Article Text
PDF
Abstract
Objectives To estimate the economic impact of failure to find and treat tuberculosis disease and prevent tuberculosis infection from progressing to active disease.
Design Estimating the economic cost of not finding and treating a patient suffering from tuberculosis.
Setting Estimation methodology is developed in the Indian context, as informed by local costs and reported tuberculosis epidemiology.
Participants No individual participants were included.
Primary and secondary outcome measures The primary outcome measure is the total cost of patients with drug-susceptible and drug-resistant tuberculosis who are and are not found and treated by tuberculosis programmes, including costs for medications, lost productivity, healthcare services and furthered transmission. We calculate the economic burdens by varying the number of individuals a person sick with tuberculosis infects (10 or 15 people) and the risk of progression to tuberculosis disease if infected (5 or 8%). The secondary outcome measure is the amount saved by finding a patient early or who would not have otherwise been found. All costs are presented in US dollars (exchange rate: 72 Indian rupees/1 US$).
Results By finding and treating a patient early before furthered transmission occurs—or stopping progression of tuberculosis infection to tuberculosis disease with preventive therapy—the Indian health system can save US$5502 to US$15 825 and US$5846 to US$25 575, for each individual with drug-susceptible and drug-resistant tuberculosis, respectively, across scenarios.
Conclusions These estimates provide crude, lower bounds for the potential costs of not appropriately diagnosing and treating a single patient with active tuberculosis in a timely manner, or preventing a patient with tuberculosis infection from progressing to active disease. The actual financial burden on society is far higher than estimated using this simple, short-term cost-effective analyses. Our results highlight the limitations of tuberculosis costing models to date, and demonstrate the importance of accounting for airborne transmission of tuberculosis.
- Tuberculosis
- HEALTH ECONOMICS
- Health economics
- INFECTIOUS DISEASES
- Public health
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
We have incorporated comprehensive costs associated with a patient with tuberculosis (TB), including treatment costs, productivity loss for the patient and caregivers, health service costs and future transmission risk.
We only calculate the economic costs of TB and do not include numerous social costs that affect patients with TB, like stigma and loss of familial supports.
We consider multiple scenarios of how infection is spread by patients with TB who are properly and improperly diagnosed, thereby generating a range of cost estimates.
Costing data were pulled from global, national and local reports; while this variability in cost estimates may not represent the true cost at a given time, it represents a realistic range of costs that may vary across the country.
Where multiple costs are available, or limited cost estimates are available, we considered the most conservative ones to ensure we are presenting the absolute lower bound of potential costs to be incurred.
Background
Despite being treatable since the late 1940s, over 10 million people develop disease with tuberculosis (TB) annually, 1.5 million of whom will die—more than 4000 people daily. Over 30% of people sick with TB disease are never diagnosed, and without the provision of life-saving treatment, up to two-thirds of them will die.1–3 Additionally, people sick with TB can infect up to 10–15 other individuals per year. Among those who become infected, there is about a 10% lifetime risk of progressing to TB disease without some form of preventive therapy—most of which occurs in the 2 years after infection.4
Comprehensive TB programmes that simultaneously aim to search for people sick with TB using appropriate active case-finding strategies with sensitive diagnostics,4 treat with the correct medicines in a timely manner,5–8 and prevent further spread of disease by offering preventive treatment to infected people who have not progressed to disease,9–12 have been shown to rapidly bring down rates of TB, and are critical to the goal of eliminating TB. Despite considerable success of this comprehensive search-treat-prevent approach in high-income communities and countries, this has not been the norm in lower-resourced settings.13 Instead, resource-constrained settings were told to focus primarily, or even exclusively, on individuals sick with pulmonary TB who passively report to health providers, often long after spreading disease to their families and communities. The result has been a minimal reduction of TB transmission globally.14
The strategy for addressing the TB pandemic globally has now changed, most notably with the United Nations High Level Meeting commitments to a set of comprehensive search-treat-prevent interventions as a path to TB elimination.15 Despite these commitments, many programmes have yet to implement robust programmes for highly-effective community-based case-finding, diagnosis and care for both TB disease and infection, largely due to having to work within predetermined, limited budgets designed around minimalist interventions over short timescales.16
This lack of investment in an approach that is known to stop transmission at each point of the search, treat and prevent care cascades, means that many existing cases are not found, many future cases (through early case detection and preventive treatment) are not prevented and transmission continues. While it is already recognised that this comes at a great cost of preventable death and suffering to families and communities, we have yet to fully understand the cost of inaction—to people, programmes, communities and the countries in which they live—of not treating an active case of TB disease or an infection that will progress to disease in high-TB-burden settings. What is the true economic impact of not implementing a standard of care for TB that has been well understood for 60 years?
