Background
The 2010 WHO Global Status report on non-communicable diseases (NCDs) showed that they are now the most important cause of mortality worldwide. Indeed, more than 36 million people died from NCDs in 2008, mainly cardiovascular diseases (48%), cancers (21%), chronic respiratory diseases (12%), and diabetes (3%). Nearly 80% of these deaths occurred in low- and middle-income countries (LMICs), where, on average, they now exceed communicable diseases as the major cause of disease burden
Considerable literature exists on the impact of NCDs on households in high-income countries
Framework for the analysis of economic impacts of NCDs on households (modified from McIntyre et al.
Framework for the analysis of economic impacts of NCDs on households (modified from McIntyre et al.
We conducted a literature review to present existing evidence on the financial burden from NCDs in low- and middle-income settings, at the individual and household level. The aim is to provide accurate and relevant information on this important issue to policymakers, and determine where further research is needed.
Methods
We performed a literature search with Cabdirect, Sciencedirect and Web of Knowledge, using combinations of the following key words: “Non-communicable disease”, “chronic illness”, “diabetes”, “cardiovascular disease”, “cancer”, and “chronic respiratory disease” with “cost”, “impoverish”, “financial burden”, “health expenditure”, “expense”, “out-of-pocket”, “health spending”, “catastrophic expenditure”, “catastrophic expense”, and “catastrophic spending”. A total of 8,966 results (including duplicates) were obtained. After duplicate removal, titles and abstracts of the remaining papers were reviewed to assess their relevance according to the following inclusion criteria: i) papers in English or French; ii) from 1990 onwards; iii) covering at least one low-, lower-middle- or upper-middle-income countrya
Studies reviewed.
Click here for file
Results
Overview of the methods used in the literature
The studies found in the literature reflect the diversity of methods used to assess household financial burden from NCDs. The methodological differences in the studies inherently prevent a formal meta-analysis from being performed. However, at the same time, these differences offer opportunities to explore results through the lens of different techniques. In this section, we present a discussion on the methodologies used.
Some studies look at a specific NCD (e.g., diabetes, cancers, cardiovascular diseases), while a majority consider NCDs in general or a combination of two or more specific NCDs. We found only one previous literature review which included studies on multiple NCDs, but it includes studies from only a few countries and did not include any studies from Africa and Latin America
The original studies found also differed according to data sources and sample sizes. Some authors conducted their own surveys for the purpose of the studies, while others used data from existing surveys carried out by another entity (e.g., National Institute of Statistics, Ministry of Health, Health Insurance Plans). In these surveys, households and individuals were generally chosen randomly, through simple, stratified or cluster sampling
In the studies looking at NCDs in general, the term “chronic diseases” is frequently used, and even if the major NCDs are usually taken into account, the definitions vary from one study to another. For example, Shi et al. defined a chronic ailment as an ailment that lasts or is expected to last for at least 12 months, resulting in functional limitations or the need for ongoing medical services, and includes disability
Irrespective of the diseases considered, many studies assessing the direct costs incurred by households for the treatment of NCDs also focus on impoverishment and catastrophic health expenditure due to these expenses. Impoverishment occurs when a respondent would have had a net income above the poverty line in the absence of the expenditure on the disease, but below it after. Different poverty lines are used across studies – US$ 1 per day, US$ 1.08 per day, US$ 1.25 per day and US$ 2 per day
Catastrophic heath expenditure occurs when people spend a disproportionate amount of their income (sometimes non-food expenditure) on the condition, as described in Xu et al.
The evidence on the direct costs from non-communicable illnesses
Many of the studies assessed direct costs, which include all costs incurred by individuals and households for the treatment of NCDs. In theory, these costs should be net of any reimbursement from insurance. We present evidence on these direct costs organized by disease.
