Intestinal helminth infections, anaemia and labour productivity of female tea pluckers in Bangladesh
Gilgen Department of Public Health and Epidemiology, Swiss Tropical Institute, Socinstr. 57, 4002 Basel, Switzerland. E-mail: firstname.lastname@example.org
We conducted a randomized clinical intervention trial over 24 weeks on a tea estate in north-east Bangladesh to investigate the effect of iron supplementation and anthelmintic treatment on the labour productivity of adult female tea pluckers. A total of 553 full-time tea pluckers, not pregnant and not breastfeeding, were randomly assigned to one of the four intervention groups: group 1 received iron supplementation on a weekly basis, group 2 received anthelmintic treatment at the beginning and half-way through the trial (week 12), group 3 received both iron supplementation as group 1 and anthelmintic treatment as group 2, and group 4 was a control group and received placebos. No significant difference in labour productivity was found between the four intervention groups over the trial period. However, there was a negative association for all three worms (Ascaris lumbricoides, Trichuris trichiura and hookworms) between the intensity of helminth infections (eggs/g faeces) and all measures of labour productivity. Lower haemoglobin values and anaemia (< 120 g/l Hb) were both associated with lower labour productivity and more days sick and absent. Taller women with greater arm circumference were able to pluck more green leaves, earn higher wages and were absent less often.
Intestinal parasitic infections, particularly from Ascaris lumbricoides, Trichuris trichiura and the two hookworm species Ancylostoma duodenale and Necator americanus are often associated with conditions such as malnutrition, vitamin A deficiency, diarrhoea and iron deficiency anaemia (Crompton et al. 1985; Stephenson 1987). Hookworms and Trichuris may also cause anaemia by consuming blood and causing plasma leakage (Gilles 1990; Cooper et al. 1992; Solomons 1993).
Field studies conducted in the 1970s and 1980s established a relationship between poor nutritional status and functional impairment including decreased physical fitness and decreased work capacity. The conclusions of these studies indicated that reduced physical ability and ill health are impairing the capacity to perform sustained moderate to heavy labour and this reduction in work load, work pace or time spent at work reduces labour productivity considerably (Davis 1973; Satyanarayana et al. 1977; Spurr et al. 1977; Basta et al. 1979; Brooks et al. 1979; Edgerton et al. 1979; Immink et al. 1982; Wolgemuth et al. 1982; Latham 1983; Bradley et al. 1988). Non-mechanized, agricultural labour is the main source of family income in most developing countries, where lower labour productivity due to poor health has profound economic consequences. A tea plantation provides ideal conditions for studying labour productivity as work attendance and weight of the green leaves plucked by each worker are carefully recorded in order to calculate individual wage packets (Bradley et al. 1988).
Subjects and methods
Field work took place between November 1995 and March 1997 in a tea estate in north-east Bangladesh and included a 24-week stratified randomized clinical intervention trial. A total of 553 non-pregnant and non-breastfeeding tea pluckers were recruited into the study. All participants were volunteers. Full details of the project were provided and a consent form was signed by each plucker. Ethical clearance for the study was obtained through WHO. Two surveys, one at baseline (pre-intervention) and one post-intervention, were conducted involving anthropometry and analyses of a blood and stool sample. Haemoglobin and ferritin levels were measured and prevalence and egg counts of A. lumbricoides, T. trichiura and hookworms determined.
The random number generator in SPSS (Version 7.5) was used to create four groups of equal size and the process was repeated until there was no statistically significant difference between the randomized groups in mean age, years of plucking experience, productivity of the previous plucking season, haemoglobin and ferritin values and prevalence and egg counts of Ascaris, Trichuris and hookworms.
• Group 1 received weekly iron supplementation for 24 weeks (n=139),
• Group 2 received anthelmintic treatment on two occasions at the beginning of the trial and 12 weeks later (n=143),
• Group 3 received both iron supplementation as group 1 and albendazole as group 2 (n=130),
• Group 4 received placebos for both iron supplementation (weekly) and anthelminthic treatment (beginning and 12 weeks later, n=141).