An analysis of cases averted through programmatic interventions in the USA (1995–2014) showed that while each case of TB cost US$17 000 in direct costs, the societal costs, excluding productivity due to premature deaths, was an additional US$20 000. The societal cost increased to a staggering US$44 000 when premature death was accounted for. In addition to preventing TB suffering and morbidity, the economic benefit for implementation of various TB treatment and prevention measures was between US$6.7 and 14.5 billion.17 These benefits are difficult to see in 1 year—the general budget cycles for TB programmes—and can only be appreciated over time.
India is the second most populous country in the world and has a high TB-burden, with 2.64 (1.80 to 3.63) million new TB cases in 2019, of which 18.2% were not notified or diagnosed.18 TB treatment coverage is estimated to be 82.0%, and 4.7% (2.8 to 7.2) of individuals with TB were infected by drug-resistant (DR-) strains.18 India’s national TB budget was US$497 million, with around 80% of funding coming from domestic sources.18
Turning the tide on TB in India will require the implementation of a comprehensive approach to finding and treating people with TB disease and infection. While the cost of such a programme likely exceeds the current TB programme budget, any analysis of cost and effect must recognise that TB is transmitted in the air, and every missed case of disease and infection leads to new cases. It is essential to understand the cost of not implementing a comprehensive search-treat-prevent strategy to stop the spread of TB. Here, we aim to estimate what a single missed patient with TB will cost (ie, the cost of inaction), in the Indian context.
Methods
Study design
To estimate the costs of a patient with TB on society, we approximate the direct and indirect economic burdens of patients diagnosed with drug-susceptible (DS-) TB, multidrug-resistant (MDR-) TB and patients who have TB but are never diagnosed. All costs are presented in US dollars (US$). The exchange rate used is Indian rupees 72 per US$1.
Direct economic burden of patients with DS-TB
The cost of DS-TB medications assumes a standard 6-month treatment course, per WHO recommendations.19 Drug doses are based on body weight of 43 kilograms,20 requiring: 215 mg isoniazid, 430 mg rifampicin, 717 mg ethambutol and 1075 mg pyrazinamide.21 Costs of drugs are taken from a local Chennai pharmacy (August 2017).
Loss of productivity is essential to consider due to great physical and mental constraints rendering individuals with TB incapable of working at optimal levels. While the minimum daily wage is fixed at US$3.56,22 we assume a daily construction wage rate of US$5.56, which is likely still conservative even for an urban area with a lot of diversity in employment opportunities.23 We assume 25 working days a month, given the standard 6-day work week and ignoring other holidays. We assume a 9-month diagnosis delay due to not seeking appropriate specialist care and because it can take 6 months to be initiated on an appropriate treatment regimen.24 During this delay, we assume a 20% loss of productivity. Then, during the first 3-month treatment period, we assume a 40% productivity loss, while during the last 3-month period we assume a 20% loss. We assume 40 half-day hospital/clinical visits (weekly visits during treatment, including travel and visit time) leading to 50% productivity loss during this time. Finally, we assume 10% of a support staff’ time for assisting patients with TB during 6 months of treatment. To calculate lost productivity, we multiply per cent of lost productivity by working days by the daily wage rate. While we have broadly validated these assumptions through informal discussions with pulmonologists who specialise in TB, we also conduct sensitivity analysis by assuming a productivity loss of 20% during the entire TB treatment.
Individuals with TB requires specialised care; we assume four full visits with a doctor at US$2.78 each,24 frequent short visits with a doctor during treatment totalling 2.5 hours at US$5.56 per hour, 2.5 hours for other technicians at US$2.78 per hour, at least two X-rays costing US$4.86 each and at least four sputum diagnostic tests or other tests, an average of US$3.13 each.25 To calculate service costs, we multiply the number of units (hours or tests) by the unit cost.
Direct economic burden for patients diagnosed with MDR-TB
Drug-resistant strains of TB require more expensive drugs for longer duration. Despite estimates that MDR-TB treatment costs US$3000–6000,26 we use WHO’s more conservative estimate that MDR-TB can be treated for less than US$1000 per patient.27
Lost productivity is assumed to be more substantial for a patient with MDR-TB. We assume the same 9-month diagnosis delay24 at 20% lost productivity during this time. We assume 70% and 30% lost productivity during the first and second treatment years, respectively. We assume 80 half-day long hospital/clinical visits throughout treatment, leading to 50% productivity loss. We assume 10% of support staff time over 2 years of treatment. We assume 300 working days during each year of treatment.
