M I N I S T E R I A L C O N F E R E N C E A M S T E R D A M , 22–24 M A R C H 2000

THE ECONOMIC
IMPACTS O F
TUBERCULOSIS

the
Stop TB
Initiative
2000 Series

the Stop TB Initiative

is a partnership hosted b y APartnership
the World Health Organization

WHO/CDS/STB/2000.5
Global
for
Original:English
Distr.: Limited Action
 Distr.: Lim

the
Stop TB
Initiative
The economic
impacts of
tuberculosis

the Stop TB Initiative

is a partnership hosted by
the World Health Organization
Project Manager and Series Editor:
Heidi Larson

Writer: Dennis A. Ahlburg, PhD

Editorial Assistant: Karen Reynolds

Series Designer: Anne Guilloux

The author has benefited from comments from Maarten Bosman, Chris Dye,
David Evans, Arata Kochi, Heidi Larson, Paul Nunn, and Vikram Pathania
and from discussions with Richard Bumgarner, David Evans, Dean Jamison,
Arata Kochi, Heidi Larson, Paul Nunn, Vikram Pathania, Mario Raviglione,
and Sergio Spinaci.

© World Health Organization, 2000

This document is not issued to the general public, and all rights are reserved by the World Health
Organization (WHO). The document may not be reviewed, abstracted, quoted, reproduced or translated, in
part or in whole, without the prior written permission of WHO. No part of this document may be stored in
a retrieval system or transmitted in any form or by any means—electronic, mechanical or other—without
the prior written permission of WHO.
The views expressed in documents by named authors are solely the responsibility of those authors.
The mention of specific companies or of certain manufacturers’ products does not imply that they are
endorsed or recommended by WHO in preference to others of a similar nature that are not mentioned.
Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.

The Stop TB Initiative is a partnership hosted by

World Health Organization • 20, avenue Appia • 1211 Geneva 27 • Switzerland
Tel. +(41) 22 791 2675 • Fax +(41) 22 791 4199 • www.who.int
 The economic impacts
of tuberculosis

Table of contents
Executive summary .................................................................... 5

Section 1—The morbidity and mortality burden of tuberculosis ........................................................................................................................................................................ 6

Tuberculosis and women ..................... 7

Tuberculosis and the poor ................. 9

Section 2—The economic and social burden of tuberculosis ................................................................................................................................................................................. 11

Treatment costs borne by patients and their families ................................................................................................................................................................................. 11

Non-treatment costs borne by patients and their families ................................................................................................................................................................ 12
Costs from estimates of simulation studies ................................................................................................................................................................................................................. 13
Psychological and social costs ............................................................................................................................................................................................................................................................... 14

Section 3—Household coping strategies ................................................................................................................................................................................................................................................... 16

Section 4—Benefits and costs of new strategies of dealing with TB ....................................................................................................................................................... 18

Intervention strategies for TB control ....................................................................................................................................................................................................................................... 18

Simulation studies of alternative TB treatment regimes ......................................................................................................................................................................... 19
Obstacles to the adoption of DOTS ............................................................................................................................................................................................................................................... 22

Section 5—Conclusion .................................................. 24

References .......................................................................................... 25

Tables ............................................................................................................ 29
Acknowledgements
The author has benefited from comments from Maarten Bosman, Chris Dye, David Evans, Arata Kochi,
Heidi Larson, Paul Nunn, and Vikram Pathania and from
discussions with Richard Bumgarner, David Evans, Dean Jamison, Arata Kochi, Heidi Larson, Paul Nunn,
Vikram Pathania, Mario Raviglione, and Sergio Spinaci.
Executive summary

Tuberculosis (TB) is the largest single infectious cause of death among young people
and adults in the world, accounting for nearly two million deaths per year.

The economic impact of tuberculosis come from the size of the problem and from the
fact that in developing countries the majority of those affected are in the economically
active segment of the population.

Women who suffer from TB are often less likely to be detected and treated than men.

Although tuberculosis is commonly thought to be a disease of the poor, this is not

exclusively the case. Although the poor are more likely to suffer from the disease,
a significant proportion of those infected are literate, have considerable education,
and earn good incomes.

The substantial non-treatment costs of TB are borne by the patients and their
families. These are often greater than the costs of treatment to the health sector.
The largest indirect cost of TB for a patient is income lost by being too sick to work.
Studies suggest that on average three to four months of work time are lost, resulting
in average lost potential earnings of 20% to 30% of annual household income.
For the families of those that die from the disease, there is the further loss of about
15 years of income because of the premature death of the TB sufferer.

When a woman suffers from TB, additional losses may result. The household loses
the activities that the woman routinely performs in the household: cooking, cleaning,
childcare, and managing the activities of the household.

In addition to the economic costs, TB causes psychological and social costs. TB

patients may be rejected by family and friends or lose their jobs. In some societies,
TB patients are seen as damaged for life or unmarriageable. Such discrimination can
result in anxiety, depression, and reduction in the quality of life.

Households have developed a number of strategies for coping with the costs of
illness and death that result in actual losses being less than the potential losses.
However, some of these short-term coping strategies can have significant long-term
costs. In particular, the sale of assets can reduce the economic prospects of the
household. Reducing the food intake of children or removing them from school
can seriously undermine their health, their education, and their future prospects.

A number of studies of the DOTS strategy showed that the introduction of DOTS
considerably lowered the indirect costs of TB to patients and their families. Estimates
suggest that the introduction of DOTS could halve the current potential national
economic loss from TB. DOTS is very cost-effective and its introduction does not
imply that more funds are needed: at least some of the funding could come from
the reallocation of funds away from poorer, less cost-effective strategy.
Section 1

The morbidity and mortality burden

of tuberculosis
Although much is known about the epidemiology of tuberculosis, relatively little is known
of its economic and social impacts. The illness and death caused by the disease have far-
reaching economic, psychological, and social consequences. By and large, households do not
react passivelyto episodes of severe illness; instead, they develop coping mechanisms to lessen
the impact of illness and death. This study will explore the economic, psychological, and social
impacts of tuberculosis and the strategies that households use to cope with them.

I n 1997, there were an estimated 7.96 million new cases of tuberculosis world-
wide and 16.2 million prevalent cases (Dye et al. 1999). Tuberculosis is the largest single
infectious cause of death among adults in the world, accounting for about two million
deaths per year (Dye et al. 1999). Ninety-five percent of cases and deaths occur in
developing countries and TB accounts for approximately seven percent of all deaths
in developing countries (Murray et al. 1993, p. 239).
The economic impact of tuberculosis comes both from the size of the problem
and from the fact that in developing countries the majority of disease and death occurs
among the most economically active segment of the population: more than 75% among
those 15 to 54 years of age (Murray, 1996, p. 212). Tuberculosis accounts for almost 20%
of all deaths in this age group and 26% of preventable deaths (Murray et al. 1993, p. 241).
Of the estimated number of cases of tuberculosis worldwide, only half will be
detected (Murray et al. 1993; Raviglione and Luelmo, 1996). Without effective chemo-
therapy treatment, 50% to 60% of people with tuberculosis will die of the disease
(Murray, 1996, p.203). Those who remain untreated and are living with active tuberculosis
will infect between 10 and 15 other people every year (WHO, 1998a). For those who are
detected and receive treatment, the fatality rate is reduced to about 15% after five years,
although the death rate can be lower depending upon the characteristics of patients
and how drugs are used (Murray et al. 1993; Dolin, Ravigilone, and Kochi, 1994).
Chemotherapy treatment increases the life expectancy of an otherwise healthy person
with TB by 25–30 years (WHO, 1997, p. 166). However, treatment is often inconsistent or
incomplete and can result in multidrug-resistant TB (MDR–TB). MDR–TB is more likely to
be fatal and is as much as 100 times more costly to treat (WHO, 1998a). Indeed, the cost
is so great—US$ 250 000 per patient in developed countries and an estimated US$ 1 000
to US$ 10 000 in developing countries (Sawert et al. 1997)—that treatment is all but
impractical in most developing countries.
In many regions of the world, tuberculosis is a growing problem. Because of
a combination of economic decline, insufficient application of control measures (case
detection and chemotherapy), and the HIV/AIDS epidemic, tuberculosis is on the rise
in developing and transitional economies. Between 1993 and 1996, there was a 13%
increase in estimated tuberculosis cases worldwide, one-third of which can be attributed
to HIV (WHO, 1998a). In 1997, there were 10.7 million people co-infected with TB and HIV
(Dye et al. 1999). A person who is HIV-positive and infected with TB is 30 times more
likely to develop clinical symptoms than is an infected person who is HIV-negative,
because their weakened immune systems allow the bacteria to develop unchecked
(WHO, 1998a). A person who is HIV-positive and develops tuberculosis can expect to
survive only five to six weeks, although chemotherapy can increase such an individual’s
life expectancy by two to five years (WHO, 1997, pp. 156, 166). It has been estimated that
an additional 105 000 to 210 000 cases of tuberculosis occur in sub-Saharan Africa each
year because of the HIV/AIDS epidemic—8% to 16% of all cases—and about 250 000
in India—about 10% of all cases—(Murray et al. 1993, p. 237; WHO, 1997, p. 160). In sum,
the HIV epidemic has produced more TB cases that are more difficult to diagnose and
more expensive to treat (Raviglione et al. 1997).
In Eastern Europe and the former Soviet Union, economic dislocation has contributed
to an increase in tuberculosis. For example, the Russian Federation has reported a 69%
increase in new cases between 1991 and 1995. The Russian Federation now has the
highest TB mortality rate in Europe (Migliori et al. n.d.).
The conditions leading to an increase in tuberculosis are unlikely to abate quickly.
Consequently, WHO (1998a) estimates that by the end of the century HIV infection will
cause an additional 1.5 million cases of tuberculosis annually. Under conservative
assumptions, Murray et al. (1993) forecast a 10% increase in deaths from tuberculosis
between 1990 and 2015. About two-thirds of the increase will result from demographic
factors such as population growth and the changing age-structure of the population;
the remainder will result from increasing incidence rates, except in sub-Saharan Africa
where the proportions will be reversed due to the HIV epidemic (Dolin et al. 1994).
While tuberculosis is on the increase, economic difficulties in some countries are
putting pressure on health budgets. In these circumstances, health departments need
to use the most cost-effective treatments for tuberculosis, that is, those approaches that
provide effective treatment or a cure for the lowest cost. In many cases, the most cost-
effective treatments are not being used. Furthermore, decisions on what treatment
regimes to follow are often based only on costs to the health ministry. The costs borne
by patients have largely been ignored, even though such costs often exceed the costs
to the health ministry. For example, Saunderson (1995) found that 70% of the cost of TB
treatment in Uganda was borne by the patient and his or her family. When the costs and
benefits of investments in health are being considered, the total social costs (public costs
plus those borne by individuals), and not just the government costs, should be taken into
account in order that efficient choices in health care may be made (Weinstein et al. 1996).
If private costs are ignored, too little investment may be made and it may be allocated in
a way that does not minimize the burden of disease.

