FALLING INTO THE MEDICAL POVERTY TRAP IN SRI LANKA: WHAT CAN BE DONE? Myrtle Perera, Godfrey Gunatilleke, and Philippa Bird

Sri Lanka’s public health care system is free at the point of use and maintains a focus on equity. However, noncommunicable diseases, such as diabetes, are rapidly increasing in prevalence and posing new challenges to the system and to patients and households. In-depth interviews and focus group discussions were conducted in four districts of Sri Lanka to investigate the care-seeking experiences of diabetes patients from households at different income levels. Although health care is free, other direct and indirect costs served as deterrents to care seeking before and after diagnosis, and placed a high burden on households. The need for frequent visits to clinics with appropriate facilities for diagnosis and management of diabetes, often far from rural communities, posed high costs, in particular due to income foregone. Households employed coping strategies, but the need for frequent clinic visits posed repeated costs, which made it difficult for households to recover their economic status. Many patients, especially those from lowincome, rural households, could not maintain the management regimen, and their condition deteriorated. There is a need for specialist facilities for the diagnosis and management of diabetes at locations closer to rural areas.

The successes of Sri Lanka’s health system, given the country’s relatively low income and health expenditure, have been widely reported (1, 2). By 2004 the life expectancy at birth in Sri Lanka had reached 74 years, the highest in South Asia, and child mortality was only 14 per 1,000 live births, despite a gross national income per capita of only $1,010 (3). Sri Lanka has a public health care system that is free at the point of use for all patients. The health system has maintained a focus on equity and social justice since its development before independence in 1948, resisting the market-oriented reforms that characterized many other low-income countries in the 1980s and 1990s. The country’s health infrastructure is well developed, with three tiers of health facility and widespread community health services. International Journal of Health Services, Volume 37, Number 2, Pages 379–398, 2007 © 2007, Baywood Publishing Co., Inc.

379

380

/ Perera et al.

In comparison with other countries in the region, the direct costs of accessing health care and catastrophic costs of illness in Sri Lanka are very low (4). However, the extent of other direct costs and the indirect costs of accessing care for poor households can be significant in Sri Lanka (5, 6) and other low-income countries. In addition, although public health care is provided free, since the initiation of an adjustment and stabilization program in 1977 there has been an expansion of the private sector, accounting for 55 percent of total health expenditure in 2003 (7). Health insurance in any form is very rare, and consequently almost all private health expenditure is provided out-of-pocket. Studies have also highlighted other nonfinancial barriers to care in Sri Lanka, such as a lack of trust and a poor relationship with doctors (8). Sri Lanka’s widely accessible health care system has effectively tackled communicable diseases and maternal and child mortality. However, the fast pace of the epidemiological transition and the aging of the population in Sri Lanka are posing new challenges to the country’s health care system and to the people using the system—a situation mirrored in many countries in Asia and Africa (9). Noncommunicable diseases require sustained, long-term, or recurrent care and specialist facilities for diagnosis and treatment. The rise in diabetes in Sri Lanka over the past decade has been dramatic. The World Health Organization estimates that there were 653,000 Sri Lankans with diabetes in 2000 (10). Given the sharp increase in diabetes and Sri Lanka’s rapidly aging population, projections of future diabetes prevalence are high: the WHO predicts more than 1.5 million people with diabetes by 2030 (10). Wild and colleagues (11) estimate an increase in the prevalence of diabetes from 3.5 percent in 2000 to 6.7 percent in 2030 (an increase from 5.4% to 9.0% among people aged 20 or over). It has been predicted that the majority of people with diabetes in Asia will be middle-aged (12). The costs of diabetes to the government, due to the need for recurrent outpatient management and diabetic complications, are high and will certainly increase considerably as prevalence and people’s treatment expectations rise. The costs of diabetes to patients and their households may also be high, but have received little attention. While health care for communicable diseases is available at close proximity to the population, there is a lack of adequate facilities for noncommunicable diseases, which may force patients with diabetes to travel farther for treatment (13). Combined with the lifelong recurrent need for care and risk of diabetic complications, diabetes could be a costly and limiting disease to the patient and the household. In this article we explore the accessibility and affordability of care for diabetes patients from different types of households in Sri Lanka, using the Affordability Ladder Program (ALPS) framework (see the preceding article in this Journal issue; 14). In so doing, we aim to draw out policy suggestions to reduce the barriers to care.

