ORIGINAL ARTICLE
Relationship between cataract severity and socioeconomic status Jason D. Wesolosky, Christopher J. Rudnisky, MD, MPH, FRCSC ABSTRACT ● RÉSUMÉ Objective: To determine the relationship between cataract severity and socioeconomic status (SES). Design: Retrospective, observational case series. A total of 1350 eyes underwent phacoemulsification cataract extraction by a single surgeon using an Alcon Infiniti system. Cataract severity was measured using phaco time in seconds. SES was measured using area-level aggregate census data: median income, education, proportion of common-law couples, and employment rate. Preoperative best corrected visual acuity was obtained and converted to logarithm of the minimum angle of resolution values. For patients undergoing bilateral surgery, the generalized estimating equation was used to account for the correlation between eyes. Univariate analyses were performed using simple regression, and multivariate analyses were performed to account for variables with significant relationships (p o 0.05) on univariate testing. Sensitivity analyses were performed to assess the effect of including patient age in the controlled analyses. Results: Multivariate analyses demonstrated that cataracts were more severe when the median income was lower (p ¼ 0.001) and the proportion of common-law couples living in a patient’s community (p ¼ 0.012) and the unemployment rate (p ¼ 0.002) were higher. These associations persisted even when controlling for patient age. Conclusion: Patients of lower SES have more severe cataracts. Objet : Établissement de la relation entre la gravité de la cataracte et le statut socioéconomique (SSÉ). Nature : Observation rétrospective d’une série de cas : 1 350 yeux qui avaient subi une extraction de la cataracte par phacoémulsification effectuée par un seul chirurgien utilisant un système Alcon Infiniti. La gravité de la cataracte a été mesurée en utilisant le temps de la phaco en secondes. Le SSÉ a été mesuré selon l’ensemble des données recueillies au niveau de l’aire de recensement : revenu moyen, éducation, proportion des conjoints de fait et taux d’emploi. La meilleure acuité visuelle (AV) corrigée avant la chirurgie a été obtenue et convertie en valeurs logMar. Pour les patients subissant une chirurgie bilatérale, l’équation de l’estimation généralisée (ÉEG) a été utilisée pour rendre compte de la corrélation entre les yeux. Les analyses à variable unique ont été appliquées avec simple régression et celles à variables multiples ont servi à rendre compte des variables ayant des relations significatives (p=0,05) sur les tests à variable unique. Les analyses de sensibilité avaient pour objet d'évaluer l’effet de l’inclusion de l’âge du patient dans les analyses de contrôle. Résultats : Les analyses à variables multiples ont démontré que les cataractes étaient plus graves lorsque la moyenne de revenu était inférieure (p=0,001), et où la proportion des conjoints de fait vivant dans une communauté de patients (p-0,012) et le taux de chômage (p=0.002) étaient plus élevés. Ces associations persistaient, que l’âge soit contrôlé ou pas. Conclusion : Les patients ayant un SSÉ inférieur ont des cataractes plus graves.
There is increasing evidence that one’s socioeconomic status (SES) affects one’s health1–7; education, income, family status, and social environment all play a role in determining the likelihood of developing a medical problem and what the prognosis might be. However, it is unclear whether low SES is causative of poor health or simply a modifier; studies have demonstrated a positive correlation between lower SES and longer wait times for surgery,8 decreased utilization of diagnostic testing,9,10 and decreased angiography use after myocardial infarction.11 In ophthalmology, low SES is associated, in particular, with cataract. The Baltimore Eye Study6 found a strong correlation between SES and the presence of cortical cataracts.5 Similarly, the Singapore Malay Eye Study
found that patients with lower SES have higher odds of nuclear and posterior subcapsular cataract.12 However, these studies examined the presence of cataract, as opposed to visually significant cataracts requiring surgical removal.1,3–5,12,13 To the best of our knowledge, there are no data on how SES relates to cataract severity. In Alberta, where there is publicly provided health care, there was a negligible financial cost to a patient when they decided to undergo cataract surgery before the introduction of premium (toric, multifocal) intraocular lenses (IOLs). Therefore, a patient’s decision to have surgery should not be biased by their financial (socioeconomic) situation. The purpose of this article is to determine what the relationship is between cataract severity and SES.
From the Department of Ophthalmology, University of Alberta, Edmonton, Alta.
Can J Ophthalmol 2013;48:471–477 0008-4182/13/$-see front matter & 2013 Canadian Ophthalmological Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcjo.2013.07.001
Originally received Mar. 18, 2013. Final revision June 28, 2013. Accepted July 2, 2013 Correspondence to Chris Rudnisky, MD, Room 2316, Royal Alexandra Hospital, 10240 Kingsway Avenue, Edmonton AB T5H3V9; [email protected]
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
471
Cataract severity and socioeconomic status—Wesolosky and Rudnisky METHODS Approval for this retrospective analysis was obtained from the University of Alberta Health Research Ethics Board before chart review and data collection. Patients who underwent phacoemulsification cataract extraction by a single surgeon between January 1, 2006, and December 31, 2008, were eligible for inclusion in this study. Cataract severity was measured by a surrogate indicator: phaco time. Eyes that had their cataract removed using OZIL (Alcon, Fort Worth, Tex.) handpieces were excluded because their phaco times were significantly shorter than with the standard Infiniti (Alcon) handpiece. Preoperative best corrected visual acuity (VA) measurements were obtained using a standard Snellen chart viewed from an effective distance of 6 m. They were converted to logarithm of the minimum angle of resolution (logMAR) values for the analysis. VA measurements that had been recorded as counting fingers were converted using the method described by Holladay.14 Charts were reviewed for surgical covariates, complications, and other cataract severity identifiers such as the use of surgical aids to visualize cataract during surgery, surgical technique (chop vs divide and conquer), and the presence of weak zonules. Surgeon experience over time was considered as well; the duration of time in practice was used as an indicator of surgeon proficiency. To assign an SES level to a patient, postal code at the time of surgery was recorded; only patients who had a valid Alberta postal code were considered for the study. Statistics Canada conducts a census every 5 years and divides the country into aggregate units: census metropolitan areas (CMA), census tracts (CT), and dissemination areas (DA). Several provinces are in the midst of or have developed indices (deprivation index [DI]) that attempt to connect disparate SES factors into a single component to determine variances in health outcomes.15–19 The DA level of geographic resolution is the finest aggregate of SES data available, with each DA containing roughly 400 to 700 people. Using 2006 census information, SES was determined by linking study patients via their postal code, providing more than 2000 discrete variables for each patient record. Although many variables have been demonstrated to affect health outcomes in Canada, several broad categories have been shown to be effective predictors: income, education, and household social level.3,4,20,21 Income was defined as median income per person older than 15 years within a DA. Education level was categorized into strata: (i) less than high school, (ii) high-school graduation, (iii) apprentice or trade certification, (iv) college and some post-secondary, and (v) and postsecondary graduation. Education levels were summed for the population older than 15 years and divided by the total educated population older than 15 years to obtain a proportion. This proportion was then multiplied by each of the category levels and the resulting values summed to
472
yield a single “unified education” value for each patient. Household social-level indicators, such as the proportion of common-law relationships, single-parent families, rental units, and the unemployed were calculated for each Alberta postal code (and therefore for each study patient). As a secondary outcome, DIs were used. DIs are sexadjusted, income-corrected aggregates of several SES indicators,15 combined into a single value via primary component analysis,17 which allows many dimensions of data to be reduced to a few based on the weighted variability that each dimension comprises. These indices are useful for health planning and research because they link many well-recognized SES variables, allowing for standardized testing across multiple medical and economic fields. Recently, a DI was created for Alberta using 2006 census variables from the DA level of data, comprising 2 reduced dimension sets: material deprivation and social deprivation.22 Data from all eligible patients were entered into a Microsoft Access (14.0.6024.1000; Microsoft Inc, Redmond, Wash.) database. All calculations were performed using SAS statistical software (version 9.2; SAS Institute Inc, Cary, N.C.). Because some patients had both eyes enrolled in the study, analyses used the generalized estimating equation with SAS PROC GENMOD to account for the correlation between 2 eyes of a single patient. Univariate analysis was performed with simple regression, and multivariate analysis was performed to adjust for the effect of factors demonstrating a significant relationship (p o 0.05) on univariate testing that were not otherwise known to be related to cataract severity. Although age would be expected to be related to cataract severity, and adjusting for it may adjust for the outcome itself, it was still considered for inclusion in a multivariate model because younger patients can also present with advanced cataracts. In addition, if access barriers existed for patients with low SES, controlling for age would provide a more conservative estimate of the relationship between cataract severity and SES. As such, a sensitivity analysis was performed to determine the impact of the inclusion of age on the results. No imputation of missing data was performed.