Tuberculosis and women

The health of women has attracted increasing attention over the last decade,
particularly in the wake of the International Conference on Population and Development
held in Cairo in 1994. Much of this attention has focused on reproductive health and
maternal mortality; relatively little attention has been given to tuberculosis despite the
fact that tuberculosis is the single leading cause of death among women of reproductive
age (WHO, 1998b). Indeed, it has been estimated that tuberculosis causes more than
twice as many deaths among women of childbearing age as does maternal mortality
(World Bank, 1993; Murray and Lopez, 1996). Tuberculosis, along with several other
diseases, imposes a heavy cost in morbidity and mortality. A study by the World Bank,
WHO, and Harvard University showed that tuberculosis caused an annual loss of 8.7 million
healthy years of life among women 15 to 44 years of age, while 3.6 million healthy years
were lost from HIV and two million from malaria (WHO, 1998b).
Women may differ from men in infection and progression to disease, and in the
probability of diagnosis, registration for treatment or treatment results.
Men have a higher prevalence of tuberculosis infection than women, but the
prevalence of disease is the same or greater in women (Murray, 1991; Hudelson, 1996;
Holmes, Hausler, and Nunn, 1998). This suggests that women are more likely to progress
from infection to disease than are men, at least at some ages. Holmes et al. (1998)
surveyed prospective studies of tuberculosis-infected individuals in developing countries
and found that females in their reproductive years had a higher rate of progression to
disease, a higher case-fatality rate (deaths due to tuberculosis divided by the population
with tuberculosis), and higher mortality rates (deaths due to tuberculosis per 100 000
population). There is as yet no consensus about why these gender differences exist.
Some possible explanations are discussed in the next section.

Differences in diagnosis and treatment

Why are women more likely to become ill and to die? In some countries, women
who suffer from tuberculosis are less likely to be detected than are men. They are also
less likely to receive treatment. In eastern Nepal, Cassels et al. (1982) found that the ratio
of men to women suffering from tuberculosis disclosed with self-referral of patients to
health services was twice that of active identification of cases by mobile teams. That is,
men are more likely to seek treatment than are women. Another study carried out in
Nepal found that the delay before receiving treatment for women was twice as long as
that for men (Smith, 1994).
Explanations for these differences may include the high cost of seeking treatment
for women because of their household duties and childcare responsibilities, and/or the fact
that in certain societies women do not make the decision to seek health care no matter
how ill they are. Their husband or other relative makes such decisions (Okojie, 1990).
The effect of costs on seeking care is illustrated in another study in Nepal, which
found that women were more likely than men to attend mobile clinics, whereas men
were more likely to attend more distant—and thus more costly—health posts and clinics
(Harper et al. 1993). However, Rieder et al. (1997) point out that failure to examine women
is not the only possible reason for their lower presentation among cases (any patient
presenting for diagnostic examination with at least one positive result for acid-fast bacilli).
A study of 42 laboratories in Benin, Malawi, Nicaragua, and Senegal revealed that women
were being examined relatively frequently, but that cases were found less frequently
among women, perhaps because of a higher prevalence of non-specific respiratory
symptoms among them.
Besides contributing to the higher mortality from tuberculosis among young to
early to middle-aged women, differential access to treatment has implications for the
success of TB control programmes. Since successful control of the disease depends upon
high levels of detection and successful treatment, lower access by women to treatment
undermines successful tuberculosis control.
 Impacts on the household
Relatively little attention has been given to the costs of TB to patients and their
households. Studies that have addressed these costs have focussed on lost earnings from
illness or death. This approach underestimates the cost of TB, particularly for women.
When a woman becomes sick or dies from tuberculosis, the household suffers from
the loss of her earnings outside the household, but suffers additional losses due to the
reduction of activities that the woman routinely performs in the household. The value
of activities such as cooking, cleaning and childcare are rarely considered in studies that
attempt to estimate the costs of tuberculosis, even though they may be considerable.
In Tamil Nadu, India, for example, female TB patients reported a 50% reduction in
household work and only one-third reported that they were able to attend adequately
to the needs of their children (Ramachandran et al. 1997). Further insight into the impact
of illness on the household activities of women can be obtained from a study of another
debilitating disease, Guinea Worm. In an ethnographic study of Guinea Worm in Nigeria,
Watts, Brieger, and Yacoob (1989) found significant impacts on the immediate household,
which then radiated out to affect the extended family. They found that when women with
infants are struck down by Guinea Worm, their self-care is likely to be neglected first,
then their domestic duties and income generating activities are reduced. Women do their
utmost to continue childcare. There were cases, however, where women were unable to
care for their children, and some cases of acute malnutrition and underweight children
were observed. Some women were unable to breastfeed; in these cases the health of
the baby suffered severely.
When attempting to determine the impacts of disease on the household, care
must be taken to estimate a value for any reduction in work done within the household,
as well as for work that is done outside the household (whether paid or unpaid work).
Ignoring the economic contribution of women to the household leads to an underestimation
of the costs associated with disease.

Tuberculosis and the poor

Tuberculosis and poverty are closely linked. Given the often crowded conditions
in which poor populations live, they are more likely to contract tuberculosis. Those who
contract tuberculosis are also more likely to fall into poverty, given the economic
consequences of the disease. Both the probability of becoming infected with the disease,
and the probability of developing clinical disease are associated with malnutrition, crowding,
poor air circulation and sanitation—factors also associated with poverty. A vicious cycle
is established: poor people are malnourished and live in crowded, unhygienic conditions,
then tuberculosis flourishes. The poor receive inadequate health care in which tuberculosis
is not diagnosed rapidly. Treatment, if received at all, is often inconsistent or partial.
Resultant ill-health and death worsens poverty.
The close link between poverty and tuberculosis has led many to think that TB
is exclusively a disease of the poor, and of little economic consequence. Several recent
studies of the socioeconomic characteristics of TB patients throw doubt on this view.
Researchers in India found tuberculosis to be present among all strata of society,
although prevalence was highest among the poor (Nayyar et al. 1989 cited in Pathania,
Almeida, and Kochi, 1997; Ramachandran et al. 1997; Devi et al. n.d.; and Chakraborty,
1996). Table 1 shows that although the poor are more likely to suffer from the disease,
a significant percentage of TB patients are educated and earn good incomes*. For example,
the prevalence of TB among those whose income was less than US$ 7 per month was
about twice that of those with a monthly income greater than US$ 20 (Nayyar et al. 1989
cited in Pathania et al. 1997). In the mid-1950s in Calcutta, the poorest areas had prevalence
rates of over 50 per 1 000 compared to a rate of 2.48 per 1 000 in comparatively affluent
areas (Chakraborty, 1997). In urban areas, prevalence among those with no schooling was
four times that of tertiary graduates (Nayyar et al. 1989 cited in Pathania et al. 1997).
A study of 673 adult tuberculosis patients at 16 randomly selected government
health facilities in Thailand (district hospitals, provincial hospitals, and referral centres)
found that the majority of patients were not poor: 23% were poor, 29% had income below
the national average but above poverty, and 48% had income above the average. While
these figures are likely biased by the lower probability of the poor to seek treatment, they
do indicate that tuberculosis is not necessarily confined to the lowest economic strata.
In addition to being more likely to contract tuberculosis, the poor are less likely
to seek and receive care when ill-exacerbating the negative impacts of the disease.
Among people in Tamil Nadu suffering from chest symptoms, only 64% of illiterates
sought treatment, compared to 80% of high school graduates. The differences were
similar for those with incomes under US$ 14 per month and greater than US$ 57 per
month (Devi et al. n.d.). The poor are also two to three times more likely than other
income groups to self-medicate (Pathania, Almeida, and Kochi, 1997, p. 14). Self-
medication is likely to increase the chances of drug-resistant strains of TB emerging,
further increasing the negative impacts on the poor and the risks to others in society.
As The Economist recently stated: “TB is a relentless leveller, an equal opportunity
killer, hard-working and persistent … going about its deadly business with cool disregard
for IQ, sex, class, race, occupation or even geographical boundaries” (1999, p. 11).
Although overstated, the main point of The Economist discussion of TB is valid: those
who callously disregard TB as a disease that afflicts only the “least valuable” in society,
and so does not warrant the investment of scarce government funds, are seriously in error.
Tuberculosis strikes the working age population and strikes the most productive as well as
the less productive. One could further argue that not only measured economic costs and
benefits should be considered when deciding government policy.

