Neighborhood-level income and MRSA infection risk in the USA: systematic review and meta-analysis

Abstract

Background

The impact of neighborhood-level income on community-associated methicillin-resistant S. aureus (CA-MRSA) risk remains poorly understood, despite established associations between MRSA risk and the social determinants of health. There are conflicting findings in the existing literature and no known systematic reviews based in the U.S. Our objective was to conduct a systematic review and meta-analysis of the association between neighborhood-level income and CA-MRSA in the U.S.

Methods

We searched MEDLINE (Ovid), MEDLINE Epub Ahead of Print, In-Process, In-Data-Review & Other Non-Indexed Citations, and Daily (Ovid), Global Health (Ovid), Embase (Elsevier), Cochrane Database of Systematic Reviews (Wiley), Cochrane Central Register of Controlled Trials (Wiley), and Web of Science Core Collection from 2017 to 10 January 2021. An updated search was completed in November 2023. Eligible studies reported stratified CA-MRSA case counts and/or effect measures by neighborhood income level, reported as a categorical or continuous variable. Relevant data were extracted using Covidence following the PRISMA guidelines. A random-effects model meta-analysis was used to estimate the pooled effect measure. Three study design-specific risk of bias assessments and a quality assessment were applied using the modified Newcastle-Ottawa Quality Assessment Scale and GRADE approach, respectively.

Results

Six publications met eligibility criteria. Five found that living in a low-income neighborhood was associated with increased CA-MRSA risk. Among the four studies eligible for the meta-analysis, the pooled odds ratio for CA-MRSA infection among low vs. high-income neighborhoods (reference group) was 1.28 (95% CI: 1.13, 1.46), with statistical heterogeneity (I² 73%). Limiting to low risk of bias studies (n = 3), there was no significant relationship between low income and CA-MRSA infection (OR: 1.13, 95% CI: 0.96, 1.33) with heterogeneity of 0%.

Conclusions

Evidence supports an association between lower neighborhood income and higher CA-MRSA infection risk, albeit with considerable heterogeneity. Future studies should consider evaluating neighborhood-level income as a continuous variable, and at the block-group level to avoid exposure misclassification. Furthermore, researchers should consider adjusting for covariates that could allow for a causal interpretation of the relationship between low neighborhood-level income and CA-MRSA risk.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-22275-5.

Background

Community-acquired methicillin-resistant Staphylococcus aureus a (CA-MRSA) emerged as a global threat in the 1990s and is responsible for up to 60% of skin and soft tissue (SSTIs) seen in emergency departments in the United States (U.S.) [1] More than 15,000 severe/invasive CA-MRSA infections including pneumonia, bloodstream infections, septic thrombophlebitis, and necrotizing fasciitis [2], and more than 1,000 associated deaths occur each year in the U.S [3]. Previous studies have suggested various predisposing risk factors for CA-MRSA infections, many of which are intertwined with socioeconomic factors. These include crowded/unsanitary living conditions, incarceration history, antibiotic use, injection drug use, smoking, history of trauma, congestive heart failure, HIV, and obesity [4–6]. Socioeconomically disadvantaged and minority groups including African American, Native American, Pacific Islander, and Alaskan Native populations appear to be disproportionately impacted by CA-MRSA [4].

While individual-level socioeconomic status (SES) has been implicated as a risk factor for CA-MRSA [7–9], the specific impact of neighborhood-level income remains understudied. In the U.S., neighborhood-level income level can influence exposure to infectious agents through diverse pathways related to living conditions [10], healthcare access [10–12], and health behaviors (i.e. physical activity, tobacco use, nutrition) [13, 14]. Importantly, because neighborhood-level income is strongly correlated with multidimensional SES indices, this exposure has been proposed as a singular indicator of area-level deprivation that may be broadly applicable in different geographic areas [15]. Despite the disparate burden of CA-MRSA infections among disadvantaged and minority groups, there is a gap in understanding how neighborhood-level income contributes to this risk. This gap limits the breadth of interventions that might be designed to address this issue and highlights the need for further investigation.

