Measuring the economic burden for TB patients in the End TB Strategy and Universal Health Coverage frameworks

Tuberculosis (TB) is a disease of poverty. Ensuring access to health care without the risk of financial hardship due to out-of-pocket health care expenditures (Universal Health Coverage [UHC]) is essential for providing accessible care to underprivileged populations, but this is not enough. The End TB Strategy promotes both patient-centred TB services and social protection measures, which aim to mitigate the economic hardship faced by TB patients and their households due to direct medical and non-medical expenditures, as well as to lost income. The strategy includes a target that no families should face catastrophic total costs due to TB. The indicator linked to this target aims to capture the total economic burden linked to TB care, and thus differs from the ‘catastrophic expenditure on health’ indicator, a key component of the UHC monitoring framework aligned with the Sustainable Development Goals. Countries, and particularly high TB burden countries, are expected to conduct nationally representative TB patient cost surveys to establish baseline measurements for the catastrophic costs indicator. Findings from these surveys should also help identify entry points for developing policies to ensure better financial and social protection for TB patients. In this paper, we define the key measurable concepts for TB patient cost surveys, notably the types of costs that are captured, and related affordability measures. We discuss methods for measuring these notions in the UHC framework and contrast them with how they are measured in TB patient cost surveys. K E Y W O R D S : TB; patient costs; financial protection; affordability

Tuberculosis (TB) is a disease of poverty.Ensuring access to health care without the risk of financial hardship due to out-of-pocket health care expenditures (Universal Health Coverage [UHC]) is essential for providing accessible care to underprivileged populations, but this is not enough.The End TB Strategy promotes both patient-centred TB services and social protection measures, which aim to mitigate the economic hardship faced by TB patients and their households due to direct medical and non-medical expenditures, as well as to lost income.The strategy includes a target that no families should face catastrophic total costs due to TB.The indicator linked to this target aims to capture the total economic burden linked to TB care, and thus differs from the 'catastrophic expenditure on health' indicator, a key component of the UHC monitoring framework aligned with the Sustainable Development Goals.Countries, and particularly high TB burden countries, are expected to conduct nationally representative TB patient cost surveys to establish baseline measurements for the catastrophic costs indicator.Findings from these surveys should also help identify entry points for developing policies to ensure better financial and social protection for TB patients.In this paper, we define the key measurable concepts for TB patient cost surveys, notably the types of costs that are captured, and related affordability measures.We discuss methods for measuring these notions in the UHC framework and contrast them with how they are measured in TB patient cost surveys.K E Y W O R D S : TB; patient costs; financial protection; affordability TUBERCULOSIS (TB) REMAINS a major threat to global public health. 1Poor people in resourceconstrained settings are most at risk of the disease and its devastating economic consequences. 2In lowand middle-income countries (LMICs), health care financing is heavily reliant on out-of-pocket payments.][5] Universal Health Coverage (UHC), whereby everyone can access the quality health services they need without financial hardship, 6 has long been on the global TB control agenda.Free diagnosis and treatment have been the cornerstone of global TB control strategies since 1994. 7The DOTS strategy emphasises the use of low-cost, cost-effective tools and interventions to enable affordable access to quality TB care, which has resulted in 53 million lives saved.[10] Aligned with the Sustainable Development Goals, the World Health Organization's (WHO's) End TB Strategy has an increased focus on poverty alleviation strategies and social protection initiatives that cover costs beyond medical expenses, including income security.It also includes as a target that no TBaffected families should suffer from catastrophic total costs due to the disease. 10,11To monitor progress towards this target, the WHO Global TB Programme convened a task force of experts in 2015 to develop a field-testing protocol and survey instrument for nationally representative, health facility-based surveys of costs faced by TB patients and their households ('TB patient cost surveys'), building upon the Tool to Estimate Patients' Costs. 12After field testing, the WHO developed a handbook for TB patient cost surveys. 11Countries, and particularly high TB burden countries, are expected to adapt and implement these surveys to document the magnitude and main drivers of costs incurred by TB patients (and their households) and the proportion of TB patients who incur catastrophic costs as a result of the cost of care, and to monitor these metrics over time.Findings from these surveys should also help identify entry points for developing policies to ensure better financial and social protection for TB patients. 8s of July 2018, 11 countries had conducted a TB patient cost survey using the WHO instrument and methodology, 13 four surveys were ongoing or near completion, and 13 countries were planning and mobilising funding to conduct such surveys (Figure).
In the present paper, we describe the key notions that are measured using these TB patient cost surveys, notably the types of costs that are captured, and measures of the affordability of these costs in relation to household income, expressed as occurrence of catastrophic costs and impoverishment.We discuss the standard methods for measuring these concepts and how they have been adapted in the TB patient cost survey handbook, and conclude by highlighting areas for consideration for those implementing TB patient cost surveys.