RESULTS During the study period, 1458 cataract extractions were performed (Fig. 1). Of these, 57 were surgeries performed using OZIL (Alcon) handpieces and were therefore excluded. Eight surgeries were secondary IOL insertions (sulcus or anterior chamber IOL) and were also excluded. Of the remaining 1393 cataract extractions, 43 were missing the primary outcome (phaco time) and were also excluded. Only patients with valid and non-retired Alberta postal codes were used in this study, and no out-of-
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
Cataract severity and socioeconomic status—Wesolosky and Rudnisky
Fig. 1 — Socioeconomic status cataract study eye enrollment data. IOL, intraocular lens.
province patients were included in the data set. The missing data amounted to only 3.1% of the total records. The final study population included 1350 sequential phacoemulsification cataract extractions performed on 873 patients. The mean age was 69.6 ⫾ 11.6 years and ranged from 27 to 98 years old. Mean preoperative logMAR VA was 0.60 ⫾ 0.54. The majority of patients were female (53.6%), and the operative side was equally split at 50:50. Although sex (p ¼ 0.124) and operative side (p ¼ 0.878) were not associated with phaco time, age was (p o 0.0001); the parameter estimate on simple regression was positive (0.425 ⫾ 0.038), indicating, appropriately, that cataracts worsen with age. The average IOL power implanted was 19.9 ⫾ 3.9 D and was also predictive of cataract severity (p ¼ 0.005); with a positive parameter estimate (0.337 ⫾ 0.118), the data suggest that phaco
time was longer for more hyperopic eyes. Because phaco time is used as a surrogate for cataract severity, this suggests that in this study, cataracts were more severe in hyperopic eyes. Preoperative VA was also highly predictive of cataract severity (p o 0.0001), which helps to validate the use of phaco time as a surrogate indicator; Figure 2 charts the relationship between phaco time and VA, indicating that as vision declines, phaco time increases (p o 0.0001). A resident participated in the phacoemulsification component of the surgery in 14.4% of eyes. Mean resident phaco time was 34.9 ⫾ 18.0 seconds and significantly longer than the supervising surgeon: 26.3 ⫾ 15.4 seconds (p o 0.001). Anaesthesia was either topical (58.0%), retrobulbar (41.2%), or general (0.8%). Topical anaesthetic was not associated with cataract severity (p ¼ 0.086), but retrobulbar anaesthetics were (p ¼ 0.045). However, during the study period, the surgeon gradually transitioned from primarily retrobulbar anaesthetics earlier in his career to primarily topical anaesthetics. Because simple regression of cataract severity to length of time in practice demonstrated a significant relationship (p o 0.0001), when the same analysis was performed controlling for anaesthetic type, there was, in fact, no relationship between cataract severity and topical (p ¼ 0.473) or retrobulbar (p ¼ 0.068) anaesthetics; the estimate for general anaesthetics was null because of sparse data. Comorbidities (Fig. 3) such as patient use of Flomaxassociated floppy iris syndrome were present in 3.3% of eyes, weak zonules in 3.2%, small pupils in 2.9%, pseudoexfoliation in 2.1%, and iris synechiae in 0.6%. Surgical adjuncts (Fig. 4) included the use of vision blue in
Fig. 2 — Relationship between visual acuity and phacoemulsification time at cataract surgery. MSE, mean square error. CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
473
Cataract severity and socioeconomic status—Wesolosky and Rudnisky
Fig. 3 — Percentage of eye comorbidities associated with cataract surgery participants. FIS, floppy iris syndrome.
16.5% of eyes, iris hooks in 2.4%, and placement of a capsular tension ring in 1.1%. The cataract was hydrated (white) at the time of surgery in 0.5% of eyes. A chop technique was used in 11.5% of eyes, and the remainder were removed using a divide and conquer technique. Posterior capsular rupture occurred in 0.5% of eyes, 80% of which underwent pars plana vitrectomy. A suture was placed at the end of the procedure in 7.3% of eyes. Of all these factors, the use of vision blue (p o 0.0001), a chop technique (p o 0.0001), and weak zonules (p ¼ 0.008) were statistically significant predictors of cataract severity. Discrete socioeconomic indicators
The mean median household income for patients in the study was $61 032 ⫾ $26 824, which is approximately the same as the Alberta average median household income of $63 988 ⫾ $288.23 Median income was negatively associated with cataract severity: As income increased, cataract severity decreased (Table 1; p ¼ 0.0044). Because resident participation, surgeon experience (time in practice), IOL power, and the presence of weak zonules were all confounding factors of phaco time, an adjusted analysis was performed to control for their effects, because they are not truly associated with cataract severity. When adjusting for these factors, cataract severity decreased as median income increased, irrespective of whether age was included (p ¼ 0.001; Table 2) or not (p ¼ 0.0008). The average level of education attained overall was 2.68 ⫾ 0.49 (Table 1); 5.1% of patients lived in a community where the majority had less than a high-school education, 71.7% lived in a community where the majority had a high-school diploma, 22.8% lived where the majority had college or incomplete post-secondary education, and 0.4% lived where the majority had graduated from a post-secondary institution. There was no relationship between education and cataract severity (p ¼ 0.941). The mean proportion of common-law couples in the study was 8.7% ⫾ 5.0% (Table 1). Simple regression demonstrated a statistically significant relationship
474
between cataract severity and a higher proportion of common-law couples living in a patient’s community (p ¼ 0.049). A controlled analysis again confirmed that as the proportion of common-law couples in a community increases, so too does cataract severity (p ¼ 0.012), even if age is included (p ¼ 0.0003; Table 3). In comparison, the overall proportion of single-parent families was higher at 17.2% ⫾ 10.8%, but there was no association with cataract severity (p ¼ 0.376). Similarly, there was no association between the proportion of rental units (μ: 29.6% ⫾ 29.4%; p ¼ 0.120) in the DA. Another social parameter that was associated with cataract severity was the unemployment rate (p ¼ 0.0014) in a patient’s community, which was 5.3% ⫾ 6.5% overall. The higher the unemployment rate, the worse the cataract was (p ¼ 0.002), even after adjustment for age (p ¼ 0.006; Table 4). Deprivation indices as socioeconomic indicators
An unadjusted analysis including both social and material deprivation indices did not demonstrate a statistically significant relationship with cataract severity (p ¼ 0.051 and 0.325, respectively). However, a controlled analysis demonstrated a relationship between the material DI when age was included (p ¼ 0.022) and the social index when age was excluded (p ¼ 0.031). In both instances, as the levels of deprivation increased, so too did cataract severity.