* These were studies of the prevalence of TB in different socioeconomic groups. The Nayyar et al. study was a
population-based survey of 200 000 urban residents and 490 000 rural residents in Wardha District, Maharashtra.
The Ramachandran et al. study was based on focus group interviews with 304 newly detected sputum positive
pulmonary TB patients. They seem to have been drawn from Government clinics and hospitals and so may over-
represent the poor. The Devi et al. study was a community-based survey of 4 624 rural and urban individuals
in Tamil Nadu that yielded 689 “chest symtomatics” of whom 649 were interviewed.
 Section 2

The economic and social burden of tuberculosis

The impact of tuberculosis is most often measured as the direct costs of treatment to
the health service, that is, the costs of medicines, personnel, and facilities used. However, the
economic impacts are considerably more far-reaching. Often patients seek costly treatment from
traditional healers or the private sector before an accurate diagnosis is made. Only then, they
may shift to the public sector. The costs to patients and their families that can be quantified are
principally in the form of lost earnings from loss of work due to illness or death. Additional costs
come from food required while in hospital and the costs of travel to hospital or clinic for care.
In addition to these direct treatment and non-treatment costs, tuberculosis imposes intangible
costs in the form of pain, suffering, grief and discrimination. Table 2 details some of the
quantifiable costs of TB treatment for rural Uganda and South Africa. The total cost of treatment
varies considerably, as does the distribution between the health service and the patient.

I n order to fully appreciate the impacts of a disease such as tuberculosis on

the well-being of a nation, the costs borne by individuals, households and communities,
and not just the costs to the public health services must be examined. Similarly, to fully
evaluate the costs and benefits of alternative strategies to combat and treat disease,
one must look at the costs and benefits of each alternative approach to individuals,
households, and communities, and not just the costs to the public health service. A goal
of public policy is to maximize social well-being, subject to budget constraints. This is
sometimes forgotten in favour of minimizing the expenditures of public agencies such
as departments of health.

Treatment costs borne by patients and their family

The direct treatment costs of tuberculosis are often underestimated because only
the costs to the health system are measured. However, these costs may be the smallest
component of treatment costs, as is the case in rural Uganda reported in Table 2.
In many countries, individuals with tuberculosis seek assistance initially from
traditional healers or private sector health care providers, either because of a preference
for private sector treatment or a lack of awareness about available public TB services.
In India, where private sector care is widespread, there is a perception that private providers
are more effective, convenient, and sympathetic than public-providers (Pathania et al.
1997, p.3). However, lack of information may also play a role: only 15% of households in
one rural study were aware that the government clinic offered free TB treatment (Uplekar,
Juvekar, and Morankar, 1996). Another possibility is that “free” care at government hospitals
is not free. Thailand has an official policy of “free care” for tuberculosis, yet Karnolratanakul
et al. (1999) found that more than 20% of patients were charged for X-rays, more than 10%
for sputum examinations, and almost 10% for drugs.
 A study of tuberculosis patients in Malawi found that almost 40% had consulted
and received four weeks’ care from a traditional healer before seeking regular medical care
(Brouwer et al. 1998). Cash payments for this care were relatively low. However, in other
settings the cost of treatment before seeking help from conventional TB providers was
substantial. In four northern districts of Bangladesh, Croft and Croft (1998) found that
patients spent on average US$ 130 before reaching the TB clinic. This is almost 20% of the
annual income of an average Bangladeshi household. In India, where use of private providers
is widespread, the direct costs of diagnosis and successful treatment average US$ 100 to
US$ 150, more than half of the annual income of a daily wage labourer (WHO, 1997).
Part of the reason for the high direct treatment costs for patients and their families
is that it often takes more than a month before a diagnosis is made. Only about one-half
of patients are diagnosed at the first source of help and are thus forced to “shop around”
for a diagnosis. For example, in a study of adult tuberculosis patients in Thailand,
Karnolratanakul et al. (1999) found that the delay between the onset of illness and
diagnosis of tuberculosis was 61–76 days, even though one-third of patients had sought
care during that time at government hospitals. Other reasons for the high costs of treatment
include the extensive use of X-rays as a diagnostic tool, prescription of expensive nutritional
supplements, and collusion among providers (Pathania et al. 1997).
The most significant cost of shopping around for a diagnosis is that the delay in
treatment contributes to high early mortality and increased transmission of infection.
This results in many patients presenting at government clinics with more severe disease
that is much more costly to treat (Pathania et al. 1997, p.17).

Non-treatment costs borne by patients

and their families
Earnings lost from illness and death from TB are commonly much greater than
the direct costs of treatment. For example, 70% of the costs to patients in rural Uganda
are from lost work time (Table 2). For individuals with TB, costs arise from a loss of work
or a reduction in productivity due to illness. Further losses result from earlier mortality
due to TB.
There are two sources of information on the non-treatment costs that are borne
by patients and their families: surveys of patients that estimate their treatment costs
and costs of illness; and simulation studies that estimate both the costs of illness and
of premature death (Tables 3 and 4 respectively). The estimates of indirect costs from
these sources are best interpreted as losses of potential income or potential income
gains if the person had not been sick, and usually exceed actual lost income.

Estimates of costs from surveys of TB patients

Croft and Croft (1998) noted that a local Bangladeshi synonym for tuberculosis
is “Rajer rog”—King’s disease—since it is a disease that only Kings can afford to suffer.
Their cost estimates bear this out. By the time the patients presented for treatment at the
TB clinic, they had already incurred costs of US$ 130 for private sector treatment and lost
14 months of work time (range five days to five years) resulting in US$ 115 in lost income
(range US$ 0–500). These losses are equal to one-third of annual household income.
Even when they enter the “free” government sector, patients often have to make informal
payments when receiving treatment.
A study in Uganda found that 80% of wage earners had stopped work because
of their disease and 95% of subsistence farmers reported that production had decreased
due to their reduced capacity to work (Saunderson, 1995). The average time lost from
normal activities was 9.5 months (range week to three years); the average income lost
from inability to work was US$ 161 or 89% of GDP per capita. In South Africa, lost earnings
were even higher: US$ 272 or 16% of GDP per capita (Floyd et al. 1997; Wilkinson, 1997).
In Thailand, income reductions were much smaller: average income fell by 5% for poor
households, 2.3% for households with income between poverty and the national average,
and 3.3% for households with income above average (Karnolratanakul et al. (1999).
A study of rural and urban patients in Tamil Nadu found substantial lost earnings—about
15% of annual household income (Ramachandran et al. 1997). The higher estimates of lost
work time and lost income in the Uganda and Bangladesh studies probably reflect the
influence of outliers on the means (lost work time of three and five years respectively).
More reasonable estimates of lost work time and lost income from TB morbidity are three to
four months— about 20% of household income. The loss of potential income is determined
in large part by the initial resources of the household and the coping mechanisms employed
to deal with the impacts of the illness. These will be discussed below.
Lost earnings exceeded direct medical costs by a factor of two in Tamil Nadu
and almost three in Uganda and South Africa. These estimates, based on lost earnings,
underestimate the costs to households since they ignore the value of lost household
production, adverse impacts on the health and education of family members, costs of sub-
optimal land use, the value of lost leisure, and the pain and suffering associated with TB.
Transport costs may appear to be a small part of the total cost of treatment, but
this is not necessarily so. A survey of patients in Bombay showed that they could spend
nearly 10% of income on travelling to the clinic twice a month to collect drugs (Chakraborty,
Rangan, and Uplekar, 1995). Van der Werf et al. (1993) found transport costs to be larger
than direct costs in rural Ghana. Clearly, such expenditures could lead to defaulting from
treatment, particularly among the poor.

