The Nigeria National Blindness and visual impairment survey2005-2007
Overview
• The National Blindness and Visual Impairment Survey in Nigeria was conducted in 2005-2007
(30 months of fi eld work)
• Multi-stage stratifi ed cluster random sampling, with probability proportional-to-size procedures, was used to select a cross-sectional, nationally representative sample of the population
• A total of 13,599 persons aged 40+ were examined across the country (response rate 89.9%).
Prevalence of blindness and visual
impairment
• The prevalence of blindness in the sample using presenting vision (PVA) of <3/60 in the better eye was 4.2% (95% CI: 3.8 to 4.6%) and the prevalence of severe visual impairment (SVI) was 1.5% (95 %
CI: 1.3 to 1.7%). Using best corrected vision (BCVA) in the better eye, 3.4% (95% CI: 3.0 to 3.8%) were blind, 0.8% (95% CI: 0.7 to 1.0%) had SVI, 5.3% (95% CI: 4.9 to 5.8%) had moderate visual impairment, 4.5% (95% CI: 4.1 to 4.9%) had mild visual impairment while 86% (95% CI: 85.2 to
86.8%) were categorised as normal/near normal. Associations of blindness and visual impairment with socio-demographic factors
• The prevalence of blindness among those aged _50 years was 5.5% (476/8702) and 9.3% (476/5125) among those aged _60 years
(PVA <3/60 in the better eye)
• The prevalence of blindness (PVA) increases signifi cantly with increasing age, from 0.8% (95% CI: 0.5 to 1.1%) at 40-49 years to 23.3% (95% CI:
20.2 to 26.7%) among those aged _ 80 years (p =<0.001)
• Females had a higher prevalence of blindness than
males (4.4% vs 4.0%)( p =<0.001)
• Illiterate participants had far higher prevalence of blindness than those who could read and write (5.8% vs 1.5%)(p = <0.001)
• The South West geo-political zone (GPZ) had the lowest prevalence of blindness (2.8%; 95% CI: 2.2 to 3.5%) and the North East GPZ had the highest (6.1%; 95% CI: 4.7 to 7.9%) (p = <0.001)
• The prevalence of blindness and SVI did not differ by urban/rural place of residence (p=0.18)
• In multivariate analysis, age, gender, GPZ and literacy remained signifi cantly associated with blindness after adjusting for age and gender
differences
• It is estimated that 1,130,000 individuals aged _40 years are currently blind in Nigeria (95% CI:1.03-1.25 million). The North West geo-political
zone (GPZ) has the largest number of blind adults (28.6%) being the zone with the largest population. A further 2,700,000 adults aged _ 40 years are
estimated to have moderate visual impairment and an additional 400,000 adults are severely visually impaired. Thus a total of 4.25 million adults aged
40 years in Nigeria are visually impaired or blind
• The prevalence of blindness varied across the different ecological zones being highest in the sahel (6.6%) and the lowest in the delta (3.3%)
• The prevalence of blindness in people of all ages was estimated to be 0.78%.
Introduction
Purpose of the survey
To determine the prevalence, causes and the
magnitude of blindness and low vision in:
1. Individuals aged 40 years and above
2. Children aged 10 to 15 years old.
Specifi c objectives
To obtain information on:
1. Cataract surgical services (i.e. cataract surgical rate; cataract surgical coverage; the visual outcome of different cataract surgical techniques, including that of couching – a traditional surgical procedure)
2. The prevalence of other ocular conditions needing treatment (i.e. glaucoma, and trachomatous trichiasis)
3. The prevalence and type of refractive errors, and estimation of the need for optical services and spectacles
4. The prevalence and causes of conditions requiring low vision services (i.e. those where sight cannot be restored by optical, medical or surgery)
5. Normative data on parameters used in the diagnosis of glaucoma, and for determining optimal intraocular lens power in Nigerian eyes. (Based on a randomly selected sub-group of one in seven participants.)
6. The impact of severe visual impairment and blindness on a) quality of life and b) visual function
7. Gender and socioeconomic differences in health seeking behaviour and barriers to the uptake of services
8. The prevalence of hypertension and obesity, and household access to a protected water supply and sanitation.
Background
In 2002 the World Health Organization (WHO) revised estimates of the global magnitude and causes of blindness which revealed a paucity of
recent data for most countries in the African region1.Though Nigeria is the most populated country in Africa, with a population of 135 million, no national data on the prevalence and causes of blindness exist2. Most data used for planning eye care services are generated either from urban areas where the large eye hospitals are situated3 or from small, focal surveys4-22.These small studies indicate that blindness is likely to be a public health problem4-22 but such data cannot be extrapolated to the entire country as the population is culturally, ethnically and geographically diverse. Similarly, national survey results from other West African countries (e.g. Benin, 1990; The Gambia, 1986, 1996; Cameroon, 1996)23-26 may not be readily comparable to present day Nigeria. Differences in population size, demographic profi les, climate and eye care service accessibility and provision, contribute in determining the frequency and distribution of conditions such as trachoma and onchocerciasis as well as other
causes of visual loss (e.g. cataract, glaucoma). Nigeria is the ninth most populous country in the world and the most populated in Africa27-29. More
than 500 languages are spoken by more than 200 ethnic groups2. The population is projected to increase to nearly 210 million by the year 202527.
