Assessment of vaccination coverage, vaccination scar rates, and smallpox scarring in five areas of West Africa*

Bull. Org. mond. Santd 1973, 48, 183-194Bull. Wld Hlth Org. Assessment of vaccination coverage, vaccination scar rates, and smallpox scarring in five areas of West Africa* RALPH H. HENDERSON,1 HILLARD DAVIS,2 DONALD L. EDDINS,8 & WILLIAM H. FOEGE ' In 1966, nineteen countries of West and Central Africa began a regional smallpox eradication and measles control programme in cooperation with the World Health Organization. This paper summarizes sample survey data collected to assess the results of the programme in Northern Nigeria (Sokoto and Katsina Provinces), Western Nigeria, Niger, Dahomey, and Togo. These data indicate that the programme, which used mass vaccination campaigns based on a collecting-point strategy, was generally successful in reaching a high proportion of the population. Analysis of vaccination coverage and vacci- nation scar rates by age underlined the importance to the programme of newborn children who accumulate rapidlyfollowing the mass campaign. Ofallpersons without vacci- nation scars at the time of the surveys, 34.4% were under S years of age; in the absence ofa maintenance programme, this figure would rise to 40 % after 1 year. In 1966, nineteen countries of West and Central Africa began a regional smallpox eradication and measles control programme carried out in coopera- tion with the World Health Organization. Most of the participating countries started a 2-3-year mass vaccination campaign employing jet injectors (refer- red to as the " attack " or " phase I " portion of the programme), and planned to follow this with a " maintenance " or " phase II " portion designed to sustain the immunity levels that had been achieved during the attack phase. The vaccination programmes were based on a collecting-point procedure by which all the inhabitants of a village (or of a given quarter in the larger villages and towns) were asked to assem- ble at a central point for vaccination. The programme has had a dramatic effect on smallpox incidence (Fig. 1), which reached zero in September 1969. (During the first 6 months of 1970, 3 additional outbreaks, involving 75 cases, were * From the Smallpox Eradication Program (SEP), Center for Disease Control (CDC), Atlanta, Ga., USA. 1 Formerly Deputy Chief, Regional Office, SEP, Lagos, Nigeria. Present address: Venereal Disease Branch, CDC. Statistician, Regional Office, SEP, Lagos, Nigeria. 'Chief, Statistical Activity, Immunization Branch, State and Community Services Division, CDC. ' Director. reported in an isolated pocket of susceptible persons who escaped vaccination during the attack phase (Center for Disease Control, 1970). No cases have been reported since.) In 1968 and early 1969, staff from the Smallpox Eradication Program Regional Office, Lagos, di- rected evaluations of the attack phase programmes of several countries. These evaluations included surveys of the areas where vaccination had been under- taken, in order to determine the coverage (based on the history of vaccination by jet injector) and the proportion of persons with vaccination scars and smallpox scars (pock marks). This report summarizes the survey data from Northern Nigeria (Sokoto and Katsina Provinces), Western Nigeria, Niger, Daho- mey, and Togo (Fig. 2). Localities where vaccination had not been carried out at the time of the evaulation were excluded from the samples within these areas. Smallpox was highly endemic in all five areas before the programme began, but no cases of indigenous smallpox have been reported from them since September 1969. METHODS A random sample of at least 1 000 persons was drawn from each of the five assessed areas, using a 2994 - 183- R. H. HENDERSON ET AL. 2600- 2400 2200- 2000- ioo 14.00- 1200 1000. 600- 200 0 J F MA M J J A S 0 N 0 J F. AM J J A 50 N D J FM A M J J A S 0 N D J F MA M J I A S 0 N D 1967 1968 1969 1970 Fig. 1. Smallpox incidence in West and Central Africa, 1967-70. The shaded areas represent the range between the highest and lowest incidence reported during 1962-66. stratified two-stage cluster sampling design specifi- cally adapted for use in rural West Africa. Sample size was computed using a formula based on the binomial distribution (Cochran, 1963, p. 74): n = (z2pq)/d2,wherep represents the estimated propor- tion of the target population having the attribute being sampled and q = 1-p. The value of n was maximized by assigning a value of 0.5 to p. The precision of the estimates, d, was set at 10 %, and z (the normal deviate) was given a value of 1.96, which assigned a chance of 1 in 20 that the sample popula- tion might not be representative of the target popu- lation being surveyed. In solving this formula for n, it was shown