Scope of the review

The protocol of this systematic review was registered with PROSPERO (CRD42023385627) [10]. This study was reported following the 2020 Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline (S1 PRISMA Checklist) [11]. Differences from the original review protocol are described with rationale (S1 Table).

This systematic literature review focused on 20 diseases selected as NTDs by WHO: Buruli ulcer, Chagas disease, dengue and chikungunya, dracunculiasis (Guinea-worm disease), echinococcosis, foodborne trematodiases, human African trypanosomiasis (sleeping sickness), leishmaniasis, leprosy (Hansen’s disease), lymphatic filariasis, mycetoma, chromoblastomycosis and other deep mycoses, onchocerciasis (river blindness), rabies, scabies and other ectoparasitoses, schistosomiasis, soil-transmitted helminthiases, snakebite envenoming, taeniasis/cysticercosis, trachoma, and yaws and other endemic treponematoses [12].

Outcomes of interest of this systematic review were the prevalence and magnitude of victims who faced financial hardship caused by NTDs, including CHS, impoverishment, and coping strategies.

Search strategy and selection process

We searched three bibliographic databases, PubMed, EMBASE, and EconLit, to identify articles reporting financial hardship among patients suffering from NTDs from any country indexed from database inception to January 1, 2023. We also searched for grey literature in two databases, OpenGrey and EBSCO Open Dissertations. The search terms used were (Disease name and its synonyms) AND (catastroph* OR impoverish* OR coping OR economic consequence* OR out-of-pocket OR "out of pocket" OR ((household OR family OR patient AND (cost* OR spending OR expen*))), that was adapted to match the search techniques of each database. A full search strategy is shown in S2 Table. There was no language restriction applied in this systematic review. A supplemental search was performed by tracking citation and snowballing the eligible articles’ reference list.

Two reviewers (CP and JYC) independently performed the study selection. They screened the titles and abstracts of identified articles from database searches for relevance. Potentially relevant articles were sought for full-text articles. We requested the authors for full-text articles or reports of highly relevant articles without full-text articles, such as conference abstracts. The retrieved full-text articles were selected based on the eligibility criteria. Discrepancies arising during study selection were resolved by discussion with the third reviewer (NC).

Eligibility criteria

We included empirical studies reporting CHS, impoverishment, or coping strategies incurred by NTDs using primary data collection.

Data extraction

We developed a data extraction sheet by performing a pilot test of extracting five randomly selected articles and refining it until finalization. Two reviewers (CP and JYC) independently performed data extraction. Another reviewer (JH) checked the extracted data for correctness. Any discrepancies were resolved by discussion among reviewers.

Study findings and methodological characteristics extracted from the eligible articles are as follows: first author, publication year, NTDs, study setting, study design, sample characteristics, sample size, data collection period, data collection methods, time horizon, a perspective of the analysis, discount rate, costing year, reported currency, cost units, the definition of CHS and impoverishment, prevalence and magnitude of CHS and impoverishment incurred, economic consequences and coping strategies of financial hardship. Corresponding authors of the eligible articles were contacted to request individual patient-level data. However, we received no response.

The financial risk protection metric is intended to capture only the OOP costs for medical services (e.g., treatment and diagnosis costs). However, some studies considered certain types of direct non-medical costs (e.g., transportation, food, and accommodation costs) and indirect costs (e.g., productivity and income losses) when quantifying financial hardship. Some studies also included informal care costs, such as traditional medicine, as OOP costs [6]. Thus, our systematic review categorized costs extracted from the eligible studies as direct costs (OOP costs) and indirect costs. Direct costs were further categorized as direct medical costs and direct non-medical costs. The combination of direct costs and indirect costs was categorized as total costs.

Quality assessment

Two reviewers independently assessed the eligible articles’ quality (CP and JYC). Any discrepancies were resolved by consensus among the reviewers. To the best of our knowledge, there is no risk-of-bias assessment tool for economic burden studies. Hence, we assessed the quality of the eligible articles using the cost-of-illness evaluation checklist by Larg and Moss [13].

Data synthesis

A narrative synthesis was performed to summarize study findings, methodological characteristics, and the quality of the eligible studies. The identified countries were categorized based on the World Bank’s income levels and regions [14].