Each tea plucker had to pluck a specified number of kilograms of green leaves per day in order to qualify for the basic wages. The daily task for each type of tea bush (pruned, unpruned, young tea) varied, for pruned bushes it was 20 kg, 120 kg per week, while the equivalent for 6 days on unpruned bushes was 96 kg per week. No daily minimum was required for young tea. A plucker who picked the required amount for the type of bush would receive 22 takas/day (40 takas=1US$), the basic wage, and each extra kilogram earned a productivity bonus of 0.45 takas. Four measures of labour productivity were compiled for each worker:
• kilograms of green leaves plucked per plucking day (kg/pld),
• Average wages in takas earned per plucking day (tk/pld),
• Number of days on sick leave, and
• Number of days absent from work.
The indices kg/pld and tk/pld were computed to control for days spent on other duties, days absent and sick, as well as for differences in the type of tea bush being plucked. Although wages earned are not work output in the strictest sense, they nonetheless represent work output indirectly as they comprise both the basic wages for the required task and any bonus earned for additional work performed. Analyses of labour productivity were conducted at three time periods: during the first half of the trial period (weeks 1–10), the second half (weeks 11–20) and over the whole trial period (weeks 1–20). Weeks 21–24 of the trial coincided with the end of the plucking season and labour productivity data from the last 4 weeks were omitted from the analyses as working hours were much reduced and there was no opportunity for workers to earn extra pay. For the first half of the trial, kg/pld and tk/pld were calculated using the total amount of green leaves plucked and takas earned in weeks 1–10 divided by the total number of days on plucking duty in the same weeks. For the second half of the trial, the total amount of kilograms of green leaves plucked and wages earned were divided by the total number of days on plucking duty in this time period. Pre-trial values of haemoglobin, ferritin, worm egg counts and anthropometry were used to assess any association between pre-trial values and labour productivity of the first half of the trial. Post-trial values were used for computations of both the second half of the trial and the whole trial period.
Height, weight and mid-upper arm circumference were measured for each individual following recommended techniques (Lohman et al. 1988). The measurements for height and weight were taken with the women wearing a cotton sari, a blouse and a petticoat. Anthropometric measurements were taken using a portable anthropometer (Sieber-Hegner & Co. France SA), a battery operated, electronic balance (Soehnle, Germany) and an insertion tape. Each individual was measured by the same assistant on both occasions, pre- and post-trial. Both inter- and intra-examiner coefficients of reliability were between 0.90 and 0.99 for all parameters measured which is within the region of acceptable technical error (Ulijaszek & Lourie 1994).
Anthelminthic treatment and faecal microscopy
Two single doses of albendazole 400 mg each (Aristopharma Ltd, Dhaka, Bangladesh) were administered to each worker in groups 2 and 3; the first dose on the first day of the trial and the second dose in week 12. Group 1 and group 4 received placebos. The tea plantation management took responsibility for deworming all residents of the plantation after the trial.
Stool samples were collected at baseline, after deworming and after the trial. Samples were collected and processed on the same day using the formalin–ether concentration method. The weight of the stool amount analysed and the number of drops in the sediment were recorded in order to calculate egg count per gram (epg) of stool. Laboratory work was performed according to the WHO Safety Code (1991).
Iron supplementation and haemoglobin and ferritin determinations
One capsule containing 200 mg ferrous fumarate and 200 mg folic acid (Aristopharma Ltd, Dhaka, Bangladesh) was given once a week throughout the 24-week trial to each individual in groups 1 and 3 while those in groups 2 and 4 received a matching placebo produced and supplied by the same company. All individuals in the two groups who were not supplemented during the trial and were found anaemic post-trial received a 12-week course of iron supplementation.
Capillary blood samples were drawn for analyses of haemoglobin and ferritin values. Haemoglobin was analysed using a battery-powered HemoCue photometer. Two to three drops were collected in a sterile plastic 1.5 ml Eppendorf tube, spun and the serum samples stored at −20 °C for 4 weeks before analysis using the NovaPath Ferritin Enzyme Immonoassay Kit.
