Currently Viewing:
Supplements Managed Care Issues in Glaucoma: Emerging Trends and Treatment
Currently Reading
The Prevalence of Glaucomatous Risk Factors in Patients From a Managed Care Setting: A Pilot Evaluation
Ervin N. Fang, MD; Simon K. Law, MD, PharmD; John G.Walt, MBA; Tina H. Chiang, PharmD, MBA; and Erin N. Williams, RN
Current Management of Glaucoma and the Need for Complete Therapy
Stuart J. McKinnon, MD, PhD; Lawrence D. Goldberg, MD; Patti Peeples, RPh, PhD; John G.Walt, MBA; and Thomas J. Bramley, PhD
History of Neuroprotection and Rationale as a Therapy for Glaucoma
Leonard A. Levin, MD, PhD; and Patti Peeples, RPh, PhD
Participating Faculty
Managed Care and the Impact of Glaucoma
Claiborne E. Reeder, RPh, PhD; Meg Franklin, PharmD, PhD; and Thomas J. Bramley, PhD

The Prevalence of Glaucomatous Risk Factors in Patients From a Managed Care Setting: A Pilot Evaluation

Ervin N. Fang, MD; Simon K. Law, MD, PharmD; John G.Walt, MBA; Tina H. Chiang, PharmD, MBA; and Erin N. Williams, RN
Based on the available ages for 847 patients, the sample generally represented an elderly cohort with a mean age of 63.0 years (SD 11.9); approximately one third of these patients exceeded the age of 70 years. Specific glaucoma-related diagnoses were recorded for 1120 patients, of whom 79 (7.1%) carried a documented diagnosis of glaucoma. Glaucoma suspects comprised the majority (69.5%), followed by patients with OHT (18.6%) and those with primary open-angle glaucoma (POAG [3.2%]). Most patients (81.3%) reported no familial history of glaucoma. Hypertension and DM were the most prevalent comorbidities affecting 38.8% and 23.5% of the participants, respectively. Approximately 18% of patients with available data were of African American descent, while Latinos and Caucasians represented 25.8% and 18.6% of the cohort, respectively.

Data on ocular measurements were also inconsistently documented across patients (Table 4). However, 1178 (99%) patients had recorded measurements of IOP on at least 1 occasion. The mean IOP for both eyes was 18.0 (SD 11.9) mm Hg. On average, the vertical CDR, based on measurements for 1048 patients, was 0.52 (SD 0.18). Visual field testing revealed a mean PSD of 2.62 (SD 1.8) dB and a mean MD of −2.20 (SD 3.2) dB based on 861 and 863 documented observations, respectively. For the 227 patients with measurements of CCT, the mean value was 553.0 μm (33.6). None of the patients had pseudoexfoliation; optic disc hemorrhage and high myopia occurred infrequently.

Univariate regression analyses examining the relationships between nonocular and ocular RFs demonstrated that when age and DM were considered alone, both were significant predictors of MD, PSD, vertical CDR, IOP, and CCT (all, P <.05; each based on the F-statistic and 1 degree of freedom). Age was positively associated with these ocular parameters, whereas the presence of DM was negatively associated. None of the remaining demographic or nonocular clinical RFs shared significant associations with ocular RFs. (Because of the negligible numbers of patients with cardiovascular disease and migraine, we did not include these RFs as predictive variables.)

Based on the results of this preliminary analysis, we stratified the sample using the following criteria: age >70 or ≤70 years, DM or no DM, mean IOP >21 or ≤0.8 in both eyes, CCT ≥536 μm in either eye or >536 μm in both eyes, and mean MD <−10 dB in either eye or MD ≥−10 dB in both eyes. These criteria were selected because of their reputed association with RFs or indicators of glaucoma progression, as reported in the published literature.20-23 The criterion for disease progression as indicated by mean MD was based on expert opinion provided by practicing ophthalmologists.19 Applying cross-tabulation statistics, we found negligible, but significant, associations between age >70 years and both the presence of elevated IOP >21 mm Hg (phi –0.10; P = .004) and the presence of greater visual field loss (phi 0.16; P <.001). Coexistent DM was also weakly but inversely and significantly associated with thinner central corneas (<536 μm) (phi –0.15; P = .029) and higher vertical CDR (>0.8) (phi –0.082; P = .008) (Table 5).



