Abstract
Purpose.:
To characterize refractive error, presbyopia, and spectacle correction among adults aged ≥40 years in Timor-Leste in 2010 and examine the impact of the National Spectacle Program by comparing results to those obtained 5 years earlier.
Methods.:
A population-based, cross-sectional survey sampled 50 clusters of 45 people each. Participants with uncorrected and undercorrected refractive error (presenting visual acuity <6/18, but ≥6/18 in the better eye with pinhole), uncorrected and undercorrected presbyopia (<N8 binocular), corrected refractive error (≥6/18 in the better eye with presenting spectacles), and corrected presbyopia (binocular ≥N8 with presenting spectacles) were identified. Willingness to wear and to pay for spectacles was elicited. Data were adjusted for sex, age, and urban/rural domicile, to compare results from 2005 and 2010.
Results.:
A participation rate of 89.5% (n = 2014) was achieved. The 2010 sex-age-domicile–adjusted parameters were as follows: 3.7% (95% confidence interval [CI], 2.8%–4.5%) met refractive error need; 9.6% (95% CI, 8.3%–10.9%) unmet refractive error need; 27.6% (95% CI, 22.2%–32.9%) refractive error correction coverage; 8.6% (95% CI, 7.4%–9.8%) met presbyopia need; 41.8% (95% CI, 39.7%–44.0%) unmet presbyopia need; and, 17.0% (95% CI, 14.7%–19.3%) presbyopia correction coverage. Refractive error correction coverage was significantly higher in 2010 than in 2005 (8.0%; 95% CI, 4.5%–11.6%), but presbyopia correction coverage was unchanged. Almost all the 2010 sample (99.5%) were willing to wear spectacles if needed. Sex-age-domicile–adjusted willingness to pay at least US$1.00 for spectacles was 56.9% (95% CI, 54.7%–59.1%) in 2010, significantly greater than in 2005 (47.2%; 95% CI, 44.5%–49.9%).
Conclusions.:
The National Spectacle Program has not produced uniform improvement across all service indicators. Evidence-based enhancements may now be initiated.
Since the Vision 2020 Right to Sight Initiative was established in 1997,
1 there has been increasing recognition of the contribution that uncorrected refractive error makes to the burden of global vision loss.
2 To a lesser extent, the same has been true of presbyopia.
3 –10
A 2005 cross-sectional eye health survey of adults aged ≥40 years found a high prevalence of blindness and low vision in Timor-Leste, with cataract and refractive error the predominant causes.
11 The same survey found that uncorrected presbyopia may also have been present in up to a third of this adult population.
10
Drawing on this and other related research, Timor-Leste's first National Eye Health Strategy (NEHS) was developed for the period 2006 to 2011.
12 Following this, the nascent National Spectacle Program (NSP) was aligned with the NEHS, with goals that included improving the equity, affordability, and sustainability of refraction and spectacle-dispensing services.
13 Implemented by a local nongovernment organization, Fo Naroman Timor-Leste (FNTL), with support from the Ministry of Health, the strategies of the NSP included increasing outreach refraction services, introducing subsidized prices for spectacles, and undertaking health education relating to the use of eye services and spectacles.
13 The Ministry of Health established eight permanent eye clinics throughout rural Timor-Leste by 2010, and while these participated in the NSP, dispensing levels were low, and outreach services to the subdistrict level were conducted regularly from only one of them.
The majority of Timor-Leste's 1.1 million people reside in rural areas, and 49.9% live below the poverty line.
14 In 2005, half of the persons surveyed were not willing to pay the NSP's lowest price for spectacles (US$1.00),
10 and so another willingness-to-pay survey was undertaken in 2006 to explore lower price points. The latter survey validated US$0.10 as the price most people were willing and able to pay.
15 Subsequently, FNTL has, with reasonable success, implemented a pricing structure based on this finding.
Beyond the NSP, a private sector has emerged in the capital, Dili, since independence in 2002, with three general retailers selling ready-made spectacles by 2006 and an Indonesian optical retail chain establishing a store in 2009. The cheapest price of these retailers (US$7.00) remains beyond the reach of many, and the private sector has not yet established a presence outside Dili.