Medical Poverty Trap in Sri Lanka

/ 381

METHODS The results presented here are taken from a large study that examined access to, use of, and quality of health care in Sri Lanka, using the ALPS framework (14). Diabetes was one of ten noncommunicable diseases included in the study. Qualitative data on access to care for diabetes patients were collected through in-depth interviews and focus group discussions between June 2002 and December 2003. In total, 49 patients with diabetes (from 49 separate households) were interviewed in depth to provide detailed case narratives and cost data. Interviews were also conducted with other members of patients’ households. The sample was recruited from contacts established through medical personnel at private and public health centers in four districts of Sri Lanka. Two of these districts were predominantly rural (Moneragala and Anuradhapura), one was mainly urban (Galle), and one was the capital (Colombo), which is urban and has the best health care facilities in the country. Households were classified as remote or proximate to appropriate health care facilities, and household income was classified as high (11 households), middle (20 households), or low (18 households). Interviews were conducted in patients’ homes in Sinhalese and, when permission was given, were tape-recorded. Otherwise, the interviewers took detailed notes. The recordings and notes were transcribed and translated into English for analysis. Separate focus group discussions were carried out with patients, community members, and health service personnel, in community and hospital settings: 8 community focus group discussions with 6 to 10 people each; 16 focus group discussions with patients and medical staff in public and private clinics; and 22 individual discussions with medical personnel, both allopathic and indigenous practitioners, on hospital premises. The focus group discussions were disease specific, and this article presents data from the discussions on diabetes. A four-ladder version of the Affordability Ladder framework was developed to organize and analyze the data (Figure 1). Data were arranged by ladder to aid analysis, allowing patients’ experiences to be traced along each of the four ladders. The first, an illness or reference ladder, documented the progression of the illness. The illness cycle was demarcated into clearly defined stages of “acute,” when aggravation of symptoms required inpatient treatment in a hospital, and “management” stages in between, when the illness was managed through a regimen of treatment and monitoring with regular outpatient visits. An expenditure ladder recorded the direct and indirect costs incurred at each stage of the illness. The household economy ladder documented the changes in the household economy after a member of the household was diagnosed with diabetes and throughout the duration of the illness. Lastly, the impact ladder traced the economic and social consequences of the illness for the patient, the

382

/ Perera et al.

Figure 1. The four ladders framework for analysis of qualitative data.

household, and other family members. We present case studies of the experiences of patients in Boxes 1 through 6, with names changed to protect anonymity. An economic analysis was conducted to determine the costs to households of care for diabetes. Cost data collected in the in-depth interviews were used to calculate direct and indirect costs and costs as a proportion of the household income for patients in different types of household. To improve the recall accuracy of the data, only the costs that patients recalled for the last three years were used in the analysis. RESULTS No Care and Delayed Diagnosis This study was based on interviews with diabetes patients—persons with diabetes who had received a medical diagnosis, as described above. People with diabetes who had never had contact with professional providers of care could not be identified, thus no assessment could be made of the extent to which people with diabetes did not receive a diagnosis and did not receive any form of treatment. However, this study does provide insight into when and why there were periods in which diabetes patients from different types of households did not receive any health care.

Medical Poverty Trap in Sri Lanka

/ 383

An early diagnosis was the most significant precondition for controlling the disease effectively. However, in low-income households in remote areas, patients reported living with symptoms for up to two to three years before finally accessing care, when symptoms became acute, as reported by Hamina (Box 1) and Sanduli (Box 2). By this stage, emergency care at a public hospital was required, or in some cases at a private care facility because the patient was too ill to endure the queues and delays at the public hospital. Those who had waited until an emergency to seek care often suffered complications and required ongoing specialist care at a higher-level hospital. In contrast, diagnosis of diabetes before the condition became severe often allowed patients to manage their condition at lower-level facilities, at reduced cost and inconvenience, as was the case for Seetha (Box 3). Common reasons for delaying care seeking until an emergency included the distance to a hospital and the costs of care and working time lost, as noted by a woman from a low-income rural household: It is a long way to the town to go to the nearest hospital. I cannot go alone and we are very busy with some kind of work to earn a daily income. We cannot just drop all that and rush to the hospital unless we feel very ill.

The majority of low-income diabetes patients in rural locations were engaged in agriculture and manual labor. Symptoms such as aches and fatigue were

Box 1 Hamina’s Experience Woman from a Low-Income Household in a Rural Area Hamina, a female patient from Moneragala, was 56 years old when her diabetes was detected. She had dealt with her symptoms for about two years without realizing that she had diabetes. Her knowledge of the disease was limited, and she attributed her usual bouts of tiredness to the hard agricultural work she had to do. After relying on self-care for a while, she took Ayurvedic treatment. On one occasion she fainted and decided to go to the local district hospital, which was only 4 kilometers from her home. However, the hospital lacked the facilities necessary for her treatment and she was transferred to Badulla hospital, a provincial hospital 52 kilometers from her home. She was an inpatient there for 23 days. After she was discharged, she was instructed to report regularly to the clinic at the provincial hospital. Because the hospital was so far from her house, Hamina had financial difficulties in maintaining the regimen of regular visits. Although she attended clinics regularly, her eyesight started to fail and she had to seek treatment for a wound that did not heal. She seemed to have accepted the local district hospital as the more convenient and less costly choice, even though she was not receiving treatment for her failing eyesight and the provincial hospital in Badulla would have provided more appropriate care.