All continuous variables were first checked for normality and those showing significant skewness were, in the main, normalized using either log(10) transformation for egg counts or square root transformation for ferritin. Where normalization was not possible, both parametric and non-parametric tests were used. Statistical analyses were performed with the statistical package SPSS (Version 7.5 for Windows).
The 553 tea pluckers were on average 39.6 years old (minimum 14, maximum 66) and they had a mean of 17.9 years of plucking experience (minimum 0, maximum 51).
The initial prevalences were 47.6% for A. lumbricoides, 56.8% for T. trichiura, and 74.4% for hookworms. Only 10.7% of the pluckers were free of infection as defined by absence of worm eggs from their stool samples; 25.7% of the pluckers had a single infection, 33.2% double infection and 26.6% were infected with all three worms.
The geometric mean egg counts (intensities of infection) for the whole sample were 20.8 epg stool for Ascaris, 14.8 for Trichuris and 57.7 for hookworm. Geometric mean egg counts of infected women (positives only) increased to 592.9 for Ascaris (minimum 14 epg, maximum 14 410 epg), 114.3 for Trichuris (minimum 15 epg, maximum 4496 epg) and 230.7 for hookworms (minimum 14 epg, maximum 6458 epg).
Prevalence of anaemia (haemoglobin < 120 g/l) was 85.7% at baseline; the mean haemoglobin level was 98.4 g/l (minimum 38 g/l, maximum 149 g/l) and serum ferritin 27.9 mg/l (minimum 1 mg/l, maximum 102 mg/l).
Mean values for anthropometric variables were: height 147.9 cm (minimum 134.9 cm, maximum 165.7 cm), weight 37.5 kg (minimum 22.4 kg, maximum 62.6 kg), BMI 17.1 (minimum 11.0, maximum 24.1) and the average mid upper-arm circumference was 21.2 cm (minimum 15.5 cm, 27.4 cm maximum). The four intervention groups were not statistically different from each other at baseline in any of the above listed parameters.
Treatment with albendazole achieved a reduction in both the prevalence and intensity of helminth infections. The cure rate after treatment was 98.2% for Ascaris, 33.5% for Trichuris and 88.1% for hookworms. The fall in mean egg counts after treatment was 93.3% for Ascaris, 67.9% for Trichuris and 97.2% for hookworms.
Iron supplementation led to a marked increase in haemoglobin concentration in the supplemented groups, averaging 5.5 g/l in the group who received iron only and 7.8 g/l in the group receiving both iron and albendazole. The haemoglobin increase was smaller in the group who received anthelmintic treatment only (1.9 g/l) while the control group showed a slight decrease in haemoglobin (0.2 g/l) (Table 1).
Haemoglobin values (g/l) pre- and post-intervention by treatment group
There were no significant differences in the number of days plucked on pruned, unpruned and young tea bushes between the four intervention groups. Overall, more tea was plucked during the first half of the trial and more pluckers were sick, absent, and on leave in the second half of the trial (Table 2). There was no significant variation between the four intervention groups in labour productivity during the trial period. Comparison of work output between the first half and the second half of the trial within each group showed that work output (kg/pld, tk/pld) was much lower in the second half (Table 3).
Summary of labour productivity data of the first and the second half of the trial (whole study population, n
Green leaves plucked and wages earned per plucking day in the first and second half of the trial by intervention group
Regression analyses showed that both kilograms of green leaves per plucking day and wages earned per plucking day declined the more days the plucker was absent; there was no association with the number of days sick. Pluckers in the lowest wage quartile of quantity of green leaves plucked per day were absent from work for 7.8 days on average; whereas the absenteeism in the highest quartile of pluckers averaged 2.3 days (P < 0.001), suggesting that pluckers did not make up for lost pay on the days they did work.
Helminth infections and labour productivity
Higher Trichuris egg counts were associated with less wages earned per plucking day. Higher Trichuris egg counts also showed a significant positive association with days sick and days of absence from work (Table 4). Higher egg counts of hookworms (infected individuals only) were associated with a higher number of days sick. Individuals with 1000 or more Ascaris epg earned significantly lower daily wages and were also absent more often (Table 4).