Diabetes
Diabetes is a leading NCD and 16 studies included in this review looked at the direct costs incurred for both outpatient and inpatient services. All studies, except one, relied on convenience samples, so the results need to be interpreted carefully. Overall, the studies found that varying shares of household income are allocated to paying for diabetes care. This ranges from as low as 5% of income for a rural low-income population in India to up to 24.5% for a low-income group in Madras (India)
Medications are frequently found to be the largest component of expenditure on diabetes
The presence of complications and the duration of the illness are usually associated with an increase of the direct costs. For example, Khowaja et al. found that in Pakistan, the direct cost for patients with co-morbidities was 45% higher than the direct cost for patients without co-morbidities
Some studies looked at coping strategies used by households to pay for these direct costs. In India, the majority of patients (89%) used their household income to fund the monitoring and treatment of their diabetes, while household savings were used by 22% of retired patients and by 19% of those in the lowest income bracket. When faced with hospitalization, 56% of patients had to dip into their savings or borrow in order to fund the costs
Five studies examined spending on cardiovascular diseases. In a study using data from a household survey in Kazakhstan, people with cardiac problems were found to pay on average 24% more for health care than people with other health problems
Out-of-pocket payments for the treatment of cardiovascular diseases also lead to significant costs for households. Up to 71% of patients who had experienced an acute stroke were found to face catastrophic health expenditureb in China, while 37% of them fell below the poverty line of US$ 1 per day after paying for their health care
In a study covering 35 states and union territories in India, Rao et al. investigated the coping strategies used by households to deal with expenses incurred for hospitalizations due to cardiovascular diseases
Cancers also represent an emerging health problem in LMICs and seeking health care for these diseases can have a significant effect on families’ welfare. We found three papers which focus specifically on the direct cost from cancers. In a study using data from a randomized household survey in Pakistan, 27.1% of those who sought care for cancer at private facilities were found to finance their care through unsecured loans, while 7.1% relied on assistance from others
Two studies using convenience samples also shed some light on components of spending on cancer care. Indeed, Zhou et al. found that health insurance facilitates the financial access of treatment for patients suffering from oesophageal cancer in China, particularly for purchasing drugs
The financial burden from other NCDs, such as epilepsy, cirrhosis, chronic obstructive pulmonary disease (COPD), rhinitis and depressive disorders, is also estimated in some studies. Even if they are not as studied as the major NCDs presented previously, these types of illnesses can also exert a considerable pressure on household finances. For example, a study from Mumbai (India) based on a random sample of households found that the share of the annual personal income spent on outpatient care for allergic rhinitis was 1.7% when treatment was sought in public facilities. Similarly, care for COPD represented 13.3% of annual personal income among those using private facilities. With hospitalization at public facilities, out-of-pocket payments for COPD represented up to 62.3% of the annual personal income compared to 50.7% for hospitalization in private facilities
Coping strategies used to pay for care associated with these NCDs are similar to those used to cope with more documented NCDs. In Pakistan for example, Mahmood and Ali Mubashir using a random sample found that 22.9% of patients with circulatory diseases (heart diseases, rheumatic fever and blood pressure) who visited private doctors/clinics for treatment financed care through unsecured loans, while 8.8% relied on assistance from others
We found a large number of studies – all based on randomized household surveys – looking at NCDs in general, instead of focusing on specific illnesses. Some studies highlight the association of having a household member suffering from a chronic disease with a significant increase in health care expenditure and a higher risk of impoverishment. In Russia, for example, each additional case of chronic disease in a household was found to increase the probability of incurring health care expenditure by 8% and the amount of healthcare expenditure by 6.2%
Undeniably, expenses incurred when seeking health care for chronic diseases represent an important financial burden for households as presented in Table
In several studies, the presence of household members with chronic ailments was also found to lead to catastrophic health expenditure and impoverishment. The probability of catastrophic expenditure was then 4.4 times higher among households having incurred expenses for treating chronically ill persons in Georgia, and up to 7.8 times higher in Burkina-Faso
As with diabetes, when households are covered by health insurance, the reimbursement rates for chronic diseases are relatively low. In Shandong and Ningxia in China, for example, only 11.16% and 8.67%, respectively, of overall medical expenditure for chronic diseases was reimbursed by the NCMS
Coping strategies documented in the literature combining chronic diseases are similar to those described in the studies on specific NCDs. In Georgia, when households were lacking financial means, the most dominant strategy was to borrow from a friend or relative (70%), followed by selling household valuables (10%) and/or household goods/products (10%)
Literature on the indirect costs due to non-communicable diseases in low- and middle-income countries
Households and individuals also bear indirect costs when they are affected by NCDs. These costs mainly include time and productivity loss by patients and caregivers because of the illness as well as income lost by patients and family members. Whereas there is no doubt that these indirect costs can pose a substantial burden on households, there are numerous methodological challenges in measuring this burden adequately; these challenges have been discussed in detail in a previous study