The country is divided into six administrative zones (geo-political zones – GPZ), one Federal Capital Territory (FCT) of Abuja and 36 States31 (Figure 1). Each State is subdivided into Local Government Authorities (LGAs), the smallest administrative division, of which there are 774 in the country31.
Nigeria has fi ve ecological zones (river delta, rainforest, transition, savannah and sahel) which are shared by 19 other countries with a total population of 345 million people in West and Central Africa. These ecological variations may have an important bearing on the prevalence and causes of blindness. Life expectancy in 2007 was 46.8 years for males and 48.1 for females 30 and 63% of the population lives in rural areas 28. Adult
literacy rate is 68% and the GDP per capita was 1,150 US$ in 200631 with 70.2% living in poverty (<1 US$ per day)2, 29. Access to eye care services is limited, especially in rural areas and amongst the urban poor. As such it
is imperative that existing resources (human, fi nancial, infrastructure and equipment) are used effectively, targeting the major avoidable causes of
blindness in order that the goals of VISION 2020 are achieved in Nigeria.
Figure 1: States and geo-political zones in Nigeria
Methodology
Details of the methods used have been published in BMC Ophthalmology 32.
Survey planning started in April 2004 and a consensus meeting was held in Nigeria which was attended by the Federal Government of Nigeria,
senior ophthalmologists experienced in population based research and international NGOs. During this meeting the fi rst draft of the protocol was written which was then extensively reviewed by technical experts. Procurement of equipment and recruitment of staff was completed and training of the team for the fi rst phase of the survey was undertaken in January 2005. Training was followed by a pilot survey in two clusters in Kaduna State.
Study protocol
Inclusion criteria: For this survey two separate age groups were recruited, namely adults aged _40 years and children aged 10 to 15 years normally
resident in the enumerated households. According to the population projections for Nigeria, there were an estimated 24.4 million persons in the adult age group (17.7%), and 20.3 million children (14.8%) aged 10 to 15 years in 200633. The majority of the population live in rural areas (68%) while 32% live in urban zones28. Children under the age of 10 years were not included as they would have required specialist equipment, training and
expertise. The age group 40 years and above was targeted as the available evidence indicates that most blindness occurs in this age category34,35. For
example, in the Bangladesh and Pakistan national surveys the prevalence of blindness was very low among participants aged 30-39 years and increased exponentially after the age of 40 years34,35. Sample size: Census data available at the time of planning was from 1991. The annual growth rate
was estimated to be 2.9%29. The target population for the survey was extrapolated from 1991 census data using annual growth rates. 23.6 million of the population were estimated to be aged 40 years and above in 2005, and the number in this age group in each of the six GPZs ranged from 16 to 30 million (Table 1). The following were used to calculate an appropriate
sample size: an assumed prevalence of blindness among those aged 40 years and above (based on previous small surveys) of 5%[12-19; 21];
relative precision: 0.5%; confi dence limits: 95%; response rate of 85% and a design effect of 1.75. The sample size was calculated to be 15,375 persons aged 40 years or above. A sample size for children was not calculated as children were eligible if they lived in the household of an eligible adult.
Sampling procedure: Multi-stage stratifi ed cluster random sampling, with probability proportional-to-size (PPS) procedures, was used to select a cross-sectional, nationally representative sample of the population. The sample was stratifi ed by place of usual residence (urban/rural). As the proportion of people living in rural areas varied by State this was taken into account during stratifi cation. In each GPZ and in the Federal Capital Territory (FCT) of Abuja, the proportion of clusters sampled was based on the proportion of
the national population living in each zone. For this survey, a rural cluster consisted of inhabited settlements with a population of <20,000 (village)
while an urban cluster comprised a ‘ward’ in a habited settlement with a population of >20,000. A total of 310 cluster sample sites were selected by
PPS, of which 226 (72.9%) were in rural areas and 84 (27.1%) were urban. This sampling strategy was advantageous in terms of time, transport, enumeration and subsequent examination in each cluster.
