The prevalence, intensities and risk factors associated with geohelminth infection in tea-growing communities of Assam, India


Rebecca Justine Traub, R. C. Andrew Thompson (corresponding author), Ian D. Robertson and Peter Irwin, World Health Organisation Collaborating Centre for the Molecular Epidemiology of Parasitic Infections, Division of Veterinary and Biomedical Sciences, Murdoch University, South Street, Murdoch, Perth, Western Australia 6150. Tel.: +61-8-93602466; Fax: +61-8-93104144; E-mail:,,,
Norbert Mencke, Bayer AG, BHC-Business Group Animal Health, 51368 Leverkusen, Germany. Tel.: +49-2173-384285;Fax: +49-2173-384956; E-mail:


Objective  To determine the prevalence, intensity and associated risk factors for infection with Ascaris, hookworms and Trichuris in three tea-growing communities in Assam, India.

Methods  Single faecal samples were collected from 328 individuals and subjected to centrifugal floatation and the Kato Katz quantitation technique and prevalence and intensities of infection with each parasite calculated. Associations between parasite prevalence, intensity and host and environmental factors were then made using both univariate and multivariate analysis.

Results  The overall prevalence of Ascaris was 38% [95% confidence interval (CI): 33, 43], and the individual prevalence of hookworm and Trichuris was 43% (95% CI: 38, 49). The strongest predictors for the intensity of one or more geohelminths using multiple regression (P ≤ 0.10) were socioeconomic status, age, household crowding, level of education, religion, use of footwear when outdoors, defecation practices, pig ownership and water source.

Conclusion  A universal blanket treatment with broad-spectrum anthelmintics together with promotion of scholastic and health education and improvements in sanitation is recommended for helminth control in the communities under study.


Direct life cycle nematode infections are among the most prevalent infectious diseases of people. The World Health Organization (WHO) estimates that infection with Ascaris lumbricoides, Trichuris trichiura and the hookworms (Ancylostoma duodenale and Necator americanus) with associated morbidity, affect approximately 250, 46 and 151 million people respectively (Montresor et al. 1998). The majority of these infections occur in developing countries where increased population density and urbanisation, poverty, inadequate sanitation and poor health awareness contribute to the increasing prevalence of infections (Chan et al. 1994; Montresor et al. 1998). Despite recent advances in epidemiological understanding of parasite population behaviour and cost-effective control strategies, the proportion of parasite populations in developing countries has remained virtually unchanged over the past 50 years (Chan et al. 1994). It could be argued that the major reason behind this fact is the relative unimportance placed on the public health impact of intestinal helminth infections in previous years as global mortality rates for intestinal helminthiasis are considered low compared with other infectious diseases. However, it has been increasingly recognized that apart from acute clinical disease (Bundy & Cooper 1989; Bahu et al. 2001; Khuroo 2001), chronic helminthiasis can lead to insidious and debilitating disease, especially in children and women of child-bearing age (Nokes et al. 1992; O'Lorcain & Holland 2000). WHO stresses the importance of soil-transmitted helminth control using a combination of chemotherapy and improved living conditions (WHO 1996; Guyatt et al. 2001). Most recent control programmes have focussed on chemotherapeutic-based options for control as they have been proven to be the most cost-effective and easy to implement (Stoltzfus et al. 1997, 1998; Guyatt et al. 2001). However, it has been increasingly recognized that for control to be both effective and sustainable, it is necessary to adopt an integrated approach (Warren 1990; Geerts & Gryseels 2000). One of the major reasons for the lack of integrated approaches is the lack of short-term benefits of the programmes. Yet, in order to prevent re-infection and minimize the potential for the development of drug resistance (Reynoldson et al. 1997; Geerts & Gryseels 2001) in the long-term, non-chemotherapeutic-based options such as health education and improved living conditions must be adopted. The first step in parasite control is to therefore conduct a baseline survey (Montresor et al. 1998) and to identify the various host and environmental risk factors associated with increased parasite prevalence and intensity within the community (Kightlinger et al. 1995, 1998).

In this study, we determine the prevalence and risk factors of significance associated with intestinal geohelminth infection (A. lumbricoides, T. trichiura and hookworms) among the three tea-growing communities of Assam. We examine the association between prevalence and intensity (eggs per gram, epg) of geohelminth infection with socioeconomic status, age, gender, worming status, religion, crowding within households, level of education, use of footwear when outdoors, individual defecation habits, water source and pig ownership.

Materials and methods

Study area and design

The tea-growing communities of rural Assam in north-east India were chosen for this project. No previous parasite surveys have been conducted in this community despite anecdotal evidence to suggest that the human population, particularly the tea-pickers, were heavily infected with intestinal parasites, especially Ascaris and hookworms. Medical practitioners of Phulbari Central hospital reported malnutrition and anaemia as commonly encountered syndromes, especially among children and women of child-bearing age. The tea estates are small and isolated with limited between-estate contact. The consent, cooperation and assistance of the managerial and medical staff of Williamson Magor and Co. was obtained to help conduct fieldwork.