Costs from estimates of simulation studies

Simulation studies base their estimates of costs on data from survey studies (Table
3), on programme records and on the judgement of experts. Estimates from three such
studies are shown in Table 4. The estimates for Indonesia and Thailand are based on data
for particular programmes. The estimates for Bangladesh are based on a model that tries
to mimic the impact of TB on the household of a riskshaw-puller.
In both Thailand and Indonesia, individuals who receive treatment are estimated to
lose about two months’ work from illness; those who do not receive treatment lose about
twelve months’ work, similar to the findings from the survey studies. The most dramatic
losses of productive life come not from illness but from death due to TB. In Indonesia,
individuals with TB but not infected with HIV lose 13 years of life; in Thailand individuals
with TB but not infected with HIV lose 15 years of life. Those co-infected with HIV die two
years earlier (Sawert, 1997; Sawert et al. 1997).
 Data for average earnings are multiplied by the time lost due to morbidity and
mortality to arrive at an estimate of the enormous potential monetary cost of TB. The
losses due to treated illness are similar to those shown in Table 3. They are about 15%
of GDP per capita. For the 50% of cases that go untreated, the potential costs are
significantly higher because of the longer time missed from work.
The economic cost of premature death from tuberculosis overshadows the morbidity
costs. For individuals not infected with HIV, death as a result of tuberculosis results in a
loss of potential income to the individual’s family of 13 times GDP per capita in Indonesia
and 10 times GDP per capita in Thailand. In the case of the rickshaw-puller, tuberculosis
also imposes significant costs: over a period of two years, the costs of illness and chance
of premature death result in a loss of income of US$ 329, or more than three times the
average per capita yearly income of the household. These estimates are larger than the
actual income lost to families and households, principally because households use a
number of coping mechanisms to reduce the economic impact of illness and death
(discussed in Section 3). However, the estimates do provide an indication of the
magnitude of the economic shock to households from illness or death due to TB.
The impacts of diseases such as tuberculosis are not confined to the infected
person. The illness or death of an adult can increase the morbidity or mortality of their
spouse; the illness or death of a mother or father can result in illness or death for their
children. Martikainen and Valkonen (1998, p. 530) concluded: “The accumulated evidence
[from research studies] suggests that excess mortality among bereaved persons persists
after controls [for other influences on mortality] are introduced.” They found that the
burden of mortality after the death of a spouse was heaviest among the poor. Parental
mortality is an important determinant of infant mortality. In Matlab, Bangladesh, a father’s
death was associated with an increase in male and female child mortality of six per 1000;
a mother’s death was associated with increases in male child mortality of 50 per 1000
and female child mortality of 144 per 1000 (Over et al. 1992). Nobel prize-winner Robert
Fogel has recently shown that the impacts of childhood malnutrition are lifelong (1994).
He established that stunting in childhood could lead to increased morbidity and mortality
at middle and later ages. Such intergenerational effects are ignored when the costs of TB
are evaluated and so the estimates of the cost of the disease are too low.

Psychological and social costs

Additional costs, both monetary and psychological, can occur due to discrimination
against those infected with TB and members of their households. Family and friends may
reject TB patients, they may receive less social support during treatment, or they may lose
their jobs. A survey of patient’s concerns in Bombay found men to be worried about the
impact of TB on their capacity to work, while women were concerned about the impact on
their marriage status and harassment by in-laws (Liefooghe et al. n.d.). Such concerns are
well founded. In a survey carried out in Tamil Nadu, Ramachandran et al. (1997) found that
while the vast majority of men reported that their family had accepted their disease, a
sizeable minority of women—especially rural women—suffered some adverse reactions or
outright rejection (see Table 5). In addition, 69% of rural women said that they could not
discuss the disease with their neighbours. In a recent study carried out in India, Rajeswari
et al. (1999) reported that 15% of rural and urban female patients faced rejection by their
families. The association between AIDS and TB increases the likelihood of rejection and
social isolation of TB patients (Saunderson, 1995).
The costs of discrimination are both psychological and economic. Discrimination,
whether experienced or expected, has been found to be associated with increased anxiety
and depression and lower life satisfaction, as well as with higher unemployment and
lower income (Rosenfield, 1997; Markowitz, 1998). In some developing countries,
discrimination against TB sufferers has taken particularly damaging forms, such as
divorce or lowered prospects of marriage. A survey carried out in West Bengal found that
almost 80% of respondents would not negotiate the marriage of their son or daughter
to an ex-TB patient (Geetakrishnan, et al. 1988). Studies in India and Pakistan found
that married women with TB were more likely to be divorced than other women and that
unmarried girls with TB would find it difficult or impossible to get married (Khanna, et al.
1977; Liofooghe et al. n.d.). Such discrimination represents significant costs because the
economic prospects for divorced or unmarriageable women in many societies are bleak.
Discrimination against those suffering from TB also has epidemiological consequences.
Social support for the patient has been found to be critical to adherence to treatment
(Pathania et al. 1997); without adherence to treatment, the likelihood of a cure is lower,
relapse and death are more likely, and further transmission of the disease occurs.
Section 3

Household coping strategies

Households face substantial immediate costs from tuberculosis: costs of diagnosis, costs of
treatment, and lost earnings and household work. Families cope with the burden of these costs in
a number of ways. They may:
• divert resources from other forms of health care, increasing the vulnerability of family members to other
diseases;
• reduce food consumption, which may increase the risk that others in the household will contract TB
or some other disease;
• take in extra household members or move into a relative’s house, increasing crowding and the risk
of further transmission;
• withdraw children from school to save on school fees, thus jeopardizing the future prospects of the child;
• increase the hours of work of household members;
• borrow or sell productive assets;
• plant crops that reduce the amount of labour required;
• expel the individual suffering from TB from the household;
• pool work and resources with others.

M any of these coping strategies reduce the future opportunities of household

members, in particular children. Furthermore, the strategies are often not abandoned
once the patient’s health improves because, in certain cases, the patient never fully
recovers his or her past productivity and the household often incurs long-term debt.
In studies of the impacts of AIDS, such coping strategies have been observed as early
as 18 months before a death and up to 30 months after a death.
The number of coping strategies open to a family depends to some extent on the
social organization of the society in which it is situated. Where extended families share
tasks and obligations and, perhaps, money, the impact of illness is likely to be more muted.
The ability of the extended family to absorb the costs of illness of a member also depends
upon the size of the group, the wealth of the group, and the number of members who are ill.
Experience with the HIV/AIDS epidemic has shown that even with extended families, the
ability of social systems to cope with illness, although impressive, is limited. In addition,
modernization has made many traditional systems less effective. For example, in the Kyela
district of Mbeya in Tanzania, the wider community no longer takes responsibility for providing
care for the seriously ill. The responsibility now rests with the immediate family (WHO, 1997)
The poor have a more limited set of coping strategies because often the only asset
they have to sell is their physical labour. This is why the illness of the head of the household
is so devastating. The poor do not have a buffer to see them through a period of reduced
income and they have little access to borrowing. The children of the poor are less likely to be
in school but, because school fees are a significant portion of household income, are more
likely to be withdrawn from school in the event of illness of a parent. A poor household is also
likely to be a member of a poor extended family, limiting the wider resources it can draw upon.
Studies of tuberculosis have observed a number of coping strategies in households
afflicted by tuberculosis. In Bangladesh, Pryer (1989) found that the initial response of
households to a large medical expense is the sale of assets, followed by taking out
consumption loans. These loans have very high interest and short repayment periods, which
make the economic recovery of the household difficult. In a study in India, 20% of rural
patients and 40% of urban patients went into debt as a result of expenses due to TB
(Upelkar et al. 1996). In Tamil Nadu, the percentage of households who became indebted
was much higher: 75% of urban households and 67% of rural households (Ramachandran
et al. 1997). The average amount borrowed was US$ 59, or 12% of annual household income.
These findings are consistent with a recent Indian study that found average household
debts of 2079 rupees (US$ 59) due to tuberculosis. This is equal to 35% of the total house-
hold cost of tuberculosis (Rajeswari et al. 1999). In northern Bangladesh, Croft and Croft
(1998) found that almost 40% of households sold assets, either livestock or land, and 14%
took out a loan. In a study in Thailand, about 20% of patients received transfer payments from
relatives, 20% used savings, about 10% took out loans, and a significant minority sold
assets. Fully 16% of the poor sold assets, mostly land, and just over 7% of the non-poor
sold assets (Karnolratanakul et al. 1999).
Another coping mechanism that reduces short-term costs but has potentially adverse
long-term consequences is the withdrawal of children from school. In the Uganda study
(Saunderson, 1995), five children from 32 families studied were withdrawn from school
because their families could not afford the school fees. In Tamil Nadu, 11% of children
whose parents became ill with TB were withdrawn from school and 8% entered employment
(Ramachandran et al. 1997). In this study, children were withdrawn from school more
often in urban areas and if the infected parent was the father (see Table 6). In a recent
study in India, 11% of children were withdrawn from school and an additional 8% took up
employment to help support their family (Rajeswari et al. 1999). This study also noted
that the care-giving activities of female patients decreased significantly as a result of
their illness. As mentioned above, this cost to households is rarely recognized in studies
of the cost of tuberculosis. In the study of Guinea Worm discussed above, school-age
daughters who were their mother’s main helper were withdrawn from school to replace
their mother’s labour in the household (Watts et al. 1989). Withdrawing children from
school lowers the child’s economic prospects and, if widespread, can have social
consequences such as higher fertility and lower production.
Many of these strategies are illustrated in a simulation study of a Chittagong,
Bangladesh, rickshaw-puller’s household (Carrin, Gray, and Almeida, 1998). Their model
combines data and expert opinion to model a typical rickshaw puller’s household and
the impacts of TB upon it. Carrin et al. showed that the impact of TB when treated
“conventionally” and with two relapses within a two-year period is a 55% reduction in
assets, a 20% reduction in food consumption which lowers work capacity and income,
and loans that are ten times the pre-TB level of household income*. However, family
members increased their work hours and increased their earnings to partially offset the lost
earnings of the household head. These coping mechanisms have long-term implications for
the household. The sale of assets makes the household more vulnerable to future shocks,
reduced food consumption increases the susceptibility to disease and, if severe, may
stunt the growth and development of children, with long-term implications for morbidity
and mortality; indebtedness can also impede the economic recovery of the household.
It must be remembered that even when coping strategies mitigate the impact of TB
on an individual, the social and economic costs are borne by the family and community
as a whole. Even if the income of the family does not fall because coping mechanisms
have been used to compensate, a member becoming ill with tuberculosis does reduce
the welfare of the family.

