Cost to government health-care services of treating acute self-poisonings in a rural district in Sri Lanka
Kanchana Wickramasinghe a, Paul Steele a, Andrew Dawson b, Dinusha Dharmaratne a, Asha Gunawardena a, Lalith Senarathna b, Dhammika de Siva c, Kusal Wijayaweera b, Michael Eddleston d & Flemming Konradsen c
a. Institute of Policy Studies of Sri Lanka, Colombo, Sri Lanka.
b. South Asian Clinical Toxicology Research Collaboration, University of Peradeniya, Peradeniya, Sri Lanka.
c. Department of International Health, Immunology and Microbiology, University of Copenhagen, Øster Farimagsgade 5, 1014 Copenhagen, Denmark.
d. Scottish Poisons Information Bureau, University of Edinburgh, Edinburgh, Scotland.
Correspondence to Flemming Konradsen (e-mail: firstname.lastname@example.org).
(Submitted: 06 February 2008 – Revised version received: 03 June 2008 – Accepted: 25 June 2008 – Published online: 29 January 2009.)
Bulletin of the World Health Organization 2009;87:180-185. doi: 10.2471/BLT.08.051920
A conservative estimate is that at least 258 000 deaths occur from pesticide self-poisoning worldwide each year,1 most of them in Asia, and this figure is greatly exceeded by the number of self-poisoned patients who seek treatment at health facilities. According to an analysis of hospital admissions in southern Sri Lanka, from 1990 to 2002 the average annual poisoning incidence rate in this southern region was 318 per 100 000, with 64% of this figure represented by self-poisonings.2 In the North Central province of Sri Lanka, the annual incidence rates in 2002 of attempted self-poisoning and fatal self-poisoning was 363 and 27 per 100 000, respectively.3 According to a community study, the high number of pesticide self-poisonings was an impulsive response to economically or psychosocially stressful events facilitated by easy access to pesticides.4
The cost to government health-care services of treating poisoned patients is probably substantial. In a Sri Lankan study of the cost of treating patients who had poisoned themselves with yellow oleander seeds (Thevetia peruviana), treatment with and without antitoxin cost US$ 691.6 and US$ 58.6 per patient, respectively.5 However, the overall costs of treating self-poisoned patients are not well documented in low-income countries, where the overwhelming majority of poisonings take place. This is unfortunate, as studies on health costs can inform health policy and guide investment and management at different levels of the health-care system to optimize the use of resources.
The objective of this study was to estimate the direct financial costs to the Sri Lankan Ministry of Health of treating self-poisoned patients in a single district, with a particular focus on self-poisoning with pesticides.
The study was carried out in the Anuradhapura district, where most of the 780 000 inhabitants depend on agriculture for their livelihood. The district is served by 1 general and 33 peripheral government hospitals, where practically all cases of acute poisoning are treated. The Anuradhapura General Hospital, with 1300 beds and limited intensive care facilities, is the largest hospital in the district. The 33 peripheral hospitals, which provide in-ward services, transfer all patients needing secondary care to the Anuradhapura General Hospital.
Selection of hospitals
The Anuradhapura General Hospital was selected as the source of detailed information on patient treatment inputs because it is the only secondary-level treatment facility in the district. The peripheral hospitals were listed according to size and distance from the Anuradhapura General Hospital. Three of them had to be excluded from the sampling owing to poor security in the areas where they were located. Of the 18 hospitals located within 40 km of Anuradhapura, 3 were selected at random for inclusion in the study; 2 of the 12 hospitals located beyond 40 km from the general hospital were also included by random selection. The decision to include only 1 of every 6 peripheral hospitals was based on the assumption that representative information on a per patient basis could be obtained at this level.
Patient treatment input costs were calculated for all intentionally self-poisoned patients admitted to the Anuradhapura General Hospital from 26 June to 26 July 2005. Costs for the 5 selected peripheral hospitals were based on all intentional self-poisoning patients admitted in July 2006. The one-year difference in the period of data collection reflected the availability of research staff.
Ongoing studies within the Anuradhapura district made it possible to obtain patient-specific transfer data on self-poisoned patients who had been transferred from all 30 peripheral hospitals in the district to a higher-level hospital facility from 1 July to 31 December 2005. Transfer costs also included the few cases transferred to the tertiary hospital outside the district.
Patient-specific inputs and costs
Using a preformatted datasheet, one study doctor, nurse or registered medical officer in the hospitals prospectively and systematically collected information on all inputs provided for each admitted patient. Research assistants cross-checked the recorded patient details against the individual in-patient records maintained by the hospitals.