Using the same costs per service, we assume a patient with MDR-TB requires eight full visits, 5 hours of a doctor’s time, 5 hours for a technician’s time, four X-rays and eight other tests for evaluation and diagnosis.
Direct economic burden for patients with TB but not diagnosed
Individuals sick with TB but who are never diagnosed may eventually seek medical care for their symptoms, or if they are misdiagnosed, they may consume other medications. While care and medication costs could be large, there is likely great individual variability. Therefore, we conservatively assume that the cost of medicine is half that for DS-TB.
Individuals not properly diagnosed are at high risk of death. We assume decreased productivity over 3 years, with 20%, 60% and 85% losses in years 1–3, respectively; only 15% of individuals spontaneously cure by year 3. We do not apply any discounting factor or wage increase during the 3-year period; they are likely to cancel each other out because the wholesale price index used in normalising money over time and wage rate inflation are likely highly correlated.28 We conservatively account for 10% productivity loss for a caregiver during this 3-year span. As is the case earlier, while we have validated these assumptions through informal discussion with pulmonologists who specialise in the treatment of TB, we also conduct a sensitivity analysis by assuming a productivity loss of 10%, 20% and 40% in the first 3 years of the disease.
Additional costs may include doctor visits for symptoms management or misdiagnoses and other non-tuberculosis-related clinical or diagnostic testing. Because these scenarios may be too varied to realistically capture, we conservatively assume a cost half that of individuals diagnosed with DS-TB.
Indirect economic burden of patients with DS-TB
Indirect costs need to be accounted for because TB is contagious through airborne transmission. People with active TB can infect 10–15 people through close contact over a year and those infected with TB have, roughly, a 10% lifetime risk of falling ill with TB disease.29 Using three conservative scenarios, we calculate indirect economic burdens. Scenario 1 assumes 10 people are infected and a 5% risk of progressing to TB disease; scenario 2 assumes 10 people and 8% risk; and scenario 3 assumes 15 people and a 5% risk. These will provide conservative lower bound estimates and slightly more realistic estimates, though certainly not an upper bound. All these three scenarios were repeated under the alternative assumptions regarding productivity loss.
Using aforementioned assumptions, we estimate the number of people that a single patient with TB will spread TB to:
(a) # new cases of DS-TB= # individuals infected × per cent risk
Each of these new cases will go on to infect more individuals, each with a risk of TB disease progression, per the three scenarios. The total number of people infected from a single TB case is:
(b) Total # of people with DS-TB= |1 + (# new cases of TB) + (# new cases of TB × # new cases of TB) + (# new cases of TB × # new cases of TB × # new cases of TB) +…|
We assume that 95% of individuals with TB will have DS-TB and 5% will have MDR-TB, per the local epidemiology.18 Assuming that 60% of people with TB receive treatment,2 we calculate the economic burden that any additional individuals infected with TB will cost the system:
(c) Total indirect cost for patient with DS-TB= (total # of people with DS-TB–1) × ((0.95×0.6×direct costs of a patient with DS-TB) + (0.05×direct cost of a patient diagnosed with MDR-TB) + (0.4×direct costs of a patient not diagnosed))
We assume that the rate at which MDR-TB spreads is the same as the rate of DS-TB (Formulas a and b). We assume that all new cases from a patient with MDR-TB will also have MDR-TB.
(d) Total indirect cost for patient with MDR-TB= (total # of people with MDR-TB–1) × ((0.6×direct costs of a patient with MDR-TB) + (0.4×direct costs of a patient not diagnosed))
Further, we assume that the 40% of patients who never get treated spread disease more than those on appropriate treatment. Patients with untreated TB become more contagious over time because the infection does not subside. We assume that an undiagnosed and untreated individual with TB actively spreads TB at a rate three times that for an individual who ultimately receives appropriate treatment. To be conservative in costs, we assume that these undiagnosed cases are DS-TB.
(e) Total indirect cost for a patient not diagnosed=3 × ((0.95×0.6×direct costs of a patient with DS-TB) + (0.05×direct cost of a patient diagnosed with MDR-TB) + (0.4×direct costs of a patient not diagnosed))
The total economic burden of TB can be estimated by:
(f) Total cost=direct cost + indirect cost,
per patient diagnosed with DS-TB, MDR-TB and not diagnosed.
