The Economic Burden of Non-Typhoidal Salmonella and Invasive Non-Typhoidal Salmonella Infection: A Systematic Literature Review

Non-typhoidal Salmonella (NTS) infection and invasive non-typhoidal Salmonella (iNTS) infection cause a significant global health and economic burden. This systematic review aims to investigate the reported economic burden of NTS and iNTS infection, identify research gaps, and suggest future research directions. Data from PubMed and Embase databases up to April 2022 were reviewed, and articles were screened based on predefined criteria. Cost data were extracted, categorized into direct medical costs (DMCs), direct non-medical costs (DNMCs), and indirect costs (ICs), and converted into US dollars (year 2022). Data primarily originated from high-income countries (37 out of 38), with limited representation from Africa and resource-limited settings. For inpatients, DMCs were the primary cost driver for both NTS and iNTS illnesses, with estimates ranging from USD 545.9 (Taiwan, a region of China) to USD 21,179.8 (Türkiye) for NTS and from USD 1973.1 (Taiwan, a region of China) to USD 32,507.5 (United States of America) for iNTS per case. DNMCs and ICs varied widely across studies. Although study quality improved over time, methodological differences persisted. This review underscores the lack of economic data on NTS and iNTS in resource-limited settings. It also highlights the need for economic burden data in resource-limited settings and a standardized approach to generate global datasets, which is critical for informing policy decisions, especially regarding future vaccines.

Various prevention strategies have been implemented to control NTS disease, targeting animal production, food industries, consumers, and national surveillance [18,19].Despite these efforts, significant under-reporting persists due to barriers to healthcare access and insufficient attention from public health authorities at national and global levels [1,2,[20][21][22].Additionally, the lack of point-of-care biomarkers and rapid diagnostic tools, such as serological or PCR-based tests, make the clinical diagnosis of NTS and iNTS infection challenging, particularly in low-resource settings such as Africa, where blood culture facilities are limited [11,23].The absence of an available vaccine against NTS infections exacerbates these challenges, contributing to a significant public health burden.
No systematic attempt has been made to estimate the per-case cost incurred due to NTS and iNTS diseases [24,25].This review aims to examine the existing studies on the economic burden of NTS and iNTS infections to date, identify gaps, and suggest future research needs.

Materials and Methods
A systematic literature review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines and its checklist (Supplementary Materials; Table S1) [26].PubMed and Embase were searched using search strategies constructed with free-texts, MeSH-Terms, and Emtree terms, adapted for each database.The search included publications from inception to April 2022, with the last search conducted on 29 April 2022.Two categories of search terms were used: (1) disease category, including terms related to non-typhoidal Salmonella and invasive non-typhoidal Salmonella, and (2) economic burden category, including terms related to the cost of illness.All search terms within each of the two categories were combined using the "OR" operator, and the two categories were combined using the "AND" operator.No filters or limits were applied, and there were no language restrictions.A detailed list of search strategies is presented in Table 1.Duplicate articles were identified and removed using EndNote software.The initial screening of titles and abstracts was conducted independently by two reviewers (SK, HLK).Studies indicating the cost of NTS or iNTS were selected for a second screening.After the initial screening, the list of studies for full-text review was shared, and discrepancies were resolved through discussion.The two reviewers independently carried out the full-text review of 77 articles using pre-determined inclusion and exclusion criteria (Table 2).The reference lists of the selected studies were further reviewed by the two reviewers.Any inconsistencies between the reviewers were resolved with the assistance of a third researcher (JSL).
The quality of the selected articles was assessed using a quality assessment tool (Supplementary Materials; Table S2) adapted from the cost-of-illness evaluation checklist proposed by Larg and Moss (2011) [27].The results of the quality assessment (high, medium, and low) are presented in Table S3.

■
The study includes the direct medical costs OR direct non-medical costs OR indirect costs of NTS-related treatment/healthcare source utilization.

■
The study includes costs incurred due to NTS disease from primary data or secondary data.

■
The study estimates the costs incurred due to NTS disease by modelling.

■
The costs are related to animal disease or food control.