384

/ Perera et al. Box 2 Sanduli’s Experience Woman from a Low-Income Household in a Rural Area

Sanduli was from a low-income household in Anuradhapura. Her condition remained undiagnosed for a long period, as she attributed her symptoms to fatigue and general weakness and used Ayurvedic treatment. When her condition became acute, she reported to the nearest hospital, only 2 kilometers from her home. However, this local hospital did not have the appropriate facilities to treat her condition properly and Sanduli was transferred to Anuradhapura hospital, the top-level medical institution in the province. She was hospitalized for 17 days and received the appropriate treatment, and was instructed to report regularly thereafter to the local hospital. Sanduli received the necessary drugs from the hospital but was unable to follow the instructions on diet and special food. She complained that she could not afford the special diet, and her family could not afford the time to prepare separate meals for her. She continued in this manner until her condition worsened and she had to be hospitalized, this time in the local hospital. Sanduli continued treatment at her local hospital, where she had to get the tests done privately because the hospital lacked facilities for diabetes. After two months she had recurrent acute episodes, for which she was again hospitalized. She then continued to attend clinics once a month at the local hospital, but after three months her condition worsened again. From then on Sanduli sought treatment from a private doctor, but could not take the prescribed drugs for financial reasons, and she again began using her local public hospital. Her condition at the time of the last visit by the researchers was unsatisfactory and further compounded by weakening eyesight.

considered a part of everyday life that had to be endured. In such cases the nature of employment “masked” the early symptoms of diabetes. Delayed diagnosis was also found, however, among non-poor people living in remote rural areas, as illustrated by the case of Jayantha (Box 4). Jayantha’s experience, although clearly a particularly severe case, illustrates the outcome of delayed diagnosis. His experience demonstrates the inadequate knowledge about the symptoms of diabetes in the population and on the part of some Ayurvedic and private allopathic practitioners. Many practitioners do not have the appropriate facilities for diagnostic tests, despite the high and rapidly rising prevalence of diabetes. Informal Care Patients commonly had used informal care before diagnosis to relieve symptoms—for example, Hamina (Box 1), Sanduli (Box 2), and Jayantha (Box 4)— and occasionally used it to manage the disease when they were no longer able to access an appropriate hospital regularly. Many patients reported buying

Medical Poverty Trap in Sri Lanka

/ 385

Box 3 Seetha’s Experience Woman from a Middle-Income Household in a Rural Area Prompt action was taken when Seetha fainted. She was taken to the Wellawaya hospital, 12 kilometers from her home. She received treatment as an outpatient. The appropriate tests were done, and Seetha followed a regular regimen of visits to the clinic and controlled her diet as advised. She continued to take her drugs regularly, making purchases from private pharmacies if and when she was unable to attend the clinic. Her health status at the time of the interview was satisfactory. Her illness was under control, and she continued to manage it with the regimen of medication and monitoring.

medication from private pharmacies to treat the symptoms of diabetes. Others used Ayurveda or traditional medications, such as this woman from a low-income rural household: Chena [slash-and-burn] cultivation is hard work, working all day in the sun. Fatigue and body pain are natural. We apply some home remedies like herbal oil to our limbs and back and continue our daily task.

A common pool of knowledge on herbal brews and applications for common ailments has been the mainstay of rural communities in dealing with health care needs and is also the historical underpinning of the indigenous Ayurveda system. However, in our study, as informal care did not adhere to the strict regimen required to manage diabetes, the patient’s condition frequently deteriorated and became acute before he or she sought professional care. Accessing Professional Care The usual health-seeking behavior in times of illness was to visit the nearest government hospital, which for most people was a rural or a district hospital, or a major hospital if they lived in close proximity to one. People from high-income households would normally access a private hospital or a private specialist. Health care could be sought at the nearest hospital for most communicable diseases. However, the scarcity of hospitals with facilities for diagnosing and testing for diabetes care made accessing an appropriate hospital essential for patients with diabetes. As appropriate facilities for blood and urine tests were available only at district-level hospitals, patients in remote areas had to travel long distances: there were substantial problems of geographic inaccessibility. Although care was free at government hospitals, patients incurred the costs of transport, meals, and tests as well as the costs of income foregone due to long travel and long waiting times at the hospital, which gave rise to problems

386

/ Perera et al. Box 4 Jayantha’s Experience Man from a High-Income Household in a Rural Area

By the time of the interview, Jayantha’s condition had seriously deteriorated due to a combination of circumstances. He first began to pay attention to his symptoms when he suffered from a wound that did not heal easily, but he continued with self-medication using indigenous home remedies and did not suspect that he could have diabetes. He then tried treatment from a private allopathic practitioner, but his condition remained undiagnosed. Diabetes was finally diagnosed, and Jayantha was admitted to Karapitya hospital, the premier teaching hospital in the district, where he was an inpatient for a month. His wound was healing well, so he was discharged and instructed to visit a rural hospital close to his home for regular dressing changes. He noticed that this place was unclean, and the dressing was done without much care for cleanliness and by untrained staff. His wound became worse. He then went to a teaching hospital in Colombo, where his leg had to be amputated. Jayantha continued treatment at the teaching hospital. However, his expenditures on health care had impoverished him to the extent that, as he reported, he could not afford to purchase a wheelchair.