The cost savings of finding a patient with TB who otherwise would not have been found are:
(g) Cost savings=Total cost of not diagnosed patient – Total cost of diagnosed patient,
for both patients with DS-TB and MDR-TB.
Patient and public involvement
Patients were not involved in this study.
Results
Direct economic burden
For patients who have been diagnosed with TB, we estimate the cost of a course of DS-TB treatment to be US$40.08, based on the prices of drugs observed in the local pharmacy. Based on assumptions about per cent decreases in productivity, we can compute lost productivity to be US$694.44. Other costs of necessary medical personnel and clinical and diagnostic testing is estimated to be US$54.17. The direct cost related to a patient diagnosed with DS-TB sums to US$788.69.
Prices of an MDR-TB treatment regimen are estimated at US$1000. Loss in productivity is estimated to be US$2472.22. Total other costs associated with medical personnel and clinical and diagnostic testing is estimated to be US$108.33. The direct cost associated with a patient diagnosed with MDR-TB can be estimated at US$3580.56.
Prices for medicines to treat symptoms associated with TB are estimated at US$22.60. The estimated cost associated with lost productivity over a 3-year span is US$2800. Other costs for medical visits and testing are estimated at US$27.08. The direct cost related to a patient not diagnosed with TB is estimated at US$2849.68.
Indirect economic burden
Scenario 1: the indirect economic burden from a patient properly diagnosed with and treated for DS-TB is US$1721; indirect costs from a patient with MDR-TB is US$3288, the indirect costs from an individual with DS-TB who goes undiagnosed is US$5162 and the indirect costs from an individual with MDR-TB who goes undiagnosed is US$9864.62.
Scenario 2: the indirect economic burden from a patient properly diagnosed with and treated for DS-TB is US$6882; indirect costs from a patient with MDR-TB is US$13 153, the indirect costs from an individual undiagnosed is US$20 646 and the indirect costs from an individual with MDR-TB who goes undiagnosed is US$39 458.50.
Scenario 3: the indirect economic burden from a patient properly diagnosed with and treated for DS-TB is US$5162; indirect costs from a patient with MDR-TB is US$9865, the indirect costs from an individual undiagnosed is US$15 485 and the indirect costs from an individual with MDR-TB who goes undiagnosed is US$29 594.
Along with these three scenarios, we also compute economic burden for three more scenarios. Scenarios 4, 5 and are similar to Scenarios 1, 2 and 3 with an assumption of lower productivity loss. In the interest of space, we do not report them here (refer to table 1).
Total cost of patients with TB (in US$)
Total economic burden
Total economic burden is the sum of the direct and indirect economic burden. In the interest of space, we refer to table 1 for the economic burden across all the six scenarios.
Cost savings
The cost that a programme would save by finding a single patient with DS-TB who would not otherwise have been found is US$5502, US$15 825 and US$12 384, respectively, for scenarios 1, 2 and 3 (table 2). The cost saved by finding a single patient with MDR-TB who would not otherwise have been found is US$5846, US$25 575 and US$18 998, respectively, for scenarios 1,2 and 3.
Amount saved with proper diagnosis, treatment and prevention
Discussion
For many years, TB programmes have been plagued by glacial diminution of TB incidence despite continuous investment in treatment programmes. The general lack of improvement is likely due to the fact that programmes have not deployed tried-and-tested approaches from the early and mid-20th century to search actively for people with TB disease and infection early using sensitive diagnostic tools, treat all forms of TB disease with the best medications as rapidly as possible, and prevent further disease through the early treatment of TB infection. The barrier that is often cited for lack of investment in comprehensive programmes is cost, with a focus specifically on the annual cost of providing high quality diagnostic, treatment and preventive care to communities struggling with TB.
In this analysis, we have reframed the question to ask what it costs a community or a country when a case of TB is missed, when the proper standard of care is not deployed. In so doing, we elucidate a methodology for calculating the cost of missing the diagnosis and treatment of people with TB disease and infection, and produce crude estimates that can be replicated by any country/community/TB programme leader. Although estimates are crude, they are comprehensive in that they take material, personnel, productivity and furthered transmission into account.