that 96 persons were needed in each of the age groups about which information was desired. Individuals within adjacent households were to be sampled, and n had to be increased to correct for a clustering effect, since the vaccination status ofmem- bers of households tends to be similar (Serfling & Sherman, 1965, p. 136). No estimates of the magni- tude of this effect in West Africa were available, and Serfling's estimates pertaining to smallpox vaccina- tion in the USA were applied. Calculations based on the age distribution of the West African population indicated that a survey designed to sample approximately 1 000 persons should include a sufficient number in each age group (with the possible exception of those aged 45 years and over) to provide results accurate to within ±10%, except for a 1 in 20 chance (Table 1, final column). This sample size provided a large enough number of 15-44-year-old persons to permit males and females to be analysed separately. The total sample population of 1 000 must be selected from at least 30 separate sites before the formula used to determine the variance of the sample estimate becomes valid (Cochran, 1963, p. 157). To increase the likelihood that the survey teams might discover unreported cases of smallpox, the number of sites was increased to 67 and at least 16 persons were selected from each. This would have given a total sample population of 1 072 if all sites had been reached. These sites were chosen by random selection of 67 villages from a list, kept at the Regional Office, of all villages known to exist in the target area. The 184 SMALLPOX VACCINATION IN WEST AFRICA 185 < J~~~~~~~~~~~Ml Niger)l Gambia *1. ~~~~~~~~~UpperVolta Porthueia GuineaP Guinea Leone Go Nigeria Ivory coast Liberia Ghana 'Central African Repubi '1'o~~~o, ~Comroan olNoth-WstemState (Solioto Province) andl Central State (Kataina Province)(3Western State Equatorial Guinea Fig. 2. Areas surveyed in West Africa. The unshaded areas in Dahomey, Niger, Nigeria, and Togo represent areas where vaccination had not been carried out. procedures used had the effect of making the chance of selection of a particular village proportional to the population of that village (Table 2). It was antici- pated that the list of villages compiled at the Regional Office would be incomplete. These villages served only as starting points at which the individual assessment teams repeated the selection process, questioning vil- lage authorities in order to compile their own lists of all villages, hamlets, and isolated compounds falling within the jurisdiction of the original village, and to select a site at random (Table 2). An individual dwelling was selected as the starting point within each site. The assessors were provided with pieces of paper on which were drawn circles divided into 100 intervals labelled sequentially at each 300 interval (from zero to 3300). Standing in the centre of the site, the assessor oriented the circle by turning the zero mark towards the east. From a random number table he then drew a number be- tween 000 and 359 that established the direction of a line from the centre to the edge of the site. From the total number of dwellings intersected by this line, one was selected using the random number table.' Within the dwelling selected, the assessor inter- 1 It should be noted that this method is biased in favour of selecting dwellings located near the centre of the village, the probability of selection being proportional to wld, where the width of the dwelling is w and d is its distance from the centre. (This bias can be minimized by counting all dwellings within a sector of prespecified size (say 300) rather than by counting along a single line.) In these particular surveys, there was no fixed relationship between the site used by the vaccination teams and the centre of the village as judged by the assessors, and most of the sampled villages were small enough to have made distance from dwelling to vaccination site per se a variable of minor importance in influencing vaccination status. It cannot be stated with certainty, however, that sociocultural patterns that influenced the distance of dwellings from the village centre did not also influence the inclination of the occupants to be vaccinated. R. H. HENDERSON ET AL. Table 1. Sample size requirements for determining smallpox immunity levels using the sample survey method Sample size Clsei Sample size needed Expected number ofAgegroup Sampledswithout Clutering with clustering persons to be obtainedAge grup neeedwitout cofficiet (col 1 x cl. 2) sample of 1 000 persons c 0-4 96 1.85 178 188 5-14 96 2.03 195 251 15-44 96 1.51 145males 197 males(age 15-39 years) 145 females 225 females > 45 96 not given 145(?) 