Report on Rapid Assessment of Avoidable Blindness (RAAB) in Singida Region, Tanzania
Dr. George Kabona
September 14th, 2017
Contents
Background
Methodology
Study design
Study area
Study population
Exclusion criteria.
Sample size
Sampling Procedure
Data Collection
RESULTS
Coverage
Blindness and visual impairment in the sample
Age and sex adjusted blindness and visual impairment
Causes of bilateral visual impairment in Singida region
Cataract related blindness and visual impairment
Cataract Surgical Coverage (CSC)
Visual outcome after cataract surgery
Location of cataract surgery
Barriers to cataract surgery
Refractive errors
Posterior segment diseases
Functional low vision (FLV) requiring low vision services
Discussion
Conclusions
Recommendations
Acknowledgements
Annex 1: List of selected villages
Background
A rapid assessment of avoidable blindness (RAAB) survey was conducted in Singida region, Tanzania in June-July 2017. This population based survey was organised by the Singida Region Health Department with the support of Sightsavers.
No population based survey on blindness and visual impairment has been conducted prior to this RAAB survey in the Singida region. The main objective of the survey was to assess the current situation of blindness and visual impairment in the Singida region and generate evidence base for proper planning of eye care services in the region. Data on disability and wealth was collected alongside the standard RAAB in order to understand how these factors may play a role in eye health and health seeking behaviour.
In this report, all data refer to the population of people aged 50 years and older; unless a specific other age group is mentioned.
Methodology
This is a rapid survey to provide the baseline data for eye care programme to enable regional health management to plan eye care services according to the actual need of Singida communities. In collaboration with partner supporting eye care programme in Singida Region, this survey was conducted under RAAB standards methodology as described inRAAB Version 5 for Windows® - October 2013.
Study design
This was a cross-sectional population-based survey. RAAB is a standardized methodology that is designed to determine the prevalence and causes of visual impairment in an area as well as other information that can help with planning eye care services in the region. The RAAB collects data on visual acuity, lens condition, principal cause of visual impairment (if appropriate), information on cataract surgery (if appropriate) and reasons for not having accessed cataract surgery (if appropriate).
In addition to the standard RAAB questionnaire an additional two sets of questions were administered to the participants to understand their disability and socio-economic status. Both tools have been validated in similar settings and are designed to be used within surveys to disaggregate data. To measure the participants’ disability status we used the Washington Group Short Set of Questions on Disability (WGSS) which are six questions examining the difficulties an individual faces in six functional domains (seeing hearing, walking, upper body movement, cognition and communication). These have previously been used in Tanzania in both the most recent census and in several standalone research studies and are available in Swahili. To measure socio-economic status we used the Equity Tool (TET), which was designed to be integrated in DHS surveys and measure a household’s asset-based wealth. Households can then be assigned to one of the national, pre-defined wealth quintiles, or can be examined in relation to other households within the same survey.
Study area
Seventy seven (77) randomly selected villages in the Singida Region
Study population
In each selected household all persons of 50 years and older residing in the household for 6 months or more over the past years, sharing meals from same kitchen.
Exclusion criteria.
All visitors/residents who live for less than 6 months, refusals
Sample size
Based on the 2012 census data population of Singida region the expected prevalence of blindness in adults aged ≥50 years is conservatively estimated at 4.0% in Sub-Saharan Africa. Allowing for a confidence of 95%, a precision of 20% (i.e. worst acceptable result of 3.2%), with a design effect of 1.5 for clusters of 50, and 10% non-response the required sample size is 3835, people age 50 and above. (RAAB 5 software).
In total 77 clusters of 50 adults aged ≥50 years will be required for this survey.
Sampling Procedure
A simple random sample of all people >50 years in Singida Region was not practical. Therefore, we used cluster sampling methodology. A list of all villages in Singida Region with their populations was obtained from the official 2012 Tanzanian census and used as the sampling frame. This information was entered in the integrated RAAB sampling software. Weenrolled50 people from each village. In order to ensure that all people have an equal chance of being selected, we select villages by a probability proportionate to size method. In order to get 3,835 people we selected 77 villages (clusters) to be surveyed.