Assam experiences a typically tropical climate with heavy monsoon rains. Phulbari, Addabarie and Balipara had a reported human population of 6531, 4839 and 2004 respectively (Census 2000). Each tea estate comprises three distinct socioeconomic groups, the executives, staff and workers. Executives make up <1% of the total population and are managerial and medical staff. This group was excluded from the study. Staff comprise approximately 5% of the total population and consist of factory workers, health workers and teachers. Most are literate and have completed a minimum of secondary schooling. They are provided with separate staff living quarters within the estate. The houses are arranged in fenced compounds. They are constructed with solid brick and cement and contain separate cooking, living, dining, sleeping and bathing areas with a verandah overlooking a backyard. Each house is provided with an enclosed Asian-style latrine connected to a septic tank. The workers make up the majority of the population and are predominantly tea-pickers. Most are illiterate. They are also provided with specific, mostly brick housing, similar in design to the staff quarters, but smaller (average compound area 2500 sq. ft), a separate enclosed area for bathing together with an Asian-style latrine connected to a septic tank. Most worker quarters are provided with running water and electricity, with only a few having to source water from open ring-wells running approximately 50 ft underground. The worker community lives in over-crowded conditions with poor standards of hygiene.

Both staff and worker communities commonly keep domesticated animals such as cattle, goats and poultry for milk, meat and eggs. A few households rear pigs, despite company regulations forbidding this practice within the estate. Each tea estate is provided with a school and a local hospital with free education, medical, surgical and dental services. A Health Officer is appointed to each estate to coordinate a biannual deworming programme with either pyrantel or levamisole, targeted at children. Medical practitioners also hold regular meetings at the Mothers’ Club aimed at educating women tea-pickers on issues such as birth and disease control. Women are prescribed iron and vitamin supplements by medical practitioners during pregnancy.


To aid cooperation, sampling was initiated by a series of preliminary lectures aimed at the assigned health workers for each worker housing division among the estates, to explain the purpose of the research project and a poster presentation demonstrating practical and simple measures of parasite control.

Households were randomly chosen from each housing division within each tea estate. In the evening, after returning from their duties, a face-to-face interview was conducted with each participant. If participants were below the age of 10 years, the interview was directed to the parent(s). The potential for interviewer bias was limited by having one interviewer (RJT) and a questionnaire with ordered and specific questions and procedures to follow. Specific data were collected firstly on household information including socioeconomic grouping, number of residents per household, toilet availability, religion and pig ownership. The second questionnaire was aimed at individuals with information relating to age, gender, overall literacy and level of education, individual defecation practices, use of footwear when outdoors, deworming status and general health. Written or verbal consent was obtained from each participant prior to participation. Faecal pots were handed to each human participant with their name and a picture drawing for those participants who were illiterate and collected the next morning. The study was approved by the Murdoch University Human Ethics Committee and the medical and managerial staff of Williamson Magor and Co. Ltd.

Faeces were stored separately in 5% formol saline and 2.5% potassium dichromate. Samples were transported to Murdoch University, Perth, for further processing.

Sampling statistics

Single faecal samples were collected from 328 people from the three tea estates (Table 1) over a 3-month period from July to September 2000. In total, 171 males and 157 females comprising 266 workers and 62 staff aged less than a year to older than 60 years participated in the study.

Table 1.  Sampling statistics: number of participants in study according to location and socioeconomic status
Tea estateWorkerStaffTotal

Parasitological techniques

Faecal samples were examined for parasite stages using a standard sedimentation in water technique followed by centrifugal flotation in saturated zinc sulphate, sodium nitrate and microscopy. Those samples positive for eggs of Ascaris, Trichuris and hookworm were then examined using the Kato Katz quantitative technique (Martin & Beaver 1968) and intensities of infection with each parasite expressed in epg of faeces.

Statistical methods

The prevalence and 95% confidence intervals (CI) were calculated for each parasite. Associations between parasite prevalence, intensity and host and environmental factors were initially made using chi-squared tests for independence, t-test and anova. Individuals were classified as ‘heavily’ infected if their egg counts were greater than the mean value of epg + 1 SD. Logistic multiple regression and linear regression was then used to quantify the association between parasite prevalence, intensity and each variable after adjusting for other variables. Only variables significant at P ≤ 0.25 in the univariate analyses were considered eligible for inclusion in the multiple regression (Hosmer & Lemeshow 1989; Frankena & Graat 1997). Pearson's correlation was utilized to determine the subfactors or variables that were highly associated with socioeconomic status. Egg intensities were transformed into common log (epg + 1) for comparison and analysis. Backward elimination was used to determine which factors could be dropped from the multivariate model. The level of significance for a factor to remain in the final model was set at 10%.