140 total 384 808 1 001 a Derived from formula n = (z2pq)/d2. See text for definitions and values of symbols. b See Serfling & Sherman (1965) p. 144. c Derived from the age distribution of the population of West Africa, as shown in the United Nations Demographic Yearbook, 1965. Table 2. Procedure for choosing villages to be sampled Village Village Cumulative Procedure name population population A 200 1-200 (1) From available census data, construct a list, similar to the one presented at the left of this table, of all of the villages known to exist in the geographical area to be sampled. B 50 201-250 (2) For these surveys it was decided to choose 67 sampling sites.a A sampling interval is C 70 251-320 obtained by dividing the total population of the area by 67 (50 000/67 = 746). D 40 321-360 (3) Select from a table of random numbers a number that falls within the sampling interval. This is the starting point. Suppose that that number is 005. The first sample of persons E 30 361-390 would then be drawn from village A ' on the list, since its cumulative population encompasses the number 5. The additional 66 villages are selected by adding the sampling F 20 391-410 interval (746) 66 times to the starting point selected from the random number table. G 100 411-510 Example: H 80 511-590 Number Village to be sampled 150 591-740 005 (from random number table) A J 100 741-840 751 (005 + 746) J K 110 841-950 1 497 (751 + 746) Q L 60 951-1 010 etc. etc. M 40 1 011-1 050 N 160 1 051-1 210 0 70 1 211-1 280 P 80 1 281-1 360 Q 180 1 361-1 540 R 190 1 541-1 730 S 40 1 731-1 770 etc. etc. etc. total - 50 000 a A minimum of 30 sample locations is required. In the West Africa assessments, additional clusters were added to look for hidden cases of smallpox. SMALLPOX VACCINATION IN WEST AFRICA 187 viewed and examined all the persons present who had spent the previous night there, and attempted to summon any absentees. If fewer than 16 persons were examined, the assessor continued walking in the direction of the line away from the centre of the village, and completed his quota by examining the occupants of additional households as needed. If he reached the edge of the village, he completed the sample by moving clockwise from the line to the next dwelling and establishing a new line back towards the centre of the village. In sampling rooms within dwellings, or dwellings within compounds, the asses- sor worked clockwise from the main entrance. Once sampling had begun in a given room all the occu- pants were sampled, even if this meant examining more than the 16 persons actually desired. Except in Western Nigeria, the surveys were di- rected at villages that between them contained 90% or more of the population of the areas being as- sessed.' The inhabitants of these villages belong, for all practical purposes, to a single social class. In West- ern Nigeria, the most highly urbanized area in West Africa, about 60% of the population reside in towns of over 5 000 persons. (In this area a separate survey of towns was conducted, which is not present- ed in detail in this report.) The surveys were conducted by assessment teams led either by an adviser from the CDC Smallpox Eradication Program or a staff member from one of the participating countries. In each of the surveys, at least one of the authors led one of the assessment teams and was responsible for training all the other team leaders and interpreters. Each person sampled was interviewed and examined by one of the assess- ment team leaders. All questions were asked in a local language in which both the assessor (or his interpreter) and the person sampled were fluent. Since the West African smallpox eradication and measles control programme is the first in this area that has used jet injectors in mass smallpox vaccina- tion campaigns, the history of having received a smallpox vaccination by jet injector was a useful Table 3. Number of persons sampled, by age and sex a Number sampled Age male female total < 1 417 1-4 1 493 5-14 756 937 1 693 15-44 1 057 2118 3175 > 45 438 635 1 073 total 7 851 a Combined results from surveys in Northern Nigeria (Sokoto and Katsina Provinces), Western Nigeria, Niger, Dahomey, and Togo. marker for assessing programme vaccination cover- age.2 In looking for smallpox vaccination scars, both arms were examined. In a few instances, a parent would indicate that his child had been vaccinated on the thigh, in which case the thigh was examined. In certain areas, many persons had been vaccinated in BCG campaigns. Since BCG vaccinations were usu- ally given on the forearm, there was little difficulty in distinguishing them from smallpox vaccination scars. Occasionally, burn scars intentionally induced as a sign of bravery, and traditional skin tattooing or scarring, were a source of confusion. A person was recorded as having