In each village, we enrolled 50 people and we selected these by the compact segment sampling method. This was done as follows:
The cluster informer visited the village few days before the team arrived and work with the village leaders who was asked to provide the names of the hamlets within the village and their populations or number of households. This information wasentered in the sampling spreadsheet and two hamlet was selected by probability proportionate to size method. The first hamlet to be selected was the cluster to be surveyed and the second hamlet to be selected was used as a spare hamlet. All households in the selected hamlet were included in the sample sequentially until 50 people aged 50+ are identified. If there were fewer than 50 people of age 50+ in this hamlet then a second hamlet (spare hamlet) was visited and sampling continues until person number fifty has been enrolled.
Data Collection
Five trained and standardized teams was used to collect data; each team comprised of an eye care personnel as examiner, an ophthalmic assistant/ophthalmic nurse, a community worker (assist within each village), a disability/equity data collector and a driver.
In a selected hamlet, door to door visit to households was done by the team, starting from one edge of the segment (selected arbitrarily) and continue systematically until the segment is completed or 50 people aged fifty years old or more have been enumerated and examined. If an individual was absent at the time of the visit they were enumerated and the house was revisited at the end of the day to carry out the examination. If they were still absent information on their visual status was collected from their family or neighbor.
Prior to examination, the information regarding survey was provided to the participants and consent sought. All consented participants underwent ophthalmic examination in their homes, including measurement of visual acuity (VA) in daylight with a tumbling-E chart, lens examination with a direct ophthalmoscope in a darkened house. Pupil dilation was only used if judged necessary by the examiner. For every individual examined the data collected was stored within RAAB app which was on an android phone, which were keyless, touch-screen, mobile devices (encrypted and password protected mobile units).
In addition, each participant enrolled onto the RAAB was administered the WGSS of questions and the TET following ophthalmic examination.
The WGSS and the TET data was collected using keyless, touch-screen, ultra-mobile devices (encrypted and password protected smartphone units). All data was backed-up daily after data has been collected, the user manually transferred the records which have been collected from the field ultra-mobile device to the cloud. The technical advisor backed up the encrypted files on a password protected external drive. This process ensured security and minimize loss of data should a field device be misplaced or broken. These devices offer a cost-effective alternative to paper-based data collection. To ensure data quality, algorithms was incorporated into the software and a midpoint review of the data was conducted.
Quality assurance:
The WGSS and TET data collectors underwent 2 days of training on the ultra-mobile data collection devices to promote standardized objective data collection and questionnaire administration. The 2 days also involved training and sensitization for administering the Washington Group questions. This cadre of staff were recruited from region where the study was taking place.
RESULTS
The field work of the Singida region RAAB survey was conducted from June 27th to July 14th, 2017. Data entry of the survey was done on small android phones.
Coverage
The survey included 3,850 people aged 50 years and older, of whom 3,722 were examined resulting in a coverage rate of 96.7%. 92people (2.4%) were absent during the visit of the survey teams, while 22 people (0.6%) refused to take part in the study and 14 (0.4%) people were not able to do the tests. (Table 1)
In order to assess whether the sample is really representative of the population in the Singida region, the age and gender distribution of the sample population is compared with the latest estimate of the actual population of the Singida region. (Table 2, 3 and Figure 1 and 2)
Table 2: Age and gender composition in the sample population
Table 3: Age and gender composition in Singida region population aged 50+
Figure 1: Proportion of males in survey area population aged 50+ and the sample
Figure 2: Proportion of males in survey area population aged 50+ and the sample
Ideally, the sample population of the survey should have the same composition by age and sex as the total population in the survey area. In this survey, males of the age group from 50 to 69 years are underrepresented in the sample population, whereas the age groups 60 to 79 are slightly overrepresented. For females, 56% of the sample were female, only 45% of the regional population is female so women are truly overrepresented the age groups from 70 to 99 years are slightly overrepresented in the sample population, while the age group of 80 to 99 years is highly overrepresented. As the consequence, the sample prevalence in both sexes does not exactly represent the actual prevalence in the region. Therefore, there is a need to have the prevalence by age and by sex adjusted for the actual populations as given by the 2012 census. This is the closest to the actual prevalence in the Singida region.