DEFINING ECONOMIC BURDEN FOR PATIENTS AND HOUSEHOLDS
At the heart of the UHC paradigm is the concept that families should not face undue financial hardship in accessing health care.This is referred to as financial protection, and it builds on the notion of affordability of care. 14,15The WHO and the World Bank track financial protection through two indicators: high (or catastrophic) health spending and impoverishment. 6atastrophic health spending quantifies the proportion of the population whose resources would be catastrophically reduced by spending on health care. 16When health care expenditures exceed a given proportion of available income (or expenditure capacity), they are considered 'catastrophic'.The impoverishment approach estimates the proportion of the population that would be pushed below a defined poverty line due to seeking and receiving care. 17Catastrophic spending and impoverishment rates are generally calculated using household level data captured through population-based surveys.

MEASURING CATASTROPHIC HEALTH SPENDING
When measuring catastrophic health spending, there are two key variables underlying this approach: 1) total household out-of-pocket payments for health care (numerator, see the following Sections on 'Measuring and valuing household costs' and 'Generating a ratio of health care costs to a measure of ability to pay'); and 2) a measure of household resources (denominator, see Section on 'Measuring ability to pay').A ratio of health care costs to a measure of ability to pay can then be generated (see Section on 'Generating a ratio of health care costs to a measure of ability to pay'), which is compared to a threshold (see Section on 'Defining thresholds for catastrophic payments').