DISCUSSION This is the first study in Canada to investigate the relationship between SES and cataract severity. Because statistically significant relationships were detected on univariate, multivariate, and exploratory sensitivity analyses using both discrete indicators and deprivation indices, this demonstrates a robust relationship between increasing cataract severity and lower SES.
Fig. 4 — Surgical adjuncts used in cataract extraction. CTR, capsular tension ring.
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
Cataract severity and socioeconomic status—Wesolosky and Rudnisky Table 1—Relationship between socioeconomic indicators and cataract severity (univariate analyses) Variable Median income (/$1000) Education Proportion of common-law couples (%) Proportion of single-parent families (%) Unemployment rate (%) Proportion of band housing or rental units (%) n
Mean ⫾ SD
Parameter Estimate
SE
95% Confidence Limit
p Value*
⫾ ⫾ ⫾ ⫾ ⫾ ⫾
–0.05 –0.08 23.11 4.11 21.53 2.92
0.02 1.01 11.42 4.64 6.75 1.88
–0.09, –0.02 –2.06, 1.91 0.722, 45.49 –4.99, 13.20 8.31, 34.76 –0.76, 6.60
0.004 0.941 0.043 0.376 0.001 0.120
61032 2.68 8.72 17.21 5.31 29.6
26824 0.49 4.97 10.78 6.54 27.4
Simple regression using the generalized estimating equation.
The use of phaco time as a surrogate for cataract severity is a novel approach, but is it appropriate? In a retrospective study, lens assessment tools that are used in major population studies24–26 are not an option. From a scholarly perspective alone, it stands to reason that denser lenses require more phacoemulsification energy to disassemble, and as such, phaco time should rise fairly linearly with cataract severity. One potential weakness of phaco time as a surrogate is that it relates more to lens density than to opacity; cortical and posterior subcapsular cataracts can cause a significant reduction in VA in the absence of nuclear sclerosis. As such, it is possible that patients underwent surgery for visually significant cataracts that are less dense, and would thus have a shorter phaco time. Although cataract type (nuclear sclerosis, cortical and/or posterior subcapsular) was not collected in this data set, we do not believe this omission affects the results or conclusions: If there were many study patients with this inverse relationship, one would expect poor correlation between preoperative VA and phaco time. However, this was not the case: VA demonstrated a very strong correlation with phaco time (p o 0.0001). As such, this would suggest that there were few patients who underwent surgery for a cataract with only cortical or posterior subcapsular components. Another approach to assessing the appropriateness of phaco time as a surrogate is to consider the relationship between it and other more easily accepted surrogates for cataract severity. VA was highly correlated to phaco time (p o 0.0001) but was not used as the surrogate because final postoperative best corrected VA was not available for all patients; preoperative VA alone would not accurately reflect true cataract severity because it would overestimate it in eyes with ocular comorbidities such as age-related macular degeneration. Vision blue is a surgical adjunct Table 2—Controlled analysis of the relationship between cataract severity and median income Variable Age Resident participation Time in practice Intraocular lens power Weak zonules Median income (/$1000)
used to improve the visualization of the capsule when the red reflex is poor, as is often the case with dense, advanced cataracts. Although it too was highly correlated to phaco time (p o 0.0001), it was not used as the surrogate because, as a dichotomous indicator, it detects only very severe cataracts and would therefore not provide any stratification of milder disease. Chop technique, which was used in this study only in eyes with palpably dense lenses (the decision to chop was made after starting to groove the lens intraoperatively), was also associated with phaco time (p o 0.0001). Like vision blue, it would not provide stratification of milder disease and would, therefore, be a suboptimal surrogate. Lastly, although weak zonules were associated with phaco time (p ¼ 0.008) and could possibly result from years of stress induced by a dense and “weighty” lens, they can also occur because of trauma, which is more commonly suffered by those of lower SES.27–29 In the end, it was included in the adjusted models to minimize bias in favour of a relationship between cataract severity and SES. Given that 3 clear indicators of cataract severity are associated with phaco time, we believe it is an excellent surrogate. Just as with our choice of outcome, justification of our socioeconomic predictors is required. Many social factors can affect medical outcomes: in-home support, geographic proximity to care centres, and access to primary care.27,30,31 Hence it is not surprising that cataract severity is not unlike other medical problems1,9,10,30 in that patients with lower SES have reduced health status and generally less favourable health outcomes. Although this study looked at only a few discrete indicators such as median income, the proportion of common-law couples in a community, and unemployment rate, there may be other related factors that are not aggregated in census data, which is a limitation encountered when using postal code Table 3—Controlled analysis of the relationship between cataract severity and the proportion of common-law couples living in a patient’s community
Parameter Estimate
SE
95% Confidence Limits
p Value
0.382 7.204
0.039 1.354
0.305, 0.459 4.550, 9.859
o0.0001 o0.0001
0.208 0.059
0.055 0.117
0.100, 0.316 –0.171, 0.289
0.0002 0.613
14.817 –0.054
5.025 0.016
4.968, 24.665 –0.086, –0.022
0.003 0.001
Variable Age Resident participation Time in practice Intraocular lens power Weak zonules Proportion of commonlaw couples
Parameter Estimate
SE
0.399 7.189 0.204 0.058 14.773 35.588
0.038 1.333 0.055 0.116 5.015 10.050
95% Confidence Limits 0.325, 4.577, 0.096, –0.169, 4.943, 16.891,
0.474 9.802 0.311 0.284 24.603 56.284
p Value o0.0001 o0.0001 0.0002 0.617 0.003 0.0003
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
475
Cataract severity and socioeconomic status—Wesolosky and Rudnisky Table 4—Controlled analysis of the relationship between cataract severity and the unemployment rate in a patient’s community Variable Age Resident participation Time in practice Intraocular lens power Weak zonules Unemployment rate
Parameter Estimate
SE
95% Confidence Limits
p Value
0.384 6.895
0.039 1.342
0.307, 0.461 4.265, 9.526
o0.0001 o0.0001
0.207 0.066
0.055 0.117
0.099, 0.316 –0.163, 0.295
0.0002 0.574
14.833 19.996
5.051 7.293
4.932, 24.734 5.702, 34.290
0.003 0.006
data. However, because medical administrative databases —or medical records, for that matter—do not carry overt SES identifiers, use of area aggregated social and material information is the only reliable and horizontally maintained (i.e., year to year) data set. Furthermore, it is impractical to collect direct SES information for the purposes of health monitoring and health care planning.7,18,32,33 Indeed, although individual SES statistics are the gold standard, area-based population information (i.e., DA/postal codes) shows similar trends as that of individual, direct metrics.21 This relationship permits the use of area-level aggregate data to determine trends that likely apply to the individual. As such, we believe the results of this study to be discretely generalizable. Increasing age (p o 0.0001), resident participation (p o 0.0001), and surgeon experience (p ¼ 0.0002) were associated with cataract severity but are not related to SES. Interestingly, the coefficient for surgeon experience, as measured by the surrogate “time in practice,” was positive, indicating that with more experience, the surgeon saw more severe cases. Inclusion of this variable in the analysis helps to reduce selection bias; although referral sources may have selected more severe cases to send to a more experienced surgeon, the relationship between cataract severity and SES is still observed after adjustment. With respect to resident participation, the observed relationship is a common sense finding: a less proficient surgeon will have greater phaco time, and controlling for this confounding relationship helps to clarify the relationship between cataract severity and SES. Because cataract is an age-related disease, it is also a common sense finding that cataract severity increases with age. Because of this, it was unclear whether age should be included in the adjusted analyses; controlling for age may control for the outcome itself, although given that traumatic cataracts are more common in younger patients and that trauma itself is more common in those of lower SES,27–29 exclusion may have induced bias. Irrespective of whether age was included in the analyses, increasing cataract severity was associated with lower SES. This suggests that age is not a consideration when discussing the impact of SES on the development of visually significant cataracts.