Statistical comparisons were performed using Statistics for Windows (Analytical software, Tallahassee, Florida), Win Episcope 2.0 (Nacho de Blas & Carmel Ortega) and Excel 97 (Microsoft).


Prevalence and intensity of parasites

The overall prevalence of Ascaris was 38% (95% CI: 33, 43) and 43% (95% CI: 38, 49) for both Trichuris and hookworms. Tables 2–4 summarize the univariate risk factors (P ≤ 0.25) associated with the prevalence of Ascaris, Trichuris and hookworms respectively, grouped by the demographic, environmental and socioeconomic proxy exposure factors and compares the prevalence of infection in each group. Tables 5–7 display the final logistic regression analysis and odds ratios for the prevalence of Ascaris, Trichuris and hookworms respectively. All three models show an overall strong model of prediction for prevalence of Ascaris (P = 0.0036), Trichuris (P = 0.0024) and hookworm infections (P = 0.0019). Heavy infections for Ascaris, Trichuris and hookworm [greater than (mean +1 SD)] were calculated at 13 600 epg for Ascaris, 876 epg for Trichuris and 320 epg for hookworms. Approximately 9% of the population harboured heavy infections with each individual parasite. Tables 8–10 summarize the univariate risk factor analysis for intensities of Ascaris, Trichuris and hookworms respectively (P ≤ 0.25). The tables also display the geometric mean values and percentage of heavily infected individuals grouped by the demographic, environmental and socioeconomic proxy exposure factors. Multiple (linear) regression analysis of Ascaris, Trichuris and hookworm egg intensities are displayed in Tables 11–13 respectively. These multiple regression models showed an overall weak model of prediction for Ascaris, Trichuris and hookworm with low r2 values.

Table 2.  Factors associated with the prevalence of Ascaris having a P-value of ≤0.25
FactorPrevalence (%)anova/t-test (P-value)
 <6 people/household29.2 
 ≥6 people/household48.0 
Anthelmintic usage0.0781
 Dewormed at least once/3 years32.9 
 Rarely or never dewormed42.4 
Level of education (school-aged population)0.0040
 Wears sandals when outdoors33.0 
 Barefoot when outdoors40.5 
Pig ownership0.0046
 Owns pigs68.4 
 Does not own pigs35.9 
Socioeconomic status0.0000
Defecation site0.0179
 Indoor latrine33.0 
Table 3.  Factors associated with the prevalence of Trichuris having a P-value of ≤0.25
FactorPrevalence (%)anova/t-test (P-value)
Socioeconomic status0.0000
Level of education (school-aged population)0.0000
 <6 people/household35.4 
 ≥6 people/household52.7 
 Wears sandals when outdoors26.3 
 Barefoot when outdoors52.9 
Pig ownership0.0012
 Owns pigs78.9 
 Does not own pigs41.1 
Defecation site0.0000
 Indoor latrine32.7 
Water source0.0034
 Tap (running)40.1 
Table 4.  Factors associated with the prevalence of hookworms having a P-value of ≤0.25
FactorPrevalence (%)anova/t-test (P-value)
 <6 people/household34.3 
 ≥6 people/household54.0 
Anthelmintic usage0.0203
 Dewormed at least once/3 years36.7 
 Rarely or never dewormed49.4 
Level of education (school-aged population)0.0000
 Wears sandals when outdoors27.1 
 Barefoot outdoors52.4 
Pig ownership0.0059
 Owns pigs73.7 
 Does not own pigs41.4 
Socioeconomic status0.0000
Defecation site0.0031
 Indoor latrine37.4 
Table 5.  Unweighted logistic regression and odds ratios for the prevalence of Ascaris
Predictor variablesβP-valueOdds ratio (95% CI: lower, upper limit)
  1. P-value: 0.0036; d.f.: 322.

Age−0.0200.0130.98 (0.96, 1.00)
Household crowding0.7060.0042.03 (1.95, 3.28)
Level of education−0.2440.0580.78 (0.61, 1.01)
Anthelmintic usage−0.6300.0150.53 (0.32, 0.88)
Religion – Hindu0.9300.0022.53 (1.42, 4.53)
Table 6.  Unweighted logistic regression and odds ratios for the prevalence of Trichuris
Predictor variablesβP-valueOdds ratio (95% CI, lower, upper limit)
  1. P-value: 0.0024; d.f.: 322.