smallpox scars if he had on his face at least five pock marks of 2 mm or more in diameter. RESULTS General A total of 7 851 persons were examined in the five sample surveys (Table 3). The proportion of sample locations that were selected but not sampled was small (Table 4). In both Northern and Western Nigeria rains had made certain villages inaccessible to the assessment teams. Since the majority of 1 Data derived from available census data in the countries surveyed, 1961-63. "To determine vaccination coverage during the pro- gramme, all persons were asked " Have you ever been vaccinated?" Those who responded affirmatively were then asked " When was your last vaccination? " and " How (by what method) were you vaccinated?" If the person did not understand what was meant by the third question, the assessor was then instructed to ask " Were you vaccinated with the jet injector or with the needle? " There were usually words in the local language for both "jet injector" and " needle," the latter term being used to denote the device used in performing multiple puncture, multiple pressure, or " scratch " vaccinations. If the person interviewed was still uncertain about the meaning of the question, a pantomime was performed, first making the motion of pushing a foot pedal while bringing the right hand towards the person's left arm as though holding an injector, and then making the motions associated with a " scratch " vaccination. In the vast majority of cases, the answers given were un- equivocal. R. H. HENDERSON ET AL. Table 4. Proportion of selected sample locations a where sampling was completed No. of No. of Sample Area sample sample locationslocations locations completed selected completed (%) Northern Nigeria Sokoto Province 67 63 94 Katsina Province 67 64 96 weighted average b - - 95 Western Nigeria area 1 67 65 97 area 2 67 37 55 weighted average c - - 90 Niger 67 66 99 Dahomey 68 68 100 Togo 67 67 100 a 16 persons were interviewed and examined at each sample location. b Sokoto Province contained 63 % and Katsina Province 37 % of the population residing in the total assessed area. c Area 1 contained 82 % and area 2 contained 18 % of the popula- tion residing in the total assessed area. vaccinations were performed during the dry season in these areas, the bias introduced by these missing sites is thought to be insignificant. The percentage age distribution of the sample population was similar to that of the population of West Africa (compiled from the United Nations Demographic Yearbook, 1965), except for a slight under-representation in the sample of males and an over-representation of females in the 15-44-year age group and children aged 0-4 years. The groups over- represented in the sample are those that tended to be in the village during the visit of the assessment teams. In computing the combined results from all five areas (Tables 6-8), the results from each area were weighted in proportion to the population of the area (Table 5). Small and large villages are analysed separately in Tables 6-8 since small villages appear to pose a particularly difficult problem for the vaccination teams-they tend to be the least acces- sible and, in areas in which the teams do not visit every village, their inhabitants must usually walk the greatest distances to be vaccinated. The percentage of the surveyed population residing in small villages (arbitrarily defined as having a population of less than 500 persons) varied from a maximum of 61 % in Niger to a minimum of 21 % in Dahomey (Table 5). Population with a history of vaccination byjet injector A history of vaccination by jet injector was given by 77.2% of the surveyed population, the highest proportion (88.4 %) being recorded in Northern Nigeria and the lowest (60.0%) in Western Nigeria (Table 6). Age-specific rates showed a similar pattern in all areas the lowest proportion of persons with a history of jet injector vaccination was observed in children less than 1 year of age and the next lowest proportion in persons 45 or more years of age. The Table 5. Population residing in assessed areas, weight given to each area in computa- tion of combined results, and proportion of assessed population residing in small (population <500) villages Weight given to each Proportion of assessed Area Population residing in area in computation population residingassessed area a of combined results b in small (pop. < 500)% villages Northern Nigeria(Sokoto and Katsina Provinces) 6.8 x 10 6 41.7 26 Western Nigeria 4.4 x 10 6 27.0 51 Niger 2.2 x 10 6 13.5 61 Dahomey 1.5 x 106 9.2 21 Togo 1.4 x 106 8.6 37 total 16.3 x 106 100.0 38 a Data derived from available census data in the countries surveyed, 1961-63. b The weights assigned to each area are proportional to the population residing therein. 