■
The study includes costs incurred due to NTS with other diseases combined such as other foodborne diseases, typhoid, and paratyphoid (costs not specific to NTS).
■ Full text is written in languages other than English.
Relevant data were extracted from each included study using a standardized template in a Microsoft Excel workbook.Data extraction was initially performed by SK and HLK and cross-checked for accuracy.The extracted variables included the title of the article, authors' names, publication year, cost year (study period), study location, disease indication, serovars, currency of the cost, cost item and description, cost type, cost perspective, cost source, duration of illness (number of visits), and other cost-related information identified in the articles.Cost types were categorized into direct medical costs (DMCs), direct nonmedical costs (DNMCs), and indirect costs (ICs).DMC refers to costs incurred by patient management such as hospital stays, physician consultations, laboratory tests, and medications.DMC was further classified into DMC inpatient (DMC IP) and DMC outpatient (DMC OP).When DMCs were not specific to costs by inpatient or outpatient healthcare services, they were categorized as DMC non-specific (DMC NS).DNMC includes costs not directly related to healthcare services, such as transportation, food, and accommodation.IC encompasses expenses borne by patients or their families due to work absenteeism related to the illness.Data costs related to outbreak control and the value of lost lives were extracted but not included in the review results.All cost components were calculated as per-case costs whenever possible, by dividing aggregated costs for a group of patients by the number of patients.All costs were converted to US dollars (USD) and adjusted to 2022 values using the official exchange rate and the GDP Deflator provided by the World Bank [28].