* Since one could argue about the precise values assigned to variables and parameters, Carrin et al. attach probability
distributions to a number of variables and parameters
Section 4

Benefits and costs of new strategies

of dealing with tuberculosis
Is there a rationale for government intervention in tuberculosis control? A rationale for
public intervention is the presence of what economists call market failures. Examples include:
a widespread lack of health insurance; externalities, such that others benefit if an individual
remains free of tuberculosis or is effectively treated if infected, and others bear some of the
costs if the individual does not (in such circumstances there is insufficient incentive for the
individual to invest optimally in healthcare); and providers, who generally have more information
about tuberculosis than do patients, and thus can exploit patients, leading to an inefficient
use of resources. All of these conditions are met in the case of tuberculosis, and government
intervention is justified.

G iven that intervention is justified, what form should this intervention take?
To answer this question, the public and private costs and benefits of alternative
interventions should be considered. To date, emphasis has often been given to the costs
and benefits to the public health system of different strategies to combat and treat
tuberculosis. The costs and benefits to individuals have either been ignored or only
partially considered. However, incorrect decisions can be made unless all costs and
benefits are considered.

Intervention strategies for TB control

Chemotherapy can have a substantial impact on the incidence of tuberculosis.
There is, however, debate on what form intervention should take (Sawert, 1996; Zwarenstein
et al. 1998 and comments in Lancet, 1999). For instance, should short-course treatment
replace conventional treatment? Should treatment be ambulatory or should it involve
hospitalization? How critical is direct observation of treatment rather than self-supervision?
Should fees be charged and, if so, how much?
In many countries, conventional treatment involves an initial, often lengthy, period
of hospitalization followed by a period of outpatient follow-up. Drop-out rates are often
high and cure rates are often low. In sharp contrast is Directly Observed Treatment,
Short-course (DOTS), which the World Health Organization has determined to be the best
approach for the delivery of TB chemotherapy. DOTS is a form of chemotherapy that lasts
for six to eight months and combines five elements: political commitment, microscopy
services, drug supplies, monitoring systems, and direct observation of treatment.
This strategy produces cure rates of up to 95% even in the poorest developing countries,
whereas non-DOTS treatments achieve a cure rate of only about 40% (WHO, 1998a).
In addition, it is estimated that about 15% of those undergoing non-DOTS treatment die
compared to as few as 5% of those on DOTS (Carrin, Gray, and Almeida, 1998, p. 10).
DOTS has been shown to lead to an eventual decrease in cases (after an initial increase
through improved case-finding) and a decrease in overall provider costs, especially if
ambulatory strategies are used (Styblo, 1991; Barnum, 1986).
 Short-course therapy is highly cost-effective: in a low-income country setting, it costs
between US$ 1 and US$ 4 per discounted year of life saved (Jamison and Mosley, 1991).
Indeed, short-course chemotherapy is the most cost-effective of all health interventions
available (World Bank, 1993). Dye et al. (1998) have shown that the impact of DOTS is
greatest where detection occurs quickly where DOTS replaces a poor programme under
which cure rates are low and the incidence rate has been falling slowly or not at all, and
where it is introduced in a young population.
Despite the cost-effectiveness of DOTS, only 16% of people with TB receive
treatment consistent with DOTS (calculated from data in WHO, 1998c and Ravaglione
et al. 1995). Why is this the case? Part of the reason may be that despite efforts by WHO
to inform policy-makers of the benefits of DOTS, many are still unaware of the evidence
on the effectiveness of alternative approaches to TB. Perhaps policy-makers are aware,
but those that implement policies are unaware of DOTS. Resistance to change and
distrust of new approaches may also inhibit the adoption of DOTS. Certainly, some
decision-makers believe that their existing approach is “better” than DOTS. It is my
contention that the failure to include the costs and benefits to individuals and house-
holds of alternative approaches to TB contributes to the belief that current strategies
are “better” than DOTS.
We now turn to a consideration of the evidence on the questions posed above
concerning alternative strategies to deal with tuberculosis.

Simulation studies of alternative TB treatment regimes

Short-course treatment and conventional treatments
Several recent simulation studies have evaluated the costs and benefits of short-
course chemotherapy versus conventional treatment. Sawert et al. (1997) investigated
replacing the current TB programme in Thailand (a detection rate of 50% and a treatment
completion of 60%) with a new programme that would detect 70% of cases and 85%
would successfully complete treatment. They found that the new programme would
initially have higher costs than the current programme because of a greater number of
cases detected, but that the benefits become greater the longer the evaluation period
considered. The major impact of the new programme on costs is in lower indirect costs to
patients and their families. Over a period of 20 years, a more effective programme would
lower these costs by US$ 2.5 billion (range US$ 65 million to US$ 7.0 billion) or 32% of
the projected indirect costs under the current programme. This is an underestimate of
indirect cost-savings because it ignores all savings other than lost earnings. Sawert et al.
argue that the situation in Thailand is similar to that in many other developing countries,
so that their findings are likely to apply to those countries. This assumption is supported
by evidence from rural Uganda (Table 7). Although DOTS may increase costs to the health
service (depending upon the programme it is replacing), it dramatically reduces costs to
the patients and is a cheaper treatment regime overall.
In 1994, Indonesia decided to replace its current TB programme with the DOTS
strategy. An evaluation of a DOTS trial project in East Java, and comparison to other areas
where conventional approaches were in use, was carried out. It was estimated that the
adoption of DOTS would result in initial cost increases because of equipment and training
costs and higher case detection. However, the DOTS strategy would save 12 million
disability—adjusted years of life over 20 years at a cost of US$ 10 per year initially and
US$ 5 per year later in the programme. A simulation showed that TB would cost Indonesia
US$ 13 billion over 20 years under the current policy. The introduction of DOTS would
halve these losses. It was estimated that every dollar invested in the DOTS programme
would lead to a US$ 55 return to the country.
Similar results were found in South Africa, where TB incidence is one of the highest
in the world. Wilkinson, Floyd, and Gilks (1997) carried out a study of direct and indirect
costs in Hlabisa, northern KwaZulu-Natal, where the number of TB cases admitted to
hospital increased three-fold between 1991 and 1996, partly as a result of HIV. They
compared DOTS with the standard treatment of four months in hospital followed by four
months of outpatient care. The cost of DOTS to both the patient and to the health system
was one-third that of the standard treatment. The long hospitalization under the standard
treatment not only increases costs substantially but it also reduces compliance and thus
the cure rate. Under standard treatment, only 17% of cases were cured compared to 81%
under DOTS (Wilkinson et al. 1997, p. 453). DOTS is almost 30 times as cost-effective as
the treatment commonly used in sub-Saharan Africa and was also three times more cost-
effective than treatment with a two-month hospital stay or with treatment in a sanatorium.
It was estimated that even the most expensive drug treatment was equivalent in cost to
just 1.3 days in hospital. In Malawi, Sawert (1996) estimated that a new ambulatory
treatment for smear-negative patients would reduce direct costs by 45% and for smear-
positive patients by 70%.
Further evidence of the cost-effectiveness of treatment regimes for TB comes from
studies in Eastern Europe and the former Soviet Union. These countries have suffered
an extraordinary demographic crisis, most notably dramatic falls in life expectancy.
In the Russian Federation, life expectancy fell by six years from 1989 to 1995 (Becker
and Bloom, 1998) and reported cases of tuberculosis rose dramatically. The probable
causes are the sharp drop in real incomes, increases of psychological stress, and the
decline in medical services accompanying the economic transition.
Economic problems are exacerbated by the expensive treatment regimes that are
standard in eastern Europe and the Russian Federation, particularly the extensive use
of long stays in hospital. Meerding (1997) compared a WHO strategy in Armenia that
consisted of 2.2 months hospitalization during the intensive phase and four to five months
outpatient follow-up with the standard treatment of four to six months hospitalization
followed by 12 to 14 months of sanatorium and ambulatory treatment. The costs per
cured patient under the WHO regime were US$ 176 and US$ 280 under the standard
approach. The inclusion of indirect patient costs would have increased the cost differences
even more. Migliori et al. (n.d.) carried out a similar comparison of alternative treatment
regimes in the Russian Federation and reached similar conclusions: costs per cure under
the new treatment were US$ 1 626 and under the old US$ 6 293. The WHO regime with
short-course chemotherapy reduced costs by 62%; a reduction in the percentage of
patients hospitalized from 100% to 10% reduced costs a further 13%.