To date, US studies investigating associations between neighborhood-level income and higher CA-MRSA risk have found inconsistent results. For example, a 2006 surveillance study in Brooklyn, New York reported that community-acquired MRSA infections were more prevalent in hospitals serving populations with lower median household income [16]. Conversely, a 2012 study of all New York City boroughs reported that neighborhood-level income was not a significant predictor of CA-MRSA hospitalization [17]. In Atlanta, Georgia a study found that individuals with CA-MRSA SSTI were more likely to have an income <$20,000 compared to non-CA-MRSA SSTI [18]. Additionally, a U.S. study using three different types of income measurements (i.e., low-income households, persons living under the poverty level, and degree of income inequality) found that all three were associated with a higher incidence of CA-MRSA, while high-income households were associated with lower CA-MRSA infection incidence [3]. The conflicting findings across these studies underscore the need for a systematic review and meta-analysis (SRMA) to clarify the relationship between neighborhood-level income and CA-MRSA risk. While some studies have found a clear association between lower neighborhood income and higher CA-MRSA risk [3, 16, 18] others have reported no significant relationship [17]. A comprehensive SRMA could help reconcile these discrepancies and assess the consistency of results across different studies. Therefore, our objective was to investigate the impact of neighborhood-level income on individuals’ risk of community-acquired MRSA infections in the U.S., to inform the development of future interventions to mitigate its spread and reduce health disparities. We chose to focus on MRSA infections, rather than MRSA colonization more broadly, as MRSA infections have clear public health implications while MRSA colonization often does not result in symptomatic infection.

Methods

This SRMA builds off a previously published scoping review [19] on the association between SES and individuals’ risk for colonization or infection with pathogens that are increasingly antimicrobial-resistant (AMR), including S. aureus. The protocol for this SRMA was registered in the PROSPERO database (https://www.crd.york.ac.uk/prospero/) as CRD42024515938.

Search strategy and eligibility criteria

For the aforementioned scoping review, study authors conducted a comprehensive search of the scientific literature in MEDLINE (Ovid), MEDLINE Epub Ahead of Print, In-Process, In-Data-Review & Other Non-Indexed Citations, and Daily (Ovid), Global Health (Ovid), Embase (Elsevier), Cochrane Database of Systematic Reviews (Wiley), Cochrane Central Register of Controlled Trials (Wiley), and Web of Science Core Collection for eligible studies that reported race, ethnicity, or SES for individuals infected/colonized with seven pathogens of interest, including S. aureus. A librarian from Tufts University School of Medicine translated the MEDLINE search strategy (Table S1) for each of the listed databases, and all databases were searched from inception through January 2022, except for MEDLINE Epub Ahead of Print, In-Process, In-Data-Review & Other Non-Indexed Citations and Daily for which the search covered 2017 through 10 January 2021. References were collected and deduplicated before export to Covidence for screening. An updated search with a slightly modified search strategy was completed in November 2023 (See Supplementary Material), through which new references were collected and included for screening. Studies were eligible for inclusion here if they reported CA-MRSA case counts or effect measures stratified by neighborhood income level, regardless of subjects’ age or gender. For this SRMA, we also searched the included papers’ reference sections for additional studies related to the research question.

Screening, data extraction, and synthesis

Two independent reviewers screened titles, abstracts, and full text, and extracted data in Covidence following PRISMA guidelines. Conflicts between reviewers were resolved by a third reviewer. A customized extraction form was created in Covidence to capture relevant study data from eligible studies. The data was then exported for further analysis. Extracted data included: study design, author, year, country, inclusion criteria, exclusion criteria, study definition of community-acquired, method of recruitment, total study population, funding source, method of exposure measurement, type of CA-MRSA-infection (e.g. SSTI), total count of patients with and without CA-MRSA (or MSSA) infections living in low-income neighborhoods, total count of patients with and without CA-MRSA (or MSSA) infections living in high-income neighborhoods, adjusted effect measure by income level (e.g., odds ratio, rate ratio), and the confidence interval. If studies reported raw numbers for more than two income levels, then the levels were collapsed to high versus low using the median as the cut-off. For studies that did not report raw numbers but reported effect measures for each income level, the effect measures for the lowest and highest income levels were extracted. For papers that did not report an effect measure, the raw totals for each income group were extracted to calculate effect measures for studies. Extracted data from all included studies are summarized in tables and figures using R version 4.3.1²⁰ and ggplot2 package [21].

