Impact of community-based presumptive chloroquine treatment of fever cases on malaria morbidity and mortality in a tribal area in Orissa State, India

The strategy was tested in 378 villages under Borigumma Community Health Centre (CHC) in southern part of Koraput district (17° 50' and 20°30'N and 81° 27' and 84°10'E), Orissa State, India (Figure 1). The CHC is the largest in the state of Orissa catering to the basic health care needs of 125,439 population. The topography and dynamics of malaria transmission in the area have already been described [4]. Villages situated on hilltops and at the foot of hills are hyper-endemic for malaria and experience perennial transmission. Villages situated on plain lands and riverbanks are hypo-endemic for malaria and experience seasonal transmission in rainy season [4, 5]. The houses have thatched roofing and mud plastered walls and are usually in-groups to form a village. Majority of the population is of tribal aborigines, who are economically poor. The men mainly cultivate paddy and women collect forest products, usually sal leaves and firewood. The health care facilities are poor and the first " medical" assistance is from traditional healers [6]. Under the National Anti-Malaria Programme, apart from routine surveillance and treatment, residual insecticide spraying was done in areas where annual parasite incidence was >10. Villages adjacent to high endemic villages, under Dasmantpur CHC with a population of about 66,500 and villages adjacent to low endemic villages under Kotpad CHC with a population of about 76,400 were selected as corresponding check areas for comparison.

The major malaria species of the study area is P. falciparum (85–90%)[5]. The parasite is sensitive to chloroquine in the study area as evidenced by both in vivo and in vitro methods [6, 7].

One volunteer from each village was selected for distribution of chloroquine and the selection was made either by villagers or head of the village. The volunteers were imparted training on diagnosis based on the symptoms of malaria and on chloroquine administration as prescribed in the National Anti-Malaria Programme through group discussion sessions and demonstrations.

Volunteers were advised to provide chloroquine to those patients who approach them for treatment and to fill up a 'fever treatment sheet' indicating the name, age, sex, number of days suffering from fever and the number of chloroquine tablets given. Those volunteers with no educational background were supplied with pre-packed chloroquine tablets in colored disposable plastic pouches for different age classes. The number of suspected malaria cases treated by age class was counted from number and type of pouches available with the volunteer at any point of time. Chloroquine was provided free of cost to patients. The volunteers were neither paid nor remunerated in kind.

The fever treatment sheets were collected from the volunteers after replenishment of chloroquine tablets through nine field workers located at different centres of the CHC area at fortnightly interval. Some Anganwadi workers of Integrated Child Development Scheme (ICDS) of the state Government also served as chloroquine distributors in addition to village volunteers to have better accessibility of all sections of people of the village.

The impact was evaluated based on the changes observed in fever days (the average number of days a patient suffered from fever before treatment), fever incidence (number of fever cases/year/1,000), malaria parasite incidence (number of microscopically proved malaria cases/year/1,000) and parasite prevalence (proportion of persons harbouring malaria parasite) in the community. Fever incidence in experimental area was calculated from the total number of cases receiving treatment from DDCs or Anganwadi centres. Malaria incidence was determined from the fortnightly fever surveillance carried out from 47 randomly selected villages (10% of the total population) in the experimental area. The finger prick blood slides were stained with Giemsa and 100 fields in each thick film were examined with 100× magnification [8]. Surveillance records of the CHC were used to determine fever and malaria incidence in the check area. Cross sectional sample blood surveys were carried out following identical methodologies in both the experimental and check areas at pre-intervention and six monthly during intervention (cold and dry season) to know the impact on the malaria parasite prevalence in the community. The villages to be sampled were selected following random number sampling design so as to cover 10 % of the total population from each endemic zone.

The number of fever cases not taking treatment from volunteers was identified through door-to-door surveys in randomly selected villages every fortnight. The patients were interviewed to know the reason for not seeking treatment from DDCs.

Comparison of impact of fever incidence, parasite incidence and parasite prevalence was made between 1st, 2nd and 3rd year of operation in the experimental villages and between the experimental and check CHCs. The odds ratio test was used to compare the change in fever and parasite incidence between experimental and check villages during 1^st and after 3^rd year of intervention.

Information on the deaths attributed to malaria was collected from the records available at the check and experimental CHCs. However, there is a wide spectrum severe and complicated manifestation of falciparum malaria reported from this area [9].

Table 1 shows the number of village volunteers and Anganwadi workers as chloroquine distributors in the two endemic areas of the study CHC. A total of 411 volunteers (281 village volunteers and 130 Anganwadi Workers) participated in chloroquine distribution in the 378 villages. Out of 281 village volunteers, 88.3% was males. The majority of volunteers (68%) was either small farmers or agricultural labourers earning less than Rupees 1,000/- (approximately 20 US$) per month. The remaining included petty shop owners, students, government employees, contractors, community leaders, former village health guides, Ayurvedic practitioners and a traditional healer. A total of 24 (8%) had no formal education. Among the literates, 35% had elementary education and the remaining had education up to 10^th standard or more. Most of the volunteers were family heads (75%).

