Estimation of incidence of poliomyelitis by three survey methods in different regions of the United Republic of Cameroon*

Bulletin ofthe WorldHealth Organization, 61(3): 501-507 (1983) © World Health Organization 1983 Estimation of incidence of poliomyelitis by three survey methods in different regions of the United Republic of Cameroon* DAVID L. HEYMANN,' VIRGINIA DAVIS FLOYD,2 M. LICHNEVSKI,3 GEORGES KESSENG MABEN,4 & FLAURIBERT MVONGO5 Surveys were conducted in one urban and two rural regions of the United Republic of Cameroon to estimate the annual incidence of paralytic poliomyelitis. Three different survey methods were used: a review of hospital and clinic registers, a school survey, and a house-to-house survey. The house-to-house survey identified the highest number of lame children andgave estimates ofincidence ofbetween 18.8 and32.6per 100 000population in the three regions. The estimates of incidence obtained by the two other survey methods in the urban region did not differ significantly from that obtained by house-to-house survey but, in the rural regions, were significantly lower. It is concluded that house-to-house surveys are a sensitive method ofidentifying lame children in both urban and rural regions. School surveys and review of hospital and clinic registers, while equally sensitive in urban regions, are less sensitive in rural regions and may significantly underestimate the annual incidence ofparalytic poliomyelitis. These limitations should be borne in mind when using the survey methods. Reports from many developing countries indicate a low incidence of paralytic poliomyelitis in rural regions of low population density (1, J. John, un- published data, 1981). Estimates of incidence in these countries are usually based on routine hospital and health centre reporting. Other methods of estimating the incidence of paralytic poliomyelitis, as used in Burma and Egypt, have indicated that the apparently low incidence in rural regions is possibly a result of under-reporting of disease, rather than low rural attack rates (2, 3). In 1978, 231 cases of paralytic poliomyelitis were reported in the United Republic of Cameroon (4), of which 72 (31 %o) were in Yaound6, the capital city. On the basis of this report, the annual incidence of polio- * Requests for reprints should be addressed to International Health Program Office, Centers for Disease Control, Atlanta, GA 30333, USA. ' Medical Epidemiologist, Centers for Disease Control (CDC), Atlanta, GA, USA. Formerly assigned to Organisation pour la Lutte contre les Endemies en Afrique Centrale. 2 Medical Officer, CDC, Atlanta, GA, USA. Assigned to Palmetto Medical Center, Palmetto, GA, USA. 3Formerly Medical Officer, Expanded Programme on Immu- nization, World Health Organization. Currently with the Depart- ment of Communicable Diseases, Riga Medical Institute, Riga, USSR. 4 Assistant Director, Department of Public Health, United Republic of Cameroon. s Surveillance Officer, Department of Public Health, United Republic of Cameroon. myelitis is 19.0 per 100 000 population in Yaounde, and 2.3 per 100 000 in the rest of the country (using population figures from the 1976 census). To investi- gate this reported difference in incidence, we studied the prevalence of flaccid paralysis in three regions of different geography and population density. In each of these regions, we estimated the prevalence of lower extremity flaccid paralysis and the annual incidence of paralytic poliomyelitis by review of hospital and health centre registers, a house-to-house lameness survey, and a school lameness survey. BACKGROUND Paralytic poliomyelitis has been well studied in Yaounde, which is situated at a latitude of 4°N in the tropical rain forest of the south of the United Republic of Cameroon. Serosurveys conducted in 1973 showed that over 90%0 of unimmunized children had antibodies to poliovirus types 1, 2, and 3 by the age of 4 years (5). Clinical observations at the Central Hospital during the early 1970s showed that 78% of children with flaccid paralysis had onset of paralysis before the age of 2 years (6). The annual incidence of poliomyelitis during the early 1970s was estimated to be 48 per 100 000 population, with most cases occur- ring between January and June (6). 4306 -501- D. L. HEYMANN ET AL. The only estimates of incidence of poliomyelitis in the United Republic of Cameroon outside Yaounde have been obtained from passive health centre report- ing which, as in 1978, has suggested a low incidence in rural regions of low population density. METHODS Survey regions One urban and two rural regions were chosen for study. The regions differed in topography, popu- lation density, and climate (Fig. 1). Yaounde, with a population of 313 700 is located at an altitude of 1295 m in mountainous tropical rain forest. The population density is over 90 inhabitants per kM2, and the mean daily temperature is 24 IC. The primary school attendance rate in Yaounde is 85%; routine poliomyelitis immunization was begun in 1976. The Bamenda