Household crowding0.4730.0561.61 (CI, 0.99, 2.61)
Level of Education−0.3390.0120.71 (CI, 0.55, 0.93)
Barefoot when outdoors0.7360.0072.09 (CI, 1.22, 3.56)
Indoor defecation−0.6560.0330.52 (CI, 0.28, 0.95)
Tap or running water−0.8980.0120.41 (CI, 0.20, 0.82)
Table 7.  Unweighted logistic regression and odds ratios for the prevalence of hookworm
Predictor variablesβP-valueOdds ratio (95% CI: lower, upper limit)
  1. P-value: 0.0019; d.f.: 323.

Household crowding0.5190.0341.68 (1.04, 2.71)
Level of education−0.4980.0000.61 (0.47, 0.79)
Barefoot when outdoors0.7000.0092.01 (1.19, 3.41)
Anthelmintic usage−0.5710.0180.56 (0.35, 0.91)
Table 8.  Factors associated with the eggs per gram (epg) intensity of Ascaris with a P-value of ≤0.25
FactorGeometric mean (epg)anova/t-test (P-value)Prevalence of heavy infection (%)Chi-square (P-value)
Age (years)0.0121 0.0002
  0–24.43 0 
  3–5131.7 32 
  6–1051.3 19 
 11–1521.4 3 
 16–2067.8 16 
 21–3014.9 5 
 31–405.6 2 
 41–5013.4 3 
 >505.0 7 
Crowding0.0007 0.0393
 ≤6 people/house9.4 5.6 
 >6 people/house40.6 12 
Religion0.0005 0.2348
 Hindu25.1 9.7 
 Other5.4 5.6 
Level of education0.0109 0.0415
 None36.1 13.4 
 Primary16.9 6.7 
 Secondary12.2 5.3 
 Tertiary2.9 0 
Footwear0.1574 0.0935
 Wears footwear outside12.2 5.1 
 Barefoot outside23.1 10.5 
Health0.0012 0.0069
 Excellent5.9 2.3 
 Average21.7 9.2 
 Poor83.3 18.6 
Pig ownership0.0004 0.0043
 Yes390.3 26.3 
 No15.2 7.4 
Sex0.0795 0.1547
 Male12.8 6.4 
 Female27.3 10.8 
Socioeconomic status0.0000 0.0303
 Staff1.6 1.6 
 Worker32.5 10.2 
Defecation site0.0004 0.0000
 Outdoors51.8 19.3 
 Indoor latrines10.6 2.8 
Table 9.  Factors associated with the eggs per gram (epg) intensity of Trichuris with a P-value of ≤0.25
FactorGeometric mean (epg)anova/t-test (P-value)Prevalence of heavy infection (%)Chi-square (P-value)
Age (years)0.0135 0.0827
  0–21.6 0 
  3–536.5 18.2 
  6–1026.6 16.6 
 11–1513.1 2.5 
 16–2010.2 9.7 
 21–307.3 3.3 
 31–405.9 5.0 
 41–504.6 3.1 
 >507.7 0 
Religion0.0226 0.1086
 Hindu11.78 8.40 
 Other5.44 3.33 
Footwear0.0000 0.1401
 Wears footwear outside3.9 4.2 
 Barefoot outside15.7 8.6 
Pig ownership0.0005 0.1227
 Yes78.2 15.7 
 No8.3 6.47 
Socioeconomic status0.0000 0.0163
 Staff1.2 0 
 Worker15.45 8.64 
Defecation site0.0000 0.0015
 Outdoor32.88 13.15 
 Indoor latrine5.0 3.74 
Heath status0.0002 0.1647
 Excellent3.88 5.75 
 Average11.08 6.12 
 Poor29.5 14.0 
Water source0.0002 0.0130
 Ring-well40.0 15.90 
 Tap (running)7.6 5.6 
Crowding0.0031 NS
 ≤6 people/house6.4 6.18 
 >6 people/house15.4 8.00 
Level of education0.0000 NS
 None16.4 8.95 
 Primary13.3 7.86 
 Secondary4.9 2.23 
 Tertiary1.6 6.8 
Table 10.  Factors associated with the intensity of eggs per gram (epg) hookworm infection with a P-value of ≤0.25
FactorGeometric mean (epg)t-Test (P-value)Prevalence of heavy infection (%)Chi-square (P-value)
Age (years)0.0191 0.1068
  0–22.02 0 
  3–52.4 4.54 
  6–1015.6 8.33 
 11–154.9 0 
 16–209.3 9.68 
 21–307.8 13.1 
 31–405.9 11.7 
 41–5010.0 3.1 
 >5030.4 22.7 
Level of education0.0000 0.0355
 None13.7 13.4 
 Primary8.4 9.0 
 Secondary4.5 3.9 
 Tertiary1.2 0 
Footwear0.0000 0.1642
 Wears footwear outside3.5 5.9 
 Barefoot outside11.4 10.48 
Pig ownership0.0009 0.0534
 Yes43.4 21.05 
 No6.7 8.09 
Socioeconomic status0.0000 0.0065
 Staff1.2 0 
 Worker11.3 10.9 
Defecation site0.0005 0.0445
 Outdoors14.0 13.15 
 Indoor latrine5.3 6.54 
Health status0.0000 0.0839
 Excellent3.5 3.45 
 Average7.6 10.20 
 Poor33.5 13.95 
Crowding0.0013 NS
 ≤6 people/house5.1 8.42 
 >6 people/house11.8 9.3 
Table 11.  Unweighted least squares linear regression for the intensity of Ascaris
Predictor variablesβSEP-value
  1. R2 = 0.1209 (F = 8.85, P = 0.0000). Dependent variable: log(egg counts + 1).