of economic inaccessibility. When a diabetes patient who reported to an “inappropriate” hospital required urgent hospitalization in a major hospital, she or he was transported by ambulance at government expense; such was the case for Hamina, whose story is typical of the experiences of patients from low-income rural households (Box 1). The high-income patients and some middle-income patients chose to access a private hospital for outpatient care, as was their usual course of action, to circumvent the inconveniences of public hospitals and to adhere to norms of social class and status. When the costs of repeated visits to manage the condition became unaffordable they reverted to using public hospitals. The story of Sena (Box 5) illustrates the problems of a middle-income patient who was caught between the inconveniences of public hospitals and the costs of private treatment. The gradual erosion of resources compelled even those with relatively high incomes to shift to public health care. Only in Colombo did all income groups have a choice of easily accessible, appropriate public hospitals, which were well equipped to provide drug treatment for diabetes. Low-income patients in remote areas were the least likely to access appropriate care. Quality of Care The high-level public hospitals provided a complete package of medication, tests, and special food supplements for diabetes patients, free of charge. However,

Medical Poverty Trap in Sri Lanka

/ 387

Box 5 Sena’s Experience Man from a Middle-Income Household in an Urban Area Sena lived in the suburbs of Colombo, within very easy access to both public and private health care facilities. He had had diabetes for nearly 15 years at the time of the interview. Immediately after fainting, he had gone to the public National Hospital, where his diabetes was diagnosed and he was asked to follow a regimen of medication and tests with regular visits to the clinics. He was seriously dissatisfied with the treatment he received. He reported that doctors had no time to discuss his condition with him or give any explanation that would help him to manage his illness. Other staff members were also not attentive. Sena stopped going to the public hospital and received treatment in the private sector. He found the service much better and more convenient, in terms of both the time spent and the attention he received from the doctor. When his condition deteriorated he had to receive treatment at the National Hospital again, but was again unsatisfied with the long queues and “unkindness” of some of the doctors and other staff. From then on Sena alternated treatment between the private sector, where he was more satisfied with the service, and the public sector when he could not afford private treatment and his condition worsened. When his wife became ill with cancer he abandoned his treatment. At the time of the interview Sena’s condition was uncontrolled.

there were instances in lower-level government hospitals when both inpatients and outpatients were requested to obtain tests and drugs from private facilities outside the hospital, as experienced by Sanduli (Box 2) and Seela (Box 6). In the management stage, the lengthy and inconvenient procedures in public hospitals (Figure 2) deterred patients from adhering to the treatment regimen. Furthermore, the distances to appropriate, higher-level hospitals posed a high and recurrent cost burden on patients, who were compelled to “get their clinics changed” to more convenient, but also less appropriate, rural or district hospitals. The absence of monitoring at lower-level hospitals and patients’ inability to purchase the drugs not supplied by the public hospitals led to the aggravation of symptoms and the recurrence of acute stages of the disease. Several patients who could not afford the costs of going to a distant government hospital chose to receive tests at closer, private medical centers (some of which were unregulated and of questionable quality), and purchased drugs, often of unknown efficacy and dose, from pharmacies. Others shifted to Ayurvedic medicines, borrowed medication from friends, or interrupted their treatment regimen for long periods. Some saved drugs for when their condition became acute, not understanding the need for continued medication for controlling the disease. A male patient from a rural area explained:

388

/ Perera et al. Box 6 Seela’s Experience Woman from a Low-Income Household in a Rural Area

Seela and her family depended on a seasonal income from chena cultivation and a small store. The store brought in a cash income of Rs 3,500 a month, and chena brought in about Rs 6,000 every three months if crops were good. The family managed its basic needs until Seela’s hospitalization costs of Rs 3,200. Immediately after that, she had to find about Rs 100 a month for clinic visits. But she had to close down the boutique, which dried up the family’s cash income, and chena cultivation was uncertain. As a result Seela was depending on Samurdhi (government aid for very-low-income households), and spent Rs 25 to go to the closest district hospital, which did not give her all the required drugs. She had no money to buy the extra drugs. When her husband also fell ill, her married daughter, Rupa, and her family moved in with them. The son-in-law continued agricultural work while the daughter cared for the parents. But Seela’s granddaughter had to be kept away from school because Rupa was too busy to take her regularly. Rupa’s hopes for her child’s education were fading away. The household recovered its economic status when Seela’s son started working as an agricultural laborer, which he combined with chena cultivation. But Seela mourned that her son had given up his education at Advanced Level, just before sitting the exam, which would have opened doors to employment or higher education. Seela’s second daughter also left school and got a low-paid job in a garment factory.