Measuring and valuing household costs
While the UHC indicator uses household surveys to capture health care expenditures (medical costs) for all conditions, the TB indicator aims to capture instead the total economic burden related to one diagnosed health condition only, i.e., TB.The UHC indicator focuses on direct out-of-pocket medical costs only.
TB patient cost surveys measure three types of cost: direct medical costs, direct non-medical costs and income loss (indirect costs or opportunity costs).Direct medical costs represent the money actually spent out of pocket by the patient on medical services such as prescribed medications, consultation fees, hospitalisation and laboratory tests.These costs are the same as the direct medical costs measured in the UHC framework.
Patients (and their carers) often incur other direct costs associated with the utilisation of health care, such as transport costs to and from the health facility, and costs for accommodation and food, which are referred to as direct non-medical costs.Direct costs are valued by asking patients to recall their actual expenditure.
When seeking care and when sick, individuals also incur costs associated with lost productivity due to illness/disability and time spent seeking care, or looking after a patient instead of working (i.e., carers).These opportunity costs are referred to as indirect costs in the End TB monitoring framework.Two approaches are typically employed to value indirect costs to households: the human capital approach and the output-based approach. 18he human capital approach involves valuing an individual's time by multiplying the number of hours spent seeking and receiving care/caring for by their reported or estimated hourly wage rate. 19If based on reported income, this method can have equity concerns, as it then implicitly values the time of more productive (higher income) individuals more highly and does not take into account the value of time lost by individuals who are performing unpaid work or are unemployed or retired. 20This can be corrected by using a standard estimated income for these individuals (e.g., the mean for the lower quintiles based on national statistics or the minimum civil servant wage).
The output-based approach considers reported changes in income/production. 21This approach is recommended by the WHO for settings predomi-nantly characterised by formal economies, where individuals can reliably report income in monetary terms.
The WHO's generic instrument for TB patient cost surveys collects data that allow the valuation of both indirect costs using the human capital approach and the output-based approach (Table 1). 13The End TB Strategy indicator is generally computed on the basis of the output-based approach, while the human capital approach is used in sensitivity analyses.3][24] To date, researchers have generally employed the human capital approach to value productivity losses associated with TB, with varying levels of precision in the estimations of time and income.However, more than a third of studies included in one recent systematic review that presented indirect costs did not clearly explain the methods that were used to calculate them. 25he economic burden of illness can be measured at the individual level, but it is perhaps more practical to look at the economic impact on the whole household, particularly as other household members also contribute to direct expenditures and may take time off work to care for the ill person or take their children out of school to contribute to the household income. 26The affordability of TB costs is also analysed at the household level due to the impact that TB potentially has on households, as discussed below.
Measuring the ability to pay Ability to pay is usually measured in terms of income, consumption or expenditure.Income refers to earnings from employment and sale of assets and receipt of transfers.Consumption refers to spending on resources (goods and services) consumed by the household.Expenditure excludes consumption that is not based on market transactions (e.g., home production), and refers to goods or services that are purchased but not immediately consumed by the household. 27hile reported income is the gold standard measure of ability to pay, in low-income settings, where employment is mainly outside the formal sector and income is hard to measure reliably, consumption expenditure is often believed to be a more valid measure of economic resources than income.][30][31] In the UHC framework, consumption expenditure is often used rather than income to measure catastrophic expenditure and impoverishment. 6It can be argued that deducting food spending from consumption (non-food expenditure) can better capture a household's ability to pay for health expenditures. 6Alternatively, no deduction for necessities is made.
TB patient cost surveys capture either income or Measuring the economic burden of TB patients consumption expenditure, or both.The TB indicator is computed using the measure of income that is more robust in the specific country setting.For countries collecting more than one measure, the more robust will be used for main analysis and the alternative measures in sensitivity analysis.
Generating a ratio of health care costs to a measure of ability to pay When computing catastrophic spending within the UHC monitoring framework, the numerator is restricted to direct medical costs, 32 and does not measure direct non-medical and indirect costs, as UHC is mainly about moving towards progressive and equitable health care financing, and national financing schemes (tax or insurance-based) covering direct medical costs.The End TB monitoring framework, on the other hand, is designed to also collect data that can guide policies on patient-centred service delivery models that can reduce both direct and indirect costs, as well as social protection schemes for income security and social support.A key element of innovation of the End TB Strategy 'zero catastrophic costs' indicator is thus that the numerator comprises direct medical, non-medical and indirect costs.In TB care, indirect costs have been found to account for a sizeable proportion of total costs (on average 60% of total costs, range 16-94%) in LMICs; 33 these are therefore important elements for capturing all care-related expenditures and the economic impact on TB patients, from the onset of symptoms to the end of anti-tuberculosis treatment.The denominator is further defined as annual household income or annual household consumption expenditure, as outlined in the section 'Measuring ability to pay'. 34The resulting ratio is then compared to the thresholds defined below to determine whether spending is catastrophic.

Defining thresholds for catastrophic payments
The catastrophic payment threshold is set as a proportion of income, i.e., households should not spend more than a pre-defined proportion of their income on health care.When a household's health care payments exceed that pre-defined threshold, they are defined as catastrophic. 16The choice of the threshold has so far been arbitrary.Various thresholds have been used in the literature: 10%, 35 15% of household annual income, 36 or 40% of household non-food expenditure. 32,37The WHO and the World Bank now track catastrophic spending on the basis of out-of-pocket expenditures exceeding 10% or 25% of household total income or consumption. 6or global monitoring of the End TB Strategy 'zero catastrophic costs' indicator, in 2017, the WHO chose to use a threshold of 20% of annual household income, 13 which was set through expert opinion voting in the task force.This is the threshold that is currently used by national TB programmes (NTPs) that implement TB patient cost surveys whose results are annually reported to the WHO. 1,38Countries that conduct national TB patient cost surveys are encouraged to undertake sensitivity analyses whereby the 20% threshold is altered so that the proportion of patients facing catastrophic costs can be assessed at different thresholds, and potentially inform a review of the threshold in the future (Table 1).
The threshold can be used to help define two measures of catastrophic health spending, in both the UHC and the End TB Strategy framework.The catastrophic payment headcount measures the incidence of catastrophic health care costs (i.e., the number, or proportion, of individuals who have been exposed to catastrophic expenses).The catastrophic payment gap (or excess) measure is used to assess the intensity or severity of catastrophic spending by looking at the extent to which health care costs exceed the pre-defined threshold (Table 1). 16he proportion of patients incurring catastrophic costs due to TB is derived from the number of TB patients with catastrophic costs divided by the number of all TB patients treated at NTP facilities.This means that the sampling frame is notified patients on treatment, rather than all people with TB in the community or households in a country.This is selected for practical reasons, as the only available sampling frame is notified TB patients, and household surveys would require a large sample size to include a sufficient number of prevalent TB cases.