Household crowding0.4280.1880.0233
Level of education−0.1550.0940.0989
Religion – Hindu0.6250.2010.0021
Pig ownership0.9260.3970.0203
Table 12.  Unweighted least squared linear regression for the intensity of Trichuris
Predictor variablesβSEP-value
  1. R2 = 0.1644 (F = 12.67, P = 0.0000). Dependent variable: log(egg counts + 1).

Household crowding0.2210.1280.0292
Level of education−0.1140.0690.1026
Barefoot when outdoors0.2650.1400.0595
Indoor defecation−0.5100.1480.0006
Tap or running water−0.5410.1830.0003
Table 13.  Unweighted least squares linear regression for the intensity of hookworm
Predictor variablesβSEP-value
  1. R2 = 0.1349 (F = 12.67, P = 0.0000). Dependent variable: log(egg counts + 1).

Household crowding0.2010.1130.0754
Level of education−0.2470.0590.0000
Barefoot when outdoors0.3360.1210.0060

Socioeconomic status

A number of environmental and behavioural factors were found to be highly correlated with an individual belonging to the staff or worker community (Table 14). For this reason, socioeconomic status was dropped from multiple regression models.

Table 14.  Environmental and behavioural factors highly correlated with socioeconomic status
FactorPrevalence staff (%)Prevalence workers (%)Pearson's correlations (P-value)
>6 people/household19.451.90.0000
Educated (primary and above)8951.50.0000
Use of footwear when outdoors72.627.40.0000
Outdoor defecation6.541.40.0000
Ring-well water source016.50.0005

Socioeconomic status was the strongest univariate predictor of prevalence and intensity of infection with all three geohelminths. Apart from a 1.6% (95% CI: 0.0, 4.7) prevalence of heavy infections with Ascaris, no staff members recorded heavy infections with either Trichuris or hookworm.


Figures 1–3 show the prevalence and average egg intensities (epg) of Ascaris, Trichuris and hookworm in different age groups. The prevalence of Ascaris peaked in the 3–5 [50% (95% CI: 29.1, 70.9)] and 16–20 [54% (95% CI: 37.3, 72.4)] year age groups. Average egg intensities for Ascaris was heaviest in the 3–5 year age group with 31.8% (95% CI: 12.4, 51.3) of this group carrying heavy infections. The prevalence of Trichuris peaked in the 3–5 [59.1% (95% CI: 38.5, 79.6)] and 6–10 year [56.3% (95% CI: 42.2, 70.3)] age groups. Both these groups also recorded the heaviest egg intensities with heavy infections in 18.2% (95% CI: 2.1, 34.3) of the 3–5 year old group and 16.7% (95% CI: 6.1, 27.2) of the 6–10 year old group. Hookworm infections peaked in the 6–10 year [60.4% (95% CI: 46.6, 74.3)] age groups. Maximum egg intensity of hookworm was recorded in the 16–20 year age group.

Figure 1.

Prevalence and intensity (eggs per gram, epg) of Ascaris among different age groups.

Figure 2.

Prevalence and intensity (eggs per gram, epg) of Trichuris among different age groups.

Figure 3.

Prevalence and intensity (eggs per gram, epg) of hookworm infections among different age groups.

Multifactorial risk factor analysis showed age to be a significant risk factor for the prevalence of Ascaris, with prevalence decreasing by a factor of 1.02 for every year of life (P = 0.0126). Multiple regression for intensity of infection showed increasing age to be a negative predictor for the intensity of Ascaris (P = 0.0104) and positive predictor for the intensity of hookworm (P = 0.0063). Age was not found significant for predicting the prevalence or intensity of Trichuris infections.


Although the overall prevalence, intensity and percentage of heavy infections with the three geohelminths were higher in females than in males, gender was not shown to be a significant risk factor. An obvious over contribution of females to the egg intensities was observed in the 16–20 year old groups for hookworms (females = 246 epg vs. males = 114 epg) and Ascaris (females = 6700 epg vs. males = 2790 epg) and the 11–15 and 21–30 year old groups (average females = 351 epg vs. males = 87 epg) for Trichuris.