Results
The literature search yielded 2839 articles after removing duplicates, of which 77 were identified for full-text review based on their titles and abstracts.Among these 77 articles, 34 met the inclusion and exclusion criteria.Additionally, three articles [29][30][31] and one database [32] were identified through bibliography searches and included in the final review (Figure 1).
Among the 38 studies, 4 articles contained data related to the economic burden of iNTS [30,[33][34][35].Evidence for NTS disease came from 11 countries: Australia, Canada, Hong Kong SAR, China (hereinafter Hong Kong), the Netherlands, Poland, Spain, Sweden, Taiwan, a region of China (hereinafter Taiwan), Türkiye, the United Kingdom (UK), and the United States of America (US).Evidence for iNTS was found in three countries: Spain, Taiwan, and the US.The income level of each country, classified by the World Bank (2021), was compared to assess the distribution of data [36].Only one study [37] was conducted in an upper middle-income country (Türkiye) in 2009, while all other studies were from high-income countries.There were no studies from lower middle-income and low-income countries for both NTS and iNTS diseases.Geographically, no studies were conducted in Africa, and only two studies were conducted in Asia (Taiwan [5,35] and Hong Kong [38]).
Publication years ranged from 1978 to 2021.All studies used economic evaluation methods from various perspectives, such as societal, the healthcare system, and the patient.More than half (n = 25, 66%) of the studies used a societal perspective, encompassing DMC, DNMC, and IC.Eight [5,35,37,[39][40][41][42][43]] out of thirty-eight studies used the healthcare system perspective, focusing on government and healthcare facility costs.Five studies used the patient perspective, which focused on the medical expenses incurred by the patient and their families, excluding indirect costs [29,33,34,38,44].
Among the 38 studies reviewed, 13 aimed to estimate the economic burden of NTS outbreaks [20,22,37,[39][40][41][45][46][47][48][49][50][51], of which five were in hospital settings [37,[39][40][41]47].Todd et al. estimated the economic burden of multiple Salmonella community outbreaks in the UK, US, Canada, Sweden, and Australia across different years [20], and Scharff et al. did so in the US from 1994 to 2009, using national surveillance data [51].None of the outbreak studies included iNTS cases.Economic evidence related to iNTS was derived from either hospital databases (US and Spain) or national claims databases (US and Taiwan) [30,[33][34][35].A summary of the extracted components from the included studies is provided in Table S3 in Supplementary Materials.Among the 38 studies, 4 articles contained data related to the economic burden o iNTS [30,[33][34][35].Evidence for NTS disease came from 11 countries: Australia, Canada Hong Kong SAR, China (hereinafter Hong Kong), the Netherlands, Poland, Spain, Swe den, Taiwan, a region of China (hereinafter Taiwan), Türkiye, the United Kingdom (UK) and the United States of America (US).Evidence for iNTS was found in three countries Spain, Taiwan, and the US.The income level of each country, classified by the World Ban (2021), was compared to assess the distribution of data [36].Only one study [37] was con ducted in an upper middle-income country (Türkiye) in 2009, while all other studies wer from high-income countries.There were no studies from lower middle-income and low income countries for both NTS and iNTS diseases.Geographically, no studies were con ducted in Africa, and only two studies were conducted in Asia (Taiwan [5,35] and Hon Kong [38]).
Publication years ranged from 1978 to 2021.All studies used economic evaluation methods from various perspectives, such as societal, the healthcare system, and the pa tient.More than half (n = 25, 66%) of the studies used a societal perspective, encompassin DMC, DNMC, and IC.Eight [5,35,37,[39][40][41][42][43]] out of thirty-eight studies used the healthcar system perspective, focusing on government and healthcare facility costs.Five studie used the patient perspective, which focused on the medical expenses incurred by the pa tient and their families, excluding indirect costs [29,33,34,38,44].
Among the 38 studies reviewed, 13 aimed to estimate the economic burden of NT outbreaks [20,22,37  The reported DMC IP was the primary cost driver for the economic burden of NTS and iNTS illnesses, ranging from USD 545.9 (Taiwan) to USD 21,179.8(Türkiye) for NTS illnesses and from USD 1973.1 (Taiwan) to USD 32,507.5 (US) for iNTS illnesses per case.Considering the significant inflation surge in Türkiye in 2022, around 72% [36], the second highest DMC IP for NTS illness was USD 19,788.7 from the US.The DMC OP reported was lower than the DMC IP, ranging from USD 17.4 (UK) to USD 739.8 (US) for NTS illnesses, with only one study from the US [30] estimating DMC OP for iNTS as USD 469.4 per case.The IC ranged from USD 188.9 in Australia to USD 3028.8 in the UK per case for NTS illnesses.The IC for iNTS illnesses, as reported by Adhikari et al. (2001) from the US, was USD 1768.8 [30].DNMCs were assessed in 14 studies, ranging from USD 4.3 in the Netherlands to USD 18,261.2 in Türkiye [20,31,37,40,41,[45][46][47]49,[51][52][53][54][55].Five studies (13%) did not report costs by components but aggregated total costs [19,39,50,56,57].Costs reported from each study are presented in Table 3 (NTS illnesses) and Table 4 (iNTS illnesses).When costs could not be categorized into DMC, DNMC, and IC, total costs were presented as total costs (TC) in the tables (all costs are expressed in US dollars at 2022 value).DMC r = a reported value retrieved from the literature as the mean of direct medical costs per case, which has been adjusted to USD in the year 2022 value.DMC c = a calculated value as the mean of DMC per case (the total amount of DMC incurred divided by the total number of cases).DMCs IP = direct medical costs incurred from inpatient healthcare.DMCs OP = direct medical costs incurred from outpatient healthcare.DMCs NS (not specified) = costs are direct medical costs that are not categorized into direct medical costs for inpatient or outpatient healthcare services.DNMCs = direct non-medical costs.ICs = indirect costs.TCs = total costs.* Studies including cost data from more than two countries.† Costs using a different currency from one of the original countries (i.e., USD reported from Taiwan) were inflated using the original countries' GDP deflator.§ GDP per capita (current USD, 2022) from the World Bank [28].‡ Percentage of Canada's GDP per capita (current USD) in 2022.++ Percentage of China's GDP per capita (current USD) in 2022.
To better understand the economic burden relative to each country's economic context, we calculated the healthcare costs as a percentage of the country's GDP per capita [28].For NTS illnesses, the DMCs IP as a percentage of GDP per capita ranged from 0.3% (UK) to 198.4% (Türkiye) (Table 3).This significant variation indicates the differing economic impacts on patients in different countries, with Türkiye experiencing a disproportionately high economic burden due to its lower GDP per capita and exceptionally high inflation rate in 2022 [36].For iNTS illnesses, the DMCs IP as a percentage of GDP per capita ranged from 0.6% (US) to 42.6% (US) (Table 4).The high percentages for iNTS reflect the severe financial impact of invasive infections, particularly in high-income countries like the US, where healthcare costs are substantial.Countries with national healthcare systems, such as Taiwan and the UK, generally report lower out-of-pocket expenses for patients compared to those without such systems, impacting the overall economic burden experienced by individuals.Regarding the study quality of the 37 articles, 20 studies were rated as high quality [5,22,30,35,37,38,[42][43][44]49,50,[53][54][55]57,58,[60][61][62][63], 12 as medium quality [29,31,33,34,41,[46][47][48]51,52,56,59], and 5 as low quality [19,20,39,40,45].Study quality was assessed using a set of questions regarding the analytical framework, methodology and data, and analysis and reporting [27].The quality assessment tool was not applicable to the one database established by the Economic Research Service (ERS) at the U.S. Department of Agriculture (USDA) due to its structure [32].Detailed information is available in Supplementary Materials (Table S2).It is noteworthy that the quality of cost-of-illness studies has improved during a span of 44 years (1978-2021), particularly in analysis and reporting, despite methods still being varied among the recent studies.
Data on the duration of illness were also collected, as shown in Table 5. Hayes et al. reported the number of outpatient visits for NTS illness management to be 1.6 per case [46].No other study clearly indicated the number of outpatient visits related to NTS illness management, and none did so for iNTS illness.Hospital bed-days or length of hospitalization was reported to be 7.5 days (95% Confidence Interval 6.0-8.9) in 16 studies for NTS illness per case and 12.6 days (95% CI 8.6-16.5) in three studies for iNTS illness per case.Roberts and Sockett, Herrick et al., and Ailes et al. presented the duration of NTS illness to be 12, 2, and 4 days (median) per case, respectively, without specifying the type of healthcare resource (inpatient or outpatient) [22,49,56].

