Socioeconomic consequences of blinding onchocerciasis in west Africa.

Socioeconomic consequences of blinding onchocerciasis in West Africa T.G. Evans1 Onchocerciasis or river blindness, a major cause of irreversible blindness among adults, has been the focus of international disease control efforts for over 20 years in West Africa. This paper employs the international classification of impairment, disability and handicap (ICIDH) to interpret results from a field study to assess the socioeconomic consequences of onchocerciasis in Guinea in 1987. In a sample of 136 blind, 94 visually impaired and 89 well-sighted persons, decreasing visual acuity is strongly asso- ciated with mobility, occupational and marital handicaps. Individual, household and disease correlates were explored. The implications of these findings for the ICIDH concept of handicap are discussed with particular emphasis on the need to extend analysis beyond the individual when assessing the socioeco- nomic consequences of disabling disease. Introduction This article considers the impact of decreased visual acuity, including irreversible blindness on indivi- duals in areas of highly endemic onchocerciasis or river blindness. In the early 1970s, river blindness gained widespread attention as the likely cause of abandoned villages in the fertile river valleys of West Africa. As the disease was a significant constraint to economic development and also a public health problem, it became the focus of a massive control effort through the Onchocerciasis Control Programme (oCp).a Although much attention was given to control- ling the disease few attempts were made to discover the socioeconomic consequences for the blinded or visually impaired individuals. Young children guid- ing frail, blind men is the popular image portrayed by the media, often accompanied by anecdotal refer- ences to "the living dead"b or the burden of support- ing blind men who are unable to cultivate (1).cd However, some studies indicate that the blind are I Assistant Professor of International Health and Economics, Harvard Center for Population and Development Studies, 9 Bow Street, Cambridge, MA 02138, USA. a Onchocerciasis in the Volta River Basin Area. Report prepared for the governments of Dahomey, Ghana, Ivory Coast, Mali, Niger, Togo and Upper Volta. Unpublished WHO document, WHO/OCP/73.1, 1973. b S6rie F. Social and economic importance of onchocerciasis. Unpublished WHO document, ONCHO/WP/75.1, 1975. c Asafu-Adajaye JBB. Some comments by the delegation of Ghana on the report by the OCP Economic Review Mission. Unpublished paper presented at a meeting of the Economic Review Mission in Ouagadougou, 7-8 September 1978. d Pairault C, Kientz AK. Bemakaha (Pays Senoufo de C6te dlvoire): un village de colonisation spontanee en zone oncho- cerquienne, approche sociologique. Unpublished paper prepared for the OCP, Ouagadougou, OCP/EPI/76.10, 1976. Reprint No. 5625 able to cultivate albeit at reduced productivity (2),e while in a sociological investigation of 17 blind persons in Burkina Faso, one man was quoted as saying that nothing had changed since becoming blind except for not seeing anymore.! As in so much of the socioeconomic research on tropical diseases, the absence of a framework with which to assess the disease impact has inhibited the development of an understanding of what it means to be blind in West Africa. Missing from the literature are studies which describe the socioeconomic profile of the blind population, e.g., the proportion of the blind who are engaged in agriculture, compared with a sighted population.9 In the last 15 years a number of models have been proposed to describe the consequences of chronic disabling conditions. Rosenfield and col- leagues (3), influenced greatly by the then recently published intemational classification of impairment, disability and handicap (ICIDH) (4), proposed that the impact of tropical disease could be viewed in terms of a sequence of impairment, disability, and the social and economic consequences of disability or handicap. More recent developments of the e Bradley AK. The effects of disease on rural economy, social structure and settlement: a case study of onchocerciasis on the Hawal Valley, Nigeria. PhD thesis, University of Liverpool, 1975. f Andre F. Le statut 6conomique et social des aveugles de Wayen, village Mossi. Etudes et documents provisoires, Centre d'Economie et de Recherche pour le Developpement internatio- nal, Clermont Ferrand, 1980. 9 A number of studies in the USA have attempted to develop a socioeconomic profile of the visually impaired and blind (see Kir- chener C. Data on blindness and visual impairment in the US. New York, American Foundation for the Blind, 1985), and the disabled in general (Nagi SZ. An epidemiology of disability among adults in the United States. Milbank quarterly, 1976, 54: 439-467), or more recently in Laplante MP. The demographics of disability. Milbank quarterly, 1991, 69: 55-77. Bulletin of the World Health Organization, 1995, 73 (4): 495-506 © World Health Organization 1995 495 T.G. Evans ICIDH model have suggested that it is the interaction between the "disabled person on the one hand and the community or the environment on the other which gives rise to handicap" (5, 6), handicap being defined as "the disruption of an individual's per- formance of life habits... resulting on the one hand from impairments and disabilities, and on the other from obstacles attributable to environment factors" (Fig. 1) (7).h Following a description of onchocerciasis and the field study undertaken in areas of highly endemic onchocerciasis in Guinea, West Africa, this article identifies the handicaps arising from decreased visu- al acuity, and describes key factors associated with variations in handicap for the affected individuals. These findings are discussed with particular empha- sis on the concept of handicap and the implications for the ICIDH model. The disease Onchocerciasis, or river blindness, is caused by the development