Summary table of statistically significant associations between predictors and labour productivity (regression analyses)
Helminth infections and iron deficiency anaemia
There was a significant negative association between hookworm egg counts and ferritin level (P < 0.05) but no association was found with Ascaris or Trichuris. For Trichuris there was a negative association between intensity of infection and haemoglobin concentration (P < 0.05).
Iron deficiency anaemia and labour productivity
Anaemic workers (haemoglobin <120 g/l, post-trial) plucked a daily average of 26.7 kg of green leaves during the whole trial period, whereas non-anaemic workers plucked an average of 28.5 kg. Anaemic individuals earned only 25.9 takas per plucking day compared with non-anaemic individuals who earned an average of 27.0 takas per day. On average, the number of days absent from work because of sickness for an anaemic worker was 1.9 days, while for a non-anaemic worker it was 1.6 days. Anaemic individuals were absent from work for 2.7 days on average compared with 1.8 days for non-anaemic individuals (Table 4).
Pre-trial haemoglobin values showed a significant association with the kilograms of green leaves plucked and wages earned in the first half of the trial. The degree of iron deficiency anaemia post-trial also showed a significant negative association with daily wages, and a significant positive relationship with the numbers of days sick and days absent in the second half of the trial. Iron-deficient individuals (ferritin <12 mg/ml) tended to pluck less, earn less and were absent more often than non-iron deficient individuals although the results were not statistically significant.
Nutritional status and labour productivity
There was a positive, highly statistically significant association between height, weight (not BMI) and mid upper-arm circumference and green leaves plucked per day (Table 4). However, a multiple regression analysis including all three anthropometric variables revealed that height was the most significant predictor for labour productivity, followed by mid upper-arm circumference. On average, workers in the shortest height quartile plucked 25.0 kg of green leaves per day during the peak plucking season; the tallest quartile plucked on average 31.3 kg per day. The tallest and heaviest quartile were absent significantly less often than the smaller and lighter pluckers (Table 4).
Sequential (hierarchical) multiple regression was used to test whether there was any association between helminth infections and labour productivity after removing the effects of a number of independent variables (age, years of plucking experience, anthropometric variables, haemoglobin and ferritin values). The results obtained confirmed the results of the bivariate analyses.
Trichuris egg counts were negatively associated with daily wages in the first half of the trial (P < 0.05). There also was a positive association between egg counts, the number of days sick (P < 0.01) and the number of days absent from work (P < 0.05). Individuals with more than 1000 Ascaris epg earned significantly less in the first half of the study (P < 0.01) and were absent more often (P < 0.01) during the whole trial period than those with lower egg counts.
Higher hookworm egg counts showed a significant negative association with ferritin values (P < 0.01), and individuals with higher hookworm egg counts had significantly more days of sick leave in the second half (P < 0.01) and during the whole trial period (P < 0.01).
Haemoglobin values showed a significant positive association with work output throughout the first half (P < 0.001) as well as throughout the whole trial period (P < 0.05); the higher the haemoglobin concentration, the higher the work output.
Height was a highly significant predictor of labour productivity (P < 0.001), followed by mid upper-arm circumference (P < 0.01).
Changes in labour productivity
Comparison of work output between the first and second half of the trial showed that work output was much lower in the second half. While the trial was begun at the height of the plucking season, the end of the trial coincided with its tail end, when much fewer leaves were available for plucking. Unfortunately, these last weeks of the trial was the time when the effects of the intervention would be expected to be most pronounced.
There was no significant difference in labour productivity between the four intervention groups over the trial period. Group 1 (iron only) plucked most and earned the highest wages in the whole trial period followed by group 4 (control) and then the two dewormed groups, group 2 (albendazole only) and group 3 (iron and albendazole). However, this group order in labour productivity did not correspond with the mean post-intervention values for any of the significant predictors of labour productivity, such as haemoglobin, intensity of helminth infections, height, mid upper-arm circumference, age and experience.
In their pilot study Bradley et al. (1988) found that the group receiving iron only plucked the greatest amount of green leaves before, during and after the project, while the group who received iron and albendazole improved their work output during the project. However, Bradley et al. did not calculate the amount of leaves plucked per plucking day and no attempt was made to correct for days absent or on other duties. Neither is there any information on helminth infections. In both Bradley's and our study, it was iron supplementation which seemed to have had the greater effect on labour productivity than deworming.