476
Although the relationship between median income and unemployment rates and cataract severity are more easily understood, the strongest predictor, as indicated by the regression coefficients, was the proportion of common-law couples in a patient’s community. This finding, although at first perplexing, requires more study; it is our hypothesis that this measure is a surrogate for other, more directly related, unmeasured factors, such as domestic violence, smoking, alcohol abuse, and substance abuse. This study provides strong statistical evidence for a connection between cataract severity and SES; as SES decreases, cataract severity increases. This result should be considered by government and health policy experts when discussing the suitability of private health care as a solution to contemporaneous health care issues: the least advantaged require the most help and yet are least able to pay for it.
Disclosure: The authors have no proprietary or commercial interest in any materials discussed in this article. Acknowledgements: Dr. Chris Rudnisky would like to acknowledge Dr. Ichiro Kawachi, whose outstanding and captivating Summer Session Society and Health Lectures at Harvard University were the inspiration and motivation behind this research study. REFERENCES 1. Lewallen S. Poverty and cataract—a deeper look at a complex issue. PLoS Med. 2008;5:e245. 2. Pampalon R, Hamel D, Gamache P. Health inequalities in urban and rural Canada: comparing inequalities in survival according to an individual and area-based deprivation index. Health Place. 2010;16:416-20. 3. Perruccio AV, Badley EM, Trope GE. A Canadian population-based study of vision problems: assessing the significance of socioeconomic status. Can J Ophthalmol. 2010;45:477-83. 4. Meddings DR, Hertzman C, Barer ML, et al. Socioeconomic status, mortality, and the development of cataract at a young age. Soc Sci Med. 1998;46:1451-7. 5. Tielsch JM, Sommer A, Katz J, Quigley H, Ezrine S. Socioeconomic status and visual impairment among urban Americans. Baltimore Eye Survey Research Group. Arch Ophthalmol. 1991;109:637-41. 6. Tielsch JM, Sommer A, Witt K, Katz J, Royall RM. Blindness and visual impairment in an American urban population. The Baltimore Eye Survey. Arch Ophthalmol. 1990;108:286-90. 7. Sit AJ. Socioeconomic factors and vision health in Canada. Can J Ophthalmol. 2010;45:441-2. 8. Shortt SED, Shaw RA. Equity in Canadian health care: does socioeconomic status affect waiting times for elective surgery? CMAJ. 2003;168:413. 9. Demeter S, Leslie WD, Lix L, MacWilliam L, Finlayson GS, Reed M. The effect of socioeconomic status on bone density testing in a public health-care system. Osteoporos Int. 2007;18:153-8. 10. Demeter S, Reed M, Lix L, MacWilliam L, Leslie WD. Socioeconomic status and the utilization of diagnostic imaging in an urban setting. CMAJ. 2005;173:1173-7. 11. Alter DA, Naylor CD, Austin PC, Chan BTB, Tu JV. Geography and service supply do not explain socioeconomic gradients in angiography use after acute myocardial infarction. CMAJ. 2003;168:261. 12. Wu R, Wang JJ, Mitchell P, et al. Smoking, socioeconomic factors, and age-related cataract: The Singapore Malay Eye Study. Arch Ophthalmol. 2010;128:1029-35.
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
Cataract severity and socioeconomic status—Wesolosky and Rudnisky 13. Younan C, Mitchell P, Cumming R, Rochtchina E. Socioeconomic status and incident cataract surgery: the Blue Mountains Eye Study. Clin Experiment Ophthalmol. 2002;30:163-7. 14. Holladay JT. Proper method for calculating average visual acuity. J Refract Surg. 1997;13:388-91. 15. Pampalon R, Hamel D, Gamache P, Raymond G. A deprivation index for health planning in Canada. Chronic Dis Can. 2009;29:178-91. 16. Krefis AC, Schwarz NG, Nkrumah B, et al. Principal component analysis of socioeconomic factors and their association with malaria in children from the Ashanti Region, Ghana. Malar J. 2010;9:201. 17. Vyas S, Kumaranayake L. Constructing socio-economic status indices: how to use principal components analysis. Health Policy Plan. 2006;21:459-68. 18. Tello JE, Jones J, Bonizzato P, Mazzi M, Amaddeo F, Tansella M. A census-based socio-economic status (SES) index as a tool to examine the relationship between mental health services use and deprivation. Soc Sci Med. 2005;61:2096-105. 19. Frohlich N, Fransoo R, Roos N. Health service use in the Winnipeg Regional Health Authority: variations across areas in relation to health and socioeconomic status. Healthc Manage Forum. 2002;15:9-14. 20. Odoi A, Wray R, Emo M, et al. Inequalities in neighbourhood socioeconomic characteristics: potential evidence-base for neighbourhood health planning. Int J Health Geogr. 2005;4:20. 21. Pampalon R, Hamel D, Gamache P. A comparison of individual and area-based socio-economic data for monitoring social inequalities in health. Health Rep. 2009;20:85-94. 22. Pampalon R, Hamel D, Gamache P. 2006 Alberta Deprivation Index Variables. Microsoft Excel; 2011. 23. Household income groups (22) in constant (2005) dollars and household type (11) for the private households of Canada, provinces, territories, census metropolitan areas and census agglomerations, 2000 and 2005 - 20% sample data. StatsCan; 2006.
24. West SK, Rosenthal F, Newland HS, Taylor HR. Use of photographic techniques to grade nuclear cataracts. Invest Ophthalmol Vis Sci. 1988;29:73-7. 25. Ortiz D, Alio JL, Ruiz-Colecha J, Oser U. Grading nuclear cataract opacity by densitometry and objective optical analysis. J Cataract Refract Surg. 2008;34:1345-52. 26. Huang W, Chan KL, Li H, Lim JH, Liu J, Wong TY. A computer assisted method for nuclear cataract grading from slit-lamp images using ranking. IEEE Trans Med Imaging. 2011;30:94-107. 27. Roos LL, Walld R, Uhanova J, Bond R. Physician visits, hospitalizations, and socioeconomic status: ambulatory care sensitive conditions in a Canadian setting. Health Serv Res. 2005;40:1167-85. 28. Dannenberg AL, Parver LM, Fowler CJ. Penetrating eye injuries related to assault: the national eye trauma system registry. The National Eye Trauma System Registry. Arch Ophthalmol. 1992;110:849-52. 29. Parver LM, Dannenberg AL, Blacklow B, Fowler CJ, Brechner RJ, Tielsch JM. Characteristics and causes of penetrating eye injuries reported to the National Eye Trauma System Registry, 1985-91. Public Health Rep. 1993;108:625-32. 30. Dunlop S, Coyte PC, McIsaac W. Socio-economic status and the utilisation of physicians' services: results from the Canadian National Population Health Survey. Soc Sci Med. 2000;51:123-33. 31. Wong ST, Regan S. Patient perspectives on primary health care in rural communities: effects of geography on access, continuity and efficiency. Rural Remote Health. 2009;9:1142. 32. Duncan GJ, Daly MC, McDonough P, Williams DR. Optimal indicators of socioeconomic status for health research. Am J Public Health. 2002;92:1151-7. 33. Frohlich N, Mustard C. A regional comparison of socioeconomic and health indices in a Canadian province. Soc Sci Med. 1996;42: 1273-81.
