Demographic and Socioeconomic Factors Associated with Fungal Infection Risk, United States, 2019

Diagnosis disproportionately affected minority and low-income populations, underscoring the need for broad public health interventions.

F ungal pathogens cause millions of deaths and tens of millions of infections globally every year (1). Fungal infections are primarily opportunistic, causing moderate to severe disease in immunocompromised patients. Fungal infections also are associated with increased illness rates and substantial healthcare costs, resulting in $6.7 billion in hospitalization costs in the United States in 2018 (2). In addition, fungal infections doubled the average length and cost of hospital stays and risk for death among patients with >1 associated risk condition (2). Despite the considerable medical and economic burden of fungal infections, standardized diagnostic and treatment guidelines are lacking.
Interest in the effects of race and ethnicity and socioeconomic status on fungal infections and associated patient outcomes has increased (18,19), especially because diagnosed fungal infections have increased since 2010 (3). Previous studies documented the relationship between health disparities and fungal infections (18,19), but not as a main analytic focus, and studies across multiple fungal pathogens are lacking. We describe diagnosed fungal infections and associated risk conditions by key demographic variables, including race and ethnicity and socioeconomic status.

Data Sources
We used hospital discharge data from the National Inpatient Sample (NIS), Healthcare Cost and Utilization Project (HCUP), from the Agency for Healthcare Research and Quality (20). NIS is the largest database of US hospitalization data, covering >96% of the population (20). HCUP data comprise hospitalizations, rather than unique patients. We use the term patient to refer to inpatient status; we acknowledge that a specific patient might be included >1 time in our analyses. For total population per income quartile, we used 2006-2010 American Community Survey (21) results to estimate population percentages, then adjusted these to the 2019 population.

Element Descriptions
We used codes from the International Classification of Diseases, 10th Revision (ICD- 10), to identify at-risk patients and invasive and noninvasive fungal infections, as previously described (2) ( Table). We defined fungal infections and associated risk conditions when relevant ICD-10 codes were recorded as any diagnosis in the hospitalization record. Sex, race, and ethnicity data were provided by patient records in NIS. HCUP excludes the data for sex when the state level patient record identifies the patient as both nonfemale and nonmale. Ethnicity took precedence over race in the HCUP database when both were provided as distinct values in the patient record.
The HCUP dataset predefines each annual income quartile (Q) according to estimated median household income in US dollars of residents living within a patient's postal code. For 2019, Q1 was $1-$47,999, Q2 was $48,000-$60,999, Q3 was $61,000-$81,999, and Q4 was >$82,000. We defined insurance type by the expected primary payer type to which the hospital visit was billed in the HCUP NIS dataset.
Fungal infections cause substantial rates of illness and death. Interest in the association between demographic factors and fungal infections is increasing. We analyzed 2019 US hospital discharge data to assess factors associated with fungal infection diagnosis, including race and ethnicity and socioeconomic status. We found male patients were 1.5-3.5 times more likely to have invasive fungal infections diagnosed than were female patients. Compared with hospitalizations of non-Hispanic White patients, Black, Hispanic, and Native American patients had 1.4-5.9 times the rates of cryptococcosis, pneumocystosis, and coccidioidomycosis. Hospitalizations associated with lower-income areas had increased rates of all fungal infections, except aspergillosis. Compared with younger patients, fungal infection diagnosis rates, particularly for candidiasis, were elevated among persons >65 years of age. Our findings suggest that differences in fungal infection diagnostic rates are associated with demographic and socioeconomic factors and highlight an ongoing need for increased physician evaluation of risk for fungal infections.

Results
Nearly 60,000 invasive fungal infections were reported during US hospitalizations in 2019, ≈10% of all diagnosed fungal infections among hospitalized patients. Another 391,000 noninvasive infections, primarily dermophyte, also were diagnosed.

Fungal Infections and Risk Conditions by Race and Ethnicity
Overall, risk conditions and fungal infections were diagnosed among racial and ethnic subgroups at rates generally consistent with the current racial and ethnic composition of the United States; most (65.9%) cases were diagnosed in non-Hispanic White patients. However, we noted deviations that highlight racial and ethnic health disparities.