In India, one study suggests that the indirect cost for diabetes patients and their caregivers was 28.76% of the total treatment cost. It was claimed that loss of income of the patient comprised the greatest portion of indirect costs (60.54%), followed by loss of income of caregivers (39.46%)
The loss of income borne by patients suffering from NCDs is mainly due to self-reported absenteeism from usual economic activity. In fact, the treatment of NCDs usually requires repetitive visits to health facilities in addition to the inability to work due to their poor health. This can lead to additional losses of working time both for patients and caregivers. In the literature, the mean loss of working time reported by patients was found to vary from 2.8 ± 1.7 hours per visit for diabetes in Pakistan to 58 ± 105 days per year for epilepsy in India
However, time costs are not limited to patients, but also affect caregivers. In Buenos Aires (Argentina) for example, Arrossi et al. found that in 45% of households with a member suffering from cervical cancer, at least one member reduced his/her working hours
Some other forms of indirect costs due to NCDs were found in the literature; these generally concern households’ livelihood and welfare. The study on cervical cancer in Buenos Aires (Argentina) by Arrossi et al. examined these and also found that due to a loss of income, there were delays in payments for essential services such as telephone or electricity and as a result 43% of households had the service cut
There were also significant effects on self-reported daily food consumption, which was reduced in 37% of households, while 38% of households reported that they sold property or used savings to offset income loss. Some impacts on education were found and school absences were more prevalent in 28% of households. There were also problems to pay for education in 23% of households. Furthermore, 45% of patients were cared for by one or more informal caregivers that did not live with them and one-third of these caregivers’ households reduced their daily consumption of food and 26% had delays in payments of essential services such as electricity or telephone services. It should be noted that these are the types of welfare losses have shaped the concept of catastrophic health expenditure.
There were also direct impacts on employment and at least one member stopped working in 28% of households affected by cervical cancer. Several interviewees who stopped working expressed the hope of going back to their jobs after treatment, fearing at the same time that this would no longer be possible. Similarly, a study from Bangalore (India) by Rayappa et al. found that only 33.4% of diabetes patients worked and among those working, 23% experienced problems at their job, affecting their productivity and at times requiring changing work to a less strenuous job (5.9%) or giving up the job (14.7%)
Discussion
This literature review has presented the available evidence on the household financial burden related to NCDs in LMICs. However, before discussing its most important results, it is important to highlight some of the methodological issues in many of the studies that were included. First, the heavy reliance on convenience samples taken from people who are seeking and obtaining treatment, often at hospitals, will almost certainly result in an upward bias in costs for the average person with the condition. The people who do not seek treatment or who seek treatment at a lower level of care, implying lower costs, have no chance of being selected.
Second, self-reported costs, even from random samples of patients, are likely to be biased upwards when there are no controls. Some of the people with the condition would have incurred some health expenses in any case and this can only be captured by including controls without the condition
This issue is particularly important when considering indirect costs. It is clear that the method of asking people how many days they could not work overestimates the true loss in work time from a disease because many of the people, particularly in low-income countries, would not have been working on those days, or for all of those days, in the absence of the disease
The substantial variations in study designs and definitions described earlier also make comparisons tricky and meta-analysis infeasible. There is considerable heterogeneity in objectives and the methodologies used in the papers. While we have more confidence in the studies relying on randomized samples, we present more details about each study in file 1: Table S1 to give readers further information and to allow them to consider possible generalizations of the results. Taking into consideration the methodological issues highlighted here and in earlier sections, we can still conclude that NCDs already impose substantial financial costs on some of their sufferers in lower-income countries. As a result, the cost of obtaining treatment for NCDs is also becoming a cause of impoverishment and financial catastrophe in these countries. While this is not particularly surprising given the growing burden of disease associated with these conditions, it has not been documented before.
Again not surprisingly, complications related to the severity of illness were found to increase the household financial burden, both for the patient and for caregivers. Health promotion, prevention and early treatment would reduce some of these costs although each country would need to choose the appropriate mix of prevention and treatment according to their relative costs and impact. We also found strong evidence that costs could be reduced by more rational use of medications for NCDs. The costs of medication for all the different types of NCDs considered here accounted for the highest proportion of the direct costs; where addressed, originator brand medicines were frequently used instead of available generics and costs were then substantially higher than they needed to be. While many LMICs already have strategies to promote the rational use of medicines, there is still some way to go particularly in promoting the use of lower cost generics.