Household crowding

On average, each household consisted of 6.75 members, with 46% of households containing more than six members. After adjusting for other factors, individuals living in households with more than six members were 2.0 (95% CI: 2.0, 3.3) times more likely to be infected with Ascaris (P = 0.0041), 1.6 (95% CI: 1.0, 2.6) times more likely to be infected with Trichuris (P = 0.0561) and 1.7 (95% CI: 1.0, 2.7) times more likely to be infected with hookworms (P = 0.0338) than individuals living in households with six or fewer members. Multiple regression showed that individuals living in households with more than six members also contributed to higher intensities of infection with Ascaris (P = 0.0233), Trichuris (P = 0.0292) and hookworms (P = 0.0754).


Overall 62.5% of individuals older than 6 years were either attending or had completed at least primary school. There was a significant correlation between having no formal education and gender (P < 0.0001, χ2 = 17.99), with 53% of males attending or having received formal education compared with only 29% of females. There was also a significant relationship between the level of education (none, primary, secondary or tertiary) and gender (P = 0.0001, χ2 = 20.58). Thirty-one percentage of males and 23% of females interviewed were or had attended primary school, 29% of males and 16.5% of females secondary school and 10.5% males and 7% of females tertiary school.

After adjusting for other variables, individuals were 1.3 times (95% CI: 1.0, 1.6) less likely to be infected with Ascaris (P = 0.0581), 1.4 (95% CI: 1.1, 1.8) times less likely to be infected with Trichuris (P = 0.0121) and 1.6 (95% CI: 1.3, 2.1) times less likely to be infected with hookworm (P = 0.0093) with increasing levels of education. When other variables were taken into consideration, an individual's level of education remained a strong predictor for the intensities for all three geohelminths.


Hindus constituted the majority of the population (72.6%) followed by Christians (22.3%), Buddhists (4%) and Muslims (1.2%). After adjusting for other variables, Hindus were 2.5 (95% CI: 1.4, 4.5) times more likely to be infected with Ascaris (P = 0.0017) than individuals of other religions. Hinduism also remained significant in the final regression model for Ascaris egg intensity (P = 0.0021).


Most of the population (64%) admitted to walking barefoot when outdoors. After adjusting for other variables, individuals walking barefoot outdoors were 2.0 (95% CI: 1.2, 3.4) times more likely to harbour hookworm (P = 0.0002) and 2.1 (95% CI: 1.2, 3.6) times more likely to harbour Trichuris (P = 0.0069). Egg intensities of individuals walking barefoot outdoors were also significantly higher for both hookworm (P = 0.0060) and Trichuris (P = 0.0595) on multiple regression.

Defaecation practices

Although 80.2% of individuals had access to a functional indoor toilet, only 65.2% of the population used these toilets. The remainder would either defecate in their backyard or in a nearby field. However, regardless of an individual's preferred defecation site while at home, the tea-pickers had no option but to defaecate in the tea gardens during working hours. Outdoor defecation was most common among children under 10 years of age (69.5%). Common reasons for defecating outdoors included dysfunction of the septic tank, defecating outdoors being perceived to be more hygienic and in the case of children, the possibility of accidents while using the open latrine unattended. Outdoor defecation only remained a significant risk factor for the prevalence (P = 0.0334) and intensity (P = 0.0006) of Trichuris in the multifactorial model.

Anthelmintic treatment

The appointed Health Officer was responsible for treating children at and below the age of 12 years with either pyrantel or levamisole at recommended dose rates (10–15 mg/kg for pyrantel and 3–5 mg/kg for levamisole, as a single administration) on a biannual basis in each tea estate. Albendazole was only prescribed by medical practitioners on an individual case basis if clinical symptoms of parasitism were apparent. According to their parents or guardians, 83% of children below the age of 3 years and 71% of children below the age of 10 years received anthelmintic treatment biannually. Forty percent of the remaining population reported taking anthelmintic treatment at least once every 3 years. Multivariate analyses showed that the use of anthelmintics at least once every 3 years was significant in reducing the prevalence of Ascaris 1.9 (95% CI: 1.2, 3.1) times (P = 0.0146) and hookworms 1.8 (95% CI: 1.1, 2.9) times (P = 0.0177), but had no effect on the prevalence of Trichuris. Anthelmintic usage had no significant effect on the intensity of infection with any of the geohelminths.

Pig ownership

Although pig rearing is forbidden, 19 individuals belonged to households that kept domesticated pigs. Whilst pig-ownership did constitute a univariate risk factor for both the prevalence and intensity with all three geohelminths, it only remained a significant predictor for the intensity of Ascaris (P = 0.0203) in the final regression model. There was a significant association between individuals who owned pigs and defecation practices: 89% of individuals in households that reared pigs defecated outdoors (P = 0.0000, χ2 = 26.63).