188 SMALLPOX VACCINATION IN WEST AFRICA Table 6. Proportion (%) of population with a history of vaccination (1) Weighted average of all assessed areas by village size and by sex Age Village population < 500 Village population > 500 Total(years) male female total male female total male female total < 1 24.3 36.8 32.9 1-4 82.4 81.1 81.8 5-14 88.3 82.5 84.5 90.7 89.0 89.8 89.4 86.1 87.5 15-44 71.3 66.4 68.8 79.4 82.5 81.0 76.8 77.7 77.5 > 45 65.4 62.8 64.0 69.5 67.8 68.3 69.0 65.5 66.9 total a 72.0 79.5 77.2 (2) Results by individual area and by village size a Village Village population population Total < 500 > 500 Northern Nigeria (Sokoto and Katsina Provinces) 81.1 90.9 88.4 (92.9) Western Nigeria 56.3 63.8 60.0 Niger 80.1 77.4 79.0 Dahomey 63.9 72.4 70.6 Togo 74.0 83.7 80.1 a Age adjusted by assigning the following weights to the individual age groups: < 1 year, 4.3; 1-4 years, 14.5; 5-14 years. 25.1; 15-44 years, 42.0; > 45 years, 14.0. These weights total 99.9. They are taken from the percentage age distribution of the population of West Africa (United Nations Demographic Yearbook, 1965). b In Northern Nigeria, two of the assessment teams (one working in Sokoto and one working in Katsina Province), did not record the populations of the villages sampled. The figures in parentheses are derived from the total population sampled, while the other figures are derived from the 86 sample sites for which the village size was known. highest vaccination coverage was observed in persons 5-14 years of age, and intermediate levels ofcoverage were observed in the age groups 1-4 and 15-44 years. No consistent differences between coverage rates in males and females were observed. Small villages had lower vaccination coverage rates than did large ones (Table 6). This finding was significant at the 0.01 % level by the X-square test in four of the five areas surveyed. The one exception was Niger, where no overall differences were ob- served. The differences were most pronounced in three groups: 15-44-year-old females (vaccination coverage 16.1 % less in small villages), infants less than 1 year old (vaccination coverage 12.5% less in small villages), and 15-44-year-old males (vaccina- tion coverage 8.1 % less in small villages). No major differences were observed in the age groups 1-4 and 5-14 years. Population with vaccination scars A smallpox vaccination scar was observed in 82.4% of the surveyed population (Table 7). The highest rate was observed in Togo (88.2°/) and the lowest in Western Nigeria (76.4%Y.). Age-specific rates follow the pattern described for vaccination cover- age: the lowest vaccination scar rates were observed in the age groups under 1 year and 45 years and over, the highest rates were observed in the 5-14-year age group, and the age groups 1-4 and 15-44 years had intermediate values. Vaccination scar rates were less consistently re- lated to village size than were vaccination coverage rates, and no significant differences are seen in the combined results (Table 7). Within individual areas, Dahomey and Togo showed lower vaccination scar rates in small villages than in large villages; no differences were observed in Northern Nigeria and 189 R. H. HENDERSON ET AL. Table 7. Proportion (%) of population with smallpox vaccination scars (1) Weighted average of all assessed areas by village size and by sex Age Village population < 500 Village population > 500 Total (years) male female total male female total male female total < 1 20.0 34.3 30.4 1-4 79.6 78.5 79.0 5-14 84.8 89.6 87.3 92.8 92.2 92.4 90.3 90.8 90.5 15-44 90.8 87.9 88.7 84.5 85.4 84.3 87.7 86.8 86.6 > 45 77.3 77.7 77.3 76.5 72.3 73.9 76.7 74.1 75.0 total a 82.4 81.8 82.4 (2) Results by individual area and by village size a Village Village population population Total < 500 > 500 Northern Nigeria (Sokoto and Katsina Provinces) 86.5 86.7 86.7 (84.3) b Western Nigeria 80.3 72.4 76.4 Niger 78.0 79.8 78.7 Dahomey 76.0 81.3 80.2 Togo 83.0 91.2 88.2 a See footnotes to Table 6. b Idem. Niger, and Western Nigeria showed higher vaccina- tion scar rates in small villages. Population with smallpox scars (pock marks) Smallpox scars (pock marks) were observed in 12.8% of the surveyed population (Table 8). The highest rate (20.4 %) was observed in Northern Nige- ria. Western Nigeria (9.3%) and Dahomey (8.3%) had the next highest rates, and Niger (5.5%) and Togo (3.1 %) had the lowest rates. Applying the X- square test, these differences are significant at the 0.01 level. In most areas, few scars were observed among those less than 5 years of age. Rates increased with age thereafter, reaching maximum values in the age groups 15-44 years and 45 years and over. While higher rates were observed in all areas for 15 44- year-old males than for females, the difference was negligible