region, at an elevation of 2550 m, is located on a high grassland plateau with a population of 33 700 and a population density of 30-45 inhabi- tants per km2. The mean daily temperature in the Fig. 1. Map of the United Republic of Cameroon, show- ing the three regions where the surveys were carried out. Bamenda region is 21 'C. The primary school attend- ance rate is 61 o and routine poliomyelitis immu- nization was begun in 1976. The Es6ka region, at an elevation of 200 m, is non- mountainous equatorial rain forest with a population of 28 900 and less than 20 inhabitants per km2. The mean daily temperature is 27 'C. The school attend- ance rate in the Eseka region is 61% and routine polio- myelitis immunization was begun in 1979. Hospital and clinic register review All hospitals, maternal and child welfare clinics, and health centres in each region were visited. A physician and nurse systematically searched the registers for 1975-78 at each centre, and recorded the names of children diagnosed as having poliomyelitis, by year of onset. Names were cross-checked between centres to prevent duplication. The total number of children with a diagnosis of poliomyelitis was divided by 4 to give the annual average. This was then divided by the 1976 population of the survey region to give an estimate of the annual incidence. House-to-house lameness survey House-to-house lameness surveys of children 5-11 years of age were conducted according to the recom- mendations of LaForce et al. (7). A two-stage cluster sampling method was used for children in Yaounde. The first-stage sampling units were 30 quartiers (neighbourhoods) chosen at random from the 1976 neighbourhood population lists; the second-stage sampling units were households clustered about a starting household, which was selected at random by direction and house number from the geographic centre of the quartier, as described elsewhere (4). In each second-stage unit, households were surveyed until a total of 150 children, born during the years 1967-74, had been registered. In the Bamenda and Eseka regions, all villages were sampled giving a total population of approximately 30 000 in Bamenda and 28 000 in Eseka. All house- holds in each village were visited to ensure that the entire population had been surveyed. A survey team was formed in each of the three regions. Each team consisted of a physician and a nurse trained in the examination of lame children, and a group of interviewers. Village and quartier chiefs were notified of the survey 2 weeks in advance by official government circular. The day prior to the survey, the chief was visited and the purpose of the survey explained. The chief was requested to ensure that the inhabitants would remain at home on the following day until they had been visited by the survey team. Interviewers arrived early in the morning of the day of the survey, and recorded the name and age of all children aged 5-11 years. The interviewer then 502 POLIOMYELITIS IN THE UNITED REPUBLIC OF CAMEROON asked if any of the children were lame, their names were noted, and a return visit was arranged with the lame child and parent for the following day. At the return visit, made by the physician or the nurse, the child was examined and the parents were questioned regarding date of onset of paralysis and place of residence at time of onset. The child was observed walking across the room. Muscle tone in the involved extremity or extremities was assessed by passive range of motion, with the child seated with feet on the floor. Muscle mass was estimated by pal- pation, and mid-calf and mid-thigh measurements were taken. Deep tendon reflexes were examined and sensation was tested for ability to differentiate sharp and dull ends of a pin. The degree of disability was classified as follows: Type I: flaccid paralysis of the lower extremity, sensation intact, able to walk without assistance; Type II: flaccid paralysis of the lower extremity, sensation intact, able to walk with assistance (pros- thesis, cane); Type III: flaccid paralysis of the lower extremity, sensation intact, unable to walk. If the lame child was not present at the first return visit, a second visit was arranged. Children absent at the second visit were not included in the analysis. The prevalence of lower extremity flaccid paralysis was calculated by dividing the number of children with flaccid paralysis and intact sensation by the total number of children surveyed. This figure was then divided by 5 (the maximum number of years each child was considered to be at risk for poliomyelitis infection) to give an estimate of the annual incidence of poliomyelitis with residual lower extremity para- lysis among the children at risk. Finally, this incidence was multiplied by 0.17 (the proportion of the popu- lation under 5 years of age) to obtain the annual inci- dence of poliomyelitis with residual lower extremity paralysis. School lameness survey The surveys of schoolchildren aged 5-11 years followed the recommendations of LaForce et al. (7). In