Quality assessments

To assess the risk of bias (ROB) a modified Newcastle-Ottawa Quality Assessment Scale (NOS) was used by author SBl in collaboration with co-author EA to create three quality assessments with three levels each (low, moderate, high bias). Assessments for cohort, case-control, and cross-sectional studies were adapted to better fit this systematic review, as suggested by NOS developers [22–25]. Each study was evaluated using the adapted assessment appropriate for their design (Tables S2-S4), each of which included three domains with 1–4 multiple questions (Table 1) (Table S5). Two reviewers (SBl and co-author SBa) independently completed each paper’s risk of bias assessment. Each indicator had a comment box for the two reviewers to record justifications, notes, or questions. SBl completed consensus, and total scores in each domain were compared between reviewers and categorized as high, moderate, or low ROB [26]. The quality of evidence was assessed using a modified GRADE approach for observational studies [27–30]. Two reviewers (SBl and EA) independently rated the methodological quality, inconsistency across studies, imprecision, indirectness of the evidence/applicability. Any disagreements were resolved through consensus. Each domain was graded as not serious, serious, or very serious depending on criteria outlined in the modified GRADE approach for observational studies (Table S6).

Table 1.

Characteristics of studies included in a systematic review of neighborhood-level income and the risk for CA-MRSA infection

Authors	U.S. Region	Study Design	Study Duration	Study Population	Exposure Measurement	Outcome Ascertainment	Outcome	Number of Participants (n)	Funding Source
Ali et al., 2019	Southeast3	Case Control	8 years	Pediatric	Poverty-to-Income Ratio1 calculated by dividing a family’s income by the poverty threshold. Calculated at the block group level based on where a patient resided at the time of the hospital visit	Patient Medical Record	S. aureus resistant to methicillin	39,371	Academia; Federal
Beresin et al., 2017	Midwest4	Retrospective Cohort	3 years, 6 months	Adult & Pediatric	Median household income2 calcuated at the Zip Code level using ACS 5-year estimates from 2007-2011, obtained through the American FactFinder repository	Patient Medical Record	S. aureus resistant to methicillin	215,218	Federal; Non-profit
Farr et al., 2013	Northeast5	Cross sectional	1 year	Adult & Pediatric	Zip Codes were aggregated into 42 New York City United Hospital Fund (UHF) neighborhoods, and the average median household income at the Zip Code leveli was calculated for each neighborhood using 2000 US Census Data	Patient Medical Record	MRSA hospitalizations as the outcome, which included hospitalizations with diagnosis codes such as S. aureus pneumonia, S. aureus septicemia, or S. aureus infection	645	None
Immergluck et al., 2019	Southeast3	Retrospective Cohort	8 years	Pediatric	Proportion below poverty was calculated at the block group level using ACS data. Poverty to-Income Ratio2 was calculated by dividing 2010 median family income at the block group level by the 2010 poverty threshold for a family of four (both derived from ACS). Block groups were then designated as very low-income, low-income, moderate-income, or high-income	Patient Medical Record	S. aureus SSTI resistant to methicillin	10,642	Academia; Federal; Non-profit
Ray et al., 2013	West6	Retrospective Cohort	3 years	Adult & Pediatric	Income quintiles2 were calculated at the block group level using 2006-2010 ACS 5-year estimates of median household income. SSTI were classified based on the quintile of the block group median income	Patient Medical Record	S. aureus SSTI resistant to methicillin	376,262	Industry
See et al., 2017	Southeast3	Retrospective Cohort	2 years	Adult	Proportions of low-income and high-income households at the census tract level were based on variables from the Harvard Public Health Disparities Geocoding Project, where low income was defined as the % of households with income <$25,000 and high income was defined as % of households in an area with income >= $200,0002	Patient Medical Record	Invasive CA-MRSA infection defined as the isolation of MRSA from a normally sterile body site (e.g., blood, cerebrospinal fluid, internal body fluid	2521	Federal

