Health Care–Associated Infections Among Hospitalized Patients With COVID-19, March 2020-March 2022

Key Points

Question

What is the incidence of health care–associated infections (HAIs) among hospitalized patients with COVID-19 vs those without COVID-19?

Findings

In a cross-sectional analysis of more than 5 million hospitalizations between 2020 and 2022, the occurrence of central line–associated bloodstream infection, catheter-associated urinary tract infection, and methicillin-resistant Staphylococcus aureus bacteremia was between 2.7- and 3.7-fold higher in the COVID-19 population. Occurrence of these HAIs among patients hospitalized without COVID-19 did not increase significantly from baseline during this same time frame.

Meaning

Despite the strain on the health care system, the increase in HAIs was not observed in the non–COVID-19 population, suggesting that key safety processes were maintained and patients with COVID-19 may require additional protective care to prevent HAIs.

Abstract

Importance

The reported incidence of many health care–associated infections (HAIs) increased during the COVID-19 pandemic; however, it is unclear whether this is due to increased patient risk or to increased pressure on the health care system.

Objective

To assess HAI occurrence among patients admitted to hospitals with and without COVID-19.

Design, Setting, and Participants

A cross-sectional retrospective analysis of inpatients discharged both with and without laboratory-confirmed COVID-19 infection was conducted. Data were obtained between January 1, 2019, and March 31, 2022, from community hospitals affiliated with a large health care system in the US.

Exposure

COVID-19 infection.

Main Outcomes and Measures

Occurrence of central line–associated bloodstream infection (CLABSI), catheter-associated urinary tract infection (CAUTI), methicillin-resistant Staphylococcus aureus (MRSA) bacteremia, and Clostridioides difficile infection as reported to the National Healthcare Safety Network.

Results

Among nearly 5 million hospitalizations in 182 hospitals between 2020 and 2022, the occurrence of health care–associated infections (HAIs) was high among the 313 200 COVID-19 inpatients (median [SD] age, 57 [27.3] years; 56.0% women). Incidence per 100 000 patient-days showed higher HAIs among those with COVID-19 compared with those without. For CLABSI, the incidence for the full 9 quarters of the study was nearly 4-fold higher among the COVID-19 population than the non–COVID-19 population (25.4 vs 6.9). For CAUTI, the incidence in the COVID-19 population was 2.7-fold higher in the COVID-19 population (16.5 vs 6.1), and for MRSA, 3.0-fold higher (11.2 vs 3.7). Quarterly trends were compared with the same quarter in 2019. The greatest increase in the incidence of HAI in comparison with the same quarter in 2019 for the entire population occurred in quarter 3 of 2020 for CLABSI (11.0 vs 7.3), quarter 4 of 2021 for CAUTI (7.8 vs 6.8), and quarter 3 of 2021 for MRSA (5.2 vs 3.9). When limited to the non–COVID-19 population, the increase in CLABSI incidence vs the 2019 incidence was eliminated, and the quarterly rates of MRSA and CAUTI were lower vs the prepandemic 2019 comparator quarter.

Conclusions and Relevance

In this cross-sectional study of hospitals during the pandemic, HAI occurrence among inpatients without COVID-19 was similar to that during 2019 despite additional pressures for infection control and health care professionals. The findings suggest that patients with COVID-19 may be more susceptible to HAIs and may require additional prevention measures.

This cross-sectional study examines the incidence of health care–associated infections in hospitalized patients with COVID-19 vs those without COVID-19.