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
477
Relationship between cataract severity and socioeconomic status Jason D. Wesolosky, Christopher J. Rudnisky, MD, MPH, FRCSC ABSTRACT ● RÉSUMÉ Objective: To determine the relationship between cataract severity and socioeconomic status (SES). Design: Retrospective, observational case series. A total of 1350 eyes underwent phacoemulsification cataract extraction by a single surgeon using an Alcon Infiniti system. Cataract severity was measured using phaco time in seconds. SES was measured using area-level aggregate census data: median income, education, proportion of common-law couples, and employment rate. Preoperative best corrected visual acuity was obtained and converted to logarithm of the minimum angle of resolution values. For patients undergoing bilateral surgery, the generalized estimating equation was used to account for the correlation between eyes. Univariate analyses were performed using simple regression, and multivariate analyses were performed to account for variables with significant relationships (p o 0.05) on univariate testing. Sensitivity analyses were performed to assess the effect of including patient age in the controlled analyses. Results: Multivariate analyses demonstrated that cataracts were more severe when the median income was lower (p ¼ 0.001) and the proportion of common-law couples living in a patient’s community (p ¼ 0.012) and the unemployment rate (p ¼ 0.002) were higher. These associations persisted even when controlling for patient age. Conclusion: Patients of lower SES have more severe cataracts. Objet : Établissement de la relation entre la gravité de la cataracte et le statut socioéconomique (SSÉ). Nature : Observation rétrospective d’une série de cas : 1 350 yeux qui avaient subi une extraction de la cataracte par phacoémulsification effectuée par un seul chirurgien utilisant un système Alcon Infiniti. La gravité de la cataracte a été mesurée en utilisant le temps de la phaco en secondes. Le SSÉ a été mesuré selon l’ensemble des données recueillies au niveau de l’aire de recensement : revenu moyen, éducation, proportion des conjoints de fait et taux d’emploi. La meilleure acuité visuelle (AV) corrigée avant la chirurgie a été obtenue et convertie en valeurs logMar. Pour les patients subissant une chirurgie bilatérale, l’équation de l’estimation généralisée (ÉEG) a été utilisée pour rendre compte de la corrélation entre les yeux. Les analyses à variable unique ont été appliquées avec simple régression et celles à variables multiples ont servi à rendre compte des variables ayant des relations significatives (p=0,05) sur les tests à variable unique. Les analyses de sensibilité avaient pour objet d'évaluer l’effet de l’inclusion de l’âge du patient dans les analyses de contrôle. Résultats : Les analyses à variables multiples ont démontré que les cataractes étaient plus graves lorsque la moyenne de revenu était inférieure (p=0,001), et où la proportion des conjoints de fait vivant dans une communauté de patients (p-0,012) et le taux de chômage (p=0.002) étaient plus élevés. Ces associations persistaient, que l’âge soit contrôlé ou pas. Conclusion : Les patients ayant un SSÉ inférieur ont des cataractes plus graves.
There is increasing evidence that one’s socioeconomic status (SES) affects one’s health1–7; education, income, family status, and social environment all play a role in determining the likelihood of developing a medical problem and what the prognosis might be. However, it is unclear whether low SES is causative of poor health or simply a modifier; studies have demonstrated a positive correlation between lower SES and longer wait times for surgery,8 decreased utilization of diagnostic testing,9,10 and decreased angiography use after myocardial infarction.11 In ophthalmology, low SES is associated, in particular, with cataract. The Baltimore Eye Study6 found a strong correlation between SES and the presence of cortical cataracts.5 Similarly, the Singapore Malay Eye Study
found that patients with lower SES have higher odds of nuclear and posterior subcapsular cataract.12 However, these studies examined the presence of cataract, as opposed to visually significant cataracts requiring surgical removal.1,3–5,12,13 To the best of our knowledge, there are no data on how SES relates to cataract severity. In Alberta, where there is publicly provided health care, there was a negligible financial cost to a patient when they decided to undergo cataract surgery before the introduction of premium (toric, multifocal) intraocular lenses (IOLs). Therefore, a patient’s decision to have surgery should not be biased by their financial (socioeconomic) situation. The purpose of this article is to determine what the relationship is between cataract severity and SES.
From the Department of Ophthalmology, University of Alberta, Edmonton, Alta.
Can J Ophthalmol 2013;48:471–477 0008-4182/13/$-see front matter & 2013 Canadian Ophthalmological Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcjo.2013.07.001
Originally received Mar. 18, 2013. Final revision June 28, 2013. Accepted July 2, 2013 Correspondence to Chris Rudnisky, MD, Room 2316, Royal Alexandra Hospital, 10240 Kingsway Avenue, Edmonton AB T5H3V9; [email protected]
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
471
Cataract severity and socioeconomic status—Wesolosky and Rudnisky METHODS Approval for this retrospective analysis was obtained from the University of Alberta Health Research Ethics Board before chart review and data collection. Patients who underwent phacoemulsification cataract extraction by a single surgeon between January 1, 2006, and December 31, 2008, were eligible for inclusion in this study. Cataract severity was measured by a surrogate indicator: phaco time. Eyes that had their cataract removed using OZIL (Alcon, Fort Worth, Tex.) handpieces were excluded because their phaco times were significantly shorter than with the standard Infiniti (Alcon) handpiece. Preoperative best corrected visual acuity (VA) measurements were obtained using a standard Snellen chart viewed from an effective distance of 6 m. They were converted to logarithm of the minimum angle of resolution (logMAR) values for the analysis. VA measurements that had been recorded as counting fingers were converted using the method described by Holladay.14 Charts were reviewed for surgical covariates, complications, and other cataract severity identifiers such as the use of surgical aids to visualize cataract during surgery, surgical technique (chop vs divide and conquer), and the presence of weak zonules. Surgeon experience over time was considered as well; the duration of time in practice was used as an indicator of surgeon proficiency. To assign an SES level to a patient, postal code at the time of surgery was recorded; only patients who had a valid Alberta postal code were considered for the study. Statistics Canada conducts a census every 5 years and divides the country into aggregate units: census metropolitan areas (CMA), census tracts (CT), and dissemination areas (DA). Several provinces are in the midst of or have developed indices (deprivation index [DI]) that attempt to connect disparate SES factors into a single component to determine variances in health outcomes.15–19 The DA level of geographic resolution is the finest aggregate of SES data available, with each DA containing roughly 400 to 700 people. Using 2006 census information, SES was determined by linking study patients via their postal code, providing more than 2000 discrete variables for each patient record. Although many variables have been demonstrated to affect health outcomes in Canada, several broad categories have been shown to be effective predictors: income, education, and household social level.3,4,20,21 Income was defined as median income per person older than 15 years within a DA. Education level was categorized into strata: (i) less than high school, (ii) high-school graduation, (iii) apprentice or trade certification, (iv) college and some post-secondary, and (v) and postsecondary graduation. Education levels were summed for the population older than 15 years and divided by the total educated population older than 15 years to obtain a proportion. This proportion was then multiplied by each of the category levels and the resulting values summed to
472
yield a single “unified education” value for each patient. Household social-level indicators, such as the proportion of common-law relationships, single-parent families, rental units, and the unemployed were calculated for each Alberta postal code (and therefore for each study patient). As a secondary outcome, DIs were used. DIs are sexadjusted, income-corrected aggregates of several SES indicators,15 combined into a single value via primary component analysis,17 which allows many dimensions of data to be reduced to a few based on the weighted variability that each dimension comprises. These indices are useful for health planning and research because they link many well-recognized SES variables, allowing for standardized testing across multiple medical and economic fields. Recently, a DI was created for Alberta using 2006 census variables from the DA level of data, comprising 2 reduced dimension sets: material deprivation and social deprivation.22 Data from all eligible patients were entered into a Microsoft Access (14.0.6024.1000; Microsoft Inc, Redmond, Wash.) database. All calculations were performed using SAS statistical software (version 9.2; SAS Institute Inc, Cary, N.C.). Because some patients had both eyes enrolled in the study, analyses used the generalized estimating equation with SAS PROC GENMOD to account for the correlation between 2 eyes of a single patient. Univariate analysis was performed with simple regression, and multivariate analysis was performed to adjust for the effect of factors demonstrating a significant relationship (p o 0.05) on univariate testing that were not otherwise known to be related to cataract severity. Although age would be expected to be related to cataract severity, and adjusting for it may adjust for the outcome itself, it was still considered for inclusion in a multivariate model because younger patients can also present with advanced cataracts. In addition, if access barriers existed for patients with low SES, controlling for age would provide a more conservative estimate of the relationship between cataract severity and SES. As such, a sensitivity analysis was performed to determine the impact of the inclusion of age on the results. No imputation of missing data was performed.