¹Data calculated from U.S. census

²Data obtained from the American Community Survey (ACS)

³Southest region consists of Alabama, Florida, Georgia, Kentucky, Maryland, Mississippi, North Carolina, South Carolina, Virginia, and West Virginia

⁴Midwest region comprises Illinois, Indiana, Iowa, Kansas, Michigan, Minnesota, Missouri, Nebraska, North Dakota, Ohio, South Dakota, Wisconsin

⁵Northeast region includes Maine, New Hampshire, Vermont, Massachusetts, Rhode Island, Connecticut, New York, New Jersey, Pennsylvania

⁶Western region consists of Arizona, California, Colorado, Idaho, Montana, Nevada, New Mexico, Oregon, Utah, Washington, Wyoming

Data analysis

All included studies were summarized in narrative form and in a summary table that reports study population characteristics, study design, exposure, and outcome. There were sufficient studies to conduct a meta-analysis of neighborhood income-level as a categorical exposure (n = 4) [17, 18, 31, 32] but not as a continuous exposure (n = 2) [3, 33]. A random-effects model was used for this meta-analysis to allow for variation in the “true effect” from study to study; this was especially important as we pooled studies comprising different age groups and genders [34, 35]. We examined the effect of living in a low- versus high-income neighborhood (reference group) on the risk of CA-MRSA infection versus no CA-MRSA infection for all studies (n = 4). A sensitivity analysis was also performed for the low ROB studies (n = 3) using a random-effects model. All analyses were completed using the RStudio (version 4.4.1) along with the R packages ‘epitools’ [36] and ‘meta’ for the calculation of effect measures and to perform the meta-analysis of binary outcome data [37], respectively. The calculated and extracted odds ratios along with their confidence intervals were uploaded to RStudio and meta-analyses of binary outcomes were performed using the R package meta [38]. A forest plot was generated for each analysis to present the results of the random effect model and 95% confidence intervals, prediction interval, and the I² variable for quantifying heterogeneity. I² statistics were further classified as follows: >25% as low; >50% as moderate; and > 75% as high.

Results

Study characteristics

We identified six eligible studies among 101 studies included in the aforementioned scoping review and updated search (Fig. 1). We first considered studies that reported income as a SES indicator. Most studies were ineligible due to income not being reported (n = 83) or wrong pathogen (n = 8). Two of 10 remaining studies were excluded because the outcome measured was CA-MRSA colonization rather than infection. An additional two studies were excluded because individual-level income was reported as the exposure rather than neighborhood-level income.

Fig. 1.

PRISMA flowchart for selection of articles in the systematic review and meta-analysis

Systematic review

All six eligible studies were conducted in the U.S. between 2002 and 2019. All studies ascertained CA-MRSA infections using electronic health records; most investigated MRSA SSTI [18, 31–33] although See et al. [3] investigated invasive CA-MRSA infections, defined as the isolation of MRSA from a normally sterile body site (e.g., blood, cerebrospinal fluid, internal body fluid) and Farr et al. [17] examined MRSA hospitalizations, which included hospitalizations with diagnosis codes for S. aureus pneumonia, S. aureus septicemia, or S. aureus infection. Four studies were retrospective cohort studies [3, 18, 31, 33], one was cross-sectional [17], and one was case-control [32]. Two studies focused on pediatric patients [18, 32] and one on adults [3]; three included both pediatric and adult patients [17, 31, 33]. Four studies used the American Community Survey (ACS) to define patients’ neighborhood-level income [3, 18, 31, 33] typically extracted at the census tract or census block group level. The other two studies used U.S. Census data to define neighborhood-level income at the census tract level but did not describe the specific dataset used [17, 32]. Detailed exposure definitions are reported in Table 1.