The Timor-Leste Eye Health Survey 2010 (TLEHS2010) provided an opportunity to examine the impact of NSP services correcting refractive error and presbyopia during the 5 years since the 2005 survey.
Using the most recent national census data (2004), the survey sample frame was stratified by the four administrative regions (eastern, western, central, and the capital, Dili) and included 173,000 people aged ≥40 years living in Timor-Leste's 12 mainland districts. The enclave of Oecusse district in West Timor was not included because of logistic constraints. Urban residents were considered those living in the Dili administrative region.
Rapid assessment of avoidable blindness (RAAB) software was used to calculate the sample size. On the basis of an anticipated target population prevalence of blindness (presenting visual acuity <6/60 in both eyes) of 7.5%,
11 absolute precision of ±1.5% (20% relative difference), with 95% confidence, a design effect of 1.6, and a response rate of 80%, the sample size was determined to be 2250 persons. Fifty clusters of 45 people were needed, and these were systematically selected from the 424 sucos (villages) within the 12 included districts by using probability proportionate to size sampling.
The selection of individuals within clusters was through the random-walk method. The team leader identified each cluster's initial household by spinning a bottle. Thereafter, consecutive households were approached sequentially in the direction the bottle had indicated. Eligible people were enumerated by trained local fieldworkers until that cluster's 45 participants were enrolled. If an eligible person was absent, with no prospect of returning during the team's visit, the absentee's demographic and socioeconomic data were elicited from a knowledgeable relative or other adult.
Participants attended a central facility, typically a community hall, for administration of the questionnaire and examination.
An interview-based questionnaire, developed in English, translated into Tetun, the local language, and back translated to ensure veracity, was used to collect data.
Participants provided demographic and socioeconomic data, including self-reported literacy (ability to read and write) and the main source of household income (which was categorized from the open-ended response into paid employment/subsistence farming and unemployed categories). Willingness to wear spectacles if necessary to improve vision was assessed, as was willingness to pay for them, using US$1.00, US$0.25, and US$0.10 price points and the binary-with-follow-up technique. To check for internal response validity, participants were then asked the maximum price they were willing to pay.
Presenting visual acuity was measured in daylight in each eye separately by three trained field workers using a single Snellen 6/18 tumbling E on one side of a card and a 6/60 optotype on the other, in accordance with the RAAB protocol.
16 Testing started with the 6/18 E at 6 m, progressing to the 6/60 optotype at 6 then 3 m if at least four of five different orientation showings were not identified at the preceding level. If visual acuity was <6/18 in either eye, the procedure was repeated with a pinhole.
Presenting near visual acuity was assessed binocularly using a logMAR tumbling-E chart (Low Vision Resource Centre, Hong Kong) held 40 cm from the participant's face. The distance of 40 cm was maintained with a string attached to the top of the chart at one end and held taught against the participant's forehead at the other. The smallest line with at least four of the five optotypes read correctly was recorded as near visual acuity. Measurement occurred undercover in daylight, as the lack of electricity at most study sites meant that a lamp could not be used to standardize illumination. Although unavoidable, these lighting conditions may have led to pupillary miosis and artificially good near vision in some participants.
Comprehensive external and intraocular examinations were performed on all participants by an ophthalmologist. This included magnified assessment of the anterior segment and dilated indirect ophthalmoscopy.
Blindness was determined to be presenting visual acuity worse than 6/60 for an eye or in the better eye for a person. Low vision was defined as presenting visual acuity of at least 6/60, but worse than 6/18. Impaired vision equated to presenting visual acuity worse than 6/18 (combination of low vision and blindness).
Uncorrected refractive error was considered to be present when an eye presented with visual acuity worse than 6/18 and then improved to at least 6/18 with pinhole. This included undercorrected error where the criteria were fulfilled despite the wearing of spectacles at presentation. Whether presenting with or without spectacles, failure to attain 6/18 with pinhole occasioned a nonrefractive error diagnosis determined by the presence of ocular findings sufficient to explain the vision loss.
