A Systematic Review and Meta-Analysis on Catastrophic Cost Incurred by Tuberculosis Patients


 Background: One of the World Health Organization End Tuberculosis (TB) Strategy is to reduce the proportion of TB affected families facing catastrophic costs (CC) to 0% by 2020. CC is defined if total cost related to TB management exceeded 20% of annual pre-TB household income. This study aimed to estimate the pooled proportion (PP) of TB affected households who suffered from CC. Method: A search of the online database through September 2020 was performed. Of 5114 articles, 29 articles were included in meta-analysis. We used R software to estimate the PP at 95% confidence intervals (CIs) using the fixed/random-effect models. Result: The PP of patients faced CC was 43%. Meta-regression revealed that country, drug sensitivity and HIV co-infection were the main predictors. CC incurred by drug sensitive, drug resistant and HIV coinfection patients were 32%, 80%, and 81% respectively. Lower CC incurred by active than passive case finding; 12% versus 42%. Direct cost represented 55% (95% CI 43-66) of the total cost. About 45% of TB-affected household faced catastrophic health expenditure at cut-off point of 10%. Conclusion: There is still a significant proportion of TB patients facing CC, which represent a main obstacle against TB control.PROSPERO registration: CRD42020221283

MDR-TB) that does not respond to at least Isoniazid and Rifampicin, the 2 most powerful anti-TB drugs [6] [7].

The nominator of catastrophic cost is the summation of direct and indirect costs.The direct cost includes either medical cost (consultation fees, diagnostic tests and treatment) or non-medical cost (transportation, accommodation, increased food needs).Indirect cost includes lost wages due to unemployment; time spent away from work and associated loss of productivity.Moreover, patients also incur large costs in the pre-treatment phase to cover consultations and laborat

y tests, symptomatic treatment, antib
otics trial, and hospitalization [8].An important segment of the nancial hardship is dissaving which means reduced nancial strength of a household or engage the household in damaging nancial coping strategies.This will reduce the nancial capac ty and their coping with the nancial shocks and cast them into the poverty trap .[9] Dissaving can take many forms like taking out a loan, taking children out of education, selling assets, reducing consumption to below basic needs to cope with health-related expenditure [8][9][10].

Consequently, WHO developed the TB patient cost survey to properly assess the total costs and proportion of patients facing catastrophic c st.This tool provide a standardized methodology for cross-sectional surveys in TB affected countries [11].Many studies used this cost survey to report catastrophic cost, catastrophic health expenditure, or hardship nancing incurred by TB patients [12][13][14].Some literatures calculated catastrophic cost or drug sensitive, MDR or HIV co-infection [14][15][16].Other studies estimated compared this cost considering adoption of different case nding strategies (ACF versus PCF) [17,18].In response to this reported catastrophic cost, the Global TB Program endorses social protection initiatives to complement Universal health coverage (UHC) initiatives [19,20].Examples of social protection interventions include cash transfers, food assistance, disability grants and health insurance.Those global nancial supports already exist in most countries, but may not be fully implemented [7].

the end, keepi
g in mind that COVID-19 pandemic may reverse the achieved progress in the TB control as many countries directed their resources toward pandemic containment.In addition, there are no published systematic reviews that report the pooled proportion of patients suffering from catastrophic cost; we aimed to perform this systematic review and meta-analysis to estimate the proportion of catastrophic cost among TB patients and their households in attempt to support the ongoing TB control programs.


Method

This syste atic review and meta-analysis was conducted according to the Preferred Reporting Items of the Systematic Reviews and Meta-Analyses (PRISMA) guidelines [21].

Data source and search strategy EMBASE, Scops, EBSCO, MEDLINE central/PubMed, ProQuest, Scielo, SAGE, Web of science, and Google scholar databases were searched for articles without timeframe, geographical or language restr ctions up to November 20 th , 2020 by two authors ( ShA & NZ) then revised by (RMG& SA).Highly focused and sensitive search strategies were developed by RMG after the approval of PubMed Help Disk.The search terms include ("tuberculosis "OR "Mycobacterium tuberculosis" OR "Koch's disease" ND "catastrophic cost").References from relevant studies were screened for supplementary articles.


Study selection and data extraction:

We aimed to include observatio al studies, which eported the proportion of patients suffering from catastrophic cost during the intensive ( rst 2 or 8 mo

hs of treatment in
S or MDR respectively) or the continuation phases of TB treatment.

The primary endpoint of interest was the proportion of TB affected patients and their households who face catastrophic cost.It was de ned as the total direct and indirect costs due to TB reaches or exceed 20% of the patie t or household's annual income [5] .Furthermore, CTC was assessed among patients according to their drug sensitivity as DS or MDR (with or without HIV), and strategy of case nding (ACF versus PCF).