Four of the six studies reported adjusted effect measures describing the relationship between neighborhood-level income and CA-MRSA. Most adjusted for patient race and ethnicity, age, and sex. Among these four studies, three found that living in a low-income neighborhood was associated with an increased risk of CA-MRSA infection [3, 31, 33]. One study found that for each 10% increase in the percentage of households within a census tract with low income, there was a 1.35-fold predicted increase in the CA-MRSA rate [3]. Beresin et al. [33] found the average zip code level median household income was significantly lower (p < 0.05) for CA-MRSA cases compared to those without MRSA infections in Illinois. One California study found that patients living in neighborhoods with the lowest quintile of income had 42% higher odds of CA-MRSA SSTI (versus methicillin-susceptible S. aureus (MSSA) SSTI) than patients living in neighborhoods with the highest quintile of income [31]. Farr et al. conducted a sex-specific analysis and did not identify a significant association between neighborhood income distribution and CA-MRSA for either males or females in New York City [17].

We directly derived odds ratios for two studies that did not report effect measures for the relationship between neighborhood income and CA-MRSA [18, 32]. The derived odds ratio for Ali et al. [32] indicated that block groups in Atlanta, GA with more than 15% of households with a poverty-to-income ratio below 1 had 42% higher odds of CA-MRSA SSTI (termed community onset-MRSA in the study), relative to block groups with fewer than 15% of households with a poverty to income ratio below 1. The derived odds ratio for Immergluck et al. [18] indicated that the odds of MRSA SSTI (versus MSSA SSTI) were 7% greater in block groups in Atlanta, GA where more than 18% of households were living below the poverty level compared to block groups with fewer than 18% of households living below the poverty level.

Quality assessment outcomes

Three of the four cohort studies were determined to have a low risk of bias (ROB) after completing the modified NOS assessments [3, 18, 33], while Ray et al. had a moderate ROB [31]. The single case-control study was found to have a moderate ROB [32]. The cross-sectional study had a low ROB [17]. Completed risk of bias assessments can be found in the supplementary materials (Table S5). In general, we noted a risk of exposure misclassification in several studies because the median household income of patients’ neighborhoods was defined at relatively broad geographic scales (ZIP code) [17, 33] and because it was unclear if or how authors validated patients’ residential addresses to ensure they were patients’ address at the time of the encounter, rather than merely their most recent address on file, which is common in the electronic health record [3, 17, 31, 33]. The quality of evidence was rated to be moderate indicating we have moderate confidence in the observed effect of neighborhood level-income and CA-MRSA risk. The completed GRADE assessment can be found in the supplementary materials (Table S7). Inconsistency across studies, imprecision, and indirectness of the evidence/applicability areas were rated serious while the methodological quality area was rated not serious.

Meta-analysis findings

For the four studies [17, 18, 31, 32] with categorical and not continuous income measures included in the meta-analysis, the pooled odds ratio for CA-MRSA risk among individuals living in low-income neighborhoods compared to high-income neighborhoods (reference group), was 1.28 (95% CI 1.13, 1.46), with high statistical heterogeneity (I² 73%) (Fig. 2). In a sensitivity analysis limited to studies with low ROB [17, 18], there was no statistically significant relationship between living in a low-income neighborhood and CA-MRSA infection risk (1.13 95% CI 0.96, 1.33) with heterogeneity of 0% (Fig. 3).

Fig. 2.

Meta-analysis of the association between neighborhood-level income and CA-MRSA infection risk among included studies (n = 4). Note: Farr 2013 separated their analysis for male and female gender study populations: “Farr 2013-M” and “Farr 2013-F indicate effect measures for the male study group and female study population, respectively

Fig. 3.

Meta-analysis of the association between neighborhood-level income and CA-MRSA infection risk among low risk of bias studies (n = 2). Note: Farr 2013 separated their analysis for male and female gender study populations: “Farr 2013-M” and “Farr 2013-F indicate effect measures for the male study group and female study population, respectively