Met refractive error need was defined as the percentage of people who presented with refractive error correction and achieved at least 6/18 in the better eye. Unmet refractive error need was defined as the percentage of people who had uncorrected (including undercorrected) refractive error that would improve with appropriate correction to at least 6/18 in the better eye (ascertained by improvement with pinhole to at least 6/18 in this study). Refractive error in need of correction was the combination of met and unmet need. Refractive error correction coverage was defined as the percentage of vision-impairing refractive error that was corrected to 6/18 or better, being the percentage of met need over the combined met and unmet refractive error need.
Participants able to read binocular N8 without correction were determined not to have presbyopia. Met presbyopia need was defined as the percentage of people who presented with presbyopia correction and achieved binocular N8 using it. Unmet presbyopia need was defined as the percentage of people who had either undercorrected or uncorrected binocular near vision worse than N8 that would improve with correction to at least N8. Presbyopia in need of correction was the combination of met and unmet need. Presbyopia correction coverage was defined as the percentage of presbyopia that was corrected to N8 or better binocularly, which is the percentage of met need over the combined met and unmet presbyopia need.
The assumptions relating to refractive error and presbyopia requiring correction, unmet and met need, and correction coverage made in this study were the same as those used and reported previously for Timor-Leste and elsewhere.
7,10,17,18
Data were deidentified and entered into a specifically designed database, with subsequent extensive random checking for entry integrity. Before analysis, missing and outlier data were checked against the survey forms.
Point prevalence estimates and 95% confidence interval (CI) were calculated assuming a normal distribution for met need, unmet need, and correction coverage for refractive error and presbyopia. National census data were used to standardize these estimates for sex, age, and urban/rural domicile and to extrapolate findings to those aged ≥40 years across the entire country. The age-sex–standardized prevalence and coverage rates of the 2010 survey were compared with the equivalent results from the 2005 survey, based on its 95% CI.
Multiple logistic regression models were used to evaluate the independent demographic factors associated with unmet need for refractive error and presbyopia. An additional model was used to determine factors associated with unmet compared to met need. Odds ratios (ORs, with 95% CIs) showed strength of associations (PASW/SPSS Statistics 18.0; SPSS Inc, Chicago, IL). P < 0.05 or nonoverlapping CIs were considered statistically significant.
Refractive Error.
Presbyopia.
It has been suggested that the degree to which refractive error is corrected in a population may be a proxy indicator for the level of development of eye services generally.
19 If this is the case, then the 2010 Timor-Leste refractive error correction coverage (27.6%) suggests better services in that country than in Nigeria (4.4%)
20 and Pakistan (19%),
21 but poorer than in Fiji (68.3%).
7 However, Timor-Leste's NSP was prioritized within the NEHS, and actively monitored and enhanced since 2005, and so using it as a proxy may have overestimated the development of overall eye care service. Also, with different rates of literacy, varying cultural and economic demands for near vision unrelated to reading, and uneven prioritization of presbyopia within eye care services, the significance of Timor-Leste's 17.0% presbyopia correction coverage rate is unclear when compared with that for Kenya (6.3%),
3 Tanzania (6%),
6 Zanzibar (17.6%),
8 India (30%),
4 Fiji (42.2%),
7 and China (51.5%).
9 It is likely that the utility of correction coverage rates is therefore not in the comparison with that of other countries, but in monitoring change over time in a single jurisdiction. Indeed, a 5-year review of correction coverage (which requires determination of met and unmet need) is the only indicator recommended by Vision 2020 for monitoring of refractive error and presbyopia.
22 To the authors' knowledge this article is the first to report the impact of a spectacle program on the basis of such a 5-year follow-up.