in the human dermis of a long thread- like filarial worm called Onchocerca volvulus. The female worms, during their 8-12-year sexually active life, give birth to millions of embryos called microfilariae. These micro-worms are mobile in the subcutaneous layer of skin and can enter the eye. They are responsible for the resulting disease symptoms such as itching, skin lesions and, most importantly, visual impairment and blindness. The disease is transmitted by a blackfly of the genus Simulium, the females ingesting the micro- filariae when they bite an infected human. A few of the ingested micro-worms undergo morphological changes within the fly, reaching the stage of infec- tive larvae. The larvae are transmitted to an uninfec- ted human in a subsequent fly bite. These develop into adult worms, and the cycle is repeated. Infection with the parasite does not mean that all persons will lose their vision. Only a small propor- tion of those who are infected actually go blind. However, as the intensity of infection increases, i.e., the average number of parasites per person in a pop- h The appropriateness of the ICIDH conceptual framework and the term "handicap" for research into the consequences of chronic disease is a subject of ongoing debate (see Bury MR. The ICIDH: a review of research and prospects. International disabilities studies, 1987, 9: 118-121 and Verbrugge LM. Dis- ability. Rheumatic disease clinics of North America, 1990, 16: 741-761), with alternative classifications of the impact of chronic diseases employed elsewhere (see Verbrugge LM, ibid. and Nagi SZ. A comparison of conceptual frameworks of disability. Paper prepared for the Committee on Secondary Prevention of Disability, Institute of Medicine, Washington DC, 1989). Fig. 1. The interaction of disabled persons and the environment to expiain handicap. DISABLED PERSON * Impairment * Disability * Age, sex HANDICAP * Socioeconomic activity * Occupation/mobility ENVIRONMENT * Community participation * Household formation 54 * Rehabilitation programmes * Welfare systems ulation increases, there is a corresponding increase in the prevalence of blindness and severe visual impair- ment. Consequently, small communities where the intensity of infection is high are the areas where the blinding form of the disease is most concentrated. Blackflies breed in fast-moving water, and villages situated close to breeding sites suffer from the most severe effects of the disease. In these hyperendemic communities, the large majority of the population are heavily infected and the prevalence of blindness often exceeds five per cent of the population. Areas of hyperendemic onchocerciasis are found in a belt of savannah across Africa, roughly between 8 and 12 degrees north of the Equator, particularly in West Africa. Methods In 1987, a comparative study of the socioeconomic impact of ocular onchocerciasis was undertaken in Mali, where the Onchocerciasis Control Programme (OCP) had been in operation since 1976, and in Guinea where OCP had not yet begun. This article deals with results from a cross-sectional survey of blind, visually impaired and well-sighted individuals in highly endemic parts of north-east Guinea. The survey consisted of a detailed household census and inventory, and semi-structured interviews with blind and visually impaired individuals within these households. Sighted household members capable of responding to questions about the blind or visually impaired person and his/her impact on the household were also interviewed as a control group. In 1985 and 1986, OCP conducted epidemiologi- cal surveys in well-known foci (8) of onchocerciasis in north-east Guinea, assessing onchocercal infection 496 WHO Bulletin OMS. Vol 73 1995 Socioeconomic consequences of blinding onchocerciasis in West Africa and visual acuity. The results express the prevalence of infection as well as the community mean microfil- arial load (CMFL), which represents the geometric mean intensity of infection for a standardized popu- lation (9). Sjogren's hand test, an accurate and reli- able method for visual screening in illiterate popu- lations (10), was used to assess visual acuity. Blind- ness was defined as the inability to identify the direction of a large hand at 3 metres in the better eye (<3/60), while visual impairment indicated the abil- ity to see the direction of a large hand, but not a small hand, at 6 metres in the better eye (.3/60, and <6/18). Measures of infection, intensity of infection, and prevalence of blindness have been used to identify the level of endemicity in areas of onchocerciasis transmission. Although there is no single definition of hyperendemicity, it is generally agreed that vil- lages with a prevalence of infection of .60%, or a CMFL of 230mf/skin snip, or a prevalence of blindness of >3% are "hyperendemic" (11). A strong positive association exists between the intensity of infection expressed by CMFL and the prevalence of blindness (12, 13). From the OCP epidemiological survey, 14 villages in Guinea were selected from the "hyperendemic" foci of onchocerciasis. Demograph- ic and epidemiological characteristics of the sample villages are presented in Table 1. The foci straddle the Guinean and Sudanic climatic zones, the rainfall in Guinea varying between 800 and 1400 mm per annum, concentrated in a single rainy season from May to October. The actual length of the season and the amount and distribution of rainfall vary enor- mously from one year to the next. Semi-subsistence farming is the primary produc- tive activity in the sample population. Using tradi- tional agricultural methods, the wetland cultivation of rice and the dryland cultivation of millet, maize, manioc and fonioi are common throughout the area. Swamp rice is the principal crop sold for cash needs, while millet and-manioc are reserved for home con- sumption. Livestock including cattle, sheep and goats are relatively scarce owing to the inhospitable conditions in many of the riverine villages. A variety of activities such as fishing, hunting, illicit gold min- ing, and dry season migration to the neighbouring Cote d'Ivoire represent important non-farming pro- ductive opportunities, which help to bridge shortfalls in domestic food production and help young men i Fonio (Digitaria exilis), otherwise known as acha or hungry rice, is an early ripening cereal crop, generally consumed during the "soudure" or hungry season before the main millet or rice harvest (see Jardin C. List of foods used in Africa. 