MEASURING IMPOVERISHMENT
An additional measure of the affordability of care used for UHC monitoring is impoverishment, or whether health care costs push households into poverty (or more deeply into poverty).In this case, the threshold is absolute and set in terms of a poverty line.If health care payments cause household income/consumption expenditure to fall below the poverty line, they are considered 'impoverishing'.The widely used international dollar-a-day poverty line proposed by the World Bank to allow international comparability was replaced by USD 1.25/day in 2009, at 2005 purchasing power parity. 39Countries also have their own national poverty lines which may be relevant for comparing impoverishment over time within a country.
The incidence of impoverishment measures the increase in poverty due to health care spending.The poverty gap is the shortfall from the poverty line.While these are not included in the End TB Strategy monitoring, countries can include them in the analyses of TB patient cost surveys.Table 2 provides a summary of the key measures presented in this section and in the Section, 'Defining thresholds for catastrophic payments'.

TOWARDS ZERO FAMILIES FACING CATASTROPHIC COSTS DUE TO TUBERCULOSIS: AREAS FOR CONSIDERATION
The End TB Strategy target is a first important step in broadening the concept and measurement of affordability to account not only for medical costs but also for the broader economic impact of TB, including non-medical and indirect costs.
However, as described above, the application of the concepts and standard methods of financial protection requires further development in the End TB Strategy.The WHO recently published a handbook based on the experiences and data from the first round of surveys between 2016 and 2017, which provides comprehensive guidance for conducting facility-based cross-sectional surveys to assess TB patient costs. 13This would benefit from periodic methodological updates based on multicountry analyses of survey findings and strengthen collaboration with health economists, NTPs and policy makers.These updates include methods for calculating confidence intervals for key survey indicators adjusted for the sampling design, a regression-based approach for imputing missing costs, recommendations on the design of a household expenditure questionnaire (to derive a household income measure based on expenditure) and adaptation of the survey instrument to high-income settings.
There are a number of areas for consideration for those implementing TB patient cost surveys going forward, including descriptive analyses of costs that unpack direct medical and non-medical costs and indirect costs, as they can provide valuable information to identify entry points for appropriate polices and interventions to minimise these costs; the use of both the human capital and the output-based approaches to value indirect costs for comparison and correlation; and measuring and comparing income and consumption expenditure to compute financial protection measures.Approaches and metrics in addition to the standard End TB Strategy framework methodology include measuring impoverishment, computing the catastrophic payment gap and sensitivity analyses with different proportions of income thresholds (Table 1).
Finally, it is important to bear in mind that the cross-sectional study design for a TB patient cost survey recommended by the WHO inevitably focuses on the economic consequences of TB using a measure at one point in time.It therefore fails to capture the long-term economic consequences of the disease for the household, including the impact on reduced labour supply and productivity, and household resilience.Coping mechanisms were originally explored as part of the development of the TB indicator, as they were deemed to be potentially less labour intensive to collect and easier to integrate in routine surveillance.However, as coping mechanisms differ in different cultures and societies, it is difficult to consider them as a proxy for catastrophic payments.
Several research studies that have adapted the WHO generic protocol to a longitudinal design, including for long-term follow-up after anti-tuberculosis treatment, are now ongoing.These studies will be helpful for the validation and interpretation of cross-sectional TB patient cost survey data.Separate studies of non-notified TB patients, such as those in private care, are required to measure costs in situations where user charges for clinical care are often higher than in facilities linked to NTPs.However, other studies sampling people with TB who are not under treatment at the time of the study are also needed, as the current methodology only includes TB patients who remain in care.Such studies can be conducted in the context of tracing patients who are lost to follow-up (e.g., initial loss to followup or loss to follow-up during treatment) by reconnecting them with treatment and explore reasons for loss to follow-up.The assessment of costs incurred by such patients may shed light on costs related to the disease and disability that are not linked to care seeking, and costs of living with TB without getting proper care.