Blindness and visual impairment in the sample
The sample prevalence of bilateral blindness with available correction (presenting visual acuity: PVA) of <3/60 in people aged 50+ in the better eye is 4.7% (3.8-5.5); 3.9% in males and 5.2% in females. The prevalence of bilateral severe visual impairment (SVI) is 2.4% and bilateral moderate visual impairment (MVI) is 8.3%. The prevalence of functional low vision, requiring low vision services, is 1.4%. (Table 4)
Table 4: Sample prevalenceof blindness, severe (SVI), moderate (MVI), and functional Low Vision
Age and sex adjusted blindness and visual impairment
The age and sex adjusted prevalence and the estimated number of cases of bilateral blindness, SVI, andMVI are shown in Table 5 below. Accordingly, in Singida region there are an estimated 5,682 people aged 50+ who are bilaterally blind, 3,163 persons aged 50+ are severely visually impaired and another 11,301 people have moderate visual impairment. (Table 5)
Table 5: Age and sex adjusted prevalenceof blindness, SVI and MVI in adults aged 50+
The age-specific prevalence of blindness and low vision by age group clearly demonstrates the increase in prevalence of blindness and visual impairment with age as well as the fact that females are more affected than males (Table 6 and Figure 3).
Table 6: Age-specific prevalence of blindness in Singida region.
Causes of bilateral visual impairment in Singida region
Table 7 shows the causes of visual impairment in people aged 50 years and above with bilateral blindness, SVI, MVI. In people aged 50+, untreated cataract is the most common cause of bilateral blindness (PVA<3/60 in the better eye) with 70.5%, followed by non-trachomatous corneal opacity(7.5%), glaucoma (6.9%), surgical complications (4.6%), phthisis and other posterior segment diseases (both 2.9%) and trachomatous corneal opacity and other globe/CNS abnormalities (both 1.7%). For SVI, untreated cataract is still remains the main cause with 82%, followed by uncorrected refractive errors (6.7%), other posterior segment disease (3.4%) and trachomatous and non-trachomatous corneal opacity (both 2.2%) as well as cataract surgical complications, glaucoma and phthisis (all1.1%). For MVI, untreated cataract with 60.3% is still the main cause followed by uncorrected refractive errors (31.3%).
Table 7: Main causes of blindness, SVI, MVI in people aged 50+ in Singida region
Figure 3: Main causes of blindness in persons as a pie chart
Of all bilateral blindness in Singida region, 94.8% is considered to be avoidable i.e. 71.1% is considered treatable (cataract and Uncorrected refractive error); 12.1% preventable through primary eye care service provision (phthisis, trachoma, non-trachomatous corneal opacity) and a further 11.6% potentially preventable by specialised ophthalmic care services (glaucoma). Posterior segment diseases account for 10.4% of all bilateral blindness. (Table 8 & Figure 5)
Table 8: Blindness, SVI, and MVI in persons by intervention category in Singida region
Figure 4: Main category of blindness in persons
Figure 5: Major actions required to reduce blindness in Singida region, Tanzania.
Cataract related blindness and visual impairment
The age and sex adjusted prevalence of bilateral blindness due to cataract is 1.7% (95% CI: 01.2 – 2.3) resulting in an estimated 2,883 people bilaterally blind due to cataract in Singida region of whom 927 are males (1%) and 1,956 are females (2.6%). There are an estimated 17,663 eyes blind due to cataract in the region.