Fungal Infections and Risk Conditions by Income
Of 35.5 million hospitalizations in 2019, nearly one third were associated with residence in lower income areas (Appendix Table 2). Patients from Q1 postal codes had 1.6 times the hospitalization rate as patients from Q4 areas. Fungal infections were diagnosed in patients from Q1 postal codes at 1.2 (95% CI 1.2-1.2) times the frequency of patients from Q4 postal codes ( Figure 6, panel A; Appendix Table 2). Cryptococcosis was diagnosed at 2.0 (95% CI 1.8-2.1) and histoplasmosis at 1.7 (95% CI 1.5-1.8) times the rate in Q1 patients as in Q4 patients. The only fungal infection diagnosed more frequently in Q4 patients was aspergillosis (RR 1.3, 95% CI 1.2-1.4).
Despite representing 14.9% of hospitalizations in 2019, pediatric patients accounted for only 4.2% of diagnosed fungal infections and had one third the diagnostic rate (RR 0.3, 95% CI 0.3-0.3) of adult patients ( Figure 11, panel A); rates of all fungal pathogens and manifestations were reduced. Pediatric patients had >1 fungal-associated risk condition diagnosed at 0.2 (95% CI 0.2-0.2) times the rate for adult patients (Figure 11, panel B). Only the diagnostic rate for cystic fibrosis (RR 1.3, 95% CI 1.2-1.3) was higher among pediatric than adult patients.