in Northern Nigeria. Smallpox scar rates were not consistently related to village size, and no significant differences are seen in the combined results (Table 8). In Western Nige- ria, Niger, and Togo, higher smallpox scar rates were observed in large villages than in small villages, but this pattern was reversed in Northern Nigeria and Dahomey. DISCUSSION Population with a history of vaccination by jet injector The West African smallpox eradication and measles control programme has clearly been effective in reaching its target population, and has been respon- sible for the dramatic decline in smallpox incidence shown in Fig. 1. The collection-point strategy used during the attack phase was successful, and permit- ted the vaccination teams to move at a speed and 190 SMALLPOX VACCINATION IN WEST AFRICA 191 Table 8. Proportion (%) of population with smallpox scars (pock marks) (1) Weighted average of all assessed areas by village size and by sex Age Village population < 500 Village population > 500 Total(years) male female total male female total male female total < 1 0.0 2.3 1.7 1-4 2.0 1.3 1.3 5-14 1.2 4.6 3.5 6.0 5.7 5.6 4.5 5.0 4.7 15-44 23.2 19.8 20.8 25.0 17.7 20.0 23.6 17.8 19.6 > 45 16.2 19.0 18.4 29.0 20.2 24.2 24.5 19.7 21.9 total a 12.5 13.5 12.8 (2) Results by individual area and by village size a Village Village population population Total < 500 > 500 Northern Nigeria (Sokoto and Katsina Provinces) 20.4 20.4 20.4 (25.6) b Western Nigeria 7.5 11.2 9.3 Niger 4.4 7.2 5.5 Dahomey 12.7 7.1 8.3 Togo 2.6 3.3 3.1 a See footnotes to Table 6. b Idem. with an economy that would not have been possible had each dwelling been visited (Millar et al., 1971).' The proportion of persons giving a history of vaccination by jet injector showed wide differences between age groups. The lowest rates of vaccination coverage were observed in infants less than 1 year old because most of them had been born since the last visit of the vaccination teams. This finding empha- sizes the rapidity with which new susceptible persons enter the population and emphasizes the need for maintenance programmes directed at this age group. Among the age group 45 years and over (which had the next lowest coverage rate) a considerable propor- In the studies in Brazil, Millar and his co-workers were able to demonstrate that a team of three persons using one jet injector and a collection-point approach were able to perform work equivalent to that performed by a team of 38 vaccinators and their supervisors who performed multiplopressurevaccinations using a door-to-door approach. tion had already been vaccinated several times, or had actually had smallpox, and considered vaccina- tion to be unnecessary. Others felt that they were too old to be bothered with it. The 5-14-year-old child- ren were the best vaccinated group in the pro- gramme. The arrival of the vaccination teams was a significant event in the life of the community, and children of this age group apparently made every effort to be on hand to witness the proceedings. The somewhat lower coverage observed in the 1-4-year age group may have been related to their dependence on a parent or older sibling to bring them for vaccination. The 15-44-year-old persons who were not vaccinated during the programme frequently said that they were absent from the village at the time of the vaccination team's visit. The proportion of persons who gave a history of vaccination by jet injector was lower in small villages R. H. HENDERSON ET AL. than in large ones, but the reason why this difference was pronounced in certain groups and absent in others is not clear. The most pronounced differences were observed in 15 44-year-old females (the major- ity of whom are mothers) and in children less than 1 year old. Throughout most of West Africa the mother is the chief provider for the family and has the greatest demands placed upon her time. One can speculate that the distance to the vaccination site, which was generally greater for small villages than for large villages, was a more important consider- ation for her than for the other age groups in deciding whether or not to be vaccinated. Since children less than 1 year old are generally carried on their mothers' backs, one would expect factors affect- ing the vaccination coverage of mothers to have a similar effect on these infants. (The speculations re- quire confirmation by additional studies.) It is of interest to contrast the vaccination pro- gramme in Northern Nigeria, where the highest overall vaccination coverage rates were observed, with that in Western Nigeria, where the lowest rates were found. The programme in the west enjoyed several advantages over that in the north: the road network is better developed in the west and the distances the teams had to travel