SVI (BCVA<6/60 – 3/60 in better eye) due to bilateral cataract occurs in 1% (95% CI: 0.7 – 1.2), an estimated 1,650 people aged 50+ in the Singida region: 444 males (0.5%) and 1,206 females (1.6%). There are a total of 4,547 eyes (prevalence 1.4%) with SVI due to cataract: 1,826 eyes in males (1%) and 2,721 in females (1.8%).
MVI (BCVA<6/18 – 6/60 in better eye) due to bilateral cataract occurs in 3.7% (95% CI: 3.3 – 4.2), an estimated 6,260 people aged 50+ in the Singida region: 2,804 males (3.1%) and 3,456 females (4.6%). There are a total of 14,237 eyes (prevalence 4.3%) with MVI due to cataract: 6,494 eyes in males (3.6%) and 7,743 eyes in females (5.1%). (Table 9)
Table 9:The age and sex adjusted prevalence of bilateral blindness,SVI and MVI due to cataract
With an indication for cataract surgery of BCVA <6/60 an estimated 4,533 people aged 50+ require surgery in both eyes; and an estimated 22,209 eyes would require surgery.
With an indication for cataract surgery of BCVA <6/18 an estimated 10,794 people aged 50+ require surgery in both eyes; and an estimated 36,449 eyes would require surgery (Table 10)
Cataract Surgical Coverage (CSC)
The cataract surgical coverage for persons in Singida region is relatively low with only about 53.6% of people blind due to cataract having been operated upon. Males have higher coverage than females with CSC for males being 65.9% and for females 43.9%. The coverage of people with VA<6/60 (42.8%) and VA<6/18 (25.5%) due to cataract indicates that only few people with cataract are operated upon in an early stage.
In Singida region, the cataract surgical coverage for eyes is also very low and only 29.7% of all eyes blind due to cataract have been operated upon: 35.2% in males and 25% in females. Moreover, the coverage for eyes with VA<6/60 is 25.2% and for VA <6/18 is 17%.
Table 10:Cataract Surgical Coverage (CSC)
Visual outcome after cataract surgery
Of all the eyes operated for cataract, 55.2% can see 6/18 or better and 25.6% cannot see 6/60 with available correction. With pinhole correction, the results slightly improve to 62.8% good outcome and 21.2% poor outcome, illustrating the importance of adequate optical correction. Although all operated eyes in the sample population had an IOL implanted, the percentage of those with poor outcome was relatively high. Couching is not practiced in the Singida region. (Table 11 and 12)
Visual outcome of eyes operated during the last 3 years (67.4% good; 21.7% poor) is better compared to those operated 4-6 years ago (60.5% good; 20.9% poor) and also better than those operated 7 or more years ago (35.3% good and 33.8% poor.) It is expected that the post-operative visual acuity outcome would continue to improve as there are improvements over the years in eye care surgical facilities and introduction of better surgical techniques such as MSICS. (Table 13)
In general, the post-operative visual outcome in the first three years after surgery is below the recommended standards of the WHO of >80% good outcome with presenting VA. The main causes of poor outcome are surgical complications (40.4%), sequelae (34.6%), poor patient selection (17.3%), and inadequate optical correction (7.7%). Improvement of the pre-operative examination for proper patient selection could reduce the number of patients with concurrent blinding conditions who may not regain vision after cataract surgery. Patients with concurrent blinding conditions may need counselling to provide them realistic expectations about their expected visual outcome.
Furthermore, improvement in surgical skills and use of appropriate techniques such as introduction of MSICS with individually adjusted IOLs and better refraction services post-operatively are likely to improve the visual outcome considerably. Review of the surgical procedures and skill gaps of individual surgeons could lead to further improvement of the visual outcome. (Table 14)
Table 11: Visual outcome after cataract surgery (PVA)
Table 12: Visual outcome after cataract surgery (BCVA)
Table 13: VA in operated eyes in the sample by years after surgery