2nd ed. revised (FAO Nutrition Information Document Series). Rome, FAO, 1970). Table 1: Demographic and epidemiological characteris- tics of 14 study villages in Guinea Characteristic Mean number Population per village 198 Households per village 33 Persons per household 7 Prevalence of onchocerciasis 65% Community mean microfilarial load 37 Blind persons per village 11 Prevalence of blindness 6.5% Visually impaired persons per village 9 Prevalence of visual impairment 4.9% amass sufficient funds for the bride price and other expenses associated with marriage. Although the sample villages were an average of 15 km from a principal road, only 40% of them were accessible throughout the year. The villagers used an intricate network of footpaths to reach distant fields, markets and neighbouring villages as other forms of transport were scarce. None of the 14 villages sur- veyed possessed any agricultural extension services, health facilities, or rehabilitation programmes for the blind or for any other victims of disabling diseases. Two of the villages (both on main roads) had pri- mary schools and three others held weekly markets. Water from traditional wells, seasonal rivers and springs was available in all the villages year-round. In each of the 14 villages, all the blind and visu- ally impaired persons were interviewed, unless absent or deceased. A cohort of sighted persons from households with no members suffering from de- creased visual acuity was selected as a control for the blind and visually impaired. The control charac- teristics were age, sex, and household size. A total of 136 blind, 94 visually impaired and 89 sighted per- sons were interviewed. Information gathered through the semi-structured interview included (1) demo- graphic and disease status; (2) visual acuity history, self-perceived health status and mobility; (3) occupa- tional and agricultural activities; and (4) household food security. In addition, 158 sighted persons, 93 from blind households and 65 from visually impaired households, were interviewed. Having observed the onset of blindness or visual impairment, these indi- viduals, whether a spouse, parent or sibling, respond- ed to questions about the blind/visually impaired person and his/her impact on the household. The definition and values of the variables according to visual acuity are presented in Table 2. Three dependent variables were chosen to represent WHO Bulletin OMS. Vol 73 1995 497 Table 2: Variables and their values according to visual acuity Visual acuity Sighted Impaired Variable (n_89) (n=94) Blind (n=1 36) P-valuea Age (years) 51 53 55 NS Age of onset (years) NA 47 41 Sex (male:female) 62:38 78:22 74:26 Healthb No complaints 39% 34% 46% NS III 53% 54% 45% Other 8% 10% 12% Migrationc Yes 28% 21% 14% NS No 72% 79% 86% Village roled Yes 37% 25% 20% No 63% 75% 80% Household responsibility" Head 56% 57% 47% Wife of head 32% 20% 12% Other kin 11% 18% 32% Non-kin 1% 4% 11% Size of household (persons) 7 7 7 NS Household food insecurity' NS Seasonal 83% 85% 76% Chronic 17% 15% 24% Mobility Constrained: 13% 49% 83% -to compound 5% 7% 15% - to village 8% 42% 68% Not constrained 87% 51% 17% Occupational status Active 98% 62% 21% - agriculture 92% 54% 10% -other 6% 8% 11% Inactive 2% 38% 79% Marital status Married 90% 79% 57% Not married: 10% 22% 43% -single 7% 11% 14% -widowed 3% 11% 17% - divorced/inherited - - 12% a * P<0.05, ** P<0.01; P<0.001; NS = not significant. b "Health" refers to self-reported morbidity: "ill" aggregates acute and chronic morbidity regardless of sever- ity; "other" groups all the permanent impairments including the deaf, mute, and physically and mentally dis- abled, but excluding the blind or the visually impaired. c "Migration" has been defined as any departure from the village for a period greater than two weeks in duration. This includes absence for reasons of medical treatment, family visits and begging as well as more traditional migratory activities such as farm labour and petty commerce. d "Village role" refers to administrative, religious, cultural and other official positions of responsibility within the village. e "Household responsibility" describes the relationship of household members to the household head - the head and his wife or wives forming the group with household responsibility, and other kin and non-kin com- prising the group without such responsibility. f Household food insecurity: "seasonal insecurity" refers to household food shortage of up to 4 months duration; "chronic insecurity" refers to food shortage of >4 months. WHO Bulletin OMS. Vol 73 1995 T.G. Evans 498 Socioeconomic consequences of blinding onchocerclasis in West Africa various dimensions of the handicap associated with decreased visual acuity: mobility, occupation, and marital status. All subjects were asked if they walked outside their compounds, and if so, if there were areas where they were not able to do this. Responses on mobility were coded as "constrained", i.e., unable to walk beyond the confines of the village, and "not constrained". The indicator of economic consequen- ces was the occupational status, the responses being categorized as "active", i.e., having an occupation, or "inactive" with no occupation. Marital status referred to whether married or not married, i.e., single, wid- owed or "other". This last category includes divorce and the practice of wife inheritance. In Mande cul- ture, recently widowed women are inherited by their deceased husband's oldest surviving brother or, in