A dose of 200 mg ferrous fumarate (66.67 mg elemental iron) was administered weekly as there was sufficient evidence at the time of the trial that there is no significant difference between daily and weekly supplementation (Gross et al. 1994; Ridwan et al. 1996; Berger et al. 1997; Palupi et al. 1997). However, field studies (Bradley et al. 1988; Kaltwasser et al. 1991; Jacobs et al. 1993; Stoltzfus et al. 1998) showed that it can take weeks or months for anthelmintics or iron supplementation to have any significant effect on haemoglobin values. Therefore, pre-intervention data of haemoglobin, ferritin, worm egg counts and anthropometry were used to analyse labour productivity at the beginning of the trial. Post-trial data of haemoglobin, ferritin, egg counts and anthropometry were used to assess the impact of the intervention on labour productivity (second half of the trial) and to assess productivity of the whole trial period.
Helminth infection and labour productivity
The main functional consequence of helminth infections assessed in this study is the possible effect on labour productivity. Warren et al. (1993) suggested that hookworm infection had a negative effect on labour productivity which was reversed with treatment. We also found intensity of infection with all three helminths, Ascaris, Trichuris and hookworms, to be negatively associated with labour productivity. Anaemia and poor nutritional status were also reported to reduce labour productivity in studies where the participants were infected with Ascaris, Trichuris, hookworms and Schistosoma (Brooks et al. 1979; Latham 1983).
The geometric mean egg count of Trichuris infected individuals was only 114.3 epg pre-trial, which was regarded as light infection by Montresor et al. (1998). Nonetheless, egg counts of the whole study population showed a negative association with wages earned and a positive association with the number of days sick and days absent from work. These results suggest that labour productivity was lower the higher the Trichuris egg counts.
Higher hookworm egg counts were associated with more days sick, but had no direct impact on the amount of green leaves plucked or on wages earned. However, hookworm egg counts showed a significant negative association with ferritin levels, while ferritin showed a highly positive correlation with haemoglobin.
Heavy worm loads with Ascaris are associated with malabsorption, malnutrition and reduced physical fitness in adults, while worm burdens are directly related to the degree of wasting and stunting in children (Stephenson 1987; Tshikuka et al. 1997). In our study, tea pluckers with Ascaris egg counts of 1000 or more had lower daily wages and were absent for more days. Latham (1983) and Brooks et al. (1979) also found reduced labour productivity because of poor nutritional status in a population with Ascaris, Trichuris and hookworm infections. But in our study, intensity of helminth infections was not associated with a higher degree of malnutrition as measured by weight, BMI and mid upper-arm circumference.
Iron deficiency anaemia and labour productivity
Productivity studies have concluded that there is a strong association between haemoglobin and work output (Gardner et al. 1977; Basta et al. 1979; Wolgemuth et al. 1982; Bradley et al. 1988; Scholz et al. 1997). In this study, baseline haemoglobin values were correlated with the amount of green leaves plucked per day as well with daily earnings. Anaemic pluckers (haemoglobin <120 g/l) earned lower wages throughout the whole trial period.
Anaemia has generally been thought to affect the oxygen-carrying capacity of the muscles and VO2max in strenuous tasks such as sugar-cane cutting (Spurr et al. 1977), road construction (Brooks et al. 1979) and work in rubber plantations (Basta et al. 1979), as well as in physically less demanding tasks such as factory work (Scrimshaw 1984; Li et al. 1994). Gardner et al. (1977) found that individuals with haemoglobin concentrations between 110 and 119 g/l showed about a 20% decrease in work tolerance, based on heart rate or actual work performance. Scholz et al. (1997) associated haemoglobin values in less strenuous work with attention and alertness. Workers described by the foreman as `lazy', `weak' or `stupid' were almost invariably anaemic (Basta et al. 1979; Scrimshaw 1984).