Secondary outcomes were the proportion of the direct to the total cost of TB among DS or MDR, with or without HIV, catastrophic health expenditure CHE (de ned as direct cost that reaches or exceeds 40% of patients capacity to pay or 10% of their household income [22], and the different coping strategies.

Titles and abstracts were screened independently by four authors (AM, ShA, NZ, and EE), who discarded articles not pertinent to the topic.Non-observational studies, case reports, editorial, reviews, letters, and studies that estimated the direct and indirect cost of the popul

ion as a one unit not
ndividually were excluded from qualitative analyses but screened for potential additional references.Three other authors (RMG, SA & HE) solved the discrepancies on study judgements.Data extraction and analysis were performed by (RMG, AM, HE) and independently veri ed by (SA).


Data analysis:

The proportion of CTC among TB patients was pooled using the random-effects model.To ensure robustness of the model and susceptibility to outliers, pooled data was also analyzed with the xed-effects model.Heterogeneity was assessed by the Chi-squared test on N-1 degrees of freedom, with an alpha of 0.05 considered for statistical signi cance and the Cochrane-I-squared (I 2 ) statistic.I2 values of 25%, 50% and 75% were considered to correspond to low, medium and high levels of heterogeneity, respectively.

Sources of heterogeneity, for identifying possible effect modi ers on the pooled analyses, were explored using:

1-Sensitivity analysis (leave one out sensitivity analysis, GOSH sensitivity analysis, remove outliers)

2-Subgroup analysis: we categorized the catastrophic cost at 20% for ACF and PCF patients according to country where studies were conducted (inside/outside) India.


3-Met-regression:

The impact of country where the survey was conducted (high versus low incidence of TB) [23], quality of the study, sex, and population criteria (drug sensitivity, drug resistant with or without HIV) on the size effect of studies to explain the substantial heterogeneity.

The forest plot was used to visualize the degree of variation between studies.All data analysis was performed R software version 4.0.3 using Harrer hand-on guide [24].


Publication bias:

Publication bias was investigated by visual inspection of funnel plots, and by Egger's regression test.


Quality assessment

The Newcastle-Ottawa Scale (NOS) was used to assess the quality of studies.Studies were classi ed according to the NOS as: very good studies (9-10 points), good studies (7-8 points), satisfactory studies (5-6 points), and unsatisfactory studies (0-4 points).[25] Results

Search results:

The ow diagram of the selection process is shown in gure 1.In total of 5114 potentially relevant articles were found after data base search.One additional citation was found through a

rsonal search, of
his number, 1922 articles were excluded as duplicates by Endnote X8.After title and abstract screening 3041 article were excluded (201 duplicates found manually, 2840 irrelevant).Two unpublished data were included to the 152 text eligible articles to full text screening, in addition we added 2 articles were ad ed manually.A total of 29 articles were therefore reviewed in detail and included in the analysis.The main characteristics of these studies are summarized in table 1.The inter-rater agreement for inclusion was κ=0.95 and for the quality assessment was κ=0.8


Study characteristics

Qualitative synthesis included 29 studies conducted in 15 countries; six studies from India, ve from China, four from Indonesia, one study from each of the following countries (Egypt, Zimbabwe, Nepal, Lao PDR, Ghana, Pakistan, Vietnam, Cambodia, Peru, and Cavite), and two studies from each Uganda, and South Africa.Of included studies there were 5 cohort studies [12], [13], [17], [26] & [27] .One mixed methods study [28], whi e the other 23 studies were cross-sectional.

Male sex presentation ranged from 30% [29], to 77% [18].The sample size ranged from 50 [29], to 1178 [30].The tool used for estimation of the cost survey were either WHO T

44,47], MDR or HIV positive [28,32,35,47], male gender, [26, 27,44], and duration
f hospitalization [13,28,32,44,45].


Coping strategy

In response to balance the enormous nancial burden they encounter, the TB-affected families may adopt some coping strategies.Borrowing money, taking out loans, pledging gold and jewels, bringing their children out of schools or selling assets are options to compensate the income loss and the high out-ofpocke

expenses [37,45].All these approaches are referred to as "dissaving" which is the c
re of the hardship nancing dilemma.


