(6):1031–1037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Alper MG, Laubach JL. 1968. Primary angle-closure glaucoma in the american negro. Arch Ophthalmol. 79(6):663–668. [DOI] [PubMed] [Google Scholar]
- 23.Van Herick W, Shaffer RN, Schwartz A. 1969. Estimation of width of angle of anterior chamber. Incidence and significance of the narrow angle. Am J Ophthalmol. 68(4):626–629. [DOI] [PubMed] [Google Scholar]
- 24.Oh YG, Minelli S, Spaeth GL, Steinman WC. 1994. The anterior chamber angle is different in different racial groups: A gonioscopic study. Eye (Lond). 8 ( Pt 1):104–108. [DOI] [PubMed] [Google Scholar]
- 25.Thompson AC, Vu DM, Cowan LA, Asrani S. 2018. Risk factors associated with missed diagnoses of narrow angles by the van herick technique. Ophthalmol Glaucoma. 1(2):108–114. [DOI] [PubMed] [Google Scholar]
- 26.Shi Q, Zhao Y, Fonseca V, Krousel-Wood M, Shi L. 2014. Racial disparity of eye examinations among the u.S. Working-age population with diabetes: 2002–2009. Diabetes Care. 37(5):1321–1328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cassard SD, Quigley HA, Gower EW, Friedman DS, Ramulu PY, Jampel HD. 2012. Regional variations and trends in the prevalence of diagnosed glaucoma in the medicare population. Ophthalmology. 119(7):1342–1351. [DOI] [PubMed] [Google Scholar]
- 28.Benoit SR, Swenor B, Geiss LS, Gregg EW, Saaddine JB. 2019. Eye care utilization among insured people with diabetes in the u.S., 2010–2014. Diabetes Care. 42(3):427–433. [DOI] [PubMed] [Google Scholar]
- 29.Foster PJ, Buhrmann R, Quigley HA, Johnson GJ. 2002. The definition and classification of glaucoma in prevalence surveys. Br J Ophthalmol. 86(2):238–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Xu BY, Friedman DS, Foster PJ, Jiang Y, Porporato N, Pardeshi AA, Munoz B, Aung T, He M. 2021. Ocular biometric risk factors for progression of primary angle closure disease: The zhongshan angle closure prevention trial. Ophthalmology. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 1. Diagnosis, procedure, and treatment codes.
Supplementary Table 2. Sensitivity analysis A: Proportion of definite PACG without prior ANA diagnosis.
Supplementary Table 3. Sensitivity analysis A: Risk factors for definite PACG without prior ANA diagnosis.
Supplementary Table 4. Sensitivity analysis B: Proportion of PACG without prior ANA diagnosis (24- month lookback period).
Supplementary Table 5. Sensitivity analysis B: Risk factors for PACG without prior ANA diagnosis (24- month lookback period).