CONCLUSIONS
In this paper, we have described economic burden and affordability concepts and measurements that under-lie the End TB Strategy indicator of 'zero catastrophic costs' due to TB, and have highlighted the novel elements of this indicator in relation to approaches used in the UHC monitoring framework.Further findings from national TB patient cost surveys, multicountry analyses and research using alternative approaches will be important in providing further evidence to refine metrics and methodology for country-level implementation and global monitoring.
The conventional concepts and measurement of 'financial protection' of the UHC monitoring framework have been taken a step forward in the End TB Strategy to ensure metrics are able to capture the total economic burden of TB on patients and families.This approach has the potential to inform the design of financing and implementation of both health care and social protection policies that aim to prevent both direct and indirect costs of care, and ultimately ensure that TB care is truly affordable for TB patients.

R E S U M E N
La tuberculosis (TB) es una enfermedad de la pobreza.Procurar un acceso a la atenci ón de salud sin el riesgo de dificultades econ ómicas debido a los gastos directos de la atenci ón (cobertura universal de salud [UHC]) es primordial cuando se busca prestar una atenci ón accesible a las poblaciones desfavorecidas, pero no es suficiente.La Estrategia Fin a la Tuberculosis promueve servicios de TB centrados en el paciente y también medidas de protecci ón social, encaminadas a mitigar las dificultades econ ómicas de los pacientes con TB y sus hogares, debidas a los gastos directos médicos y de otro tipo y también a la pérdida de ingresos.Uno de los objetivos de la estrategia consiste en que no haya familias que tengan que hacer frente a gastos catastr óficos debido a la TB.El indicador vinculado con este objetivo busca captar la carga econ ómica total que se asocia con la atenci ón de la TB y con ello difiere del indicador de 'los gastos catastr óficos de salud', que es un componente primordial del marco de la vigilancia de la UHC, en consonancia con los Objetivos de Desarrollo Sostenible.Se prevé que los países, en especial los que soportan una alta carga de morbilidad por TB, realicen encuestas nacionales representativas sobre los costos que cubren los pacientes, con el fin de fijar las medidas de referencia para el indicador de los gastos catastr óficos.Los resultados de estas encuestas deberían contribuir adema ´s, a reconocer los puntos de acceso para formular políticas que garanticen una mejor protecci ón econ ómica y social a los pacientes con TB.En el presente artículo se definen conceptos primordiales medibles para las encuestas sobre los costos de los pacientes con TB, en especial los tipos de costos que se han de captar y las medidas conexas de asequibilidad.Se analizan los métodos de medici ón de estas nociones en el marco de la UHC y se contrastan con la forma como se miden en las encuestas de los costos de los pacientes con TB.
Measuring the economic burden of TB patients i

Figure
Figure Global implementation of TB patient cost surveys following the WHO methodology as of July 2018.Source: WHO Global TB Programme, July 2018. 12TB ¼ tuberculosis; WHO ¼ World Health Organization.This image can be viewed online in colour at http:// www.ingentaconnect.com/content/iuatld/ijtld/2019/00000023/00000001/art000...

Table 1
12mmary of recommended and additional approaches, metrics and valuation methods for TB patient cost surveys based on the World Health Organization methodology12 TB ¼ tuberculosis; USD ¼ US dollar; PPP ¼ purchasing power parity.

Table 2
40 )ary of key measures of catastrophic health spending and impoverishment for general Universal Health Coverage monitoring (source: adapted from40 )