some cases, given to destitute men who are frequent- ly victims of leprosy, poliomyelitis or blindness. Initially the data were analysed using contingen- cy tables and chi-squared tests. To control for inter- actions, i.e., the degree to which one independent variable is a proxy for another, and to identify the smallest group of variables which account for the greatest proportion of variance in the dependent vari- able, stepwise logistic regression models were derived for each dependent variable. Logistic regres- sion is a multiple regression technique for nominal data whereby the dependent variable is transformed by a logistic function. In this analysis the indepen- dent variables were introduced into the model in a stepwise fashion, entering the independent or explan- atory variables into the model in ascending order of their ability to account for variance. Variables were assessed at each step to determine whether or not their contribution remained significant, given the effect of other variables in the equation (14). Results The results are presented in two stages: (1) the handi- cap of decreased visual acuity and its determinants; and (2) the variation in socioeconomic consequences within both the blind and the visually impaired popu- lations. The first stage involves comparing the visually impaired and blind populations with a sight- ed cohort for three specific social and economic indi- cators: mobility, occupation, and marital status. Any significant deviations from the behavioural norm of the sighted cohort may imply the existence of a han- dicap due to ocular impairment. Since visual acuity is not likely to be the single variable explain- ing variation in these indicators, its importance rela- tive to other independent factors such as age and sex was assessed. Among the blind, 68% claimed they were unable to walk beyond the village, compared to 42% and 8% among the visually impaired and sighted samples, respectively (see "Mobility", Table 2). Only 10% of the blind compared to 54% of the visually impaired and 92% of the sighted described their occupation as farmer (see "Occupational sta- tus", Table 2). Also blindness and, to a lesser extent, visual impairment were associated with high rates of celibacy, widowhood, divorce and wife inheritance, and low rates of marriage (see "Marital status", Table 2). Fig. 2 demonstrates graphically the strong association between decreasing visual acuity and increasing frequencies of having a mobility con- straint and no occupation, and being unmarried. A wide variety of other independent variables were associated with each of the socioeconomic con- sequences. For example, younger age, marriage, migratory activity, and household food insecurity, in addition to better visual acuity, were all significantly correlated with unconstrained mobility and occupa- tional activity (see Table 3). Upon stepwise logistic regression, visual status was selected as an indepen- dent predictor of each of the three dependent vari- ables, mobility, occupation and marital status. The second stage of analysis involved compar- ing the determinants of socioeconomic function specific to each visual status group. Among the blind, the step 0 likelihood ratio chi-square for mobility (Table 4) indicated a number of significant associations. Unconstrained mobility among the blind was correlated with younger age, and age of onset of blindness, migration activity, occupation, household responsibility, village role, and severe or Fig. 2. Percentage of the population affected by handi- cap, according to level of visual acuity. Adapted from M. Chamie. Survey design strategies for the study of disability. World health statistics quarterly, 1989, 42: 122-140. ImmoDile Inactive unmarried M Sighted (m89) ColImpaired vision (n=94) r7 Blind (ns136) WHO Bulletin OMS. Vol 73 1995 499 T.G. Evans Table 3: Step 0 likelihood ratio chi-squared and summary of stepwise logistic regression of mobility, occupation, and marital status Mobilitya Occupationa Marital statusa Explanatory variable (approx. F-to-enter)b (approx. F-to-enter)b (approx. F-to-enter)b Visual acuity 62.4*** 102.1** 17.2* Sex 0.6 0.2 0.2 Age 31.2*** 35.5*** 0.0 HHRESP (Household responsibility) 7.8** 6.0* 21.1 Household size 0.6 0.9 0.2 Health 0.1 2.6 0.8 Migration 27.8*** 23.7*** 0.0 Household food insecurity 4.2* 12.1* 0.3 Village role - - 5.6* Mobility - 15.3* Occupational status 202.1 - 19.1 Marital status 15.7*** 19.1*** Step No. Term entered Improvement x2 P-valuea Goodness of fit X2 P-value Mobility: 0 413.9 0.000 1 Occupational status 130.9 282.9 0.712 2 Visual acuity 16.2 266.8 0.880 3 Migration 10.6 256.2 0.946 4 Age 7.9 248.2 0.973 5 Household responsibility 6.9 240.8 0.987 6 Household size 6.9 233.9 0.994 Occupational status: 0 417.6 0.000 1 Visual acuity 144.0 273.9 0.817 2 Age 20.6 253.3 0.962 3 Migration 19.7 233.6 0.996 4 Household responsibility 10.8 222.7 0.999 Marital status: 0 350.5 0.018 1 Visual acuity 32.5 318.5 0.166 2 Household responsibility 16.8 302.7 0.351 3 Sex 9.5 293.2 0.486 4 Food insecurity 4.1 289.1 0.537 a * P<0.05; ** P<0.01; *** P<0.001. b At step 0 the F-to-enter for a variable corresponds to the F statistic computed from a one-way analysis of variance (ANOVA) of that variable for the groups used in the analysis. "chronic" household food insecurity. Occupation and migration were the strongest independent predictors accounting for about 34% of the variance in mobili- ty. Village role, age and household responsibility were also selected in the stepwise model, and accounted for another 30% of the variance. A similar set of associations was found among the blind for the analysis of occupational status. Older age and later age of onset of blindness, no household responsibility, immobility, and seasonal household food insecurity were all significantly cor- related with inactivity among the blind (Table 4). Household food insecurity, mobility and age were selected as independent predictors of occupation, accounting for 37%, 20% and 5% of the