We travel 50 kilometers to the big hospital, staying overnight at the bus stand to reach the hospital at dawn to get a number [in queue to see the doctor]. At times we are told the numbers are over—come again next week. The cost apart, we cannot put up with that routine every month. We sometimes spend more on Ayurveda, but the patient does not always have to go. We also save our cultivation.

Those who used private care perceived it to be of higher quality than public care, although this was not borne out by those who resorted to public hospitals— when compelled by financial difficulties—and found them to be of equally high or higher technical quality. The key benefits of private care over public care were considered to be convenience, shorter waiting times, and better staff attitudes, as reported by Sena (Box 5). Patients were generally poorly informed about the need to make lifestyle changes for prevention or control of diabetes. Furthermore, low-income patients could not adhere to the necessary dietary changes, due to the unavailability or cost of appropriate foods, as was the case for Sanduli (Box 2). For agricultural

Medical Poverty Trap in Sri Lanka

/ 389

When patients were first diagnosed with diabetes, they were registered at special clinics held once a week at higher-level hospitals and given a card, which detailed the days they should return to the clinic for tests and dispensing of medicines. This ranged from once to four times a month, depending on the severity of the condition.

Patients referred to the “numbers war.” On returning to the hospital on the specified dates, patients were required to obtain a number, entitling them to see the doctor. On busy days the numbers ran out early, often by 6 a.m. There were several unethical practices for reserving or purchasing a number later in the day, through corrupt minor employees and local inhabitants who conspired with them to operate informal markets for numbers. Such payments could range from Rs30 to over Rs100 for each visit. On days when patients arrived and the numbers were all used up, they were instructed to return the next week. Although clinics were held weekly, patients complained that if they returned the next week on a date that was not specified on their card, they were scolded by staff and turned away. The hospital spread out patients’ visits for convenience, allocating dates for each visit. For this reason, when patients had missed an appointment they were required to wait until the next month to return to the hospital and had to cope without the necessary monitoring and medicines for the month.

Once they had obtained a number, patients had to negotiate the “queues war.” Typically, patients or the persons accompanying them had to wait in three queues: one to obtain a number, one for tests or to see the doctor, and one for medications. Often, no seating was available, and toilet facilities were not available in some clinics.

Figure 2. Pathway to diabetes care in public hospitals.

390

/ Perera et al.

communities affected by seasonal droughts, starchy foods such as yams, which are contraindicated for diabetes, are a staple of the diet in drought seasons. Among some patients who followed the dietary restrictions but could not afford compensatory supplements, acute phases were precipitated by virtual starvation as they avoided contraindicated foods. High- and middle-income patients, on the other hand, were likely to precipitate acute phases by consuming excessive sweet foods or alcohol. Burden of Payment Although public health care is universally free at the point of delivery in Sri Lanka, substantial and recurrent costs were incurred by the diabetes patients in our study, primarily due to transport costs and the indirect costs of lost working time. Patients faced recurrent costs for outpatient check-ups to manage their condition and sudden lump-sum payments for inpatient care when their condition deteriorated and they needed hospitalization. The costs of outpatient care included the recurrent direct costs of travel for patients and household members accompanying them, and expenditure on meals and refreshments. “Under-the-table” payments such as the cost of a number (for a place in a queue to see the doctor) were also reported (Figure 2). For those in remote areas that could access only lower-level, inappropriate hospitals, there were costs for drugs and tests, and patients who went to private hospitals also incurred consultation fees. Sanduli’s experience (Box 2) illustrates the problems of meeting the direct costs of the recurrent outpatient visits required to manage diabetes. The indirect costs of outpatient care in terms of income foregone were substantial, especially for patients in remote rural areas, who were required to travel to a distant hospital and to wait for many hours before receiving care. For households engaged in farming, lengthy visits to the hospital and neglect of the crop for even a few days could lead to long-term losses and initiate a chain of indebtedness and poverty. Middle- and high-income households also experienced losses in earnings from formal employment and business. Patients who were hospitalized in an appropriate hospital incurred virtually no direct costs, but the opportunity costs of lost working time could be substantial. The household also incurred the direct and indirect costs of travel by family members to visit the patient daily, which was stated to be necessary to ensure the patient’s care needs, such as clothing. A focus group member in a rural area, the young daughter of a diabetes patient from a middle-income household, explained: When someone fell ill we went to the nearest doctor or hospital, got some treatment and it was over. For these noncommunicable diseases we go to one hospital, they say it is no good, go to another. This one is very far . . . it is easy for them to say they will send the patient in an ambulance. When we

Medical Poverty Trap in Sri Lanka

/ 391

have to visit the patient we are not taken by ambulance. We have to leave home at dawn, walk, then take a bus, come back late, and get ready to go again the next day. Our cultivation is destroyed, our money is gone.