Ambulatory treatment versus hospitalization

Several studies have focused on evaluating the gains to be made in adopting DOTS
that accrue from replacing hospitalization with ambulatory care. Saunderson (1995)
compared the current TB programme in Uganda, based on an initial period of hospitalization
followed by a period of outpatient follow-up, with a DOTS programme which replaced the
period of hospitalization with ambulatory care and expanded education and supervision.
Saunderson found that cost per cure under the conventional treatment, which is widely
used in sub-Saharan Africa, was US$ 646, whereas the cost per cure for the DOTS regime
was US$ 280. Almost all of the difference is explained by lower costs to the patient,
primarily in the form of lower lost earnings because hospitalization is avoided.
DeJonghe et al. (1992) evaluated the direct cost-effectiveness of standard and
short-course treatments for smear-positive patients in three sub-Saharan countries:
Malawi, Mozambique, and Tanzania. They estimated that short-course treatment with
hospitalization is about 23% cheaper for the health service than is standard treatment.
Moving to an ambulatory short-course treatment would reduce the costs of short-course
treatment by 35% for Malawi, 65% for Mozambique, and 50% for Tanzania (Table 8).
In a survey of cost-effectiveness studies published up to 1995, Fryatt (1997) found
that ambulatory care is particularly cost-effective once the indirect costs to patients are
considered. Work by Pryer (1989) further suggests that the relative gains may be greater
for the poor since they lose more work time from TB than do those in better-off house-
holds. Obviously, a fully ambulatory treatment would reduce costs as long as the cure
rate did not decline significantly. In the three sub-Saharan countries studied, Murray
(1991) estimated that if hospitalization increases the cure rate by 15 percent, the cost
per case cured is similar to any form of chemotherapy. Ambulatory programmes obviously
require well-designed and well-implemented community-based care. Supervised
chemotherapy must be available cheaply and conveniently to ensure compliance.

Supervised versus unsupervised treatment

The findings from a recent study in South Africa question whether direct observation
rather than self-supervision is essential for effective treatment. Zwarenstein et al. (1998)
carried out a randomized controlled trial in two communities to test the effectiveness
of direct observation. They found no statistically significant difference in the success of
direct observation versus self-supervision. A number of procedural and methodological
criticisms have been raised concerning the study, such as the small sample sizes and
low cure rates in both arms of the study (see Lancet, 1999 for criticisms and a response).
Although the shortcomings of the study arguably undermine its conclusions, it does
raise the very important question of whether the components of DOTS are additive or
multiplicative; that is, is the success of DOTS greater than the impact of its component
parts? If not, then countries may implement parts of DOTS. If so, there is a very strong
argument for adopting the full DOTS approach.

Fees for treatment

The Carrin et al. (1998) study of the rickshaw puller’s household discussed above
illustrates the advantages of DOTS over conventional treatment in reducing the necessity
for compensatory household coping strategies, but the result is affected by whether
or not fees for treatment are charged. Carrin et al. show that if DOTS treatment is fully
subsidized, then household assets are not sold, loans and reductions in food are greatly
reduced, and the impacts on children’s education are smaller. When fees are imposed,
the indirect impacts under DOTS are smaller and the advantage of the DOTS strategy is
a higher probability of treatment success and reduced probability of relapse in addition
to lower treatment costs.
In pilot studies at mission hospitals in the north-west Province of Cameroon,
Proctor and his colleagues found that 10% to 30% of patients were discharged without
medication because of their inability to pay. This finding is significant because mission
hospitals and clinics are important providers of health care in many African countries.
Fully 50% to 60% of Cameroon’s health care is provided by church missions, 58% of rural
hospital beds in Zimbabwe are in church facilities, and it is estimated that about 40% of
health care in Kenya is provided by churches. If this finding holds widely, then drives for
cost recovery could undermine efforts to contain tuberculosis. Clearly, more information is
required on how the behaviour of patients is affected by the imposition of fees for treatment.
The research discussed indicates the following:
1) In certain circumstances, initial costs to the health service of short-course treatment
may be higher than those of existing treatments because of a rise
in detection rates. However, within 20 years significant cost-savings can occur
as detection and cure rates rise.
2) Short-course treatment results in significantly lower costs to individuals
and households.
3) Ambulatory care has much lower costs to individuals and households than
does hospitalization.
4) Charging fees may decrease the gains from DOTS because of reductions
in seeking treatment and drop-outs from treatment.

Obstacles to the adoption of DOTS

According to WHO, DOTS is most effective where tuberculosis incidence is high,
the population is large, existing treatment strategies are ineffective, and DOTS coverage
is currently low—that is, in many countries in the developing world (WHO, 1997, p. 159).
The cost-effectiveness analyses discussed clearly show that adoption of the DOTS
strategy would improve health and lower costs. Why then are only slightly more than
10% of TB patients receiving DOTS?
There are several obstacles to the adoption of DOTS. Foremost among them is
mobilizing commitment to DOTS among decision-makers and health professionals. Once
such commitment exists, other obstacles, such as adequate funding to ensure a steady
supply of drugs, diagnostic equipment, trained staff, and recording and reporting systems,
can be faced. In many countries, a lack of commitment among health professionals may
reflect a lack of knowledge about TB. For example, a study in Bombay found that private
practitioners who treat TB use a variety of regimes that include 80 different drugs, even
though only four of these drugs are effective against tuberculosis when used in the
prescribed combination (Yesudikan, 1994). Although some of these regimes are efficient,
many are not. Another study of specialist physicians in a large university teaching
hospital in Karachi, Pakistan, found that few physicians follow the WHO guidelines for
tuberculosis treatment. Particularly alarming, given the emergence of drug-resistant TB
due to incomplete of inappropriate drug therapy, is the fact that only 30% of patients
treated completed treatment as prescribed (Arif et al. 1997). Karnolratanakul et al. (1999)
also noted insufficient awareness of tuberculosis and lack of diagnostic skills on the part
of government health staff in Thailand. They concluded that intensive training activities
should help to overcome some of the problems found among medical staff. In some
cases, adoption of new approaches to treatment is due to resistance to change among
medical personnel, heavy case-loads, and insufficient funding. These problems seem
to be particularly severe in parts of the former Soviet Union (Gleissberg et al. 1999).
 Expanding DOTS to the 84% of TB sufferers who do not currently receive it does not
necessarily entail the provision of more assistance to developing countries. Funds (both
public and private) are currently expended on ineffective TB programmes. For instance, in
South Africa and Eastern Europe, 30% of the TB budget is spent on residential care. Much
can be gained from a reallocation of these resources.
Further funds could come from reallocations within the health budget from uses
with poorer cost-effectiveness than TB. Such reallocation could occur from other parts
of the government budget. An ad hoc committee of WHO concluded that middle-income
countries do not need further funding, but that low-income countries would need support
in the introduction and expansion of DOTS; the least-developed countries would also
need additional funds for the maintenance of the programme (WHO, 1998c).
Section 5

Conclusion

T uberculosis accounts for nearly two million deaths per year worldwide. In many
regions it is a growing problem, partly because of HIV, partly because of insufficient
application of control measures, and partly because of the economic decline. Since more
than 75% of infections and deaths are among the most economically active age group in
the population (those 15 to 54 years of age), tuberculosis can have far-reaching economic
and social consequences for those infected and for their households and communities.
Although the poor are disproportionately affected by tuberculosis, the disease is not
exclusively a disease of the poor.
Decisions on what treatment regimes to follow are often based only on costs to
the health ministry. The costs borne by patients have largely been ignored, even though
such costs are often larger than the direct costs to the health ministry. Ignoring these
costs leads to an underestimation of the total costs of tuberculosis and, because these
indirect costs vary across different treatment regimes, can lead decision-makers to make
poor choices in health care. Although they are difficult, if not impossible, to quantify,
psychological and social costs such as discrimination, and attendant anxiety, and depression
add further to the costs of tuberculosis.
Households attempt to cope with the large immediate costs of tuberculosis by
diverting resources from other forms of health care, reducing other forms of consumption,
withdrawing children from school, borrowing or selling assets, and the like. Some of
these short-term coping strategies can have significant long-term costs for household
members. Such costs are rarely taken into account in studies of the impacts of tuberculosis.
Conventional treatments for tuberculosis are often expensive and have low cure
rates. They tend to impose particularly large indirect costs on individuals and households.
A considerable amount of evidence exists that short-course treatment, especially if
ambulatory strategies rather than hospitalization is used, are much more cost-effective
than conventional treatments. The cost-savings to individuals are particularly notable.
Why then are there more people with tuberculosis today than at any other time in
history, despite the fact that a highly effective treatment has been available for over four
decades and a highly cost-effective treatment strategy has been available for a decade?
Reichman (1997) places the blame at the feet of the medical sector and politicians.
Perhaps their reluctance to change current practice is understandable, although wrong.
Changes in the design of treatment delivery do not always dramatically reduce costs to
the health department; and it is these costs that health professionals and politicians care
about most. But as has been shown in this study, the treatment of tuberculosis, particularly
with the DOTS regime, substantially reduces the costs of the illness to patients and their
families. Once these costs are taken into account, there is a compelling case for treatment
for tuberculosis to be given a high priority by decision-makers and for DOTS to be adopted
as the treatment regime.
 References

Arif K, Ali SA, Amanullah S, Siddiqui I, Khan JA, and Nayani P. “Physician compliance with national
tuberculosis treatment guidelines: A university hospital study”. International Journal of Tuberculosis
and Lung Disease, 1997, 2:225–230.