Introduction

The occurrence of health care–associated infections (HAIs) increased during the COVID-19 pandemic. An analysis of data submitted to the National Healthcare Safety Network (NHSN), comparing the HAI incidence by quarter in 2020 to 2019, demonstrated substantial increases in the occurrence of central line–associated bloodstream infection (CLABSI), catheter-associated urinary tract infection (CAUTI), ventilator-associated adverse events, and methicillin-resistant Staphylococcus aureus (MRSA) bacteremia, while the occurrence of Clostridioides difficile (CDIFF) infection showed a modest decrease.¹ A follow-up analysis demonstrated that HAI rates continued to increase through the third quarter of 2021.² Similarly, a report from the VA Health Care system documented increases in the occurrence of CLABSI and ventilator-associated adverse events and a simultaneous decrease in CDIFF during the pandemic, although there was no reported change in CAUTI or MRSA CLABSIs.³ In the HCA Healthcare (HCA) system in 2021, data from 82 hospitals indicated an increase in HAIs as well as clusters of infection, and these increases correlated with surges in volume of patients with COVID-19 at the hospital level.⁴

While this evidence indicates that the occurrence of HAIs increased during the COVID-19 pandemic, it is unclear whether this was due to the increased strain on available hospital resources or increased susceptibility of patients with COVID-19. It has been speculated that the stress of the COVID-19 pandemic on hospitals, particularly nursing and infection prevention resources, led to an inability to maintain strict attention to known processes that prevent HAIs.⁵ Yet reports also suggest that the enhanced infection prevention controls implemented in response to COVID-19 were associated with local reductions in HAIs.^6,7 The change to optional reporting of health care–associated infections to the Centers for Medicare & Medicaid Services in early 2020 further complicate the accurate assessment of HAI rates.⁸

If overall changes to the hospital system are the primary source of the increase in HAIs, one would expect that rates increased among all hospitalized populations during the COVID-19 pandemic. However, to our knowledge, none of the published analyses to date have evaluated the occurrence of HAIs specifically among inpatients with a diagnosis of COVID-19 vs the occurrence among those who are hospitalized for other reasons. In this retrospective analysis, we assessed HAI occurrences between January 1, 2019, and March 31, 2022, for inpatients admitted to hospitals in a large health care system. We compared quarterly HAI trends between patients with and without COVID-19.

Methods

This retrospective cross-sectional analysis was conducted using data collected during the course of clinical care at 182 inpatient facilities in 21 states affiliated with the HCA system. The HCA system maintains a clinical data warehouse that includes information about coded diagnoses, laboratory events, and events reported to the NHSN. In accordance with HCA institutional policy, this research was determined to be exempt from institutional review board oversight. This determination was made using a decision support system based on a wide variety of interconnected decision trees and regulations to determine whether an activity must or may not be reviewed by an institutional review board before engaging human participants. This system was developed by and is overseen by institutional experts and has been recognized as a best practice in human research protection.⁹ This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

We used these centralized data resources for the study period of January 1, 2019, through March 31, 2022, to identify the aggregate occurrence of HAIs and the COVID-19 patient population. The HAIs (CLABSI, CAUTI, MRSA, and CDIFF) were defined by the NHSN and detected using standard HAI surveillance by trained infection preventionists.¹⁰ These 4 HAIs were selected for analysis as they are required to be reported to the NHSN and underwent consistent surveillance during the pandemic, resulting in reliable data that were applicable to all hospitalized patients. All HCA facilities continued to report HAIs to the NHSN throughout the pandemic.

Using the same data set, demographic characteristics (age, sex, race and ethnicity) were obtained. Race and ethnicity were included to evaluate potential demographic differences between those hospitalized with COVID-19 and the rest of the hospitalized population. The total patient days of discharges were calculated and separated into those with an accompanying COVID-19 diagnosis and those without. The diagnosis of COVID-19 was based on laboratory-confirmed SARS-CoV-2, via either molecular or antigen testing, during the encounter or within 30 days of admission without laboratory-confirmed negative tests before admission.

Case mix index was calculated from the total Medicare Severity Diagnosis Related Group weights of discharges divided by the total number of discharges per quarter.¹¹ Length of stay was calculated based on documented day of admission through day of discharge.

Statistical Analysis

Event rates were calculated per calendar quarter for each of the 4 HAIs for the full population and for both the COVID-19 and non–COVID-19 populations separately. Incidence is reported as events per 100 000 patient-days for each HAI.