Data on the price of drugs, the cost of treatments and laboratory tests, staff salaries and allowances, utilities, patients’ meals, ambulance fuel and maintenance costs, the cost of cleaning wards and annual hospital administrative costs for 2005 were collected from the Medical Supply Division of the Ministry of Health. Provincial health budget data were obtained from the Finance Commission of the Ministry of Finance. All prices excluded government value-added tax and were collected in United States dollars (US$). At the time, US$ 1 equalled 100 Sri Lankan rupees.
Secondary hospital treatment costs
Treatment costs included four categories: ward care staff costs, laboratory investigation costs, drug costs and non-drug treatment costs. The total treatment cost was disaggregated based on the type of poison.
Total capital and maintenance costs for the Anuradhapura General Hospital were calculated by summing all annual cost items provided by hospital administrators. Per patient capital and maintenance costs were calculated by dividing total annual capital and maintenance costs by the total number of admissions (including for self-poisoning) over the year. Depreciation of hospital buildings, ambulance and other capital items were not included in the estimations.
The total per patient cost to the government of treating self-poisoned patients at the general hospital was calculated by summing per patient treatment and per patient capital and maintenance costs.
Peripheral hospital treatment costs
The per capita cost of treating self-poisoned patients in the 5 peripheral hospitals was calculated the same way as for the general hospital. However, due to the lack of reliable records on capital and maintenance costs, these were excluded from the analysis. Because peripheral hospitals lack many facilities such as intensive care units (ICUs), the capital and maintenance costs per patient were assumed to be smaller than the cost of other items.
Transfer costs were calculated based on patient transfer data collected from 30 peripheral hospitals and included the cost of the ambulance fuel. Ambulance capital and maintenance costs were already included in the capital cost estimations of the general hospital.
The number of self-poisoned patients admitted to all the hospitals during 2005 was available from information generated by a prospective study.3 The cost to the Anuradhapura district was compared with the health budget allocation for the Sri Lankan North Central province (Anuradhapura being one of its two districts), with a population of 1.2 million. Since the most recent information on the total number of poisonings in Sri Lanka was from 2004, this year was used for national cost estimates.
Treatment at the general hospital
Over the 31-day study period in 2005, 179 self-poisoned patients were admitted to the Anuradhapura General Hospital. Of these patients, 165 were discharged after treatment, 8 were transferred to other hospitals outside the district for further treatment and 6 died. The average length of admission in the hospital was 73 hours and 4 minutes.
Of the 179 self-poisoned patients, 107 had suffered self-poisoning with pesticides, 56 with yellow oleander seeds (Thevetia peruviana) and 16 with medicine.
The average total cost of treating a self-poisoned patient was US$ 31.83, and of this amount, only US$ 0.19 corresponded to capital and maintenance costs. The average total cost of treatment was highest for self-poisoning with pesticides (US$ 49.12), followed by medicine overdose (US$ 9.27) and ingestion of yellow oleander seeds (US$ 5.25). The average total cost of treating pesticide self-poisoning was greatest for the 37 patients who had ingested organophosphate pesticides (US$ 87.44), followed by the 11 who had ingested carbamate (US$ 51.80). It was lowest for the 26 who had ingested a herbicide (US$ 10.49) (Table 1).
Table 1. Estimated costs of treating 179 self-poisoned patients admitted to the Anuradhapura General Hospital in Sri Lanka from 26 June to 26 July 2005
The cost of ward staff input was the highest expenditure category, with a mean of US$ 13.08 per self-poisoned patient (range: US$ 0.11–563.75), followed by the cost of drugs with a mean of US$ 11.96 per patient (range: US$ 0.00–1032.77), non-drug treatment with a mean of US$ 4.90 (range: US$ 0.00–132.39) and investigation with a mean of US$ 1.75 (range: US$ 0.00–68.4). Nursing and medical staff respectively took up 65% and 29% of the total cost of ward staff input.
Although only 10 of the 179 patients received intensive care, this small group made up 75% of the total cost of treating all self-poisoned patients at the general hospital. The very high cost involved in treating ICU patients was mainly attributable to the costs of drugs and nursing input. The 10 cases treated in the ICU during the month of observation were more than the monthly average of 6.5 and 8.0 ICU cases seen throughout the rest of 2005 and 2006, respectively.
The highest treatment cost for any single patient during the study period (US$ 308.86) was seen in a case of chlorpyrifos self-poisoning. This included care in the ICU. The average total treatment cost for the eight other chlorpyrifos poisoning cases that were not transferred to the ICU was US$ 10.7.