We conducted a detailed cost analysis, calculating costs for inpatient and outpatient care for diabetes for each income group (Table 1). The daily direct and indirect costs of a period of inpatient care were equivalent to an average 211 percent of daily income for low-income households, compared with 131 percent for middle-income and 57 percent for high-income households. Private care costs far exceeded money earned—624 percent of a household’s daily income, on average. The need for regular monitoring through outpatient visits to the clinic at least once, and in some cases two or four times, a month posed a significant and recurrent financial burden on households in all income groups (Table 2). The direct and indirect costs of an outpatient visit to a public hospital amounted to 345 percent and 150 percent of the daily household income for low- and middle-income households, respectively, and amounted to 408 percent of a household’s daily income, on average, for those accessing private care. The recurrent costs of special foods also eroded the household resources, requiring 8 percent and 5 percent of the household’s monthly income for low- and middleincome households, respectively (Table 3).

Table 1 Cost of inpatient care for diabetes at public hospitals by income group and at private hospitals for all patients Total cost/patient, Rs/day Public care Low-income group (n = 16) Middle-income group (n = 13) High-income group (n = 4) Private care, all income groups (n = 4)

Direct cost/patient, Rs/day

Direct cost as % total cost

Total cost as % household income/day

285

167

59

211

378

232

61

131

467

277

59

57

2,068

1,927

88

624

Note: Household income was calculated from the mean income of the patients included in the table for each income group. Costs were calculated for patients who had incurred inpatient costs in the past three years. Rs, Sri Lankan rupees.

392

/ Perera et al. Table 2 Cost of an outpatient visit for diabetes at public hospitals by income group and at private hospitals for all patients Total cost/patient, Rs/day

Public care Low-income group (n = 13) Middle-income group (n = 12) High-income groupa Private care, all income groups (n = 4)

Direct cost/patient, Rs/day

Direct cost as % total cost

Total cost as % household income/day

418

341

70

345

385

342

59

150

—

—

—

—

1,496

1,484

89

408

Note: Household income was calculated from the mean income of the patients included in the table for each income group. Rs, Sri Lankan rupees. a Ten high-income patients were concurrently receiving treatment from government clinics but getting tests in the private sector; they were incurring high travel costs by private vehicle. These patients were not included since they would distort the total average.

Table 3 Mean cost of special food during non-acute stages of diabetes, by income group Cost of special food, Rs/month

Cost of special food as % household income/month

Low-income group

375

8

Middle-income group

550

5

High-income group

—

—

Note: Costs are for 10 patients who accessed government hospitals. Rs, Sri Lankan rupees.

Although the magnitude of costs in terms of rupees (Rs) was sometimes considerable, it was more often the recurrent small sums, which formed a substantial proportion of a low and at times uncertain income, that challenged the household economy. Seela, a low-income patient from Moneragala, described the erosion of her family’s resources as a result of her diabetes (Box 6).

Medical Poverty Trap in Sri Lanka

/ 393

Economic Consequences of Paying The burden of health care costs on households must be considered in the context of typical livelihoods. In the rural areas studied, a large proportion of the population depended on agriculture, cultivating chena (slash-and-burn) crops or a rice paddy in small plots. Although irrigated agriculture removed the uncertainty of cultivation for some, most agriculture depended on rain and was therefore unreliable. The labor of every member of the household was essential for cultivation. Household income fluctuated from year to year, and cash came in only at harvest time, approximately two to four times a year. When cash was required, most households borrowed small sums from the grocery store or pawned assets such as gold jewelry until the harvest, when debts were cleared and a new cycle of credit was started. In urban areas daily labor was also essential for the majority of people, who depended on daily wage labor and small businesses. The households in our study dealt with the costs of managing diabetes in various ways. Many pawned or sold assets, sometimes income-earning assets such as land or businesses. Pawning brought in only about one-third of the market value, and when households had to sell assets, including income-generating assets, they often accepted a below-market price in order to raise cash quickly for emergency care. Loans were taken out, often at extremely high interest, from local lenders or the village store, and middle- and high-income households could obtain loans from their workplaces or banks. Some households had taken children out of school so that they could work in paid employment. Some short-term relief was provided by government poverty-alleviation programs such as the scheme known as Samurdhi, by employers, and by family members. In remote areas, households received a great deal of active community support in the initial stages of the illness, including help with agricultural labor, provision of food, and accompanying and visiting the patient during hospital stays. This provided crucial support for the short term, but was not sustained to relieve the long-term pressures from diabetes. None of the households cited insurance as a means of paying for medical costs. The majority of households reported drops in household earnings throughout the course of the illness. Income-earning potential was often diminished as income-earning assets were sold, the patient (and, in severe cases, a family caregiver) was removed from household income-earning activities, and continued interruptions to cultivation once or several times every month disrupted the household’s cultivation beyond recovery. Seven of the 18 low-income households had experienced reductions in household income throughout the course of the illness (reductions ranged from 14% to 100%). Among the middle- and highincome households, several had suffered reductions in income, but many had taken remedial action to increase their incomes. Many of the households would have been able to recover from the costs of short-term illnesses; but, given the