Anaemic pluckers in this study were also found to be absent more often from work which resulted in them not receiving any wages or rations for that day. Basta et al. (1979) showed that lower haemoglobin levels were associated with increased incidents of infectious diseases such as diarrhoea, influenza and bronchitis. Such diseases would cause a plucker to remain away from work but perhaps not to seek medical attention and thus to be recorded absent rather than to be signed off sick. Anaemic pluckers did not seem able to make up for lost days in working harder on the days they did work, although Brooke et al. (1988) suggested over- or under-compensatory behaviour of individuals when confronted with diminished working potential.
Nutritional status and labour productivity
A total of 82% of the tea pluckers had a BMI below 18.5, the cut-off point suggested for chronic energy deficiency (Shetty & James 1994). However, François and James (1994) warned that a cut-off point of 18.5 was the outcome of a series of compromises and is likely to categorize as underweight a proportion of a supposedly normal population. Nonetheless, BMI categories and the resulting grade of chronic energy deficiency proposed by Shetty and James (1994) suggests that 32% suffered from grade 1 chronic energy deficiency (BMI 17–18.45), 24% from grade 2 (BMI 16–16.99) and 26% from grade 3 (BMI below 16), whereas only 4.5% had a BMI of more than 20.
Most studies found an association between nutritional status and labour productivity. However, the same variables were not found to be important in all cases. Both variation in weight (Wolgemuth et al. 1982; Latham 1983; Torún et al. 1989) and in height (Satyanarayana et al. 1977; Spurr et al. 1977; Immink et al. 1984) have been compared for assessment of productivity. Spurr et al. (1977) and Immink et al. (1984) concluded that tall cane cutters in Guatemala were significantly more productive than short workers and that there may be a critical minimum adult height below which productivity is significantly reduced. In work where height is especially important, such as sugar-cane cutting, shorter people will be disadvantaged. Their choice of work and employment prospects are limited or labour productivity will be reduced in work where height is crucial.
Among tea pluckers it was not BMI which was associated with labour productivity, instead height was found to be the single most significant predictor for kilograms of green leaves plucked and wages earned, followed by mid upper-arm circumference. The average height of the tea pluckers was 147.9 cm; mature tea bushes were allowed to grow to a height of approximately 70–120 cm. Only young leaves from the top of the bush were plucked, which required the pluckers to work all day with elevated arms. Taller individuals with larger arm circumference plucked significantly more green leaves every day while the shortest women plucked the smallest amount.
The association between status of health and labour productivity may be complicated by several other factors such as motivation, economic pressures and incentives. Kennedy and Garcia (1994) suggested that the relationship may be reversed, in that increased income improves nutritional status. Hewner and Sun (1995) named motivation as the single most important factor which had a positive association with labour productivity. In a tea garden, labourers identify strongly with the tea estate and its total productivity. The labourer status is inherited and jobs are kept in families for generations and the right to live on the estate is dependent on employment of at least one family member. The social network is very strong too, as colleagues are also neighbours and often family or caste members. Perhaps the increment of 0.45 takas/kg green leaves plucked above the daily requirement was not sufficient motivation to work at maximum capacity every day.
All results from this study seem to point in the same direction: haemoglobin values increased in the supplemented and in the dewormed groups, while prevalence and intensity of infection with all three helminths was reduced in the dewormed groups. Both prevalence and the degree of anaemia and the intensity of helminth infections had a negative effect on labour productivity. Yet, no significant difference in labour productivity was observed between the four intervention groups. It might not only take months for haemoglobin levels to increase significantly, but also additional time to have a measurable effect on labour productivity. The effects of iron supplementation were contradicted by the seasonal nutritional stress during the rainy season which resulted in a decrease in weight and ferritin concentrations. Furthermore, the reduction in yield of the tea bushes in the last weeks of the trial meant that the pluckers were not under as much pressure as at the beginning of the trial, so that differential haemoglobin concentration or differences in intensity of helminth infections, were not as strongly associated with labour productivity any more.
Sincere thanks go to the tea pluckers and to Mr AMSO Subhan of Duncan Brothers Limited (Bangladesh) for their support for this research project. This study was funded by UNICEF and The Nestlé Foundation, Lucy Cavendish College, the Cambridge Philosophical Society and The Board of Graduate Studies of Cambridge University.