To address the second research question, we investigated the relative contribution of nonocular RFs and ocular RFs in predicting visual field loss and optic nerve damage (as measured by MD and CDR, respectively) by analyzing 2 separate multiple regression models using stepwise methods. In both models, based on the results of the prior analyses, we assigned nonocular RFs, age and diabetes, and ocular RFs, mean IOP, and mean CCT as explanatory variables while controlling for mean MD and mean vertical CDR in the alternate model. Both final regression models were determined to be adequately fitted.

In Model 1, age, mean CCT, and mean vertical CDR were significant predictors of MD, cumulatively accounting for 11.2% of the variance in visual field measurement. Of the 3 explanatory variables, the extent of visual field loss, as measured by mean MD, was most sensitive to variances in age. This was evident in comparing the standardized beta coefficients (Table 6), which showed that increasing age by 1 SD (ie, ~12 years) in the study sample worsened the mean MD by 0.34 SD (1.8 dB). By comparison, increases in the mean CCT and vertical CDR by 1 SD worsened the mean MD by 0.12 SD and 0.09 SD, respectively. Because of the inverse relationship between mean CCT and mean MD, indicating worsening visual field defect with thicker central corneas, we evaluated the independent variables in model 1 for colinearity, since correlations between predictor variables may affect the direction of relationships between independent and dependent variables. In this analysis, we observed a moderate, significant association between mean CCT and vertical CDR (Pearson’s r = –0.522; P <.001). Although knowledge of the mean CCT and vertical CDR significantly improved the overall predictive value of the model, these ocular RFs jointly accounted for approximately 1% of the variance in mean MD.



In Model 2, age, mean CCT, mean IOP, and mean MD were collectively responsible for 28.9% of the variance in mean CDR, with all variables contributing significantly to the predictive merits of the model (Table 6). Among the RFs, mean CCT yielded the greatest influence over the extent of optic nerve injury. The standardized beta coefficient for this variable indicated that increasing the mean CCT by 1 SD (33.6 μm) was inversely associated with a decrease in the mean vertical CDR by 0.50 SD (0.9). Compared with the prior model, age was a less reliable predictor, accounting for 0.3% of the variance in the mean vertical CDR.

5-Year Risk of Conversion to Glaucoma for OHT Patients
Medical records for 42 of the 220 (19.1%) cases with an ICD-9 diagnosis of OHT rendered sufficient data for the risk conversion analysis. The average age of this subset was 60.7 (SD 9.2) years with 9 of the 42 (21.4%) patients older than 70 years. The mean IOP for both eyes was 21.9 (SD 3.1) mm Hg. The mean vertical CDR, CCT, and PSD were 0.53 (SD 0.18), 551 (SD 31.9) μm, and 2.3 (SD 1.4) dB, respectively. Compared with the 155 and 220 OHT subjects who had available data on age and mean IOP, respectively, the subset appeared to be similar in age (mean age 60.7 years for subset vs 62.1 for sample; P = .326) and mean IOP (21.9 mm Hg for subset vs 21.0 for sample; P = .079).

Applying the risk scoring system for the OHTS predictive model,4,8 the mean composite score for these 42 patients was 9.7 (SD 3.2), signifying a 15% cumulative 5-year risk for developing glaucoma. Overall, scores ranged from 1.0 to 16.0. More than 71% of the sample harbored a 5-year risk of ≥15%; for 9 (21.4%) of these patients, the cumulative probability of conversion to glaucoma was ≥33%. The Figure illustrates the distribution of patients grouped by risk scores and the associated 5-year risk for glaucomatous progression.