Water source

The majority of people had access to running tap water for the household, but 13.4% had access to water solely from an open ring-well that ran approximately 50 ft deep. Water source was found to be a significant risk factor for both prevalence and intensity of Trichuris infections. After adjusting for other variables, individuals using ring-well water were 1.9 (95% CI: 1.2, 3.6) times more likely to harbour Trichuris (P = 0.0122) and to have significantly higher intensities of infection with Trichuris (P = 0.0003) than individuals using running water in their households.

Health status

Twenty-seven percent of individuals reported excellent health status, 60% average and 13.2% poor health status. The intensity of Ascaris (P = 0.0012), Trichuris (P = 0.0002) and hookworm (P < 0.0001) infections increased significantly with declining health status.


Infection with the geohelminths Ascaris, Trichuris and hookworm were found to be endemic among this remote tea-growing community in Assam. The most striking observation from this study was the strong association between socioeconomic status and the prevalence and intensity of all three geohelminths. Subfactors that significantly correlated with socioeconomic status, such as household crowding, literacy, use of footwear, defecation practices and sources of drinking water, were identified as being important determinants for the difference in prevalence and intensity of infections observed in these two groups.

Our study was consistent with previous findings (Anderson 1986; Haswell-Elkins et al. 1987) showing that in an endemic community, the age at which predisposition to heavy infections occurred was the 3–10 year old groups for Ascaris and Trichuris and early adulthood (16–20 year olds) for hookworms. It was noteworthy that apart from a single 9-year-old child belonging to a staff household, no other member of the staff community, regardless of age, was observed to harbour ‘heavy’ infections with either Ascaris, Trichuris or hookworm. It would appear that children from lower socioeconomic status harboured significantly higher intensities of infection due to other exposure and behavioural subfactors strongly correlated with socioeconomic status rather than age alone (Holland et al. 1988).

Although gender was not a significant risk factor for either prevalence or intensity of geohelminth infection, females were found to harbour over-proportional intensities of Ascaris and hookworm between the 16–20 year old age groups and the 21–30 year old age groups for Trichuris. This finding could be partially explained by the difference in gender roles. Males in their early adult life were more likely to adopt duties confining them to indoor environments (e.g. factory workers) whereas girls were more likely to commence duties as tea-pickers. The tea gardens were a common place of defecation for female workers during working hours and therefore the contamination of soil from these areas with geohelminth eggs would constitute a significant risk for infection (see below).

Religion appeared to play a significant role in determining the prevalence and intensity of infection with the geohelminths. Hindus were found to harbour a higher prevalence and intensity of geohelminths, in particular Ascaris. Studies correlating ethnicity and parasitism have been conducted in Malaysia (Kan et al. 1989) suggesting that different ethnic groups have different patterns of infection with geohelminths. Although all workers in this tea community were primarily of Indian ethnicity, religion was shown to be the main determinant of the cultural, dietary and behavioural factors that could have contributed to the differences observed.

An individual's education level or in the case of a child, maternal education, has been found to be one of the most important risk factors for parasitism in numerous studies (Toma et al. 1999; Phiri et al. 2000). We confirmed this finding and found that as level of education increased the likelihood of infection with each geohelminth declined significantly.

There is sufficient data to show that individual behaviour and exposure factors may enhance the risk of infection. The density of people in a house was shown to positively influence the frequency of exposure to infective stages of Ascaris (Haswell-Elkins et al. 1989), Trichuris (Narain et al. 2000) and hookworm (Olsen et al. 2001b). Our study confirms results showing household crowding to be a significant determinant of both prevalence and intensity of all three geohelminths.

Although it is argued that in highly contaminated environments, focal transmission in defecation grounds may not be the single most important site for geohelminth infection (Feachem et al. 1983), we found it to be a significant univariate risk factor for infection with all three geohelminths and a multivariate risk factor for Trichuris. In this community children were at particular risk with nearly 70% of children under 10 defecating outdoors in either the back yard or nearby field. At this age, non-supervised children may be more prone to defecating in sites that are already polluted by faeces and therefore be exposed to more frequent and heavy infection than adults (Hominick et al. 1987). A gross underestimation of defecation outside a latrine by female workers is also expected. When in the fields, regardless of defecation practices while at home, female tea-pickers had no option but to discretely defecate in the furrows between the tea trees. This environment of damp alluvial soil, shaded by tea trees, provides a perfect environment for development of geohelminth larvae and would constitute a significant source of infection and re-infection.

The association between walking barefoot while outdoors and a higher prevalence and intensity of hookworm infection would be expected and explained due to the cutaneous penetration of larvae as an infection source. However, why the same correlation has been shown true for Trichuris infection remains unclear.