Pooled proportion of catastrophic cost at 20% among different subgroups 1.3.1 Pooled proportion of catastrophic cost at 20% among TB drug sensitive

The pooled proporti

of patients facing catastrophic cost was 39%, 95CI (28-51%), the report
d heterogeneity was 99%.After removing outliers, the pooled proportion of 11 studies recruited 3492 patients dropped to 32%, 95% CI [29 -35].The pooled prevalence of DS-TB patients facing catastrophic costs ranged from 24%, 95%CI [19 -30] in the study of Gadallah,2018 [27] to 42%, 95% CI [35 -49] in the study of Rebecca L.Walctt, 2020 [13].The heterogeneity of the included studies was a follows; I 2 = 70%, P < 0.01.(Table .The proportion of the mean direct cost to the mean total cost addressed in 6 studies, the pooled proportion of direct to total cost at catastrophic cost of 20%

was not calculated as the heterogeneity was high.The proportion was variants, two studies reported similar propor

g, 2018 [37
.Two other extreme values reported, 33% by Fuady 2018 [41] and 65% reported by Muttamba, 2020 [30].


Pooled proportion of direct cost in MDR

The proportion of the mean direct cost to the mean total cost at 20% addressed in 7 studies, ranged from 26% in Chittamany, 2020 [33] to 93% in Yang, 2020 [32].Low proportions were observed in Fuady, 2018 [41], Tomeny, 2020 [15], and Collins Timire, 2020 [47] with proportion of 32%, 34% and 49% respectively, while high proportion also reported in Muttamba, 2020 [30], with 66% and in Nhung, 2018 [37] with 68%.The pooled proportion of mean direct to total cost was di cult to assess because of the heterogeneity which wasn't explained even after a meta-regression performed.

1.1.3Pooled proportion of direct cost to total cost in case of active case nding (ACF)

The pooled proportion of the mean

rect cost to the m
an total cost was addressed in 3 studies, the pooled proportion of mean direct to mean total cost was 25%, 95%CI [16-37%], I 2 =83%.After conducting leave one out sensitivity analysis, the Suman Chandra Gurung, 2019 [39], was removed, the pooled proportion dropped to 29%, 95%C1 [20-41%] I 2 =55%.(Table .2)

1.1.4Pooled proportion of direct cost to total cost in case of passive case nding (PCF)

The pooled proportion of the mean direct cost to the mean total cost addressed in 4 studies [17,18,39,45], the pooled proportion of mean direct to mean total cost was 37%, 95%C1 [31-42%] I 2 =0%.(Table .2)


Proportion of direct cost to total cost in case HIV and TB co-infection

The proportion of the mean direct cost to the mean total cost addressed in 2 studies.Don Mudzengi , 2017[16] and his team showed that the proportion of mean direct cost to the mean total cost was 30% among HIV and TB co-infection patients, while a higher proportion reported in Chittamany, 2020 [33] with 59%.As we couldn't pool the study because of the high un-explained heterogeneity.

The pooled proportion of the mean direct cost to the mean total cost addressed in 14 studies, the pooled proportion of mean direct to mean total cost was 55%, 95%CI [43-66%], I 2 = 99%.After conducting outliers removal, study Mihir P. Rpan, 2020 [46] was excluded, the pooled proportion dropped to 51%, 95%CI

[43-66%], I 2 = 96%.(Table .2)


Catastrophic Health Expenditure at 10% & Capacity to Pay at 40%

In this study, we have found that there are six studies that also calculated the CHE 10% and the CTP 40%, in addition to their results regarding the CTC 20%.


1 Pooled proportion of CHE at 10%:

The pooled proportion of the CHE at 10% were studied also among the studies which they calculated CTC 20%.Three studies [ 2,27,32]


Discussion

Compared to the unknown data on the proportion of TB-patient affected household facing catastrophic cost in 2015, the GDGs goals set that 0% of household affected by TB have faced these costs by 2020 [51].To the best of our knowledge, this is the rst article that pooled of the proportion of TB patients or their households who suffered from catastrophic cost.In this meta-analysis 29 surveys conducted in 22 countries recruiting DS-TB, MDR-TB with or without HIV recruited through ACF, PCF.The quality score of the included studies ranged from 3-10.The proportion of patients facing catastrophic cost at a cut-off point 20% was 43%, (32%, 95%CI [29][30][31][32][33][34][35] among DS and 80% 95%CI [74-85%] among MDR).TB co-infected with HIV faced the highest catastrophic cost 81%, 95%CI [78-84].Catastrophic cost was variables according to the strategy of case nding (ACF 12%95%CI [9-16%], versus PCF 42% 95%CI [35-50%]).The direct cost including medical and non-medical cost represented 51%, 95%CI [43-59%] of the total cost.Among drug sensitive and drug resistant TB, the proportion of direct cost to the total cost ranged from (33-65%) [15,30,33,37,41,47] and (26%-93%) [15,30,32,33,37,41,47] respectively.ACF incurred lower catastrophic than PCF 29%, 95%C1 [20-41%] versus 37%, 95%C1 [34- 0%].The direct cost to the total cost among TB and HIV co-infected patients ranged from 30% [16]-59% [33].The CHE was 50%, 95%CI [47-54%], and 70%, 95%CI [64-76%] at 10% of household yearly income and 40% of their capacity to pay respectively.