Yaounde, schools were selected randomly from the Ministry of Education lists of primary school en- rolment for 1978-79, until a total of 15 000 students were available for examination. The overall ratio of private to government schools was maintained in the sample. In Bamenda and Eseka, all schools were surveyed to give approximately 10 000 and 8000 students, respectively. In Bamenda and Eseka, where only 61%o of the chil- dren attend school, the survey was designed to cover the school catchment area. Children in school were thus surveyed for lameness and, in addition, were asked to identify lame children aged 5-11 years who were siblings or neighbours and were not enrolled in school at the time of the survey. A physician and nurse team was formed and trained in each region, and made two visits to each school. At the first visit, the purpose of the study was explained to the school director, who was asked to fill in a standard enrolment form giving the total number of students born between 1967 and 1974. All children in each class were asked to walk past the survey team, and the lame students were identified. In Bamenda and Eseka, all students were requested to give the name of any lame sibling or neighbour who did not attend school. A return visit was then arranged with the school director, who was requested to ensure that all lame children attending school or living in the school catch- ment area were present and accompanied by a parent. At the second visit, parents were interviewed and children examined as in the house-to-house survey. Paralysis was listed as type I, II, or III. A further visit was made to cover those not present at this exami- nation. In Bamenda and Eseka, the survey team also asked village chiefs, village elders, and traditional healers in each school catchment area to supply the names of lame children residing in their village. Names ob- tained in this manner were cross-checked with those found during the school survey to avoid dupli- cation. The prevalence of lower extremity flaccid paralysis was calculated by dividing the number of children with flaccid paralysis and intact sensation by the total number of children surveyed. In Yaounde, the school enrolment figure for 5-11 -year-olds was used as the total. In Bamenda and Eseka, the enrolment figure for 5-1 1-year-olds was divided by 0.61 (the school attendance rate) to give the number of children born between 1967 and 1974 living in the school catchment area. The annual incidence of paralytic poliomyelitis with residual lower extremity paralysis was calculated as described previously. RESULTS Hospital and clinic register review The annual incidence of paralytic poliomyelitis, estimated by hospital and clinic register review, was 33.0 per 100 000 population in Yaounde, 4.4 per 100 000 in Bamenda, and 6.9 per 100 000 in Eseka (Table 1). The estimate of annual incidence in Yaounde is significantly higher than those in Bamenda and in Eseka (X2 test, P < 0.001). 503 D. L. HEYMANN ET AL. Table 1. Annual incidence during 1975-78, estimated clinic registers of paralytic poliomyelitis by review of hospital and No. of Annual Population poliomyelitis incidence Region (1976 cases (per 100 000 census) 1975-78 population) Yaound6 313 700 414 33.0 Bamenda 33 731 6 4.4 Es6ka 28 839 8 6.9 House-to-house lameness survey The prevalence of lower extremity flaccid paralysis, as estimated by house-to-house lameness survey, ranged from 5.5 to 9.6 per 1000 children (Table 2). The annual incidence of paralytic poliomyelitis with residual lower extremity paralysis, estimated from these figures, ranged from 18.8 to 32.6 per 100 000 population in the three regions studied. These rates were not significantly different. The 95% confidence interval for the estimated incidence in Yaounde was 25.5-30.1 per 100 000 population. School lameness survey The prevalence of lower extremity flaccid paralysis, as estimated by school lameness survey, ranged from 2.2 to 8.7 per 1000 children (Table 3). The annual inci- dence of paralytic poliomyelitis with residual lower extremity paralysis, estimated from these figures, was 29.7 per 100 000 population in Yaounde, 7.6 per 100 000 in Bamenda, and 13.5 per 100 000 in Eseka. The incidence in Yaounde was significantly higher than those in Bamenda and Eseka (X2 test, P < 0.001). The 95% confidence interval for the estimated incidence in Yaounde was 27.1-32.3 per 100 000 population. Fig. 2 summarizes the survey results in all three regions. Table 2. Annual incidence of paralytic poliomyelitis with residual lower extremity paralysis during 1967-74, estimated by house-to-house lameness surveys Residual paralysis Estimated annual incidencea No. of children No. of Prevalence Per 1000 Per 100 000 Region 1966-74 cases per 1000 children population children under 5 years Yaound6 4534 37 8.2 1.6 27.8 Bamenda 4332 24 5.5 1.1 18.8 Es6ka 4057 39 9.6 1.9 32.6 a These figures have not been corrected for upper extremity paralysis, death, or spontaneous recovery of cases. Table 3. Annual incidence of paralytic poliomyelitis with residual lower extremity paralysis during 1967-74, esti- mated by school lameness surveys Residual paralysis Estimated annual incidence" N o. of children School No. of Prevalence Per 1000 Per 100 000 R oiNo. catin schment ratendanc cases per 1000 children populationRegion ch ldren atchm rate cide nesurveyed are children undersurveyed a (% er5 years Yaound6b 9391 85 82 8.7 1.8 29.7 Bamenda 8503 13 393 61 31 2.2 0.4 7.6 Es6ka 6307 10 339 61 41 3.9 0.8 13.5 a These figures have not been corrected for upper extremity paralysis, death, or spontaneous recovery of cases. b The catchment area was not surveyed. 