Barro RJ. Determinants of Economic Growth. Cambridge, MA: The MIT press, 1997.

Becker, C and Bloom DE. “The demographic crisis in the former Soviet Union: Introduction”. World
Development, 26:1913–1919.

Bloom DE, Bennett N, Mahal AS, Noor W. “The impact of AIDS on Human Development”. Unpublished
working paper. Department of Economics, Columbia University, 1996.

Bloom DE and Williamson JG. “Demographic transitions and economic miracles in emerging Asia”.
Unpublished paper. Harvard Institute for International Development, 1998.

Brouwer JA, Boeree MJ, Kager P, Varkevisser CM, and Harries AD. “Traditional healers and pulmonary
tuberculosis in Malawi”. International Journal of Tuberculosis and Lung Disease, 1998, 2:231–234.

Carrin G, Gray E, and Almeida J. “Coping with ill-health in a rickshaw puller’s household in Chittagong,
Bangladesh”. Macroeconomics, Health, and Development Series Technical Paper No. 30, Geneva: WHO,
1998.

Cassles A, Heineman E, LeClerq S, Gurung PK, and Rahut CB. “Tuberculosis case finding in eastern Nepal”.
Tubercle, 1982, 63:175–185.

Chakraborty AK. “Prevalence and incidence of tuberculosis infection and disease in India: A comprehensive
review.” Unpublished project report sent to WHO, New Delhi, 1996.

Chakraborty AK, Rangan S, and Uplekar M. Urban Tuberculosis Control: Problems and Prospects. Bombay:
The Foundation for Research in Community Health, 1995.

Croft RA, and Croft RP. “Expenditure and loss of income incurred by tuberculosis patients before reaching
effective treatment in Bangladesh”. International Journal of Tuberculosis and Lung Disease, 1998, 2:252–254.

Devi S, Ganapathy S, Sivasubramanian, Mathew R, Charles N, and Sundaram V. “Behaviour pattern of ‘chest
symptomatics’in urban and rural population”. Report submitted to WHO Regional Office for South-East
Asia, n.d.

Dholakia R. “The potential economic benefits of DOTS strategy against TB in India”. TB Research, The Global
TB Programme, Geneva, WHO, 1997.

Dolin PJ, Raviglione MC, and Koch Ai. “Global tuberculosis incidence and mortality during 1990–2000”,
Bulletin of the World Health Organization, 1994, 72:213–220.

Dye C, Garnett GP, Sleeman K, and Williams BG. “Prospects for worldwide tuberculosis control under WHO
DOTS strategy”. The Lancet, 1998, 352:1886–1891.

Dye C, Scheele S, Dolin P, Pathania V, and Raviglione MC. “Global burden of tuberculosis: Estimated
incidence, prevalence and mortality by country in 1997.” Unpublished paper. Geneva: WHO Global
Surveillance and Monitoring Project, 1999.

Economist. “The white death”. The Economist, 1999, 350:11.

Floyd K, Wilkinson D, and Gilks C. “Comparison of cost effectiveness of directly observed treatment (DOTS)
and conventionally delivered treatment for tuberculosis: Experience from rural South Africa”. British Medical
Journal, 1997, 315:1407–1412.
Fogel R. “Economic growth, population theory and physiology.” American Economic Review, 1994,
89:369–395.

Fryatt RJ. “Review of published cost-effectiveness studies on tuberculosis treatment programmes.”

International Journal of Tuberculosis and Lung Disease, 1997, 1:101–109.

Geetakrishnan K, Pappu KP, and Roychowdhary K. “A study of knowledge and attitudes towards
tuberculosis in a rural area of West Bengal”. Indian Journal of Tuberculosis, 1988, 35:83–89.

Gleissberg VG, Maximova ZD, Golubchikova VT, Wares DF, and Banatvala N. “Developing nursing practice as
part of the collaborative TB control program, Tomsk, Siberia.” International Journal of Tuberculosis and
Lung Disease, 1999, 3:878–885.

Harper I, White A, Fryatt B, and Gurung BB. “Tuberculosis case finding in East Nepal: An evaluation of
microscopy camps performed by the British Nepal Medical Trust.” Biratnagar. Unpublished paper, 1993.

Holmes CB, Hausler A, and Nunn P. “A review of sex differences in the epidemiology of tuberculosis”.
International Journal of Tuberculosis and Lung Disease, 1998, 2:96–104.

Hudelson P. “Gender differentials in tuberculosis: The role of socioeconomic and cultural factors”. Tubercle
and Lung Disease, 1996, 77:391–400.

Jamison, Dean and Mosley WH. “Selecting disease control priorities in developing countries: Health policy
responses to epidemiological change”. American Journal of Public Health, 1991, 81:15–22.

Jamison DT, Lau LJ, and Wang J. “Health’s contribution to economic growth, 1965–90”. In Health, Health
Policy, and Economic Outcomes, Final Report, Health and Development Satellite, WHO Director-General
Transition Team. Geneva. WHO, 1998.

Karnolratanakul P, Sawert H, Kongsin S, Lertmaharit S, Sriwonga J, Na-Songkhla S, Wangmanee S,

Jittimanee S, and Payanandana V. “Economic impact of tubercilosis at the household level”. International
Journal of Tuberculosis and Lung Disease, 1999, 3:1–7.

Khanna BK, Srivastava AK, and Ali M. “Drug default in tuberculosis”, Indian Journal of Tuberculosis, 1977,
24:121–126.

Lancet, The. “Correspondence”. The Lancet, 1999, 353:145–148.

Liefooghe, Michiels RN, Habib S, Moran I, and De Muynck A. Perceptions, social consequences and stages
of illness behaviour: A focus group study of tuberculosis patients in Sialkot, Pakistan. Publication No. 21.
Antwerp, Instituut voor Tropishe Geneeskunde, n.d.

Leslie J. “Women’s nutrition: The key to improving family health in developing countries?”, Health Policy
and Planning, 1991, 6:1–19.

Martikainen Pand Valkonen T. “Do education and income buffer the effects of death of spouse on
mortality?”. Epidemiology, 1998, 9:530–533.

Meerding WJ. “Economic evaluation of tuberculosis control in Armenia: Cost-effectiveness analysis of the
WHO strategy.” Unpublished paper, Department of Public Health, Faculty of Medicine and Health Sciences,
Erasmus University, Rotterdam, 1997.

Migliori GB, Khomenko AG, Punga VV, et al. “Cost-effectiveness analysis of different policies of tuberculosis
control in Ivanova Oblast, Russian Federation.” Working paper, n.d.

Murray CJL. “Social, economic and operational research on tuberculosis: Recent studies and some priority
questions”. Bulletin of the International Union on Tuberculosis and Lung Disease, 1991, 66:149–156.
Murray CJL. “Epidemiology and demography of tuberculosis.” in Timaeus IM, Chackiel J, and Ruzieka L (eds)
Adult Mortality in Latin America, Oxford: Clarendon Press, 1996, pp. 199–216.

Murray CJL and Lopez A. The Global Burden of Disease. Cambridge, MA: Harvard School of Public Health,
1996.

Murray CJL, Styblo K, and Rouillon A. “Tuberculosis in developing countries: Burden, cost and intervention”.
Bulletin of the International Union Against Tuberculosis and Lung Disease, 1990, 65:6–26.

Murray CJL, Styblo K, and Rouillon A. “Tuberculosis.” In Jamison DT, Mosley WH, Measham AR, and
Bobadilla JL (eds.). Disease Control Priorities in Developing Countries. Oxford: Oxford University Press,
1993, 233–259.

Okojie CEE. “Gender inequality of health in the third world”. Social Science and Medicine, 1990,
39:1237–1247.

Over M, Ellis RP, Huber JH, and Solon O. “The consequences of adult ill-health” in Feachem RGA, Kjellstrom
T, Murray CJL, Over M, and Phillips MA (eds), The Health of Adults in the Developing World. New York:
Oxford University Press, 1992, pp. 161–208.

Pathania V, Almeida J, and Kochi A. “TB patients and private for-profit health care providers in India”. TB
Research No. 1 (WHO/TB/97.223), The Global TB Programme, WHO, Geneva, 1997.

Proctor MH. “WHO’s DOTS strategy”. The Lancet, 1999, 352:755.

Rajeswari R, Balasubramanian R, Muniyandi M, Geetharamani S, Thresa X, and Venkatesan P. “Socio-

economic impact of tuberculosis on patients and family in India”. International Journal of Tuberculosis
and Lung Disease, 1999, 3:869–877.

Ramachandran R, Balasubramanian R, Muniyandi M, Shanmugham G, Xavier T, and Venkatesan P.

“Economic impacts of tuberculosis on patients and family.” Tuberculosis Research Centre, Indian Council
of Medical Research, Chennai, South India, 1997.