Discussion

To the best of our knowledge, this is the first SRMA of the effect of neighborhood-level income on CA-MRSA infection risk in the U.S. Overall, five of the six included studies observed an association between living in a low-income neighborhood and a higher risk of CA-MRSA infections. Meta-analysis of the four studies with categorical income measures indicated an association between living in a low-income neighborhood and heightened CA-MRSA infection risk, albeit with considerable heterogeneity. Our results align with previous systematic reviews examining the relationship between poverty and colonization and infection with AMR organisms [39–41], including CA-MRSA. Living in lower-income neighborhoods could plausibly increase the risk of CA-MRSA infections for several reasons, including limited access to healthcare services, which can push individuals to seek unreliable treatments, increasing improper diagnosis and vulnerability [10–12]. Crowded or unsafe living environments (e.g. poor housing structure or water shut-offs) can also make proper hygiene difficult [42]. Low-income neighborhoods often lack the resources to promote healthy behaviors related to diet, smoking, and exercise, which contribute to chronic illnesses that increase infection risk [13, 14, 43]. Additionally, chronic stress from socioeconomic hardships weakens the immune system, compounding the risk of chronic diseases and infections, including CA-MRSA [10, 13]. Identifying which of these mechanisms may be underpinning the associations we report here could be helpful for addressing long-standing disparities in CA-MRSA infection risk.

We observed substantial variability in effect sizes across the four studies included in our meta-analysis (I² 73%) which could be attributed to differences in study design, covariates that were adjusted for, the comparator group for included case counts or effect measures, or methods of measuring neighborhood-level income. The primary aims of each study varied, so the variables used to statistically adjust the association between income and CA-MRSA infections often differed. For example, Farr et al. [17], the single paper that reported no significant association, used a multivariable logistic regression model that included individual factors such as age, HIV status, and diabetes and neighborhood-level factors such as HIV prevalence, men-who-have-sex-with-men proportion, and emergency department usage rate. Ray et al. [31] adjusted for different covariates in their multivariable logistic regression model, including gender, age, race/ethnicity, type of SSTI, and diabetes status. Meanwhile, we manually calculated unadjusted effect measures for Immergluck et al. [18] and Ali et al. [32] All included studies focused on assessing the risk of CA-MRSA; however, the comparator groups differed. In the meta-analysis, two studies compared CA-MRSA to no CA-MRSA, while the other two used MSSA as the comparator group. The differences in comparator groups could have also contributed to the observed heterogeneity. MSSA represents a pathogenic infection, whereas “no MRSA” includes a mix of MSSA infected and non-infected individuals. This variation could lead to underlying differences in measured associations with neighborhood income. Additionally, the lack of a uniform definition for neighborhood-level income across studies (Table 1), could have also contributed to the substantial heterogeneity value. Two of four studies utilized ACS data but Immergluck et al. [18] used the median household income values from ACS to calculate the proportion below poverty within blocks, while Ray et al. [31] used ACS data to define census block income quintiles. Farr et al. [17] and Ali et al. [32] both used U.S. census data but defined neighborhood-level income as ZIP code median household income and poverty-to-income ratio, respectively. In our subgroup analysis of low ROB studies, we found no statistically significant association between income and CA-MRSA risk. This could suggest that studies with a higher risk of bias might drive the observed effect in the overall meta-analysis. However, we could not conduct further subgroup analyses or meta-regression to investigate causes of heterogeneity due to the limited number of studies.

The sensitivity analysis among low ROB studies generated a heterogeneity of 0%, although due to the p-value of 0.39, it cannot be concluded that there is truly no heterogeneity present. One reason why we may have observed 0% heterogeneity among low ROB studies is because there were limited studies available. Although there were three effect measures included in the analysis, two came from one paper that had stratified their analysis by sex. The two effect measures were extremely similar in value and statistical significance, which could have decreased the heterogeneity value. While the effect measures exhibited some degree of similarity, it’s essential to note that there were only three studies included, which may contribute to increased bias in the calculation of the I [2] statistic due to the smaller sample size [44].