Between 2005 and 2010 the refractive error correction coverage for Timorese rural dwellers increased fourfold (
Table 4). Also by 2010, although the change from 2005 did not achieve statistical significance, the proportion of urbanites who required and used appropriate spectacles had increased to 51.9%. Overall, the refractive error correction rate increased significantly, from 8.0% to 27.6%. Given the dearth of other active providers in rural areas, FNTL was likely responsible for these changes. As the major implementing organization for the NSP, FNTL increased output from 1200 spectacles dispensed per annum in 2005 to approximately 9500 in 2010 and raised the proportion dispensed in rural areas from 35% to 75%. It also introduced higher quality spectacles
23 and applied a cross-subsidizing pricing and dispensing model, to address uptake barriers of cost
3,4 and poor quality
9,24 that had been experienced elsewhere and previously in Timor-Leste.
Presbyopia correction coverage rates were not significantly different in 2005 and 2010 for both the urban and rural dwellers and overall (
Table 3). This finding may be the result of a low imperative to seek presbyopia correction from FNTL's more accessible services, reflecting the smaller contribution made by near compared with distance vision to vision-specific quality of life
25 in Timor-Leste.
There was a significant increase in unmet need for rural inhabitants. In part, this may be due to an increase in the number of rural people with undercorrected presbyopia who had not updated their spectacles as corrected near vision deteriorated beyond N8. This result does not necessarily equate to a lesser need for presbyopia correction among rural dwellers. In a country such as Timor-Leste, where illiteracy is common (>40%), presbyopia spectacles are likely often needed for near tasks other than reading. This may explain why, when compared with literacy, illiteracy was no more likely to be independently associated with unmet presbyopia (3.1 times) than with unmet refractive error (3.5 times). Therefore, although N8 was used to determine the presence of presbyopia in both the 2005 and 2010 surveys, this may not be an appropriate threshold in the Timor-Leste context.
The proportion of participants willing to pay at least US$1.00 for spectacles was significantly higher in 2010 than 5 years earlier (
Table 5). Also, the median maximum price people were willing to pay increased from US$0.25 in 2006 to US$1.00, and the proportion willing to pay at least US$0.25 (81.6%) was similar to that willing to pay US$0.10 in 2006 (82.9%). In addition to the economic improvements experienced by some over the 5 years, these increases may be related to greater perceived value of spectacle correction for visual functioning and quality of life, as has been found elsewhere.
8,26 If the NSP were to apply the same rationale to price-setting that followed the 2006 survey
15 and expect to subsidize spectacle dispensing for up to 20% of service users, the findings of the TLEHS2010 suggest that the subsidized price could now be increased from US$0.10 to US$0.25.
Correction of the existing distance (17,700) and near (77,000) vision loss due to refractive error and presbyopia among Timorese aged ≥40 years would require 94,700 spectacles. Then, assuming that each pair remains functional for 3 years and that the population age structure remains constant, approximately 39,100 spectacles would have to be dispensed annually to replace the spectacles in use and maintain near 100% correction coverage rates. This approximates the 37,000 annual dispensing target calculated in 2005.
10 Although the 9500 spectacles dispensed in 2010 by the NSP indicate progress toward this steady state, it is evident that the gap remains large.
Being of the female sex was not independently associated with unmet compared with met refractive error and presbyopia need in Timor-Leste. However, rural refractive error correction coverage rates were significantly lower than occurred among urban dwellers (
Table 3). To maintain the former and redress the latter and other inequities found, the NSP should continue to recognize disadvantage and implement remedial strategies.
Comparable survey methodologies and analysis of the same parameters for 2005 and 2010 have permitted assessment of the impact of Timor-Leste's spectacle services in the intervening 5 years. Unique to date, this assessment fulfilled the Vision 2020 recommendation for 5-year correction coverage rate monitoring.
22 However, more important, although significant service enhancements have occurred since 2005, with greater accessibility and better outcomes for individual users, this study highlights that improvement has not been uniform across all service indicators for all population subgroups. With the second Timor-Leste NEHS in preparation, an opportunity now presents itself to mandate further evidence-based development of the NSP.
Supported by The Fred Hollows Foundation New Zealand.
Disclosure:
J. Ramke, None;
G. Brian, None;
T. Naduvilath, None
The authors thank the TLNEHS2010 survey team for their assistance.