RESULTS During the study period, 1458 cataract extractions were performed (Fig. 1). Of these, 57 were surgeries performed using OZIL (Alcon) handpieces and were therefore excluded. Eight surgeries were secondary IOL insertions (sulcus or anterior chamber IOL) and were also excluded. Of the remaining 1393 cataract extractions, 43 were missing the primary outcome (phaco time) and were also excluded. Only patients with valid and non-retired Alberta postal codes were used in this study, and no out-of-
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
Cataract severity and socioeconomic status—Wesolosky and Rudnisky
Fig. 1 — Socioeconomic status cataract study eye enrollment data. IOL, intraocular lens.
province patients were included in the data set. The missing data amounted to only 3.1% of the total records. The final study population included 1350 sequential phacoemulsification cataract extractions performed on 873 patients. The mean age was 69.6 ⫾ 11.6 years and ranged from 27 to 98 years old. Mean preoperative logMAR VA was 0.60 ⫾ 0.54. The majority of patients were female (53.6%), and the operative side was equally split at 50:50. Although sex (p ¼ 0.124) and operative side (p ¼ 0.878) were not associated with phaco time, age was (p o 0.0001); the parameter estimate on simple regression was positive (0.425 ⫾ 0.038), indicating, appropriately, that cataracts worsen with age. The average IOL power implanted was 19.9 ⫾ 3.9 D and was also predictive of cataract severity (p ¼ 0.005); with a positive parameter estimate (0.337 ⫾ 0.118), the data suggest that phaco
time was longer for more hyperopic eyes. Because phaco time is used as a surrogate for cataract severity, this suggests that in this study, cataracts were more severe in hyperopic eyes. Preoperative VA was also highly predictive of cataract severity (p o 0.0001), which helps to validate the use of phaco time as a surrogate indicator; Figure 2 charts the relationship between phaco time and VA, indicating that as vision declines, phaco time increases (p o 0.0001). A resident participated in the phacoemulsification component of the surgery in 14.4% of eyes. Mean resident phaco time was 34.9 ⫾ 18.0 seconds and significantly longer than the supervising surgeon: 26.3 ⫾ 15.4 seconds (p o 0.001). Anaesthesia was either topical (58.0%), retrobulbar (41.2%), or general (0.8%). Topical anaesthetic was not associated with cataract severity (p ¼ 0.086), but retrobulbar anaesthetics were (p ¼ 0.045). However, during the study period, the surgeon gradually transitioned from primarily retrobulbar anaesthetics earlier in his career to primarily topical anaesthetics. Because simple regression of cataract severity to length of time in practice demonstrated a significant relationship (p o 0.0001), when the same analysis was performed controlling for anaesthetic type, there was, in fact, no relationship between cataract severity and topical (p ¼ 0.473) or retrobulbar (p ¼ 0.068) anaesthetics; the estimate for general anaesthetics was null because of sparse data. Comorbidities (Fig. 3) such as patient use of Flomaxassociated floppy iris syndrome were present in 3.3% of eyes, weak zonules in 3.2%, small pupils in 2.9%, pseudoexfoliation in 2.1%, and iris synechiae in 0.6%. Surgical adjuncts (Fig. 4) included the use of vision blue in
Fig. 2 — Relationship between visual acuity and phacoemulsification time at cataract surgery. MSE, mean square error. CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
473
Cataract severity and socioeconomic status—Wesolosky and Rudnisky
Fig. 3 — Percentage of eye comorbidities associated with cataract surgery participants. FIS, floppy iris syndrome.
16.5% of eyes, iris hooks in 2.4%, and placement of a capsular tension ring in 1.1%. The cataract was hydrated (white) at the time of surgery in 0.5% of eyes. A chop technique was used in 11.5% of eyes, and the remainder were removed using a divide and conquer technique. Posterior capsular rupture occurred in 0.5% of eyes, 80% of which underwent pars plana vitrectomy. A suture was placed at the end of the procedure in 7.3% of eyes. Of all these factors, the use of vision blue (p o 0.0001), a chop technique (p o 0.0001), and weak zonules (p ¼ 0.008) were statistically significant predictors of cataract severity. Discrete socioeconomic indicators
The mean median household income for patients in the study was $61 032 ⫾ $26 824, which is approximately the same as the Alberta average median household income of $63 988 ⫾ $288.23 Median income was negatively associated with cataract severity: As income increased, cataract severity decreased (Table 1; p ¼ 0.0044). Because resident participation, surgeon experience (time in practice), IOL power, and the presence of weak zonules were all confounding factors of phaco time, an adjusted analysis was performed to control for their effects, because they are not truly associated with cataract severity. When adjusting for these factors, cataract severity decreased as median income increased, irrespective of whether age was included (p ¼ 0.001; Table 2) or not (p ¼ 0.0008). The average level of education attained overall was 2.68 ⫾ 0.49 (Table 1); 5.1% of patients lived in a community where the majority had less than a high-school education, 71.7% lived in a community where the majority had a high-school diploma, 22.8% lived where the majority had college or incomplete post-secondary education, and 0.4% lived where the majority had graduated from a post-secondary institution. There was no relationship between education and cataract severity (p ¼ 0.941). The mean proportion of common-law couples in the study was 8.7% ⫾ 5.0% (Table 1). Simple regression demonstrated a statistically significant relationship
474
between cataract severity and a higher proportion of common-law couples living in a patient’s community (p ¼ 0.049). A controlled analysis again confirmed that as the proportion of common-law couples in a community increases, so too does cataract severity (p ¼ 0.012), even if age is included (p ¼ 0.0003; Table 3). In comparison, the overall proportion of single-parent families was higher at 17.2% ⫾ 10.8%, but there was no association with cataract severity (p ¼ 0.376). Similarly, there was no association between the proportion of rental units (μ: 29.6% ⫾ 29.4%; p ¼ 0.120) in the DA. Another social parameter that was associated with cataract severity was the unemployment rate (p ¼ 0.0014) in a patient’s community, which was 5.3% ⫾ 6.5% overall. The higher the unemployment rate, the worse the cataract was (p ¼ 0.002), even after adjustment for age (p ¼ 0.006; Table 4). Deprivation indices as socioeconomic indicators
An unadjusted analysis including both social and material deprivation indices did not demonstrate a statistically significant relationship with cataract severity (p ¼ 0.051 and 0.325, respectively). However, a controlled analysis demonstrated a relationship between the material DI when age was included (p ¼ 0.022) and the social index when age was excluded (p ¼ 0.031). In both instances, as the levels of deprivation increased, so too did cataract severity.