were shorter, making team supervision easier and vehicle break- down less frequent. Villages were more accessible in the west, and the programme could afford to visit every village. In the north, the residents of smaller villages were asked to assemble for vaccination at a central larger village. The vaccination programme in Northern Nigeria succeeded despite its logistic problems because of the influence of the traditional authorities, the Emirs, with their highly organized and centralized system of local government. Information about the programme was effectively transmitted to the population con- cerned, who-at the request of the Emir that they be vaccinated-often walked several miles to the vacci- nation site and waited for hours for the arrival of the vaccination teams. The task of motivating the village population of Western Nigeria was much more difficult. Tradi- tional government in the West is not nearly as centralized as in the North, and no authority ap- proaching that exercised by the Emirs could be used to the advantage of the programme. Advance publi- city was not very effective, so that many persons were unaware that the vaccination teams were coming, and, even when informed, many villagers chose to be absent from the village when the teams arrived. Occasionally, entire villages in the west fled from the vaccination teams. (Fewer problems were encoun- tered in vaccinating the population of towns with over 5 000 inhabitants, in which 60% of the popu- lation of Western Nigeria reside, where 83.1% of those examined gave a history of having been vac- cinated by jet injector.) Population with a smallpox vaccination scar The age distribution of smallpox vaccination scar- ring in the population was similar to that observed with the percentage of persons who gave a history of jet-injector vaccination. The lowest vaccination scar rates were observed in those aged under 1 year, the highest rates were observed in the 5-14-year age group, and the remaining groups had intermediate values. It is not surprising that, in the age group >4 years, the proportion of persons with vaccination scars was greater than that of persons who had been vaccinated by jet injector. This reflects the influence of past vaccination campaigns. Of some concern, however, is the fact that in children aged under 4 years the proportion of those who had a vaccination scar was 2-3% smaller than that of those who had a history of vaccination by jet injector. Analysis by individual area indicated that this discrepancy was observed only in Northern Nigeria and Niger. The cause could have been that some parents indicated to the assessors that their children had been vaccinated when in fact they had not or that some of the vaccinations performed by the teams were ineffective. The data gathered during the surveys do not permit an objective choice to be made between these two alternatives, but the impres- sion of the assessors was that parents were being truthful concerning the vaccination histories of their children and that this finding might be related to the performance of the vaccination teams. The best method for resolving this question is of course to read take rates 6-8 days after vaccination; and this is now being done with increasing frequency. Analysis of the population still susceptible to smallpox after the end of the mass vaccination programme indicated that the 0-4-year age group accounted for over one-third of the total number (Table 9). (Smallpox-susceptible persons were de- fined as those lacking a vaccination scar. The results would have been similar had smallpox scarring also been taken into consideration, since the majority of those with pock marks also had evidence of a vaccination scar.) This, of course, was the situation that existed at the time of the surveys. Without a 192 SMALLPOX VACCINATION IN WEST AFRICA Table 9. Contribution of each age group to total pool of persons susceptible to smallpox (A) (B) Pecnae(C)Age Percentage age distribution Proportion (% Percentage contribution of(years) of population of West Africa susceptible to smallpox a each age group to the totalsusceptible pool b < 1 4.3 69.6 17.0 1-4 14.5 21.0 17.4 5-14 25.1 9.5 13.6 15-44 42.0 13.4 32.1 > 45 14.0 25.0 19.9 total 99.9 17.6 c 100.0 a Smallpox-susceptible persons were defined as persons who lacked a smallpox vaccination scar. The percentages are derived from Table 7. b [(A x B)/E(A x B)] x 100. cZ(Ax B). maintenance programme, the 0-4-year age group would increase its contribution to the total suscep- tible pool by some 7% during the first year and, thereafter, by a percentage that would diminish slightly each year as the total number of susceptible persons continued to grow. It is the task of the maintenance phase to mount a programme of continuous