Treatment at peripheral hospitals
During July 2006, 40 intentionally self-poisoned patients were admitted to the 5 selected peripheral hospitals. Of the 40, 9 were discharged after treatment and 31 were transferred to another hospital. There were no deaths that month. The average duration of admission in the hospital was 8 hours and 26 minutes. Of the 40 self-poisoned patients admitted to the 5 peripheral hospitals, 29 had ingested pesticides, 8 had ingested yellow oleander seeds (Thevetia peruviana) and 3 had taken a medicine overdose.
The average total cost of treating a self-poisoned patient was US$ 3.33 at peripheral hospitals. The average total cost of treatment was highest for self-poisoning with pesticides (US$ 3.70) – US$ 3.96 for organophosphate, US$ 3.77 for carbamate and US$ 3.50 for herbicides – followed by ingestion of oleander seeds (US$ 2.34) and medicines (US$ 2.45) (Table 2).
Table 2. Estimated costs of treating 40 self-poisoned patients admitted to 5 peripheral hospitals in the Anuradhapura district in Sri Lanka from 1 to 31 July 2006
The cost of ward staff input was the highest expenditure category with a mean of US$ 1.65 per self-poisoned patient (range: US$ 0.34–4.08), followed by drugs (mean: US$ 0.67; range US$ 0.00–3.98), laboratory investigations (mean: US$ 0.59; range: US$ 0.00–2.68) and non-drug treatments (mean: US$ 0.02; range: US$ 0.00–0.80). Overall, nurses, medical doctors and ward attendants made up 53% of the total cost of treating a case of self-poisoning, with an approximately equal share from each staff category.
From July to December 2006, 816 self-poisoned patients were admitted to the district’s 30 peripheral hospitals, with a transfer rate of 57%. The Anuradhapura General Hospital received 95% of these transfers. The average cost per transfer, which comprised primarily the cost of fuel, was US$ 14.03. Assuming the same patient load3 and the same transfer rate for the previous 6 months of 2006, total admissions of self-poisoned patients for all of 2006 were estimated to be 1632. Among them there were 930 transfers, 473 of them patients self-poisoned with pesticides.
Self-poisoned patients in the district, 2006
During 2006, 1908 patients were admitted to the Anuradhapura General Hospital following self-poisoning. Using average total costs by type of poison calculated from the detailed observational studies of 179 persons in 2005, we estimated the total treatment cost for the year to be US$ 58 114.06 (Table 3).
Table 3. Estimated costs of treating 1908 self-poisoned patients in the Anuradhapura General Hospital in Sri Lanka in 2006
In 2006, an estimated 1632 self-poisoned patients were admitted to all peripheral hospitals in the Anuradhapura district. As seen from Table 4, the total cost of treating all self-poisoned patients in peripheral hospitals in the district was estimated at US$ 5438.46.
Table 4. Estimated costs of treating 1632 self-poisoned patients in all peripheral hospitals in the Anuradhapura district in Sri Lanka in 2006
The estimated transfer cost for all 930 self-poisoned patients in the Anuradhapura district for 2006 amounted to US$ 13 047.
The total cost of treating self-poisoned patients in the Anuradhapura district amounted to US$ 76 599. Of this amount, US$ 53 834 were attributable to self-poisoning with pesticides.
Estimate for all poisoned patients in Sri Lanka
In 2004, the Medical Statistics Unit of the Ministry of Health recorded a total of 48 098 poisoned patients admitted to health-care services throughout Sri Lanka. Based on our estimates of the total per patient treatment cost for the different types of poisonous substances and for different levels of health-care services, and assuming a transfer rate of 57%, the annual cost of treating poisoned patients in Sri Lanka for 2004 was estimated at US$ 866 304. This included a cost of US$ 527 452 for secondary and tertiary treatment services, US$ 87 727 for peripheral services and US$ 251 125 for cost of transfers. These figures apply to all poisoned patients, not only self-poisoned ones, as the cause of poisoning could not be obtained from the existing national data.
We estimated the total cost of treating self-poisoned patients in the Anuradhapura district during 2006 to be US$ 76 599. This represents 2.8% of the total government health expenditure for all health services of US$ 2 704 020, including all preventive, curative and administrative activities in the North Central province.
This study clearly shows that the average cost of treating cases of self-poisoning with pesticides is much higher than that of treating patients self-poisoned with other substances. This is partly due to the very high cost of care in ICUs and to the higher numbers of general ward staff needed for the care of patients self-poisoned with pesticides. We found the average cost of treating poisoned patients to be higher for organophosphates than for other pesticides. A shift to the use of pesticides whose ingestion is linked to a low likelihood of ICU admission would produce substantial savings. However, the pesticide-specific costs in this study are based on a small number of cases, which makes it difficult to generalize.