DISCUSSION This is the first study in Canada to investigate the relationship between SES and cataract severity. Because statistically significant relationships were detected on univariate, multivariate, and exploratory sensitivity analyses using both discrete indicators and deprivation indices, this demonstrates a robust relationship between increasing cataract severity and lower SES.
Fig. 4 — Surgical adjuncts used in cataract extraction. CTR, capsular tension ring.
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
Cataract severity and socioeconomic status—Wesolosky and Rudnisky Table 1—Relationship between socioeconomic indicators and cataract severity (univariate analyses) Variable Median income (/$1000) Education Proportion of common-law couples (%) Proportion of single-parent families (%) Unemployment rate (%) Proportion of band housing or rental units (%) n
Mean ⫾ SD
Parameter Estimate
SE
95% Confidence Limit
p Value*
⫾ ⫾ ⫾ ⫾ ⫾ ⫾
–0.05 –0.08 23.11 4.11 21.53 2.92
0.02 1.01 11.42 4.64 6.75 1.88
–0.09, –0.02 –2.06, 1.91 0.722, 45.49 –4.99, 13.20 8.31, 34.76 –0.76, 6.60
0.004 0.941 0.043 0.376 0.001 0.120
61032 2.68 8.72 17.21 5.31 29.6
26824 0.49 4.97 10.78 6.54 27.4
Simple regression using the generalized estimating equation.
The use of phaco time as a surrogate for cataract severity is a novel approach, but is it appropriate? In a retrospective study, lens assessment tools that are used in major population studies24–26 are not an option. From a scholarly perspective alone, it stands to reason that denser lenses require more phacoemulsification energy to disassemble, and as such, phaco time should rise fairly linearly with cataract severity. One potential weakness of phaco time as a surrogate is that it relates more to lens density than to opacity; cortical and posterior subcapsular cataracts can cause a significant reduction in VA in the absence of nuclear sclerosis. As such, it is possible that patients underwent surgery for visually significant cataracts that are less dense, and would thus have a shorter phaco time. Although cataract type (nuclear sclerosis, cortical and/or posterior subcapsular) was not collected in this data set, we do not believe this omission affects the results or conclusions: If there were many study patients with this inverse relationship, one would expect poor correlation between preoperative VA and phaco time. However, this was not the case: VA demonstrated a very strong correlation with phaco time (p o 0.0001). As such, this would suggest that there were few patients who underwent surgery for a cataract with only cortical or posterior subcapsular components. Another approach to assessing the appropriateness of phaco time as a surrogate is to consider the relationship between it and other more easily accepted surrogates for cataract severity. VA was highly correlated to phaco time (p o 0.0001) but was not used as the surrogate because final postoperative best corrected VA was not available for all patients; preoperative VA alone would not accurately reflect true cataract severity because it would overestimate it in eyes with ocular comorbidities such as age-related macular degeneration. Vision blue is a surgical adjunct Table 2—Controlled analysis of the relationship between cataract severity and median income Variable Age Resident participation Time in practice Intraocular lens power Weak zonules Median income (/$1000)
used to improve the visualization of the capsule when the red reflex is poor, as is often the case with dense, advanced cataracts. Although it too was highly correlated to phaco time (p o 0.0001), it was not used as the surrogate because, as a dichotomous indicator, it detects only very severe cataracts and would therefore not provide any stratification of milder disease. Chop technique, which was used in this study only in eyes with palpably dense lenses (the decision to chop was made after starting to groove the lens intraoperatively), was also associated with phaco time (p o 0.0001). Like vision blue, it would not provide stratification of milder disease and would, therefore, be a suboptimal surrogate. Lastly, although weak zonules were associated with phaco time (p ¼ 0.008) and could possibly result from years of stress induced by a dense and “weighty” lens, they can also occur because of trauma, which is more commonly suffered by those of lower SES.27–29 In the end, it was included in the adjusted models to minimize bias in favour of a relationship between cataract severity and SES. Given that 3 clear indicators of cataract severity are associated with phaco time, we believe it is an excellent surrogate. Just as with our choice of outcome, justification of our socioeconomic predictors is required. Many social factors can affect medical outcomes: in-home support, geographic proximity to care centres, and access to primary care.27,30,31 Hence it is not surprising that cataract severity is not unlike other medical problems1,9,10,30 in that patients with lower SES have reduced health status and generally less favourable health outcomes. Although this study looked at only a few discrete indicators such as median income, the proportion of common-law couples in a community, and unemployment rate, there may be other related factors that are not aggregated in census data, which is a limitation encountered when using postal code Table 3—Controlled analysis of the relationship between cataract severity and the proportion of common-law couples living in a patient’s community
Parameter Estimate
SE
95% Confidence Limits
p Value
0.382 7.204
0.039 1.354
0.305, 0.459 4.550, 9.859
o0.0001 o0.0001
0.208 0.059
0.055 0.117
0.100, 0.316 –0.171, 0.289
0.0002 0.613
14.817 –0.054
5.025 0.016
4.968, 24.665 –0.086, –0.022
0.003 0.001
Variable Age Resident participation Time in practice Intraocular lens power Weak zonules Proportion of commonlaw couples
Parameter Estimate
SE
0.399 7.189 0.204 0.058 14.773 35.588
0.038 1.333 0.055 0.116 5.015 10.050
95% Confidence Limits 0.325, 4.577, 0.096, –0.169, 4.943, 16.891,
0.474 9.802 0.311 0.284 24.603 56.284
p Value o0.0001 o0.0001 0.0002 0.617 0.003 0.0003
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
475
Cataract severity and socioeconomic status—Wesolosky and Rudnisky Table 4—Controlled analysis of the relationship between cataract severity and the unemployment rate in a patient’s community Variable Age Resident participation Time in practice Intraocular lens power Weak zonules Unemployment rate
Parameter Estimate
SE
95% Confidence Limits
p Value
0.384 6.895
0.039 1.342
0.307, 0.461 4.265, 9.526
o0.0001 o0.0001
0.207 0.066
0.055 0.117
0.099, 0.316 –0.163, 0.295
0.0002 0.574
14.833 19.996
5.051 7.293
4.932, 24.734 5.702, 34.290
0.003 0.006
data. However, because medical administrative databases —or medical records, for that matter—do not carry overt SES identifiers, use of area aggregated social and material information is the only reliable and horizontally maintained (i.e., year to year) data set. Furthermore, it is impractical to collect direct SES information for the purposes of health monitoring and health care planning.7,18,32,33 Indeed, although individual SES statistics are the gold standard, area-based population information (i.e., DA/postal codes) shows similar trends as that of individual, direct metrics.21 This relationship permits the use of area-level aggregate data to determine trends that likely apply to the individual. As such, we believe the results of this study to be discretely generalizable. Increasing age (p o 0.0001), resident participation (p o 0.0001), and surgeon experience (p ¼ 0.0002) were associated with cataract severity but are not related to SES. Interestingly, the coefficient for surgeon experience, as measured by the surrogate “time in practice,” was positive, indicating that with more experience, the surgeon saw more severe cases. Inclusion of this variable in the analysis helps to reduce selection bias; although referral sources may have selected more severe cases to send to a more experienced surgeon, the relationship between cataract severity and SES is still observed after adjustment. With respect to resident participation, the observed relationship is a common sense finding: a less proficient surgeon will have greater phaco time, and controlling for this confounding relationship helps to clarify the relationship between cataract severity and SES. Because cataract is an age-related disease, it is also a common sense finding that cataract severity increases with age. Because of this, it was unclear whether age should be included in the adjusted analyses; controlling for age may control for the outcome itself, although given that traumatic cataracts are more common in younger patients and that trauma itself is more common in those of lower SES,27–29 exclusion may have induced bias. Irrespective of whether age was included in the analyses, increasing cataract severity was associated with lower SES. This suggests that age is not a consideration when discussing the impact of SES on the development of visually significant cataracts.