Comparisons are described in relation to same-quarter performance during 2019, to minimize the influence of seasonal trends. For the overall study period, an F test for a linear trend in HAI occurrence was calculated for each type of HAI for both the COVID-19 population and the population without COVID-19. Additionally, an exact binomial test for significance was calculated for each HAI each year, comparing the HAI rates in the COVID-19 population with those in the non–COVID-19 population.

Results

There were 313 200 patients with COVID-19 from quarter (Q) 1 of 2020, the first documented diagnosis of COVID-19 in the data set, through the first quarter of 2022. The overall inpatient population included 43.5% men and 56.0% women; median (SD) age was 57 (27.3) years. For the same period, there were 4 564 375 discharges of individuals without a diagnosis of COVID-19. This translates to a total of 28 527 585 inpatient days overall: 25 652 380 non–COVID-19 patient-days and 2 875 205 COVID-19 patient-days. Case mix index was 2.29 days and mean (SD) length of stay was 8.2 (10.8) days for patients with COVID-19 during the study; for the non–COVID-19 population, the case mix index was 1.63 and mean (SD) length of stay was 4.7 (7.5). The percentage of total discharges of patients with a COVID-19 diagnosis was 6.4% overall, with a range from 0.1% (n = 780; Q1 2020) to 10.7% (n = 61 080; Q3 2021). Additional population demographic characteristics for the study period are presented in the Table.

Table. Demographic Characteristics of Patient Population, 2019 and 2020-2022.

Characteristic	January 1 to December 31, 2019, all inpatients	January 1, 2020, to March 31, 2022
		COVID-19 inpatients	Non–COVID-19 inpatients
Total inpatients, No.	2 184 994	313 200	4 564 375
Total inpatient days, No.	12 232 982	2 875 205	25 652 380
Age, median (SD), y	57 (27.3)	64 (19.7)	55 (27.2)
Sex, %
Male	43.5	50.9	44.1
Female	56.0	49.0	55.8
Unknown	0.4	0	0.1
Race and ethnicity, %
Hispanic	16.5	24.2	16.6
Non-Hispanic
Black	14.6	16.4	15.2
White	59.6	51.6	59.6
Other	5.9	5.6	6.1
Unknown	3.4	2.2	2.5
Length of stay, mean (SD), days	4.6 (7.3)	8.2 (10.8)	4.7 (7.5)
Case mix index	1.56	2.29	1.63
CLABSI, count (rate)	996 (8.1)	731 (25.4)	1759 (6.9)
CAUTI, count (rate)	1068 (8.7)	475 (16.5)	1562 (6.1)
MRSA, count (rate)	455 (3.7)	323 (11.2)	948 (3.7)
CDIFF, count (rate)	2384 (19.5)	257 (8.9)	2481 (9.7)

Abbreviations: CAUTI, catheter-associated urinary tract infection; CDIFF, Clostridioides difficile; CLABSI, central line–associated bloodstream infection; MRSA, methicillin-resistant Staphylococcus aureus.

Occurrences of CLABSI, CAUTI, MRSA, and CDIFF per 100 000 patient-days are shown in the Figure. Discharges and HAI data from 2019 are included to allow historical comparison quarter to quarter. Broad-based trends show that HAIs among the COVID-19 population were consistently higher than for the non–COVID-19 population (Figure). The greatest difference was for occurrence of CLABSI, where the incidence for the full 9 quarters was 3.7-fold higher among the COVID-19 population (25.4 vs 6.9). For CAUTI, the occurrence was higher in the COVID-19 population by 2.7-fold (16.5 vs 6.1), and for MRSA, 3.0-fold (11.2 vs 3.7). There was no significant difference in the CDIFF occurrence between the COVID-19 and non–COVID-19 populations.

Figure. Time Trend for COVID-19 Inpatient Volume and Hospital-Acquired Infection (HAI) Occurrence.