Discussion
We analyzed rates of fungal infection diagnoses in hospitalizations on the basis of racial and ethnic  background and socioeconomic status. Our findings demonstrate that health disparities between racial, ethnic, and socioeconomic groups extend to fungal infections, especially for predisposing risk conditions. In HCUP NIS, male patients had 1.4-3.5 times the rate of invasive fungal infection diagnoses as female patients, a finding supported by existing literature (26). The influence of genetic components by sex has been postulated, as have higher environmental exposure and behavioral risks (26,27). The relationship between sex and susceptibility is more complex than our analyses can capture, but >1 risk condition for fungal infection was more frequently diagnosed among male patients.
Aspergillosis was diagnosed more frequently in non-Hispanic White and AA/PI patients than in other racial and ethnic groups. As previously described (3), invasive aspergillosis is closely associated with stem cell and solid organ transplantation, and noninvasive manifestations, including allergic bronchopulmonary aspergillosis and chronic pulmonary aspergillosis, are more often diagnosed in cystic fibrosis and tuberculosis patients; AA/PI patients have >9 times the rate of tuberculosis diagnoses as non-Hispanic White patients (28). In addition, aspergillosis is the only fungal infection diagnosed more frequently in patients from higher income areas. Higher income is associated with higher probability of receiving a transplant (29,30) and improved patient outcomes in cystic fibrosis care (31), possibly because these patients have better access to healthcare facilities and the financial capacity for regular treatment. Aspergillosis likely is more frequently diagnosed in higher income patients because of their ability to continually seek treatment for associated risk factors. Income differences also could relate to cost of living because aspergillosis is more likely to be diagnosed in urban than rural patients (32).
Candidiasis was diagnosed more frequently in Black patients. Invasive candidiasis was more frequent in male patients, fitting with previous findings (33), but noninvasive candidiasis was more frequent in female patients. Increased rates of candidiasis among senior patients compared with adult patients also is consistent with prior findings (33). All candidiasis clinical manifestations were more frequent in patients from lower income areas. Assessments of the relationship of candidiasis and income are lacking, but these diagnoses might be related to the higher frequency of diabetes in patients from low-income areas (7). This finding also might be an artifact of the relationship between low income and increased frequency of repeat hospitalizations (34). All candidiasis clinical manifestations were diagnosed moderately more frequently in rural patients.
Coccidioidomycosis and histoplasmosis are endemic infections that can affect immunocompetent persons, but severe disease is more common in immunocompromised persons. Coccidioidomycosis is endemic in the US Southwest and histoplasmosis in the Ohio and Mississippi River Valley regions. Our analysis showed coccidioidomycosis was diagnosed more frequently in Hispanic, AA/PI, and Native American adult male patients than in non-Hispanic White or Black, senior, or female patients. Environmental exposure is key in coccidioidomycosis; workers performing soil-disturbing work or exposed to dusty conditions in endemic areas are at increased risk. Black and Hispanic persons are overrepresented in lower wage, more manual labor, and higher risk occupations, including occupations with frequent dust exposure (35,36). Previous reports noted higher frequencies of coccidioidomycosis in AA/PI and Hispanic male adults residing in urban areas, but older state-level data also indicated increased rates in Black compared with non-Hispanic White male persons (36)(37)(38).
Non-Hispanic White patients had up to 3 times the rate of histoplasmosis as other racial and ethnic groups. Histoplasmosis diagnoses were higher among adult, low-income, and rural patients. These results are supported by previous reports of histoplasmosis predominantly among middle-aged adult White male persons living in rural areas (39). These demographic variables likely capture persons with environmental or occupational exposure, including persons employed in construction, agriculture, and forestry industries (40).
Historically, cryptococcosis and pneumocystosis were closely tied to HIV, which continues to disproportionately affect Black and Hispanic/Latino populations (41). We found cryptococcosis and pneumocystosis were diagnosed in Black and Hispanic patients at 2-3 times the rate for non-Hispanic White patients. HIV, cryptococcosis, and pneumocystosis frequencies also were elevated in Q1 patients and were far more frequent in adult than senior patients, fitting with previous literature (42). HIV, cryptococcosis, and pneumocystosis rates were elevated in hospitalizations billed to Medicaid or self-pay and in urban patients.
Incidence of mucormycosis, a rare and often fatal infection, has been rising (43). We found mucormycosis diagnoses were more frequent among AA/PI and Hispanic patients than among non-Hispanic White patients. The most common underlying condition for mucormycosis is diabetes mellitus (43), but diabetes was not diagnosed more frequently in AA/PI or Hispanic populations in our study. We noted no differences in mucormycosis rates by income or insurance type. Adult patients were more likely to have mucormycosis than senior patients, and we noted a slight elevation in diagnoses among urban patients.
Other fungal infections include primarily superficial cutaneous and mucosal infections, which were diagnosed more frequently in senior patients and in hospitalizations billed to Medicare, consistent with previous studies (44). Unspecified mycotic infections also were more frequently diagnosed in senior patients, which could reflect increased mortality and shorter survival times associated with an aging immune response failing to control invasive fungal infections, as previously described (45).
Our results are informative, but our data likely underrepresent the true burden of fungal disease in the United States. Evidence suggests that only half of invasive fungal infections are diagnosed before patient death (46). The sensitivity and specificity of many ICD-10 codes for fungal infections are unknown, and misclassification is possible. HCUP NIS enabled us to comprehensively study fungal infections; however, unique patients cannot be identified in NIS, so our data likely represent multiple hospitalizations per patient. Data collection also could be a limitation because race and ethnicity analyses are limited by single identifiers and failed to represent patients with multiracial or multiethnic identities. In addition, some previously studied racial and ethnic subgroups might not have been included for this variable in the NIS dataset. Finally, hospitals might have reported a private insurance payer type for patients covered by a Medicare-managed care program administered by a private insurance company, potentially underrepresenting differences between payer types.
In conclusion, we provide a comprehensive summary of fungal infections and associated risk conditions among hospitalized patients, including corresponding rate ratios by demographic and socioeconomic factors. These findings are based on bivariate analysis, but future studies could use a multivariable analysis of the potential predictive weight of demographic and socioeconomic risk factors and >1 comorbidity to measure evaluated risk for fungal infection by type. Our findings suggest that differences in fungal infection diagnostic rates are associated with demographic and socioeconomic factors. Because fungal infections increase mortality rates and healthcare costs, our results highlight an ongoing need for increased physician evaluation of risk for fungal infections, especially among minority and low-income populations that are disproportionately affected.