A total of 50 individuals aged 40 years and older were enumerated in each cluster. In small villages, if there were less than 50 eligible adults the nearest village was included and enumerated until the requisite number were identifi ed. The sample of children included those who resided in the
households of the adults who had been enumerated for the study. Information was also collected on the causes of blindness among those
aged less than 10 years and 16 – 39 years by asking the head of the household. Ethical and governmental approval: The protocol was reviewed by all the funding agencies (Sightsavers International, CBM and Velux Stiftung). Ethical approval was obtained from the London School of Hygiene and Tropical Medicine. The Federal Government of Nigeria also approved
the survey and provided logistic support and advocacy for the study. A Project Advisory Committee (PAC) was formed to guide the survey. Membership included Federal Ministry of Health offi cials, the NPPB coordinator, international NGOs based in Nigeria, leading ophthalmologists and academics, the survey team coordinator and ICEH staff. The PAC met periodically to review progress and solve outstanding problems. The SSI Nigeria Country Offi ce provided all the logistic and administrative support for the survey. Training and pilot studies There were two survey teams who worked concurrently. Each team had two ophthalmologists (designated as a community ophthalmologist and a clinical ophthalmologist) and one optometrist. These “core staff” remained practically unchanged throughout the fi eld work. Other personnel were recruited for each of the six GPZs i.e. two ophthalmic nurses, four enumerators, one liaison offi cer, and one interviewer. Each team had one cook and two drivers who also did not change throughout the survey. A central coordinating team was also drawn up consisting of a National Coordinator, one Finance Advisor and two data entry clerks. All survey team members underwent intensive training for two weeks at the start of the survey which was led by technical experts from ICEH. A detailed manual was developed and given to each team member. The manual covered details of all the methods, guidelines for completing the data collection forms, and information on the duties and responsibilities of all survey personnel. Training included diagnostic algorithms for identifying the
principal cause of visual impairment, and the importance of quality and of team work. Training was repeated before fi eld work started in each of
the six GPZs and included pilot studies in two clusters. Data from the pilot studies were analysed immediately and fed back to all the team members.
Inter-observer agreement exercises were undertaken for the ophthalmologists, optometrists and the ophthalmic nurses. Wherever agreement was below expectations, personnel were replaced if retraining did not improve agreement. Studies were done in the clinic as well as in the community so that performance could be evaluated in a ‘real-life’
scenario. Overall agreement between pairs of personnel with similar skills was good.
Survey data collection
Mapping and identification of cluster segment for survey Liaison Offi cers visited survey villages in advance where they met village elders to explain the purpose and procedures of the survey, to obtain consent and to request full participation of all eligible persons. Enumeration procedures Meticulous enumeration is of crucial importance in a cross sectional/ prevalence survey, providing the correct denominator for determining blindness and low vision rates. If a house was locked at the fi rst visit by the enumeration team, information was given to the neighbours that the team would return later in the day. Repeat visits were made the same day to gather information about the locked house.
If contact could not be established after two visits the household was categorised as a non- responding household. Typically two to three days were required for each cluster. At the examination site set up in each cluster the interviewer systematically identifi ed one out of every seven adults that reported to the examination site for a detailed eye examination for collecting
normative data (yellow card). The purpose of the normative database was to determine the distribution of ocular variables in normal adult Nigerian eyes (e.g. intraocular pressure, cup disc ratio, etc) to give a range of values which could be considered normal for this population. Oral informed consent was obtained from each participant by the enumerators and interviewer.
Personal and demographic data were recorded prior to eye examination by a trained interviewer. All subjects had their blood pressure (three readings)
height and weight measured. All participants underwent distance visual acuity measurement with a reduced logarithm of minimum angle of
acuity (logMAR) tumbling “E” chart[36,37]. The reduced LogMAR E chart was used because of ease
Results
Adult population surveyed
Response rates: The overall response rate was 89.9% (range across GPZs: 88.2 – 91.1%). This high response rate means that the fi ndings can be generalised to the Nigerian adult population across the country. The age and gender distribution of
those enumerated and examined showed that women were over-represented in the younger age group (40-49 years) but the gender differences were not signifi cant at other ages (Table 3). The mean age of those examined was 55.9 years (SD ±12.4) being signifi cantly higher for males (56.7; SD ± 12.5) than females (55.2; SD± 12.2)
(p = 0.001). The mean age of those enumerated but not examined was 51.5 years (SD ± 10.9). Among those not examined, mean age for males was 51.3 years (SD ±11.2) and 51.6 years (SD ± 10.8) for females. There were 1,523 individuals who either did not
complete the examination, refused to be examined, or were not available (76.5%). The age and gender of those enumerated and those examined were similar (Table 3). There were marginally more younger females enumerated (40-49 years) compared to males (Table3). Similarly enumeration and examination rates were marginally lower for
older females compared to males (_ 80 years).