Pig ownership proved a significant univariate risk factor for the prevalence and intensity of all three geohelminths and also remained in the final multivariate model for intensities of Ascaris. Nearly 90% of individuals living in households that reared pigs defaecated outdoors which may reflect poorer standards of hygiene by these households. This risk could be exacerbated by the fact that pigs are coprophagous and may be acting as mechanical transporters and focal disseminators for infection with human parasitic stages (Olsen et al. 2001a; Traub et al. 2002). Moreover, the zoonotic potential of human- and pig-derived Ascaris remains unclear (Anderson 2001); hence it is possible that cross-infection in these households did occur (Anderson 1995).

Individuals who sourced drinking water from ring-wells had a greater prevalence and intensity of Trichuris infection than those who had access to running tap water. Although water source has been implicated as a risk factor for Trichuris here and in other studies (Norhayati et al. 1998; Narain et al. 2000), it seems an unlikely direct source of Trichuris infection (Feachem et al. 1983). Even if the open ring-wells were contaminated with helminth eggs, they would be expected to rapidly sediment to the bottom becoming inaccessible for collection with a bucket. Curtis et al. (1995) demonstrated that mothers from a poor community in Burkina Faso with access to tap water in the yard were more likely to use safe hygiene practices than mothers using wells in the yard (Curtis et al. 1995). It is possible that poor hygiene practices associated with access to well water are the more probable risk factor for increased Trichuris infection among this community. However, the question as to why the other geohelminths did not display a similar correlation remains unresolved.

Use of anthelmintics within the community was effective in reducing the prevalence of Ascaris and hookworms, but had no effect on the prevalence of Trichuris. The lack of efficacy of pyrantel and poor and variable efficacy of levamisole (cure rate 16–18%, egg reduction rate 73%) against Trichuris (WHO 1996) could account for this finding. However, the drugs used showed no impact on reducing the intensities of infection with either Ascaris or hookworm. The most probable reason for this finding could be the high rate of re-infection by geohelminths in this community.

The endemic geohelminth problem in this tea-growing community in Assam warrants an integrated control programme, which can only be achieved if financed and supervised by medical and managerial staff at each estate. Intensity of helminth infection and the degree of anaemia are negatively associated with labour productivity in female tea-pickers in Bangladesh (Gilgen et al. 2001). We also found that the individuals’ perception of their health status worsened with increased intensity of infection with all three parasites. It is therefore also in the interest of long-term productivity that practical and cost-effective measures be employed to tackle this problem among the workers.

The main aim of helminth control programmes is to reduce and sustain the intensity of infection below that associated with morbidity (WHO 1996). The WHO expert committee set guidelines for control programmes based on geohelminth prevalence and intensities of infection. Although the set boundaries for categorizing egg intensities into light, moderate and heavy can be strongly debated (Hall & Holland 2000), WHO propose a universal blanket treatment in areas where hookworm infections are endemic (prevalence >20–30%) and where anaemia is prevalent. To maintain low intensities, and as a result morbidities of infection with the three geohelminths, it is recommended that a more intensive anthelmintic treatment regime be administered within this community. A programme consisting of mass treatment of children aged between 3 and 15 years of age at 4–6-monthly intervals and the adult population at 12-monthly intervals with a broad spectrum anthelmintic such as Albendazole is recommended. Personnel costs for executing this programme would be low, as Health Officers are already employed by the management to execute such programmes. The cost of Albendazole based on current prices on the international market of approximately $US 0.03 (PCD 1999; Guyatt 2003) would result in an estimated cost of approximately $US 250 per tea estate per year.

Scholastic and health education through already existing programmes such the Mothers’ Club, especially among females, would by far outweigh any other form of non-chemotherapeutic intervention in this study community. In many cases, although resources allow for proper practices, belief structures, retained behavioural practices and poor awareness of disease transmission lead to non-compliance. For example, despite having a functional latrine, some people prefer to defecate or allow children to defecate outdoors, perceiving it to be a cleaner practice. Similarly, although the workers are provided with rubber thongs annually, some people walk barefoot outdoors. Improvements in sanitation by regular maintenance of septic tanks, installation of temporary deep pit latrines in plantation sites, increasing access to running tap water and stricter enforcement of pig rearing regulations would undoubtedly lead to improved living and working conditions for this community. For interventions to be effective, helminth control must first take on a high priority by medical and managerial staff, it must be implemented in an integrated fashion and be sustainable. The benefits of such a programme will not only be in the interests of the workers at a grass-root level, but the improvements to long-term health and therefore productivity at minimal cost will undoubtedly lead to economic benefits for the industry and region as a whole.


We thank Williamson Magor and Co. for permission to conduct research at their tea estates, and the entire personnel at the Phulbari, Addabari and Balipara Tea Estates for their help and cooperation during sample collection. We also thank Bayer AG, BHC Business Group Animal Health (Leverkusen, Germany) for providing Rebecca J Traub with a postgraduate scholarship through the Veterinary Trust, Division of Veterinary and Biomedical Sciences, Murdoch University.