Our findings for India—where the cost of missing a patient is, on a lower bound, three to four times the cost of treatment—demonstrates that investment in a platform for community-based TB diagnosis, treatment and prevention, would reap huge social and economic benefits (between US$1.8 and US$4.0 million per 100 patients found). We found that as much as there is productivity loss, the main driver of economic burden is the indirect effect. Compared with our lost productivity estimate of US$694.44, previous studies have estimated the productivity loss associated with TB in India to be US$674 in 2009.30 Therefore, our estimates remain conservative.
People whose TB disease is undiagnosed suffer greater debiliating effects, and also spread the disease across more number of people. Our simple simulation exercise suggests that the best way to reduce disease burden due to TB is to invest in programmes that stop transmission by searching for patients actively using sensitive diagnostics, treating people sick with TB in a timely fashion with the correct medications and providing preventive treatment to stop progression from TB infection to active disease. This will not only reduce long-term costs for TB programmes, but the added by-product of a community-based diagnostic and delivery platform would undoubtedly also serve as a mechanism for finding patients with other diseases, such as diabetes, hypertension and hepatitis C.
Our study has some limitations. First, our medicine prices were based on local pharmacy rates which might underestimate the actual cost of the drug and not demonstrate the heterogeneity in drug costs across different areas. We also chose to use the most conservative costs available for the price of MDR-TB drug costs; literature suggests that the true costs could be upwards of 6–7 times the costs used.26 For our calculations of lost productivity, we purposely use an oversimplification of the demographic breakdown of the population with TB; age, sex, occupation and other characteristics are not taken into account, despite knowledge that certain subgroups may be at higher risk of TB than others and that certain subgroups are more likely to have steady streams of income than others. Here we only calculate the economic costs of TB and do not include numerous social costs that affect patients with TB, like stigma and loss of familial supports.
We also do not account for if an individual were to die from TB the perpetual lost income to the household. In many cases, if the patient is the primary earner of the family, the loss would be substantially greater than the loss in productivity itself. When calculating other costs, we do not include costs associated with treating side effects from the toxicity of MDR-TB drugs due to such large variability likely to be experienced based on baseline health status, existing comorbidities and drug regimen composition. These costs also assume that all individuals diagnosed with TB are initiated on treatment and complete treatment, which we know is not the case. We also do not account for the costs of extra testing and misdiagnosis in the private sector. Fortunately, there are many programmes in India that demonstrate the feasibility and efficacy of community-based active TB case finding strategy in challenging settings with myriad public and private providers.31
Lastly, we do not take care-seeking behaviour into account directly, as this would take significantly more resources and research to ascertain. However, we do account for cases that are and are not notified, per national reports and delays in care seeking and diagnosis as reported in the literature.32
Although our empiric cost estimates were assessed with the best data available at the time of analysis, some data were from global reports, some from national reports and others from local health catchment areas in India. This variability in cost estimates may not represent the true cost at any given time, but will represent a realistic range of costs that may vary across the country. All of the above assumptions are made to provide a crude estimate that can inform health systems at the simplest level, and all are assumptions to ensure that the estimate is conservative, meaning we are presenting the absolute lower bound of potential costs to be incurred.
Conclusion
This study provides lower bound estimates for the economic impact of not implementing a comprehensive search-treat-prevent strategy to stop the spread of TB in the Indian context. The cost of not diagnosing and treating an individual with TB in a timely manner, and not preventing TB infection that can progress to disease, is significantly more prohibitive than an upfront investment in a comprehensive programme for TB elimination. Given our conservative approach, the economic impact that an undiagnosed, untreated or unprevented case of TB places on society is likely far higher than estimated. Our results suggest very high value-for-money for an investment in the comprehensive search-treat-prevent strategy to stop the spread of TB.
Data availability statement
All data relevant to the study are included in the article or uploaded as supplementary information.
Ethics statements
Patient consent for publication
Ethics approval
Neither ethical approval nor informed consent was required for this study, as human subjects research was not involved and data were aggregated and collected from public sources or published literature.
Footnotes
Twitter @meredithbbrooks
Contributors TN, SK and VP conceptualised the study. MBB and VP collected data and conducted analysis. MBB wrote the first draft of the manuscript; all authors revised the manuscript critically and approved the final version for submission. MBB and VP contributed equally to this paper. As guarantor of the study, MBB accepts full responsibility for the work and conduct of the study, had access to the data, and controlled the decision to publish.
Funding This work was supported by the Dubai Harvard Foundation for Medical Research (grant number: N/A), Advance Access & Delivery (grant number: N/A) and the Harvard Medical School Centre for Global Health Delivery (grant number: N/A) for salary support (MBB and SK) and conference support (MBB, VP, TN and SK).
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.