Ravigilioni MC and Luelmo F. “Update on the global epidemiology of tuberculosis”, Current Issues in Public
Heath, 1996, 2:192–197.

Raviglione MC, Snider DE, and Kochi A. “Global epidemiology of tuberculosis: Morbidity and mortality
of a worldwide epidemic”, Journal of the American Medical Association, 1995, 273:220–226.

Raviglione MC, Harries AD, Msiska R, Wilkinson D, and Nunn P. “Tuberculosis and HIV: Current status
in Africa”, AIDS, 1997, 11:S115–S123.

Rieder HL, Arnadottir Th, Tardencilla Gutierrez AA, Kasalika AC, Salaniponi FLM, et al. “Evaluation of a
standard recording tool for sputum smearmicroscopy for acid-fast bacilli under routine conditions in low
income countries”. International Journal for Tuberculosis and Lung Disease, 1997, 1:339–345.

Sachs J and Warner A. “Fundamental sources of long-run growth”. American Economic Review, 1997,
87:184–188.

Saunderson PR. “The 20 th Kellesberger Memorial Lecture: Leprosy and tuberculosis combined
programmes—An uneasy partnership”. Ethiopian Medical Journal, 1994, 32:269–280.

Saunderson PR. “An economic evaluation of alternative programme designs for tuberculosis control in rural
Uganda”. Social Science and Medicine, 1995, 40:1203–1212.

Sawert H. “Cost-analysis and cost-containment in tuberculosis control programmes: The case of Malawi”.
Health Economics (WHO/TFHE/96.1). Geneva: WHO, 1996.
Sawert H. “The economic cost of tuberculosis in Indonesia: With an assessment of costs and benefits
of an improved control strategy”. Geneva: WHO, 1997.

Sawert H, Kongsin S, Payanandana V, Akarasewi P, Nunn PP, and Raviglione MC. “Costs and benefits of
improving tuberculosis control: The case of Thailand”. Social Science and Medicine, 1997, 44:1805–1816.

Singh G, Banerji SC, and Mathur SK. “A study of defaulters in antituberculosis therapy”. Indian Journal
of Tuberculosis, 1976, 23:98–102.

Smith I. Women and tuberculosis: Gender issues and tuberculosis control in Nepal. Unpublished Masters
thesis, Nuffield Institute for Health, 1994.

Sutherland I. The Ten-Year Incidence of Clinical Tuberculosis Following ‘Conversion’ n 2,550 Individuals Aged
14 to 19 Years. Tuberculosis Surveillance Research Unit (TRSU) Progress Report, The Hague, 1968.

United Nations. World Economic and Social Sur vey 1997. New York: United Nations, 1997.

Uplekar M and Rangan S. Tackling TB: The search for solutions. Bombay: The Foundation for Research in
Community Health, 1996.

Uplekar M, Juvekar S, and Morankar S. Tuberculosis patients and practitioners in private clinics. Bombay:
The Foundation for Research in Community Health, 1996.

Van der Werf TS, Dade GK, and van der Mark TW. “Patient compliance with tuberculosis treatment in Ghana:
factors influencing adherence to therapy in a rural service programme”. Tubercle, 1990, 71:247–252.

Watts SJ, Brieger WR, and Yacoob M. “Guinea Worm: An in-depth study of what happens to mothers,
females, and communities”. Social Science and Medicine, 1989, 29:1043–1049.

Weinstein MC, Siegel JE, Gold JE, Kamler MR, and Russell LB.. “Recommendations of the panel on cost-
effectiveness in health and medicine”. Journal of the American Medical Association, 1996, 276:1253–1258.

Wilkinson D, Floyd K, and Gilks CF. “Costs and cost-effectiveness of alternative tuberculosis management
strategies in South Africa: Implications for policy”. South African Medical Journal, 1997, 87:451–455.

World Health Organization. Report from a consultation on the socioeconomic impact of HIV/AIDS on
households. UNAIDS/97.3, Geneva, WHO, 1997.

World Health Organization. “Tuberculosis”. WHO Fact Sheet No.104. Geneva, 1998a.

World Heath Organization. “TB is single biggest killer of young women.” WHO Press Release WHO/40,
May 26. Geneva, 1998b.

WHO. Report of the Ad Hoc Committee on the Tuberculosis Epidemic. Global Tuberculosis Programme
WHO/TB/98.245. Geneva: WHO, 1998c.

WHO. Global Tuberculosis Control WHO Report 1999 WHO/CDS/CPC/TB/99.259. Geneva: WHO,
Communicable Diseases, 1999.

Yesudikan OAK. “Behaviour of the private sector in the health market of Bombay”. Health Policy and
Planning, 1994, 9:72–80.

Zumla A and Grange JM. “Establishing a united front against the injustice of tuberculosis”. International
Journal of Tuberculosis and Lung Disease, 1998, 2:179–181.

Zwarenstein M, Schoeman JH, Vundule C, Lombard CJ, and Tatley M. “Randomisd controlled trial of self-
supervised and directly observed treatment of tuberculosis”. The Lancet, 1998, 352: 1340–1343.
 Tables

Table 1. Socioeconomic profile of TB patients—Tamil Nadu, India

Percentages
Years of education Male Female Total

Illiterate 25 34 29
Less than 5 33 27 30
5–8 23 20 21
More than 8 19 19 19

Income (US$) Male Female Total

Less than 14 8 13 10
14–29 24 18 21
29–57 29 41 34
More than 57 40 28 35

Source: Ramachandran et al. (1997)

Table 2. Costs of Tuberculosis Treatment—Uganda and South Africa

Rural Uganda Rural South Africa

Cost to health service (US$) 95 2,570

Cost to patient (US$) 229 371

Total (US$) 324 2,941

Source: Uganda: Saunderson (1995) — South Africa: Wilkinson et al. (1997)

Table 3. Estimated household costs of tuberculosis: Survey studies

Cost to patient Bangladesh1 Uganda2 India3 South Africa4

Direct costs (US$) 130 68 41 99
Lost work 57% 91% na na
Time loss 14 months 10 months 3 months 4 months
Lost income (US$) 115 161 89 272
Indirect cost as % of annual
household Income 15 na 14 na
Total cost as % of annual
household Income 31 na 20 na

1. Croft and Croft (1998), Northwest Bangladesh, 21 patients registered at a clinic

2. Saunderson (1995), Southwest Uganda (rural), 34 patients at large hospital
3. Ramachandran et al. (1997), Tamil Nadu, India, 304 urban and rural TB patients under treatment for 2 to 6 months
4. Floyd, Wilkinson and Gilks (1997) Hlabisa, northeastern KwaZulu-Natal, South Africa
na = not available from the study

Table 4. Estimated household costs of tuberculosis: Simulation studies

Indonesia1 Thailand2 Bangladesh3

Time loss
• Morbidity
Treated 2 months 2 months na
Untreated 12 months 12 months na
• Mortality
HIV– 13 years 15 years na
HIV + 2 years 2 years na

Income loss (US$)

• Morbidity
Treated 149 317 329
Untreated 898 1 900 na
• Mortality
HIV– 11 490 19 400 na
HIV + 1 718 3 490 na
Cost-effectiveness 10–5 ‹ 32 na
Income per capita 898 1 900 105

1. WHO
2. Sawert et al. (1997)
3. Carin, Gray and Almeida (1998), Chittagong rickshaw puller’s household. Income loss includes morbidity
and mortality risk over two-year period. Annual household income for six persons estimated at US$ 629
na = not available from the study
Table 5. Acceptance of TB patients by family—Tamil Nadu, India

Percentage
Rural Urban
Male Female Male Female
Accepted 93 82 85 68
Not well accepted 3 4 7 17
Rejected 4 14 8 15

Source: Ramachandran et al. (1997)

Table 6. Impact of parent’s TB on child’s schooling—Tamil Nadu, India

Region Sex of patient Percent stop school

• Rural Male 8
Female 4

• Urban Male 16
Female 16

Source: Calculated form data in Ramachandran et al. (1997)

Table 7. Comparison of costs of conventional and DOTS treatment—Rural Uganda

Costs per Patient

Conventional DOTS
Costs to health service 95 112
Costs to patient 229 84
• Before diagnosis 32 16
• During hospital stay 36 0
• Lost income from work 161 68

Source: Sanderson (1995)

Note:
Conventional treatment, which is widely practised throughout sub-Saharan Africa, consists of two months hospital
treatment followed by 4–10 months outpatient treatment.
The alternative treatment, DOTS, assumes ambulatory treatment with supervised drugs weekly for 8 weeks, then 4–10
months outpatient treatment. Included is increased spending on health education and supervision.
Table 8. Cost-effectiveness of TB treatments—Developing countries

Malawi Mozambique Tanzania Hlabisa,

South Africa

Short course/hospitalization 165 232 202 388

Standard/hospitalization 215 301 270 1 031

Ambulatory, Short-course 107 92 202 na

Ambulatory, standard 111 92 107 na

Source: Murray (1996) — Wlkinson et al. (1997)

 Contact: the Stop TB Initiative
20, avenue Appia • 1211 Geneva 27 • Switzerland
Tel. +(41) 22 791 2675 • Fax +(41) 22 791 4199 • www.stoptb.org