Total discharges of patients with COVID-19 by quarter (Q) and infection rate per 100 000 patient-days for central line–associated bloodstream infection (CLABSI) (A), catheter-associated urinary tract infection (CAUTI) (B), methicillin-resistant Staphylococcus aureus (MRSA) bacteremia (C), and Clostridioides difficile (CDIFF) infection (D) among all inpatients, inpatients without COVID-19, and inpatients with COVID-19.

The greatest increase in the incidence of HAI in comparison with the same quarter in 2019 for the entire population occurred in Q3 2020 for CLABSI (11.0 vs 7.3), Q4 2021 for CAUTI (7.8 vs 6.8), and Q3 2021 for MRSA (5.2 vs 3.9). When limited to the non–COVID-19 population, the same comparisons showed that any increase in HAI incidence was eliminated. In Q3 2020, the incidence of CLABSI in the non–COVID-19 population was 7.3 vs 7.3 in Q3 2019. For CAUTI, the incidence in the non–COVID-19 population was lower in Q4 2021 vs Q4 2019 (6.2 vs 6.8). The incidence of MRSA for patients without COVID-19 was also lower in Q3 2021 vs the 2019 comparator quarter (3.7 vs 3.9).

The CDIFF rate in both patients without COVID-19 and all inpatients significantly decreased by quarter (F test; non–COVID-19, β = −7.1; P < .001; all inpatients, β = −6.8; P < .001), but was not significant for the COVID-19 population (β = −3.2; P = .57). The CAUTI rate in patients without COVID-19 significantly decreased over time (β = −1.6; P = .01). There were no significant trends identified for CLABSI or MRSA for any of the populations. When comparing the occurrence of each HAI between the COVID-19 and non–COVID-19 populations, the COVID-19 population had a significantly higher rate of infection for every HAI except CDIFF and every year (exact binomial test, P < .001, for all combinations of year and infection type except CDIFF).

Discussion

In our analysis of a large multistate population of inpatients, our findings are consistent with previous reports that the occurrence of HAIs increased during the COVID-19 pandemic, reversing a multiyear national trend of improving performance regarding hospital infection.¹ However, our subanalysis revealed that this increase in the overall infection rate appeared to be entirely due to the occurrence of HAIs in the COVID-19 population. Patients without COVID-19 had rates of HAIs that would be expected based on the incidence observed before the pandemic. The differences in HAI rates between the COVID-19 and non–COVID-19 populations for CLABSI, MRSA, and CAUTI were all significant during every year since the beginning of the pandemic.

A study of HAIs during the COVID-19 pandemic in facilities within our HCA system confirmed that the relative rates of CLABSI, CAUTI, and MRSA bacteremia were associated with increasing monthly COVID-19 discharges.⁴ Between March and September 2020, there were 43% to 60% more of these infections than expected. Furthermore, clusters of hospital-onset pathogens also increased during COVID-19 surges, suggesting possible increased transmission. That study temporally linked increased HAI rates with increases in the number of patients with COVID-19 but did not distinguish among which patients (those with or without COVID-19) the HAIs were more likely to occur. Given this, that study, and others that were similar, could not discount disruptions in routine infection prevention practice as a mechanism for increased HAIs during the COVID-19 pandemic.

Previous commentary has invoked an association between the increase in HAIs during the COVID-19 pandemic and the enormous stress on hospital personnel during this time, particularly with nursing and infection prevention personnel, and the ability to maintain general infection prevention practices, such as hand hygiene, dressing care, and appropriate use of personal protective equipment.¹² The results presented herein, however, suggest that HAI prevention practices were maintained, at least for the population without COVID-19.