The weakness or non-existence of mechanisms to protect households financially from the burden of NCDs is, however, probably the most important finding in this study. In the studies that considered insurance and provided information on reimbursement rates, NCD-related treatment is generally uncommon and frequently patients and their relatives do not report that they claimed any reimbursement from insurance or employers. Likewise, none of the studies we reviewed reported a system of social security that provides compensation for loss of income incurred by patients and their families because of NCDs. Poor households are more likely to suffer disproportionally from the financial effects of this lack of social protection. To meet the costs, households reported taking unsecure loans, using savings or selling household assets, all of which can lead to longer-term problems for the household. For example, the wider literature suggests that many of the loans taken by households for health expenses are at very high interest rates that can take generations to repay
While we think that the financial costs reported in this review will overestimate the costs of a typical patient with NCDs, such that the numbers cannot be used to extrapolate the costs of NCDs to a country, they highlight the other consequences of the lack of financial risk protection in LMICs. In the random sample studies, many people with NCDs reported that they did not seek care at all because of financial reasons (Additional file
Financial difficulties: a major cause for not seeking care for NCDs.
Click here for file
Through this review, we are also able to identify areas where further research is needed. Among the four major NCDs, the financial costs from chronic respiratory diseases are very poorly documented, although they cause four times more deaths than for example diabetes, which has been researched more
Conclusions
The literature on the social, financial and economic consequences of NCDs in developing countries has not kept pace with the epidemiological evidence. It has been known for some time that the burden of disease associated with NCDs and injuries is already higher than that associated with the health conditions included in the Millennium Development Goals (HIV/AIDS, tuberculosis, malaria, and maternal, child and reproductive health), even in developing countries. Moreover, it has been well documented that the share of NCDs in the overall disease burden will continue to increase globally. Indeed, the UNs’ 2011 conference on NCDs stressed the importance of these diseases as a development issue.
The literature we reviewed sheds some light on the financial consequences of NCDs on households in LMICs. Nonetheless, there are limitations to generalization of these findings due to methodological challenges. Valid estimates of the average costs of NCDs will require random samples with controls to account for people who have costly and less costly treatments, and what would have happened in the absence of the diseases. Panel data would be ideal although these studies are more expensive than cross-sectional designs. However, importantly, this review suggests that it is equally as important to focus on people who could not seek care for NCDs due to financial reasons. Little is known about the subsequent development of disease, impacts on these people’s health and the financial, social and other consequences associated with foregone treatment.
The push to develop health-financing systems that improve financial risk protection and help achieve universal health coverage in LMICs is promising. However, policymakers need to ensure that the health as well as the financial burden from NCDs is adequately addressed in future reforms, while at the same time improve access and financial protection for all other health services needed by the population.
Endnotes
aUS$ 995 or less, US$ 996 to US$ 3,945, and US$ 3,946 to US$ 12,195, respectively.
bDefined as out-of-pocket expenses that accounted for ≥30% of the total annual household income that was reported at baseline.
Abbreviations
COPD: Chronic obstructive pulmonary disease; LMICs: Low- and middle-income countries; NCDs: Non-communicable diseases; NCMS: New cooperative medical scheme; WHO: World Health Organization.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
PS and KX conceived the project and the questions to be studied. HTK and PS conducted the literature review, produced the draft of the paper and produced the final version with the input of other authors. DBE and KX contributed to writing the paper. All authors read and approved the final manuscript.
Authors’ information
The views and opinions expressed in this article are entirely those of the authors and should not be attributed in any manner whatsoever to the organizations the authors are affiliated to. HTK is PhD Student at Aix-Marseille University (Aix-Marseille School of Economics), Marseille, France. PS is Technical Officer at the Department of Health Systems Financing, World Health Organization, Geneva, Switzerland. KX is Team Leader, Health Care Financing at WHO Regional Office for the Western Pacific Region in Manila, Philippines. DBE is the Director of the Department of Health Systems Financing, World Health Organization, Geneva, Switzerland.