394

/ Perera et al.

long-term nature of diabetes, the financial burdens continued to accumulate relentlessly, plunging households into poverty and preventing them from making recoveries. Health and Social Consequences Poor health and burdensome health care costs were found to be both cause and effect of each other (Figure 3). When the recurrent medical costs were too high and households could no longer afford appropriate treatment, the illness progressed, causing recurrent acute phases and, in turn, leading to higher health care costs. Acute phases and diabetic complications were common. Wounds and injuries are normal for agricultural populations, and were usually treated using herbal remedies at home or, at the most, by a visit to the nearest government dispensary for dressing. However, delay in treating a wound for diabetes patients can have grave consequences. Such delays had resulted in infection of the wound and amputation for several of the study participants. Failing eyesight was also common. Diabetes frequently affected the second and sometimes even the third generation. There were changes in the household living arrangements with married children, and children had to forfeit schooling or training to take up jobs, to help maintain the household economy, or to care for household members. The cascading economic deprivation that resulted affected the future prospects of young family members and even extended to married children. Children lost their

Figure 3. The illness and health care costs cycle.

Medical Poverty Trap in Sri Lanka

/ 395

parents’ investments in land and enterprises and household savings. Seela, a low-income patient who had lost the income from her boutique as a result of her illness, expressed concerns about the impact on her son and daughter (Box 6). The burden of caring for ill household members fell on women. When women became ill they tended to postpone care seeking, mainly due to the high opportunity costs of abandoning household tasks and the need to be accompanied to the clinic. Although women prepared special foods for others, they rarely prepared these foods for themselves. POLICY IMPLICATIONS Sri Lanka has developed a health care system with relatively low levels of inequities and high economic and geographic accessibility, which has been successful at tackling communicable diseases and maternal and child mortality. However, demographic and epidemiological changes are posing new challenges to the health system. Currently, infectious diseases remain the predominant cause of morbidity, but the profile of disease is changing rapidly, with a substantial increase in the prevalence of diabetes. Patients’ testimonies revealed problems of accessibility and affordability for appropriate care for diabetes, and the negative effects on households. Through these descriptions of the problems that households are facing, we can draw out several policy implications. The household perspective reveals that the government health care system is the first, final, or only resort for the majority of diabetes patients in all income groups, and therefore the government system will have to take responsibility for health care for the increasing burden of diabetes in Sri Lanka. Diabetic complications were common among the study participants, the most severe resulting in amputation and loss of sight. There is evidence from other studies that diabetes patients in Sri Lanka are particularly likely to suffer complications and that complications are often present at the time of diagnosis (15). Early diagnosis was found to be crucial to effective management of diabetes at the lowest cost and with little or no physical impairment for diabetes patients. The health facilities required to diagnose diabetes are disproportionately located in urban areas, with the result that delayed diagnosis was more common in rural areas. Patients often had to pay for tests at private facilities when they could not travel long distances for diagnostic tests in “free” government hospitals. For some of the patients, their diabetes had been detected in mobile field clinics or health camps held in remote areas, organized by nongovernment organizations in collaboration with government medical officers. This practice could be extended regularly to remote rural areas with the help of service organizations such as Lions Rotary or Jaycees, for screening for diabetes. The potential costs and benefits of implementing a screening program for diabetes in high-risk groups at the primary care level should also be considered.

396

/ Perera et al.

Our results indicate that the need for regular visits to distant hospitals served as a deterrent to adherence to the treatment regimen. To reduce the geographic and economic inaccessibility resulting from travel and opportunity costs of appropriate care, the facilities for managing diabetes must be enhanced at the local level. Managing the condition must be considered a key element of primary care. Lower-level hospitals that are well equipped to treat communicable diseases could be given basic facilities to provide adequate care to patients in the early stages of diabetes. Those with advanced diabetes could then be given specialized care at the major hospitals. We found that there were some forms of external assistance for households in economic hardship, including the government poverty-alleviation program (Samurdhi) and a charity allowance for the indigent. The Samurdhi program needs to include support for patients in beneficiary households who have chronic long-term illnesses such as diabetes, in recognition of the contribution of such illnesses to enhancing poverty. The gap in health education on noncommunicable diseases should be addressed through dissemination of information with as wide an outreach as that achieved for communicable diseases. This will require a shift in the frontiers of primary health care to cater for the noncommunicable diseases. The symptoms of diabetes, the importance of adherence to the treatment regimen, and the changes in lifestyle essential for preventing diabetes or minimizing symptoms and acute disease phases need to be comprehensively addressed. The long-term relationship between provider and patient that is critical for diabetes treatment is not facilitated by the procedures and attitudes now prevailing in hospitals. A greater concern for the rights of the patient to adequate care and information would help to encourage the recurrent visits that are required. The health system needs to reduce crowding at higher-level hospitals and increase the capacity of lower-level hospitals to care for patients at early stages of diabetes. The restructuring should provide more community-centered programs, as were available in the past for noncommunicable diseases. For this the community needs to be given a role in prevention and monitoring through civil society organizations such as patients’ associations, which to some extent can reduce the deficiencies in the health care system. CONCLUSION Sri Lanka has successfully developed a health care system with relatively low levels of inequities and high economic and geographic accessibility. The focus on public funding of inpatient care, with less emphasis on outpatient care, has led to a system in which those who can afford it opt to use private outpatient care, the poor have access to public outpatient care, and all income groups are insured by the public system against the high costs of inpatient care. However, demographic and epidemiological changes are posing new challenges to the health system. As