Interestingly, the total cost of transferring patients to the general hospital was found to be more than twice the total cost of treating patients self-poisoned with pesticides in the peripheral hospitals. Transfer costs are extremely sensitive to fuel costs, which doubled between the beginning of the study and May 2008. Moreover, the cost of treatment at peripheral hospitals was significantly lower than at secondary hospitals, although a direct comparison is prone to bias since the severity of complications among the patients admitted to the two levels of health care following poisoning may be different. The many transfers to referral hospitals is likely to reflect the lack of essential antidotes and appropriate resuscitative equipment for treating pesticide self-poisonings in smaller peripheral health facilities.6 Increased investment at the peripheral level may reduce the number of transfers and speed up the onset of appropriate treatment.
We estimated US$ 866 304 to be the total government treatment cost of treating all poisonings in Sri Lanka for the year 2004. As self-poisoning is a major problem across rural south-east Asia, the cost to government health services in the region will be substantial. This study suggests that significant financial benefits could be attained by investing in measures and policies outside the health sector aimed at preventing self-poisonings.
We are fully aware that the financial costs to government health-care services do not reflect either the overall cost to society or the psychological impact on households of self-poisoning or poisoning in general.7
The estimates presented in this study are based on several assumptions. The annual cost calculations in particular should be interpreted with caution, since even small fluctuations in the number of patients placed in the ICU have a substantial influence on the average cost of treating pesticide self-poisoning cases. Also, the lack of detailed estimates of maintenance costs across hospital departments and peripheral hospitals weakened the study. Similarly, the large proportion of cases for which the type of poison purposely ingested was unknown makes the estimates less reliable. The overall cost to national health-care services of treating pesticide self-poisoning cases will be greatly influenced by treatment capacity across the island, especially by the availability of ventilators and ICU services, the local pattern of transfers to higher level facilities, and the quality of the health information system (which is likely to underestimate the number of poisoning cases). Furthermore, although national guidelines on the management of poisoning have been in place for almost 10 years, treatment approaches are likely to vary across the country. In addition, the national estimate is influenced by the fact that the estimated cost per patient for secondary services had to be used to estimate the cost for tertiary services as well, since the per patient cost for the latter was unavailable.
Further investment in decentralized treatment capacity at peripheral hospitals is likely to reduce the overall cost to the government and optimize treatment. ■
We appreciate the support provided by the Sri Lankan Ministry of Health and Ministry of Finance.
Funding: ME is a Wellcome Trust Career Development Fellow funded by grant GR063560MA from Wellcome’s Tropical Interest Group. The study received support from the Danish International Development Agency (DANIDA) and the Wellcome Trust Grant GR071669MA.
Competing interests: None declared.
- Gunnell D, Eddleston M, Phillips MR, Konradsen F. The global distribution of fatal pesticide self-poisoning: systematic review. BMC Public Health 2007; 7: 357- doi: 10.1186/1471-2458-7-357 pmid: 18154668.
- van der Hoek W, Konradsen F. Analysis of 8000 hospital admissions for acute poisoning in a rural area of Sri Lanka. Clin Toxicol (Phila) 2006; 44: 225-31 pmid: 16749538.
- Eddleston M, Sudarshan K, Senthilkumaran M, Reginald K, Karalliedde L, Senarathna L, et al., et al. Patterns of hospital transfer for self-poisoned patients in rural Sri Lanka: implications for estimating the incidence of self-poisoning in the developing world. Bull World Health Organ 2006; 84: 276-82 doi: 10.2471/BLT.05.025379 pmid: 16628300.
- Konradsen F, Hoek W, Peiris P. Reaching for the bottle of pesticide — a cry for help. Self-inflicted poisonings in Sri Lanka. Soc Sci Med 2006; 62: 1710-9 doi: 10.1016/j.socscimed.2005.08.020 pmid: 16165259.
- Eddleston M, Senarathna L, Mohamed F, Buckley N, Juszczak E, Sheriff MH, et al., et al. Deaths due to absence of an affordable antitoxin for plant poisoning. Lancet 2003; 362: 1041-4 doi: 10.1016/S0140-6736(03)14415-7 pmid: 14522536.
- Senarathna L, de Silva D, Buckley NA, Eddleston M, Dawson AH. Effect of resource availability and medical staff workload on outcome for poisoning in rural Sri Lanka. Proceedings of the 6th annual congress of the Asian Pacific Association of Medical Toxicology, Bangkok, 12-14 December 2007.
- Cole DC, Carpio F, León N. Economic burden of illness from pesticide poisonings in highland Ecuador. Rev Panam Salud Publica 2000; 8: 196-201 doi: 10.1590/S1020-49892000000800007 pmid: 11036430.