variance, respectively. Only household responsibility and late age of onset of blindness were significantly associat- ed with "married" marital status at step 0 (Table 4). However, once household responsibility was entered in the stepwise model, females and small household size became significant predictors of "married" status. Among the visually impaired who were immo- bile, there were associations with no occupation and migration activity and older age (Table 5). Occupa- tion was the most powerful predictor of mobility (P = 0.429); while age and migration were signifi- WHO Bulletin OMS. Vol 73 1995500 Socioeconomic consequences of blinding onchocerclasis In West Africa Table 4: Step 0 likelihood ratio chi-squared and summary of stepwise logistic regression of mobility, occupation, and marital status among the blind Mobilitya Occupationa Marital statusa Explanatory variable (approx. F-to-enter)b (approx. F-to-enter)b (approx. F-to-enter)b Age of onset 5.18 8.8** 4.6* Sex 1.7 0.2 0.2 Age 7.3** 8.5** 0.3 Household responsibility 7.1* 4.2* 10.9** Household size 0.1 1.2 0.7 Health 0.1 2.9 1.5 Migration 18.6 - 1.7 Household food insecurity 5.4* 22.1*** 0.1 Village role 5.2* - - Mobility - 18.18** 1.72 Occupational status 19.4*** - 0.00 Marital status 0.01 Step No. Term entered Improvement X2 P-valuea Goodness of fit X2 P-value Mobility: 0 157.1 0.005 1 Occupational status 15.8 *** 141.2 0.200 2 Migration 8.4 ** 132.8 0.344 3 Village role 4.7 * 128.1 0.431 4 Age 6.1 * 122.0 0.558 5 Household responsibility 4.0 118.0 0.635 Occupational status: 0 139.0 0.279 1 Food insecurity 16.7 *** 122.3 0.649 2 Mobility 10.9 ** 111.3 0.853 3 Age 4.6 * 106.8 0.903 Marital status: 0 179.3 0.003 1 Household responsibility 10.4 ** 168.9 0.012 2 Sex 5.8 * 163.0 0.023 3 Household size 4.5 * 158.6 0.034 4 Age of onset 4.0 * 154.6 0.048 a * P<0.05; ** P<0.01; *** P<0.001. b See footnote b in Table 3. cant but relatively weak predictors of mobility (P = 0.098 and 0.079 respectively). Constrained mobility, older age and older age of onset of visual impairment, and poor health were significantly cor- related with no current occupation. Mobility was the most powerful predictor (P = 0.341), and when entered in the model both age and age of onset lost their significance, thereby leaving health as the only other significant predictor of occupation (P = 0.122). At step 0, only household responsibility was signifi- cantly associated with "married" status among the visually impaired (Table 5). Once household respon- sibility was entered in the stepwise model, seasonal household food insecurity and later age of onset of visual impairment were identified as significant pre- dictors of marital status. Discussion As seen in Fig. 2 and the Step 0 associations in Table 3, visual status appears to stratify the sample for each of the three measures of socioeconomic function. The selection of visual status as an inde- pendent predictor indicates that decreasing visual acuity is not simply a proxy for increasing age or poor health status, but rather explains a large part of the difference in social and economic functioning of the population. Therefore, the higher prevalence of constrained mobility, "inactive" occupational status, and "non-married" status are representative of a han- dicap associated with decreased visual acuity. Visual status, however, is not an exclusive or absolute pre- dictor of handicap. Although less powerful as predic- WHO Bulletin OMS. Vol 73 1995 501 T.G. Evans Table 5: Step 0 likelihood ratio chi-squared and summary of stepwise logistic regression of mobility, occupation, and marital status among the visually impaired Mobilitya Occupationa Marital statusa Explanatory variable (approx. F-to-enter)b (approx. F-to-enter)b (approx. F-to-enter)b Age of onset 9.9** 8.5** 0.2* Sex 0.1 - 0.0 Age 11.2** 9.7** 0.2 Household responsibility 0.0 0.2 6.2* Household size 0.3 0.4 1.5 Health 0.7 5.4* 0.1 Migration 5.1* - 1.3 Household food insecurity 0.3 0.0 2.2 Village role 0.1 - Mobility - 24.1* 0.7 Occupational status 34.1* - 1.2 Marital status 0.7 -- Step No. Term entered Improvementx2 P-valuea Goodness of fit x2 P-value Mobility: 0 104.4 0.035 1 Occupational status 24.9 79.5 0.464 2 Age 4.1 75.4 0.562 3 Migration 3.4 72.0 0.641 Occupational status: 0 107.8 0.021 1 Mobility 25.0 82.8 0.362 2 Health 4.3 77.8 0.484 Marital status: 0 80.5 0.463 1 Household responsibility 5.8 74.7 0.617 2 Food insecurity 4.4 70.3 0.721 3 Age of onset 4.3 66.0 0.782 a * P<0.05; ** P<O.01; *** P<O.001. b See footnote b in Table 3. tors, other variables such as age, sex, household responsibility, and household size do meet the entry criteria or explain the variance in economic and social functioning in the population. With the different degrees of handicap identified among the blind and visually impaired, it is interest- ing to consider how and why the levels of handicap differ within these groups. The step 0 associations for mobility and occupation among the blind (Table 4) reveal a very similar cluster of variables: early age of onset of blindness, younger age, household responsibility, and household food insecurity are linked with unconstrained mobility and occupational activity. These associations might suggest that the handicap of blindness is partly a function of age and age of onset of blindness, i.e., the younger the blind and the younger their age of onset of blindness, the greater their ability to adapt to physical impairment and to remain mobile and active. These variables, however, are either not selected or explain only a small portion of the variance in handicap upon entry in the stepwise model. The selection of (and the large P-values attributed to) household food in- security (change in P=0.37, Table 4) and occupa- tion and migration for the models of occupation and mobility, respectively, suggest that a second hypoth- esis of economic necessity may have more power in explaining the activity and mobility of the blind. In Mande culture, the household and village- level social networks function to support the elderly and disabled,J thereby alleviating the need to remain mobile and active among the older blind. Further- more, in focus group discussions the importance of the family in supporting the blind was universally stressed; the household does everything possible to i Adams A. Seasonal food insecurity in the Sahel: nutritional, social and economic risk among Bamana agriculturalists in Mali. PhD thesis, University of London, 1992. 