Catastrophic cost

In fact, the cost incurred by some patients may be catastrophic and minimal for others.This is based on the household annual income.In the current study, we have included many studies that addressed the catastrophic cost among the TB at different thresholds, points (30%, 25%, 20%, 10% and 5%).Despite absence of robust evidence on the sensitivity of the cut-off point at 20% to re ect the catastrophic cost regardless patients are drug sensitive or resistant.Fuady et al, [12]settled 15% and 30% as more consistent cut-of points for treatment adherence and success respectively.In the current work, the proportion of TBhousehold patients facing catastrophic cost was 39%, which considered very high compared to the targeted GDGs in 2020 (0)%, more efforts and activities need to be directed to reduce this cost.It is worthy to note that diagnosis and treatment are provided for free in many of the included countries under the umbrella pooled of NTP, however, the treatment related expenditure is still very high.Yadav and his group, [52] illustrated that even with free services for tuberculosis care, 21.3% of the people in their study exposed to hardship nancing, advising the need to take into consideration more innovated ways to increase the supported coverage of tuberculosis treatment in the country.The study also suggests the use of hardship nancing as an index to measure the effectiveness of tuberculosis control program in the country.It is crucial to decrease the burden of catastrophic cost among the TB patients as it results in poorer treatment outcome.Patients suffer from catastrophic cost had 2-4 times higher odds of treatment failure than those who do not [12].The latter is due to reduces access to the treating health facility, and treatment completion.Turning to the coping cost, a large proportion of household's resort to different coping strategies to confront the increased out-of-pocket costs; and to compensate the consequences of income loss.Those coping strategies include selling a property or livestock, taking loans, pledging jewels, dropping their child en out of school and cutting down their consumption to below basic needs [7].Despite pooling of these studies' outcome yielded substantial heterogeneity, the current study has found that almost 51% of heterogeneity, was mainly because of two predictors, the rst was that some studies estimated CTC of DS and patients with MDR with or without HIV together.This factor playe

a major role in the hetero
eneity, as it was clear that the CTC was dramatically higher among patients with HIV.The second predictor was the classi cation of country where the study was conducted [23].Two-third of the new cases of TB reported in eight countries of the world, with India foremost the count, followed by Indonesia, China, the Philippines, Pakistan, Nigeria, Bangladesh and South Africa.Consequently, we divided studies into studies conducted in countries with high versus low incidence.In meta-regression, the country,

here the study was conducted was
second major determinant of the different size effect.

The reported high incidence of CTC in many countries raised the need for social protection interventions.The most common social protection intervention is the cash transfer or cash assistance; it has already implemented in many countries across the world either conditionally or unconditionally [53].In such a way, it is supposed that the household can get better access to treatment and food.Other social protection interventions include disability grants, food baskets (food assistance), food or travel vouchers and social insurance [7].Many countries implemented reimbursement programs to help TB patient

to cope with the disease cost.Ho
ever, these programs prioritize poorer and MDR [54].The effect of this intervention is questionable.At a cutoff point of 20%, two studies have applied and calculated a catastrophic cost before and after reimbursement.Lue et al,2020 [42] there reported a slight change on the proportion

f CTC; befo
e reimbursement, the CTC was (22%) and declined to 19% after the reimbursement.In contrary, Fuady,2019, [55] showed a higher change in the proportion of CTC after the reimbursement.The intervention program effectively decreased CTC from 44% to 13%.With regards to cash transfer, Wing eld et al, 2016 [56] reported that the proportion of TB household suffered from CTC was 30% and 42% among intervention and control respectively.These ndings indicate that this social support is not enough to mitigate the impact of TB.Consequently, household of TB patients should receive su cient nancial support that covers the indirect cost (job lost), and direct cost (transportation, food, accommodation) [57].Of note, this social support should be proportionate to the income lost, this is due to the high variability of the pretreatment income.We speculate that development of newer treatment guidelines for TB of shorter duration would be bene cial.At the bottom, provision of free medication is not su cient to prevent the catastrophic cost.TB patients should receive transport vouchers, reimbursement schemes and food assistance to