Our meta-analysis has several limitations. First, effect estimates were calculated manually using raw numbers extracted from Immergluck et al. [18] and Ali et al. [32] and do not consider the influence of potential confounding variables or mediators like race/ethnicity, gender, age, or household crowding, many of which were controlled for in the other papers. An additional limitation is the small number of studies included, which can decrease the statistical power of meta-analysis and lower the precision of effect sizes. The few studies included limited our ability to conduct subgroup sensitivity analyses by age groups (i.e., adult vs. pediatric) or different genders. The summary measures we extracted may mask real differences in risk that exist between subgroup populations. Due to the limited literature on CA-MRSA where neighborhood-level income is the primary exposure of interest, a third limitation is that effect estimates were extracted from papers that reported this association but were not necessarily designed to investigate it. Ray et al. [31] investigated multiple demographic and clinical risk factors for MRSA SSTI simultaneously, with income being one of them. Similarly, See et al. [3] was interested in assessing neighborhood-level SES factors, among which neighborhood income was one of the factors of interest. Immergluck et al. [18], Farr et al. [17], and Ali et al. [32] examined the effects of individual and neighborhood risk factors separately and then included a subset of these covariates in a multivariable model, i.e., neighborhood-level income was just one of the many neighborhood-level risk factors assessed. The primary exposure for Beresin et al. [33] was contact with swine, and income was one of the covariates assessed. Because the authors’ statistical models were not explicitly designed to assess the association we were interested in for this SRMA, there may be additional confounding that was not adjusted for, and a causal interpretation may therefore be inappropriate. We extracted all relevant data from the studies available, given that there were no papers investigating neighborhood-level income as the sole primary exposure. Finally, we only included studies that measured neighborhood-level income as a categorical variable in our meta-analysis. Considering income as a stepwise gradient or continuous change might be useful because some researchers hypothesize there is an income-health gradient [45, 46]. We were unable to do this here because there were only two studies identified in this systematic review that examined income as a continuous variable (e.g. percentage increase of low-income households within a census tract). Future studies should consider reporting neighborhood-level income as a continuous variable so an income-risk gradient can be explored. Overall, the moderate quality of evidence grading indicates meaningful results, but due to the aforementioned limitations there is a level of uncertainty about the exact effect size [27, 28]. Strengthening methodological approaches, as outlined above, could provide higher confidence insight into this relationship. By addressing these gaps, future research can provide more knowledge and inform targeted interventions aimed at reducing disparities in CA-MRSA infection risk.

Conclusions

Living in a lower-income neighborhood was associated with an increased risk of CA-MRSA infections albeit with considerable heterogeneity and moderate confidence, suggesting further research is needed. We suggest that future studies use more comparable and clearer definitions of neighborhood-level income and consider adjusting for covariates that could allow for a causal interpretation of the relationship between low neighborhood-level income and CA-MRSA infection risk. Our findings underscore the ongoing interplay between CA-MRSA infection risk and the social determinants of health in the United States.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

SES

Socioeconomic status

U.S

United States

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

SRMA

Systematic Review and Meta-Analysis

CA-MRSA

Community-Acquired Methicillin-Resistant Staphylococcus aureus

MRSA

Methicillin-Resistant Staphylococcus aureus

CA-MRSA SSTI

Community-Acquired Methicillin-Resistant Staphylococcus aureus Skin and Soft Tissue Infection

PROSPERO

International Prospective Register of Systematic Reviews

MEDLINE

Medical Literature Analysis and Retrieval System Online

SSTI

Skin and Soft Tissue Infection

NOS

Newcastle-Ottawa Quality Assessment Scale

ACS

American Community Survey

MSSA

Methicillin-Susceptible Staphylococcus aureus

MSSA SSTI

Methicillin-Susceptible Staphylococcus aureus Skin and Soft Tissue Infection

ROB

Risk of Bias

GA

Georgia

ZIP

Zone Improvement Plan

AMR

Antimicrobial Resistance

HIV

Human Immunodeficiency Virus

Author contributions

MLN and SBl conceptualized the study. MLN, NN, and SD acquired the funding. RM performed the updated literature search. SBl, EA, SBa, SAA, and MLN reviewed titles and abstracts, reviewed full texts, and extracted data. SAB prepared the tables and figures. SAB, RM, and MLN wrote the first draft of the manuscript. All authors reviewed the manuscript.

Funding

Research reported in this publication was supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number UM1AI104681. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. S.D. received support from the Tupper Fund. M.L.N was supported by Emory University and the MP3 Initiative. Funders had no role in study design; in the collection, analysis, or interpretation of data; or in writing the manuscript. The ARLG Publications Committee reviewed the manuscript prior to submission for publication.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics of approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.