Data collected includes a survey on the effect of malaria on agricultural worker productivity and physical activity as well as the demand for malaria health insurance. A mobile health clinic was established on the plantation and used an exogenously determined ordering of workers to test and treat workers. Despite large treatment effects, we find that workers have low rates of seeking curative and preventative treatments. To better understand this puzzle, this study offers access to malaria treatment and insurance at exogenously varied prices to estimate its effect on take-up and frequency of health care. In another study phase, the study will also measure the effect of treatment on both worker productivity and physical activity. This will allow estimation of the effects of malaria on physical activity in general, and allow us to extrapolate our findings in this context to other physical occupations in endemic areas. The consequences of ill health for productivity and economic development are presumed to be severe yet the rigorous evidence base for such a linkage is small. In previous work, statistically significant and large intent to treat effects were estimated on earnings, labor supply and productivity of a curative malaria treatment at a large sugarcane plantation in Nigeria.

Data was collected in two different rounds (Round 4 and 5) using different data collection methods in each of the rounds. Round 4 Data: Survey Design This study implemented a randomized testing and treatment program within a large plantation that hires sugarcane cutters throughout the sugar cane cutting agricultural season (November-April) and pays them on a piece rate basis. Across the plantation, supervisors collect daily worker output using a standardized measuring stick. A worker’s payment is based entirely on his output. The plantation’s wage is 2.24 Naira (1 USD = 165 Naira approx.) per meter of cane cut. Workers carefully observe the recording of their output as it is the basis of their monthly payments from the plantation, often maintaining their own separate ledger. The daily output data is collected for all workers throughout the entire harvest period. The plantation also designates some workers as scrabblers who collect cut sugar cane from the fields and load on to trucks for transport to the processing factory. Since the link between their wage and productivity is not as clear as cane cutters, these workers were dropped from our analysis. We observe the entire cane cutting population of 770 workers during an eight week study period from February to April 2013. A random sub-sample of 100 workers was assigned activity trackers with three-dimensional accelerometers which measured daily steps taken, distance walked, calories burned, floors climbed, activity duration and intensity. At the end of every work week, the data recorded by these trackers was synced to a computer at the field survey office and the trackers were simultaneously charged. Activity information collected by the trackers was linked to plantation wage data and worker and health characteristics collected by a set of survey enumerators and health workers. The information on worker characteristics was administered by enumerators and included employment history, demographic information, place of living and household welfare. Then the registered health worker administered the second questionnaire by first asking a brief health history and then recording the results of two tests: a Rapid Diagnostic Test (RDT) for malaria and a blood slide which is used to microscopically verify malarial status. All workers who had a positive RDT or were parasitic positive according to the microscopy results from the collected slides were treated with the appropriate doses of Artemisinin Combination Therapy (ACT). Compliance with the treatment protocol was maximized through follow-up visits by the health workers and a small incentive (50 Naira) to return used ACT boxes to health workers who would conduct a short follow up visit. Sampling Two types of sampling were undertaken in this study round: i) Sampling of order of treatment of all 770 workers, and ii) sampling of 100 workers who carried the activity trackers throughout the study period. The sampling of treatment order followed a two stage procedure where 8 worker groups - called ‘gangs’ by workers - were first selected and then workers within a group were selected. A list of workers was obtained from the plantation before the beginning of the experiment. Selection of the order of gangs and then workers within gangs was also completed before the beginning of the experiment, so that the survey team had a predetermined number of workers from each gang to survey each day. Each week one group was provided with access to treatment implying that each worker was tested and treated if positive (and surveyed), as set out above. This process continued until all eight gangs of the plantation had been served and the entire workforce had received access to treatment. The order of testing and treatment was randomized over time. The sampling of workers assigned activity trackers was carried out such that they belonged to one of two groups. The first group consisted of 75 workers allocated the tracker randomly while the second comprised of the first 25 workers tested positive for malaria during week 1 of the health survey. Using the predetermined sampling plan which outlined order of testing for workers, the 75 workers randomly allocated trackers in week 1 were selected from the workers who were to be tested in weeks 2 and 3. Workers could not change trackers through the study period. Round 5 Data The project was implemented during the February and March 2104 harvest season on a large sugar cane plantation in Adamawa State, Nigeria. The project offered access to malaria services at the work place over a six week period. The service covered testing and treatment of malaria alone, for the worker himself only (not his family or friends), for six weeks only. Workers who received the service were entitled to two visits to a mobile health clinic installed and managed by the project and staffed with qualified health personnel. Once the worker paid for the service there was no reimbursement, including if the worker did not fall ill or if he choose not to use the service eventually. Each worker was individually approached in a predetermined random order. After conducting the baseline survey interview and explaining the service, each worker was guided through a script to elicit his willingness to pay. Following the Becker-DeGroot-Maarschak (BDM) mechanism, the worker was asked his willingness to pay for the service. Several checks and iterations made sure the worker mentioned his highest willingness to pay. It was then verified whether he had enough earnings from work at the plantation to be able to afford the stated price. Finally, the worker was asked to draw a price from a bag. If the drawn price was equal or below his willingness to pay he obtained access to the service, if it was above, he did not. Workers who chose to utilize the service were administered the Malaria Health Service questionnaire during their visits to the health team and then a follow up interview was conducted. At the end of the season, an endline labor and health survey was conducted again for the entire workforce. Wage information recorded by the plantation for each worker is also linked to data collected during this round.