476
Although the relationship between median income and unemployment rates and cataract severity are more easily understood, the strongest predictor, as indicated by the regression coefficients, was the proportion of common-law couples in a patient’s community. This finding, although at first perplexing, requires more study; it is our hypothesis that this measure is a surrogate for other, more directly related, unmeasured factors, such as domestic violence, smoking, alcohol abuse, and substance abuse. This study provides strong statistical evidence for a connection between cataract severity and SES; as SES decreases, cataract severity increases. This result should be considered by government and health policy experts when discussing the suitability of private health care as a solution to contemporaneous health care issues: the least advantaged require the most help and yet are least able to pay for it.
Disclosure: The authors have no proprietary or commercial interest in any materials discussed in this article. Acknowledgements: Dr. Chris Rudnisky would like to acknowledge Dr. Ichiro Kawachi, whose outstanding and captivating Summer Session Society and Health Lectures at Harvard University were the inspiration and motivation behind this research study. REFERENCES 1. Lewallen S. Poverty and cataract—a deeper look at a complex issue. PLoS Med. 2008;5:e245. 2. Pampalon R, Hamel D, Gamache P. Health inequalities in urban and rural Canada: comparing inequalities in survival according to an individual and area-based deprivation index. Health Place. 2010;16:416-20. 3. Perruccio AV, Badley EM, Trope GE. A Canadian population-based study of vision problems: assessing the significance of socioeconomic status. Can J Ophthalmol. 2010;45:477-83. 4. Meddings DR, Hertzman C, Barer ML, et al. Socioeconomic status, mortality, and the development of cataract at a young age. Soc Sci Med. 1998;46:1451-7. 5. Tielsch JM, Sommer A, Katz J, Quigley H, Ezrine S. Socioeconomic status and visual impairment among urban Americans. Baltimore Eye Survey Research Group. Arch Ophthalmol. 1991;109:637-41. 6. Tielsch JM, Sommer A, Witt K, Katz J, Royall RM. Blindness and visual impairment in an American urban population. The Baltimore Eye Survey. Arch Ophthalmol. 1990;108:286-90. 7. Sit AJ. Socioeconomic factors and vision health in Canada. Can J Ophthalmol. 2010;45:441-2. 8. Shortt SED, Shaw RA. Equity in Canadian health care: does socioeconomic status affect waiting times for elective surgery? CMAJ. 2003;168:413. 9. Demeter S, Leslie WD, Lix L, MacWilliam L, Finlayson GS, Reed M. The effect of socioeconomic status on bone density testing in a public health-care system. Osteoporos Int. 2007;18:153-8. 10. Demeter S, Reed M, Lix L, MacWilliam L, Leslie WD. Socioeconomic status and the utilization of diagnostic imaging in an urban setting. CMAJ. 2005;173:1173-7. 11. Alter DA, Naylor CD, Austin PC, Chan BTB, Tu JV. Geography and service supply do not explain socioeconomic gradients in angiography use after acute myocardial infarction. CMAJ. 2003;168:261. 12. Wu R, Wang JJ, Mitchell P, et al. Smoking, socioeconomic factors, and age-related cataract: The Singapore Malay Eye Study. Arch Ophthalmol. 2010;128:1029-35.
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
Cataract severity and socioeconomic status—Wesolosky and Rudnisky 13. Younan C, Mitchell P, Cumming R, Rochtchina E. Socioeconomic status and incident cataract surgery: the Blue Mountains Eye Study. Clin Experiment Ophthalmol. 2002;30:163-7. 14. Holladay JT. Proper method for calculating average visual acuity. J Refract Surg. 1997;13:388-91. 15. Pampalon R, Hamel D, Gamache P, Raymond G. A deprivation index for health planning in Canada. Chronic Dis Can. 2009;29:178-91. 16. Krefis AC, Schwarz NG, Nkrumah B, et al. Principal component analysis of socioeconomic factors and their association with malaria in children from the Ashanti Region, Ghana. Malar J. 2010;9:201. 17. Vyas S, Kumaranayake L. Constructing socio-economic status indices: how to use principal components analysis. Health Policy Plan. 2006;21:459-68. 18. Tello JE, Jones J, Bonizzato P, Mazzi M, Amaddeo F, Tansella M. A census-based socio-economic status (SES) index as a tool to examine the relationship between mental health services use and deprivation. Soc Sci Med. 2005;61:2096-105. 19. Frohlich N, Fransoo R, Roos N. Health service use in the Winnipeg Regional Health Authority: variations across areas in relation to health and socioeconomic status. Healthc Manage Forum. 2002;15:9-14. 20. Odoi A, Wray R, Emo M, et al. Inequalities in neighbourhood socioeconomic characteristics: potential evidence-base for neighbourhood health planning. Int J Health Geogr. 2005;4:20. 21. Pampalon R, Hamel D, Gamache P. A comparison of individual and area-based socio-economic data for monitoring social inequalities in health. Health Rep. 2009;20:85-94. 22. Pampalon R, Hamel D, Gamache P. 2006 Alberta Deprivation Index Variables. Microsoft Excel; 2011. 23. Household income groups (22) in constant (2005) dollars and household type (11) for the private households of Canada, provinces, territories, census metropolitan areas and census agglomerations, 2000 and 2005 - 20% sample data. StatsCan; 2006.
24. West SK, Rosenthal F, Newland HS, Taylor HR. Use of photographic techniques to grade nuclear cataracts. Invest Ophthalmol Vis Sci. 1988;29:73-7. 25. Ortiz D, Alio JL, Ruiz-Colecha J, Oser U. Grading nuclear cataract opacity by densitometry and objective optical analysis. J Cataract Refract Surg. 2008;34:1345-52. 26. Huang W, Chan KL, Li H, Lim JH, Liu J, Wong TY. A computer assisted method for nuclear cataract grading from slit-lamp images using ranking. IEEE Trans Med Imaging. 2011;30:94-107. 27. Roos LL, Walld R, Uhanova J, Bond R. Physician visits, hospitalizations, and socioeconomic status: ambulatory care sensitive conditions in a Canadian setting. Health Serv Res. 2005;40:1167-85. 28. Dannenberg AL, Parver LM, Fowler CJ. Penetrating eye injuries related to assault: the national eye trauma system registry. The National Eye Trauma System Registry. Arch Ophthalmol. 1992;110:849-52. 29. Parver LM, Dannenberg AL, Blacklow B, Fowler CJ, Brechner RJ, Tielsch JM. Characteristics and causes of penetrating eye injuries reported to the National Eye Trauma System Registry, 1985-91. Public Health Rep. 1993;108:625-32. 30. Dunlop S, Coyte PC, McIsaac W. Socio-economic status and the utilisation of physicians' services: results from the Canadian National Population Health Survey. Soc Sci Med. 2000;51:123-33. 31. Wong ST, Regan S. Patient perspectives on primary health care in rural communities: effects of geography on access, continuity and efficiency. Rural Remote Health. 2009;9:1142. 32. Duncan GJ, Daly MC, McDonough P, Williams DR. Optimal indicators of socioeconomic status for health research. Am J Public Health. 2002;92:1151-7. 33. Frohlich N, Mustard C. A regional comparison of socioeconomic and health indices in a Canadian province. Soc Sci Med. 1996;42: 1273-81.
CAN J OPHTHALMOL — VOL. 48, NO. 6, DECEMBER 2013
477