Medical Poverty Trap in Sri Lanka

/ 397

yet, infectious diseases remain the predominant cause of morbidity, but the profile of disease is changing rapidly, with a substantial increase in diabetes and other noncommunicable diseases. In this article we have considered the accessibility and affordability of care for diabetes, in the context of the rapid epidemiological transition in Sri Lanka. Patients’ testimonies have revealed problems of accessibility and affordability for appropriate care for diabetes. The negative outcomes in terms of health, the household economy, and social consequences are clearly affecting not only the estimated 653,000 people in Sri Lanka who have diabetes (11), but also other members of their families and households. Given the high and increasing prevalence of diabetes, particularly in the middle-aged group, action to relieve the burden on diabetes patients and their families is essential. The policy implications suggested above indicate the need for increased funding for the training of personnel and provision of specialist facilities for the diagnosis and management of diabetes at the local level. Increased investment in lower-level outpatient care will now be essential, given that local, government-provided services are the only feasible option for the ever-increasing number of diabetes patients in order to avoid the long-term, crippling health, economic, and social costs of this disease. Acknowledgments — This study was funded by the Rockefeller Foundation, grant 2001 HE 108, as part of the Affordability Ladder Program on equity and health sector reforms. We are grateful to Di McIntyre, Margaret Whitehead, and Göran Dahlgren for their comments on earlier drafts of this article. REFERENCES 1. Rannan-Eliya, R. P. Strategies for Improving the Health of the Poor—The Sri Lankan Experience. Paper prepared for the Health Systems Resource Centre, Department for International Development, London, 2001. 2. McNay, K., Keith, R., and Penrose, A. Bucking the Trend: How Sri Lanka has Achieved Good Health at Low Cost—Challenges and Policy Lessons for the 21st Century. Save the Children, London, 2004. 3. World Bank. Regional Fact Sheet from the World Development Indicators 2006, South Asia. 2006. http://siteresources.worldbank.org/DATASTATISTICS/Resources/ sas_wdi.pdf (September 2006). 4. Van Doorslaer, E., et al. Paying Out-of-Pocket for Health Care in Asia: Catastrophic and Poverty Impact. EQUITAP Project Working Paper 2. London, 2005. 5. Attanayake, N., Fox-Rushby, J., and Mills, A. Household costs of “malaria” morbidity: A study in Matale district, Sri Lanka. Trop. Med. Int. Health 5:595–606, 2000. 6. Russell, S. Can Households Afford to Be Ill? The Role of the Health System, Material Resources and Social Networks in Sri Lanka. Ph.D. thesis, London School of Hygiene and Tropical Medicine, University of London, 2001. 7. World Health Organization. The World Health Report 2006: Working Together for Health. Geneva, 2006.

398

/ Perera et al.

8. Russell, S. Treatment-seeking behaviour in urban Sri Lanka: Trusting the state, trusting providers. Soc. Sci. Med. 61:1396–1407, 2005. 9. Ghaffar, A., Reddy, K. S., and Singhi, M. Burden of non-communicable diseases in South Asia. BMJ 328:807–810, 2004. 10. World Health Organization. WHO Diabetes Programme Country and Regional Data. www.who.int/diabetes/facts/world_figures/en/ (September 2006). 11. Wild, S., et al. Global prevalence of diabetes: Estimates for the year 2000 and projections for 2030. Diabetes Care 27:1047–1053, 2004. 12. King, H., Aubert, R., and Herman, W. Global burden of diabetes, 1995–2025. Prevalence, numerical estimates and projections. Diabetes Care 21:1414–1431, 1998. 13. Gunatilleke, G. Equity in Health: The Case of Sri Lanka. Marga, Ministry of Health, WHO, Geneva. 1998. 14. Dahlgren, G., and Whitehead, M. A framework for assessing health systems from the public’s perspective: The ALPS approach. Int. J. Health Serv. 37:363–378, 2007. 15. Weerasuriya, N., et al. Long-term complications in newly diagnosed Sri Lankan patients with type 2 diabetes mellitus. Q. J. Med. 91:439–443, 1998.

Direct reprint requests to: Myrtle Perera Marga Institute P.O. Box 601 93/10 Dutugemunu Street Kirulapone Colombo 6 Sri Lanka e-mail: [email protected]