502 WHO Bulletin OMS. Vol 73 1995 Socioeconomic consequences of blinding onchocerciasis in West Africa ensure there is no need for blind persons to continue working. Blindness is viewed as an impairment which entitles the individual to a life of dependency on others. It is recognized, however, that there are situations where village charity and the reduced pro- ductive strength of the household are not sufficient to support a blind individual. In these cases, the blind individual's productive input, diminished as it may be, cannot be spared by the household. Hence, young blind individuals in positions of household responsibility may find that, in a less secure house- hold situation, they remain economically active and mobile despite their disability. This concurs with Carroll's analysis of the blind in which he states "only the strongest motivations of terrible need will cause the blind person to move" (15). Although their disability is not as severe as the blind, it appears that young age, young age of onset of visual impairment, occupation, migratory activity and mobility define the less handicapped visually impaired. The absence of a significant association with household food insecurity may indicate that the handicap of the visually impaired is generally not severe enough to influence the food security of the household, although the activity of the young visual- ly impaired suggests that households rely on their productive input. Furthermore, self-reported ill- health is significantly correlated with inactivity for the visually impaired but had no association with the activity of the blind. Thus, sickness appears to be linked to the economic imperative, i.e., blind persons will work whether or not they consider themselves ill, if their labour input is needed. The analysis of marital status among the blind provides further insight into their handicap. Unmar- ried status was strongly associated with a cluster of variables including no household responsibility, male sex, large household size and early onset of blind- ness. As was bome out in focus group discussions, the onset of blindness before marriage for men made it very unlikely that they would ever marry or assume positions of responsibility within the house- hold. Instead, they remain within households (usual- ly large) that can support them. Blind women are considered eligible for marriage, but their families receive a much reduced bride price.k For those blind k In her analysis of landless agricultural labourers in Tamil Nadu, India, Heyer discusses the low cost of marrying disabled persons (Heyer J. Landless agricultural labourers' asset strate- gies. Institute of development studies bulletin, 1989, 20(2): 33-40). Likewise, in an analysis of the disabled in urban Sene- gal, Fassin describes disadvantages in marriage transactions for invalid men and women (Fassin D. Handicaps physiques, pra- tiques 6conomiques et strat6gies matrimoniales au Sen6gal. Social science and medicine, 1991, 32: 267-272). and visually impaired with spouses, the large major- ity (88%) were married before the onset of decreased visual acuity. The advent of blindness was not con- sidered to pose irreconcilable problems, although it was noted that women may leave a blind husband in cases of poverty and extreme hardship.' The high percentage of blind men who are "widowed" may also be related to this phenomenon of women leav- ing blind men because of extreme hardship. The marital handicap of the visually impaired is much less severe than that of the blind, which is partly due to the fact that others may not be aware of their disability. When comparing the marriage pro- files of the visually impaired and the sighted popula- tion, the evidence of handicap lies in the greater rate of "widowed" persons (11% vs. 3%, Table 2) and to a lesser extent the "single" population (11% vs. 7%, Table 2). The greater rate of widowed status prob- ably represents the fact that the majority of elderly widowed persons in areas of hyperendemic oncho- cerciasis are likely to have some visual impairment. However, the higher rate of "single" visually impaired, relative to the sighted, is likely to be indic- ative of severe visual impairment which is socially perceptible and therefore, like the blind, a virtually insurmountable barrier to marriage. The following brief case studies illustrate some of the patterns of handicap discussed above. * Djiba is 45 years old and has been blind for 4 years. Together with his eldest son Mory, age 14, he walks to the household field each day, and cultivates until fatigue and hunger overcome his efforts. The household has no food stocks and is dependent on what Djiba's wife can procure by working. With no prospect of extended family support, and five depen- dent children, Djiba recognizes that despite his dis- ability he must continue to cultivate in an effort to provide for his starving family. * In constrast, Moriba, 67 years old, ceased to culti- vate and walk beyond the village at the onset of his blindness 7 years ago. Unlike Djiba, he has a large extended family to support him. He enjoys food security and busies himself like most other village elders with the religious and political activities of the village. * Finally, Sekou, just 26 years old, lost his sight 6 years ago. Although he is much respected for his i This theme of women leaving blind husbands due to hardship in the household has been popularized in an article on the importance of blindness rehabilitation in Burkina Faso (see Ouattara S. Quel metier pour les