504 POLIOMYELITIS IN THE UNITED REPUBLIC OF CAMEROON 40' o 30' , 30 2 g25 - 20, tu O 15 z Z 10 5 1- U, sc -C la t 0 0I x0 <nI 0 !2 la Go 0 UA to > 90 INHABITANTS/KM2 30-45 INHABITANTSIKM2 < 0 INHBITANTSKM2 POLATION DENSITY Fig. 2. Annual incidence of poliomyelitis with residual lower extremity paralysis in three regions of the United Republic of Cameroon with different population den- sities, estimated by three survey methods. Type ofparalysis and age at onset A total of 254 children with flaccid paralysis of the lower extremity were found in the school and house- to-house surveys. Of the 246 for whom age at onset of paralysis was known, over 80% had been less than 3 years old (Fig. 3). There was no significant difference in the age of onset or extent of paralysis in the three regions. In the different regions, 69-82%o of lame children were able to walk without assistance, 13-24%o could walk with either a cane or prosthesis, and 4-7% were unable to walk (Table 4). Between 17% and 237o of children were living outside the survey region at the time of onset (Table 5). 100 90 80. 70 w 60' Z 50 E 40. 30' 20' 10. 0 Hz 0 fS 0 a 0 v > 90 INHABITANTS/KM2 z 0 v 30-45 INHABITANTS/KM2 I0 a v _V z 0 V < 20 INHABITANTS/KM2 POPULATION DENSITY Fig. 3. Percentage of children with flaccid paralysis, by age of onset, in three regions of the United Republic of Cameroon, as found in the combined house-to-house and school surveys (cumulative total). 505 Table 4. Type of paralysis found among lame children in the combined school and house-to-house surveys in the United Republic of Cameroon, 1979 Region Type I Type II Type IlIl No. % No. % No. % Yaound6 98 82 15 13 6 5 Bamenda 38 69 13 24 4 7 Es6ka 66 82 11 14 3 4 Table 5. Proportion of children living outside the survey region at the time of onset of paralysis, house-to-house survey, 1979 No. of Children living outside children region at time of onset Region with flaccid paralysis No. % Yaound6 37 7 19 Bamenda 24 4 17 Es6ka 39 9 23 Time required to complete the surveys The time required to complete each survey in the three regions is shown in Table 6. The house-to-house lameness survey, which required over 1400 person- hours in each region, was the most time-consuming. The review of hospital and clinic registers was the least costly technique. More time was required for each of the survey methods in the regions oflow popu- lation density than in the region of high density. DISCUSSION The annual incidence of paralytic poliomyelitis was estimated in one urban and two rural regions of the United Republic of Cameroon. The regions had different population densities and varied from high- land savanna to tropical rain forest. The estimates obtained relate to children who were born at least two years prior to the introduction of routine polio- myelitis immunization programmes, except for that based on review of hospital registers in Yaounde. This rate reflects the incidence during the early years of the poliomyelitis control programme, which was estab- lished over a 4-year period, during which time the reported poliomyelitis incidence decreased by over 70%. cn x I.- z i 7 z v zi so ra D. L. HEYMANN ET AL. Table 6. Time required to complete each lameness survey in the United Republic of Cameroon, 1979 Hospital and clinic Region register review School survey House-to-house survey No. in Total No. in Total No. in Total team person-hours team person-hours team person-hours Yaound6 3 24 3 216 9 1458 Bamenda 3 96 3 288 9 1728 Es6ka 3 120 3 312 9 2016 The house-to-house survey seemed to be the most sensitive method since it identified the highest number of lame children, and gave the highest estimates of annual incidence in the three regions. In each region, a similar percentage of children were living outside the survey region at the time of onset of flaccid paraly- sis, suggesting that the prevalence figures have not been unduly affected by migration in any one region. There was no significant difference in the regional figures found in the house-to-house surveys, and there thus appears to be no geographical variation in the incidence of poliomyelitis in the United Republic of Cameroon, as had been suggested by routine reporting from hospitals and health centres. The school lameness survey and register review gave estimates of incidence in the rural regions that were significantly lower than the rates established by