Discussion
Given that the majority of patients treated for ICD-9 glaucoma-related diagnoses in this managed care ophthalmology practice were either glaucoma suspects or patients with OHT, our observations, albeit retrospective, imply that there may be a trend toward preventive screening for patients at risk for glaucomatous progression. Alternatively, the higher proportions of glaucoma suspects and patients with OHT may have reflected outdated diagnoses arising from the failure to update the initial recorded diagnoses. However, the documented ocular measurements (Table 4) also confirmed that, on average, patients in this sample did not exhibit clinical evidence of advanced disease. In contrast, nonocular comorbidities were prevalent, with almost one fourth of the cases having DM and over one third afflicted with systemic hypertension.

Exploring the potential relationship between nonocular and ocular RFs, we found that the results of the preliminary univariate analyses mirrored the findings of earlier studies demonstrating significant positive associations between older age and MD, PSD, CDR, IOP, and CCT.3,4,8,13,24,25 Coexistent DM, which has been associated with an increased risk of POAG with some equivocality,26-28 was weakly but inversely and significantly associated with thinner central corneas (<536 μm) and higher vertical CDR (>0.8). In other terms, patients with DM in this sample tended to have thicker central corneas and lower vertical CDR. Regarding the relationship between DM and CCT, the Barbados Eye studies observed thicker corneas among participants who had a history of DM.24 More recently, Özcura and Aydin (2007) surmised that the risk of POAG in diabetic patients is equal to or less than normal individuals because patients in the former group tend to have thicker corneas and, thus, higher IOPs than the latter, which may lead to a false diagnosis of glaucoma.28 The cross-tabulation analysis in our study also appeared to support this postulation. Compared with patients without DM, a significantly lower percentage of patients with DM had central corneal measurements <536 μm, but a greater percentage had IOPs >21 mm Hg (Table 5).

Quantifying the relative contribution of nonocular and ocular RFs in predicting MD and CDR in this sample was compromised by the large percentages of missing data, particularly for the measurements of CCT; only approximately 19% of the total sample had recorded measurements for this RF. Using the EM algorithm to replace missing data, the multivariate regression analyses identified age, mean CCT, and mean  vertical CDR as significant predictors of MD. These explanatory RFs collectively  accounted for only 11.2% of the variance in MD; however, this low predictive power was not unexpected given the fact that, by definition of diagnosis, visual field defects would not be prevalent among glaucoma suspects, who comprised the majority of the sample.

In Model 2, 4 of the 5 RFs (specifically age, mean CCT, mean IOP, and mean MD) were significant  predictors of the extent of optic disc injury. The strongest predictor was CCT, accounting for approximately 50% of the variance in vertical CDR. As in the prior regression analysis, the absence or presence of DM did not contribute significantly to the model’s predictive value.

Based on the OHTS risk scoring system, the final RF analysis showed that, on average, the patients with OHT in our sample had a 15% cumulative 5-year risk for glaucomatous progression. This risk estimate was higher than the 11.6% cumulative 5-year risk reported for the Diagnostic Innovations in Glaucoma Study (DIGS) cohort in Medeiros et al.8 In this pilot study, the risk scoring system was derived from the pooled hazard ratios for both untreated and treated patients from the OHTS and validated longitudinally in untreated patients from the DIGS. For the present study, we used the updated OHTS predictive model, which relied on the pooled data for untreated cohorts from the OHTS and the EGPS.4 We did not collect information on treatment history, which represents a significant limitation of this analysis. It might be noted, however, that the OHTS failed to identify any appreciable differences between the predictive factors for treated and untreated patients.4,8,29

 
Copyright AJMC 2006-2018 Clinical Care Targeted Communications Group, LLC. All Rights Reserved.
x
Welcome the the new and improved AJMC.com, the premier managed market network. Tell us about yourself so that we can serve you better.
Sign Up