vaccination designed to keep this pool of smallpox-susceptible persons to a minimum. Such programmes will be required until smallpox eradication programmes in other areas of the world have succeeded in eliminating this disease. Proportion ofpopulation with smallpox scars At the commencement of the West African small- pox eradication and measles control programme, little information was available concerning the occurrence of smallpox in small villages. Since that time, several investigations (Center for Disease Control, 1967, 1968, 1969; Henderson & Yekpe, 1969) have shown that it did occur, and the results of the present survey, which demonstrated little difference in the rate of smallpox scarring between the residents of small and large villages, suggest that it occurred commonly. For many years Northern Nigeria has been re- garded as a major focus for smallpox, and the fact that the smallpox scar rates observed in Sokoto and Katsina Provinces were more than double the rates observed in other areas supports this view. It is particularly interesting that the Niger population, a large proportion of whom live in a fertile strip of land bordering Northern Nigeria, had such a low rate of smallpox scars. The fact that a higher fre- quency of smallpox scars was not observed in Niger is probably related to the success of that country's past vaccination programmes, which have concen- trated for many years on protecting this border population. Western Nigeria and Dahomey appear to have had a similar smallpox problem. The Yoruba and Fon tribes in the two areas mingle across their common border, and share many beliefs, including the belief in a god of smallpox. While travellers to or from the North may have augmented the incidence of small- pox in these areas, it seems likely that Western Nigeria and Dahomey have been important foci in their own right in the past. ACKNOWLEDGEMENTS The regional smallpox eradication and measles control programme was financed partly by the United States Agency for Interational Development and partly by the World Health Organization. Technical and material assistance was provided by the Center for Disease Control, Atlanta, Ga., USA. 193 194 R. H. HENDERSON ET AL. RESUME EVALUATION DE LA PROPORTION DE LA POPULATION VACCINEE CONTRE LA VARIOLE, PRtSENTANT DES CICATRICES DE VACCINATION OU DES CICATRICES DE VARIOLE DANS CINQ REGIONS D'AFRIQUE OCCIDENTALE En 1966, on a lance un programme regional d'eradi- cation de la variole et de lutte contre la rougeole dans 19 pays d'Afrique occidentale et centrale. Les resultats relatifs a la variole de la premiere phase (phase d'attaque) de ce programme ont Wt6 appr&ci6s en 1968 et au d6but de 1969 au cours d'enquetes par sondages effectu6es au Nig6ria septentrional (provinces de Sokoto et de Katsina), au Nig6ria occidental, au Niger, au Dahomey et au Togo. Les enquetes ont montre que les campagnes de vacci- nation antivariolique, au cours desquelles on a utilis6 pour la preniiere fois les injections sous pression, avaient touch6 une forte proportion de la population. Dans l'ensemble de la region prospectee, 77,2% des personnes interrog6es declaraient avoir Wt6 vaccin6es par l'injecteur sans aiguille; la couverture 6tait la plus forte au Nig6ria septentrional (88,4%) et la plus faible au Nig6ria occi- dental (60,0 Y). La proportion des personnes porteuses de cicatrices de vaccination atteignait pour l'ensemble des r6gions 82,4% avec un taux maximal au Togo (88,2%) et un taux minimal au Nigeria occidental (76,4%). L'analyse des taux de couverture vaccinale par age a montr6 la n6cessit6 de maintenir un rythme de vacci- nation ad6quat parmi les tres jeunes enfants pour assurer la r6ussite du programme. Au moment des enquetes, les enfants ag6s de moins de 5 ans representaient 34,4% des sujets receptifs a la variole, cette proportion devant de- passer 40% un an plus tard en l'absence d'une phase d'entretien. La proportion des sujets porteurs de cicatrices de variole etait la plus dlevee au Nig6ria septentrional (20,4 %), de valeur intermediaire au Nig6ria occidental (9,3 Y.) et au Dahomey (8,3 Y), et la plus faible au Niger (5,5 ,) et au Togo (3,1 %). REFERENCES Center for Disease Control (1967) SEP Report, vol. 1, No. 3, Atlanta, Ga., USA Center for Disease Control (1968) SEP Report, vol. 2, No. 2, 3, and 5, Atlanta, Ga., USA Center for Disease Control (1969) SEP Report, vol. 3, No. 1, Atlanta, Ga., USA Center for Disease Control (1970) SEP Report, vol. 4, No. 3 and 4, Atlanta, Ga., USA Cochran, W. C. (1963) Sampling techniques, New York, Wiley Henderson, R. H. & Yekpe, M. (1969) Amer. J. Epidem., 90, 423428 Millar, J. D. et al. (1971) Trop. geogr. Med., 23, 89-101 Serfling, R. E. & Sherman, I. L. (1965) Attribute sampling methods, US Public Health Service Publication No. 1230