Discussion
At the global level, the economic burden of NTS and iNTS is predominantly informed by data from a limited number of high-income countries, mainly the US and the UK.Among the 37 articles and 1 database reviewed, 15 included data from the US and 10 from the UK.Notably, no economic evaluation was conducted in the African region, despite the high burden of NTS disease in sub-Saharan Africa [2,11,17,23].Additionally, there was no evidence from any lower middle-income or low-income countries, highlighting the scarcity of cost data for NTS and iNTS diseases in resource-limited settings, as previously reported [13,14,23].Direct comparisons of costs across studies are challenging due to the varied approaches and definitions of the costs used.Furthermore, several studies did not separately describe cost categories (DMC, DNMC, IC) and reported combined total costs instead [19,39,50,56,57].
This study has limitations.Although we employed a systematic approach to identifying literature relevant to the economic burden of NTS and iNTS diseases by searching multiple databases with different search terms, there remains a possibility of missing some literature.Costs associated with iNTS diseases may not have been separately estimated and reported if the study did not specifically aim to do so.Additionally, the diverse methods and outcome measures of cost-of-illness studies made it infeasible to conduct a meta-analysis.
To understand the global economic burden of NTS and iNTS, a consensus on standardized methods of economic evaluation across nations is needed to generate reliable data.Efforts to improve diagnostics, understand disease sources, and study routes of transmission should be combined with high-quality primary data to assess the economic burden of NTS and iNTS.A comprehensive understanding of the economic burden is essential for making evidence-based policy decisions at national and global levels.

Conclusions
The findings of this review highlight the significant lack of economic evidence for NTS and iNTS in the regions most in need, particularly Africa.This disparity underscores the disproportionate attention and resources concentrated in high-income countries, especially within a limited number of nations.Currently, there are a few Salmonella combination vaccine candidates in early clinical trial phases, including bivalent NTS vaccines (GSK Vaccines Institute for Global Health (GVGH), Boston Children's Hospital) and trivalent NTS and typhoid vaccines (University of Maryland and Bharat Biotechnology, GVGH, SK Bioscience and International Vaccine Institute) [64].McLennan et al. emphasized that a vaccine-related strategy with multivalent Salmonella vaccines for control should be applicable in LMIC settings, considering the lack of diagnostic capacity and adequate treatment options [64].To support the development of NTS and iNTS vaccines by informing investments and policy decisions, a better understanding of the economic burden of NTS and iNTS diseases, especially in endemic regions, is essential.Public health communities must make concerted efforts to identify and address the substantial burden of NTS and iNTS diseases globally.

Table 2 .
Inclusion and exclusion criteria.

Table 3 .
Healthcare costs by cost type related to non-typhoidal Salmonella diseases (USD in 2022).

Table 4 .
[28]thcare costs by cost type related to invasive non-typhoidal Salmonella diseases (USD in 2022).= a calculated value as the mean of DMC per case (the total amount of DMC incurred divided by the total number of cases).DMCs IP = direct medical costs incurred from inpatient healthcare.DMCs OP = direct medical costs incurred from outpatient healthcare.ICs = indirect costs.†Costsusinga different currency from one of the original countries (i.e., USD reported from Taiwan) were inflated using the original countries' GDP deflator.§GDPpercapita (current USD, 2022) from the World Bank[28].++ Percentage of China's GDP per capita (current USD) in 2022. c

Table 5 .
Number of outpatient visits and hospital bed-days by indication.