Statistical analysis

We performed meta-analyses to calculate the pooled prevalence of households experiencing financial hardship. However, this was possible only for studies that quantified financial hardship using the same measurement definition for a particular NTD. For example, we performed a meta-analysis to calculate the pooled prevalence of households experiencing CHS due to visceral leishmaniasis based on two studies that defined CHS as direct costs exceeding 10% of annual household income [8,15]. The remaining studies were not meta-analyzed due to the differences in the definitions of CHS. We estimated the pooled prevalence of CHS and 95% confidence intervals (CI) using a random-effects model under the DerSimonian and Laird approach [16]. Effect sizes were computed using each study’s Freeman–Tukey double-arcsine-transformed proportion. This variance-stabilizing transformation is particularly preferable when the proportions are close to 0 or 1 [17]. p < .05 was considered statistically significant in 2-sided tests.

Heterogeneity was evaluated by observing the forest plots and using the I² statistic that estimated the proportion of variability in a meta-analysis that is explained by differences between the included trials rather than by sampling error. Subgroup analyses were performed to explore possible causes of heterogeneity among study results. Publication bias was assessed using the funnel plot asymmetry test and the Egger regression asymmetry test [18]. Statistical analyses were conducted using Stata version 18.0 (Stata Corporation).

Patient and public involvement

Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.

Overall characteristics of the included studies

A total of 1,768 articles were identified from the search, of which 10 studies were included (Fig 1) [8,15,19–26]. A list of excluded studies with reasons is presented in S3 Table. These studies quantified financial hardship among 2,761 patients in five LMICs (India, Nepal, Nigeria, Sudan, and Vietnam) who had been diagnosed with six out of the WHO’s 20 NTDs, including Buruli ulcer [20], chikungunya [21,26], dengue [22], visceral leishmaniasis [8,15,25], leprosy [19,23], and lymphatic filariasis [24]. Table 1 provides a summary of the study characteristics. We found no major concern in the quality of the included studies (S4 Table)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Study selection flow.

https://doi.org/10.1371/journal.pntd.0012086.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of studies assessing financial hardship.

https://doi.org/10.1371/journal.pntd.0012086.t001

Financial hardship caused by NTDs was quantified as CHS (10 studies) [8,15,19–26], and impoverishment (1 study) [8]. All studies were conducted in LMICs with a focus on South Asia (7 studies) [8,19,21,23–26], Sub-Saharan Africa (2 studies) [15,20], East Asia & Pacific (1 study) [22]. Patients were mostly identified using a hospital-based approach (7 studies) [8,15,19,20,22,23,25], with active case-finding intervention implemented in two of those studies [20,23]. Five studies reported that patients sought informal healthcare, such as traditional healers, ayurveda, and homeopathy [19–21,25,26].

Costs captured in the financial hardship were direct medical costs (10 studies, 100%) [8,15,19–26], direct non-medical costs (9 studies, 90%) [8,15,19–21,23–26], and indirect costs (7 studies, 70%) [8,15,19,21,23,25,26], as summarized in Table 2. These costs were captured with a different timeframe, including during a disease episode [8,15,20,21,25,26], during hospitalization in an intensive care unit [22], monthly costs with a maximum recall period of 3 years [19], per one outpatient visit in the last 6 months [23], and per one hospitalization episode in the last year and per one outpatient visit in the last 15 days [24]. Abandonment of healthcare seeking due to financial burden was not reflected in the reported costs as the included studies captured only patients who sought healthcare.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Financial hardship among patients suffering from neglected tropical diseases.

https://doi.org/10.1371/journal.pntd.0012086.t002

The health insurance systems or special programs covered some of the costs. The costs for diagnosis and treatment of visceral leishmaniasis were provided free of charge to patients under the publicly financed health insurance system in Nepal [8,25] and Sudan [15]. In Nigeria, international development partners funded a special program that provided free diagnosis and treatment of Buruli ulcers, as well as accommodation, school funding, and basic allowance [20]. Additionally, the Indian government had a special program that provides financial assistance to families of patients affected by leprosy [19]. However, patients in India had to pay high OOP costs for medical services for leprosy [19,23], chikungunya [21,26], and lymphatic filariasis [24]. Similarly, patients in Vietnam also paid high OOP costs for the medical treatment of dengue [22]. For more details, refer to Table 3.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Details of costs incurred from neglected tropical diseases.

https://doi.org/10.1371/journal.pntd.0012086.t003

Financial hardship among patients suffering from NTDs

Catastrophic health spending.