aveugles? Famille et d6veloppement, 1987, 45: 30-37). WHO Bulletin OMS. Vol 73 1995 503 T.G. Evans agricultural prowess, he has little prospect of marry- ing and establishing a family. Unlike most men of his age who migrate to the COte d'Ivoire during the dry season in order to raise the funds necessary for a bride price, Sekou remains a dependent in his father's household, without a wife or a family of his own. It is useful to consider the implications of the findings of handicaps arising from decreased visual acuity, in the context of the conceptual framework presented in Fig. 1. When viewing the sample popu- lation as a whole, decreasing visual acuity is strongly correlated with more severe mobility, occupational and marital handicaps. In addition, other factors such as age, sex, household responsibility, and household size are significantly associated with variance in handicap. This demonstrates that handicap is not a direct consequence of the disability arising from vis- ual impairment, but a product of the interaction between the disability, or the disease process, and other factors. These other factors are explored in some detail when investigating why handicap varies among the blind and the visually impaired. Multivar- iate models reveal age, sex, health, and age of onset of blindness or visual impairment to be factors spe- cific to the individual which explain the variance in mobility, occupation and marital status among the blind and the visually impaired. Specific factors such as household food security, household responsibility, and household size are also identified as significant predictors of the handicap variables. For example, household food insecurity, the strongest predictor of activity among the blind (Table 4), provides evi- dence for the hypothesis that in food-insecure house- holds it may be necessary for the blind to work (the case of Djiba), while in food-secure households their productive services may not be required (the case of Moriba). Evidence that factors not specific to the "individual" (in this case the household) can play a significant role in determining the degree of individ- ual handicap emphasizes that handicap cannot be viewed solely as the product of interactions between the individual and the disease process. Rather, larger social organizations such as the household and com- munity must also enter the equation. This corre- sponds to the concept of handicap (outlined above) as an interaction of the disabled person and the envi- ronment (see Fig. 1). Similar findings in the litera- ture on rheumatoid arthritis reveal that social and family factors exert a significant influence on work disability (16, 17). Furthermore, larger forms of social organization such as the household must not only be viewed in terms of their influence on individual handicap, they must also be considered in the context of the conse- quences of an individual's disability on the house- hold. The association between household food in- security and activity among the blind (Table 4) may represent the degree to which household production capacity is compromised by the less productive blind member: the chronic food insecurity experienced by Djiba's household may be a direct result of his blind- ness. This emphasizes the necessity to extend the analysis of handicap beyond the individual to the household and community,m especially in environ- ments where the household, not the individual, is the primary economic unit. Similar concepts have been identified when considering the unremitting burden on the carers and on families resulting from chronic disease (18, 19). Another issue which emerges from this analysis is how to interpret different measures of individual handicap: does greater handicap mean an individual is worse off than someone with a lesser handicap? From the results in this study, the elderly blind who are inactive and immobile (Moriba) appear more handicapped than those who are younger, active and mobile (Djiba). However, the younger blind are also more likely to belong to food-insecure households (Djiba) and are unlikely to find a spouse (Sekou). It appears therefore that the multiple dimensions of individual handicap do not lend themselves easily to a single aggregate measure. It may be more useful from a policy perspective to view the spectrum and dynamic nature of handicaps arising from a disabling disease as it affects individuals and households at different stages of life. m See Evans TG. The impact of permanent disability on rural households: river blindness in Guinea. Institute of development studies bulletin, 1989, 20(2): 41-48, and Evans TG. The socio- economic consequences of human disease on subsistence agri- culture: the case of onchocerciasis in West Africa. Unpublished DPhil thesis, Oxford University, 1989. Acknowledgements This research was sponsored by the Edna McConnell Clark Foundation, the Royal Commonwealth Society for the Blind (Sight Savers), the CIDA Award Programme and the Rhodes Trust. Logistic support was received from the Onchocerciasis Control Programme in Mali and Guinea. The author is grateful to Rosemary Fennell, Peter Tugwell, Elisabeth Badley, and Alayne Adams for comments on earlier drafts. Special thanks are extended to the study interviewers for their excellent work and to the participants whose cooperation was indispensable 504 WHO Bulletin OMS. Vol 73 1995 Socioeconomic consequences of blinding onchocerciasis in West Africa Resume Cons6quences socio-6conomiques de l'onchocercose c6citante en Afrique de l'Ouest L'onchocercose, ou c6cite des rivieres, est une cause majeure de cecite irr6versible chez les adultes au faite de leur vie active, et contribue a I'abandon des villages dans les vallees fluviales de l'Afrique de l'Ouest. Comme cette maladie fait peser de lourdes contraintes sur le developpe- ment economique, outre le probleme de sant6 publique qu'elle pose, elle a fait l'objet d'un impor- tant effort de lutte, concretise par le lancement du Programme de lutte contre l'Onchocercose (OCP) en 1974. Cet article presente des donnees pri- maires en provenance de r6gions de Guinee ou l'onchocercose est fortement endemique, et exa- mine le modele des cons6quences de la maladie fourni par la Classification internationale des han- dicaps: d6ficiences, incapacit6s et d6savantages, pour 6valuer les consequences socio-6cono- miques de la cecite et de la perte de vision dues a cette maladie. En 1987, une enquete transversale portant sur les personnes et les m6nages frappes de cecite ou de deficience visuelle, ou ayant une bonne vue, a ete r6alis6e dans le nord-est de la Guinee. L'6chantillon de 136 aveugles, 94 d6fi- cients visuels et 89 personnes ayant une bonne vue a ete obtenu dans 14 villages d'hyperend6mie onchocerquienne identifi6s lors d'enquetes 6pid6- miologiques portant sur l'infection onchocerquien- ne et l'acuite visuelle realis6es en 1985 et 1986 par l'OCP. Lors d'entretiens avec les sujets de l'6chan- tillon et les personnes faisant partie de leur m6na- ge, des enqueteurs qualifies ont utilise des ques- tionnaires portant sur les ant6cedents d'acuit6 visuelle, 1'6tat de sante tel qu'il est perqu par l'int6resse, la mobilite, I'activit6 agricole, la securi- te des approvisionnements alimentaires du mena- ge, et les donn6es d6mographiques. Trois indicateurs socio-6conomiques specifi- ques (mobilite, profession et situation de famille) ont ete retenus pour comparer les groupes d'aveu- gles et de deficients visuels et la cohorte de sujets ayant une bonne vue. Parmi les aveugles, 68% 6taient incapables de se d6placer a pied en de- hors du village, contre 42% des deficients visuels et 8% des sujets ayant une bonne vue. Seuls 10% des aveugles contre 54% des deficients visuels et 92% des sujets ayant une bonne vue se d6cla- raient agriculteurs. La cecit6, et dans une moindre mesure la deficience visuelle, etait associee a des taux 6lev6s de c6libat, de veuvage ou de prise en charge de la veuve d'un frbre decede, et a un faible taux de mariage. Dans une analyse multi- variee, I'acuite visuelle a ete retenue comme fac- teur predictif ind6pendant pour chacune des trois variables socio-economiques, d'autres facteurs independants, comme l'age, le sexe, la position de l'interesse au sein du menage et le nombre de personnes composant le menage, satisfaisant aussi aux critbres de significativit6. La variation du handicap a l'int6rieur de chaque groupe d'acuit6 visuelle a ete etudiee au moyen de modeles de regression logistique mul- tiple pour chacune des trois variables socio-eco- nomiques. Diverses variables independantes ont pu etre associees A la variation du handicap, et ont ete par la suite interpret6es en utilisant des donnees qualitatives. Ces observations sont dis- cutees dans le cadre du concept de handicap selon la Classification internationale des handi- caps, et en soulignant le fait que des facteurs tant individuels que lies au menage ou a l'environne- ment jouent un r6le important dans le degre de desavantage associe a une maladie chronique invalidante telle que la cecite. References 1. Prost A, Prescott NM. Cost-effectiveness of blind- ness prevention by the Onchocerciasis Control Pro- gramme in Upper Volta. Bulletin of the World Health Organization, 1984, 64: 667-681. 2. Hervouet JP. Du riz et des aveugles. L'oncho- cercose a Loumana. Ouagadougou, ORSTOM, 1978. 3. Rosenfield PL, Widstrand CG, Ruderman AA. How tropical diseases impede social and economic development of rural communities: a research agen- da. Rural africana, 1980-81, 8-9: 5-20. 4. International classification of impairments, disabil- ities and handicaps. Geneva, World Health Organ- ization, 1980. 5. Chamie M. The status and use of the international classification of impairments, disabilities and handi- caps (ICIDH). World health statistics quarterly, 1990, 43: 273-280. 6. Badley E. The ICIDH: format, application in different settings and distinction between disability and hand- icap. International disabilities studies, 1987, 9: 122-125. 7. Fougeyrollas P, St-Michel G, Blouin M. Proposal for revision of the third level of the ICIDH: the hand- icap. ICIDH international network, 1989, 2: 22. 8. Knuttgen HJ, Buttner DW. [Investigations on the epidemiology and significance of onchocerciasis in Upper Guinea]. Tropenmedezin und Parasitologie, 1968, 19: 8-42 (in German). 9. Moreau JP, Prost A, Prod'hun J. Essai de normal- isation de la methodologie des enquetes clinico- parasitologiques sur l'onchocercose en Afrique de WHO Bulletin OMS. Vol 73 1995 505 T.G. Evans l'ouest. Medecine tropicale, 1978, 38: 43-51. 10. Thylefors B. Vision screening of illiterate popula- tions. Bulletin of the World Health Organization, 1977, 55: 115-119. 11. Prost A, Hervouet JP, Thylefors B. Les niveaux d'endemicite dans l'onchocercose. Bulletin of the World Health Organization, 1979, 57: 665-662. 12. Remme J et al. A force-of-infection model for onchocerciasis and its applications in the epidemi- ological evaluation of the Onchocerciasis Control Programme in the Volta River basin area. Bulletin of the World Health Organization, 1986, 64: 667-681. 13. Remme J et al. Ocular onchocerciasis and intensity of infection in the community, I: West African savan- na. Tropical medicine and parasitology, 1989, 40: 340-347. 14. Dixon WJ. ed. BMDP Statistical software manual. Berkeley, University of California Press, 1985. 15. Carroll TJ. Blindness. Boston, Little, Brown & Co., 1961: 37. 16. Yelen E et al. Work disability in rheumatoid arthritis: effects of disease, social and work factors. Annals of internal medicine, 1980, 93: 551-556. 17. Relsine ST et al. Work disability among women with rheumatoid arthritis: the relative importance of disease, social, work and family factors. Arthritis and rheumatism, 1989, 32: 538-543. 18. Anderson R, Bury MR. eds. Living with chronic ill- ness: the experience of patients and their families. London, Hyman Unwin, 1988. 19. Anderson R. The unremitting burden of carers. Brit- ish medical journal, 1987, 294: 73-74. 506 WHO Bulletin OMS. Vol 73 1995