Majority of the volunteers administered chloroquine as per the age class dosage schedule and maintained records properly. About 10% of the volunteers were replaced at different points of time as they were losing interest and were not performing well. The majority (64%) of volunteers reported having participated in drug distribution for social recognition.

The volunteers treated 88,575 fever cases during the 3-year study period, 39,301 (mean 492 ± 76 cases per annum per 1,000 population) from high endemic and 49,274 (mean 236 ± 142 cases per annum per 1,000 population) from low endemic experimental villages. In check areas, 26,657 (mean 137 ± 12) fever cases were treated from high endemic villages and 28,354 (mean 120 ± 29) from low endemic villages. About 14.5% of fever cases did not receive treatment from DDCs (Table-2) of which 17.9% did so due to mere negligence, whereas some 13.8% had no faith in the DDCs, probably because local persons manned them.

The age wise and overall annual fever incidence (AFI) in high endemic villages of the experimental area (Figure 2A) and check areas (Figure 2B) shows that the AFI was reduced by 31.7% in the 2nd year with a subsequent increase by 24.1% during the 3rd year in the experimental villages, whereas the reduction was 3.05% in the 2nd year with an increase of 20.5% during the 3rd year in the check area. The overall reduction observed in AFI at the end of 3rd year in the experimental villages was significantly higher (χ² = 520.03, P < 0.05), when compared to the check villages. The reduction in AFI during the 2nd year and subsequent increase during the 3rd year was observed in all age classes and in all months in the experimental area.

The age wise and overall annual fever incidence (AFI) in low endemic villages of the experimental area (Figure 2C) and check areas (Figure 2D) shows that in low endemic villages of the experimental area, the AFI was reduced by 48.4% during the 2nd year with a further reduction by 43.2% during the 3rd year in the experimental area, whereas there was an increase by 6.25% and 58.6% during the corresponding years in the check area. The overall reduction observed in AFI at the end of 3rd year in the experimental villages was significantly higher (χ² = 5712.75, P < 0.05) when compared to the check villages. The AFI declined in all age classes during the 2nd year of operation with further reduction during the 3rd year, except in the adult age class and in all months in the experimental area (Figure 3A).

In high endemic villages, the API reduced by 63.1% during the 2nd year followed by an increase of 66.6% during the 3rd year in experimental villages, whereas the reduction was 21.2% followed by an increase of 24% during the corresponding period in the check villages. The overall reduction observed in API at the end of 3^rd year was not significantly different (χ² = 1.46, P > 0.05)) between the experimental and check villages. No definite month-wise pattern was observed in monthly parasite incidence in the high endemic villages of the experimental area (Figure 3B).

In low endemic villages, the API was reduced by 56.8% during the 2nd year in experimental villages, whereas in the control villages the percentage of reduction was only 3.4%. However, during the 3rd year there was a marginal increase of 14% in API in the experimental villages when compared to an increase of 79.9% in the control villages. The overall reduction observed in API at the end of 3^rd year was significantly different (χ² = 40.14, P < 0.05)) between the experimental and check villages. Though no definite month-wise pattern was observed in monthly parasite incidence, there was declining trend in the low endemic villages of the experimental area (Figure 3B).

The pre-intervention parasite prevalence was 17.2% and 15.6% in high endemic check and experimental areas respectively. The survey was carried out in the month of May (dry season). There was marginal increase in the parasite prevalence during the cold season (month of December) during the first year in both the check (18%) and experimental (20.7%) areas. However, there was significant reduction (P < 0.05) in the prevalence in the experimental area both during dry and cold seasons of 2nd and 3rd year when compared with the check area (Figures 4A and 4C).

The pre-intervention parasite prevalence in dry season (month of May) was 3.2% and 3% in low endemic check and experimental areas respectively. There was a marginal increase in the parasite prevalence during the cold season (month of December) during the 1^st year in both the check (4%) and experimental (3.4%) areas. However, there was no significant reduction (P > 0.05) in the prevalence in the experimental area both dry and cold during 2^nd and 3^rd year when compared with the check area (Figures 4B and 4D).

The average number of fever days (AFD) during the pre-control period was 5.9 and 4.8 in the experimental and control villages respectively. It reduced to 1.6 in the experimental study area where as it remained at five in the check areas.

Data on deaths confirmed to be due to malaria during the study period in check and experimental CHCs are given in Table 3. During the year preceding the study, a total of nine deaths were recorded in the check area and four in the experimental area. During the three years of intervention, a total 53 deaths occurred in the check villages whereas, in contrast to three deaths in the experimental villages. The reduction could be observed in the 10–14 and >= 15 year age class. No conclusion could be drawn on the impact on younger age classes since no deaths was recorded in the experimental CHC in these age classes during the pre-intervention year.