CHS was variedly defined across studies in terms of types of costs (medical costs, medical and transportation costs, direct costs, indirect costs, or total costs), thresholds (5%, 10%, 15%, 25%, 30%, 40%, or 100%), timeframe (monthly, quarterly, or annual), household resources (income, consumption expenditure, national average annual household expenditure, or international poverty line) and perspective (household or individual). All studies used the budget share approach to quantify CHS. The most commonly used definitions of CHS caused by NTDs were direct costs of a disease episode exceeding 10% of annual household income (3 studies) [8,15,20] and total costs of a disease episode exceeding 10% of annual household income (3 studies) [8,15,25]. CHS that included only the direct medical costs was reported in two studies [8,22].

We summarized the prevalence of households experiencing CHS and the magnitude of CHS, determined as the percentage of the costs of NTDs as a share of income, in Table 4. The prevalence and magnitude of CHS varied depending on the definitions of CHS, disease duration (episodic or chronic), and thresholds used (≤10% or >10%). Overall, the direct costs of NTDs resulted in a wide range of households experiencing CHS. CHS was generally low among patients with leprosy (0.0–11.0%) [19,23], dengue (12.5%) [22], and lymphatic filariasis (0.0–23.0%) [24], and relatively high among patients with Buruli ulcers (45.6%) [20]. CHS varied widely among patients with chikungunya (11.9–99.3%) [21,26] and visceral leishmaniasis (24.6–91.8%) [8,15,25].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Summary of prevalence and magnitude of catastrophic health spending.

https://doi.org/10.1371/journal.pntd.0012086.t004

Meta-analyses were performed to pool the prevalence of CHS in studies reporting CHS using the same measurement definition in a particular CHS. This was only possible for visceral leishmaniasis, in which CHS was quantified as direct costs of a disease episode exceeding 10% of annual household income in two studies [8,15], and total costs exceeding 10% of annual household income in three studies [8,15,25].

The pooled prevalence of CHS, defined as direct costs exceeding 10% of annual household income, was 73% (95% CI; 65–80%, n = 2, I² = 0.00%), as shown in Fig 2A. Egger’s test (P = 0.80) indicated no evidence of small-study effects. Visual inspection of the funnel plot indicated no evidence of publication bias (S1A Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Meta-analyses of a prevalence of households experiencing catastrophic health spending due to visceral leishmaniasis.

https://doi.org/10.1371/journal.pntd.0012086.g002

The pooled prevalence of CHS, defined as total costs exceeding 10% of annual household income, was 74% (95% CI; 49–93%, n = 3, I² = 94.72%), as shown in S2 Fig. We explored the source of heterogeneity by visual inspection of the forest plot. We found that the source of heterogeneity was the differences in the treatment of visceral leishmaniasis, where sodium stibogluconate was used in two studies [8,15], and miltefosine in one study [25]. Therefore, we performed a subgroup meta-analysis based on different treatments, as shown in Fig 2B. We removed one study [25] from the meta-analysis to investigate the publication bias without the presence of heterogeneity. Egger’s test (P = 0.81) indicated no evidence of small-study effects. Visual inspection of the funnel plot indicated no evidence of publication bias (S1B Fig).

Coping strategies

Four studies reported coping strategies used by patients to pay the costs of NTDs. These strategies included using savings (71–100% of patients), taking out loans (32–80%), selling livestock or other assets (17–32%), or borrowing money (0–23%), as shown in Table 2. However, these studies did not distinguish between coping strategies used by patients who experienced CHS and those who did not [8,19,24,25].

Cost drivers and determinants of financial hardship

To understand the cost drivers of financial hardship caused by NTDs, we analyzed the percentage share of types of costs captured in the direct costs. The findings are presented in Fig 3. Direct medical costs were the primary cost driver in nine studies [8,19–21,23–26]. However, one study identified food and transportation costs as the main cost drivers [15].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Cost drivers of out-of-pocket costs.

Abbreviation: ENL–erythema nodosum leprosum. Tripathy et al, 2020 [24]; Tiwari et al, 2018 [23]; Chandler et al, 2015 [19]; Uranw et al, 2013 [25], Meheus et al, 2013 [15], Adhikari et al, 2009 [8], McBride et al, 2019[22], Vijayakumar et al, 2013 [26], Gopalan et al, 2009 [21], Chukwu et al, 2017 [20].

https://doi.org/10.1371/journal.pntd.0012086.g003

Determinants of CHS were assessed in one study among patients with Buruli ulcers. The study concluded that neither age, gender, rural/urban location, education, occupation, religion, nor patient income group was a determinant of CHS [20]. There was no study investigating determinants of impoverishment.