house-to-house survey. In the urban region of high population density, such a difference was not found. Low attendance at schools and the shortage of hos- pitals in the rural regions are probably responsible for the discrepancies in the incidence estimates. Over 80% of all poliomyelitis with residual lower extremity paralysis occurred among children under 3 years of age in both the urban and rural regions, a finding consistent with that of Boche et al. in the early 1970s (5), and of Guyer et al. in 1976 (8). Most children with lower extremity paralysis had good muscular function, and only 4-7% were totally incapacitated. Different periods of time and numbers of personnel were required to complete the various surveys. The register review required the least time and personnel, while the house-to-house lameness survey required the most. Surveillance in regions with low population density required more time than in the high density region, reflecting the wider dispersal of the popu- lation. Because the reliability of each survey method varies according to the geographic features of the region and the population density, care must be taken in choos- ing the method to be used. Factors such as time and personnel available to conduct the survey, as well as the required precision of results, must be considered. If precise estimates of incidence are needed, such as in determining priorities for childhood disease control programmes, the house-to-house survey must be carried out unless the region has a high population density and high primary school attendance rate. In such cases, register review or school lameness survey may be considered as alternatives. However, if inci- dence estimates are required only as indicators of trends in poliomyelitis morbidity over specified periods of time, register review and school survey should be considered in all regions with constant population dynamics, regardless of primary school attendance rates and population density. Although these survey methods may give falsely low estimates of incidence in regions of low population density, repeat estimates over time will show changing trends. Incorporation of these survey techniques into routine health information systems will provide a means for estimating baseline incidence and for periodic evalu- ation of efforts to control paralytic poliomyelitis. RlSUMIt ESTIMATION DE L'INCIDENCE DE LA POLIOMYELITE PAR TROIS METHODES D'ENQUETE DANS DIFFERENTES REGIONS DE LA REPUBLIQUE-UNIE DU CAMEROUN Trois methodes d'enquete ont e utilis&es en vue d'es- timer l'incidence annuelle de la poliomyelite paralytique dans la Republique-Unie du Cameroun. Les enquetes ont e effectuees dans une region urbaine et deux regions rurales differant par la geographie et par la densite de la population. L'enquete porte-a-porte a permis de d6couvrir 506 POLIOMYELITIS IN THE UNITED REPUBLIC OF CAMEROON 507 le nombre le plus eleve d'enfants infirmes et a fourni des estimations similaires pour les trois regions. L'incidence annuelle 6valu6e par cette methode allait de 18,8 a 32,6 pour 100 000 habitants. Les estimations de l'incidence, d'apres l'enquete scolaire sur la claudication et 1'examen des registres des h6pitaux et des dispensaires, ne diff6raient pas notablement, dans la region urbaine, de celles qui avaient ete obtenues par l'enquete porte-a-porte, toutefois, les valeurs obtenues dans les regions rurales 6taient signifi- cativement plus faibles. On en conclut que les enquetes porte-a-porte sont une m6thode sensible pour d6couvrir les enfants infirmes dans les r6gions urbaines comme dans les regions rurales. Les enquetes scolaires sur la claudication et l'examen des registres des hopitaux et des dispensaires sont d'une sensibilite egale dans les regions urbaines, mais sont moins sensibles dans les r6gions rurales et, de ce fait, sont susceptibles de sous-estimer sensiblement l'incidence an- nuelle de la poliomyelite paralytique. II faut tenir compte de ce defaut lorsqu'on utilise ces differentes m6thodes d'en- quete a l'interieur d'un systeme d'information sur la sant6. REFERENCES 1. GUILLOZET, N. Paralytic diseases of childhood in the tropics. Journal oftropicalpediatrics and environmental child health, 22: 294-303 (1976). 2. Weekly epidemiological record, 52: 269-271 (1977). 3. Weekly epidemiological record, 52: 145-146 (1977). 4. Weekly epidemiological record, 47: 361-368 (1979). 5. BOCHE, R. ET AL. La poliomyelite au Cameroun. Revue d'epidemiologie, medecine sociale et santepublique, 21: 79 (1973). 6. GUYER, B. ET AL. Surveillance de la poliomyelite de forme paralytique A Yaounde, Cameroun 1973-1975. Afrique medicale, 15: 697-704 (1976). 7. LAFORCE, F. M. ET AL. Clinical survey techniques to esti- mate prevalence and annual incidence of poliomyelitis in developing countries. Bulletin of the World Health Organization, 58: 609-620 (1980). 8. GUYER, B. ET AL. The seroepidemiology of poliovirus in Yaounde, Cameroon 1977: a survey following one year of immunization. Journal of tropical pediatrics (London), 27: 140-143 (1981).