There are several possible mechanisms by which the HAI rate increased in the COVID-19 population at the exclusion of patients without COVID-19. First, patients with COVID-19 had a longer length of stay, which increases the risk of developing and detecting HAI. However, including patient-days as the denominator in the analysis adjusts for this difference in length of stay. Second, health care personnel assigned to COVID-19 units may have experienced reduced resources or altered workflows. Both resource limitations and care of patients with COVID-19 improved over the course of the pandemic, but the retrospective nature of this data set is not able to fully account for individual decisions that could have increased HAI risk. Third, the overall degradation in performance could be fully or partially attributable to the risks associated with this new, very high-risk population: patients with COVID-19. A previous analysis suggested that the reduction in elective admissions during the COVID-19 pandemic skewed the overall population toward patients with conditions of higher acuity.¹³ In our study, the difference in HAI incidence between those with and without COVID-19, after adjusting for major confounders, suggests that the disease or its treatment preferentially increases the vulnerability of the COVID-19 population to HAIs.

A full understanding of the high vulnerability of the COVID-19 population to HAIs will be useful in guiding infection prevention practices in the future. The high occurrence of HAIs may be associated with inherent risk associated with the clinical condition of COVID-19 infection requiring hospitalization or it may be associated with care practices that introduce risk. Patient care management, particularly at the beginning of the pandemic, had elements of improvisation. Examples would include proning techniques, intravenous line extension to allow remote access, and room ventilation alterations. While HAI rates were higher in the COVID-19 population, the occurrence of CLABSI, MRSA, and CAUTI in this population decreased over the course of the pandemic from 2020 to 2022. This is likely reflective of both improving practice in the management of COVID-19 as well as the decreasing acuity and length of stay in this population over time. Additional analysis will be needed to determine whether the occurrence of HAIs continues to decrease or will stabilize in the COVID-19 population.

Limitations

There are several limitations to this study. First, we are only reporting on 4 HAIs and have not assessed other reportable HAIs, such as ventilator-associated adverse events and surgical site infection. We also did not assess other measures of infection performance, such as the occurrence of clusters or trends in antimicrobial resistance. Baker et al⁴ reported that clusters of hospital-onset pathogens increased as the COVID-19 burden increased, as did the occurrence of multidrug-resistant organisms. The degree to which these phenomena are attributable to the COVID-19 population exclusively has not been identified.

Second, the measure used for comparison herein is an unadjusted incidence rate and does not address potential factors such as variation in patient acuity, patient conditions, or patient demographic characteristics. However, for the non–COVID-19 population, the length of stay and case mix index remained relatively constant. While reductions in elective procedures may have affected the number and acuity of patients seen during the pandemic, the consistency in case mix index for patients without COVID-19 suggests there were minimal systemwide changes in the overall risk of populations served by participating hospitals during this timeframe. The COVID-19 population is clearly at higher risk of HAIs both because of a longer length of stay and higher level of acuity, although the substantially higher occurrence of HAIs, as well as the variability between infection types, suggests that there are additional factors inherent to this population that place them at greater risk of HAIs.

Third, this cross-sectional study was conducted at facilities within a single health care system in the US. While these facilities represent primarily community hospitals in urban and suburban settings across multiple states, it is possible that there are features of the health care system that limit generalizability. For instance, affiliated hospitals have the benefit of access to the expanded purchasing power of a large system, which may have helped with distribution of personal protective equipment, as well as access to clinical experts and best practices, and may have provided support to infection control practices during the pandemic. However, as HAIs during the pandemic increased in these facilities to a similar degree as in unaffiliated facilities, we are reasonably certain that the findings reported herein may extend outside our health care system.

Fourth, conclusions regarding trended multiyear observations in HAIs require that HAI surveillance is performed consistently, and if surveillance practice degraded during the pandemic, it could mask a true increase in the HAI rate. The HCA facilities all continued to report HAIs to the NHSN, even while reporting expectations were relaxed during the pandemic, and continued to perform routine audits of surveillance processes by an external vendor, thereby minimizing the risk of any inconsistencies in surveillance during the period of the study.

Conclusions

In this cross-sectional study of a large, diverse group of hospitals, we observed an increase in HAIs coincident with the increase in the volume of patients with COVID-19, and further observed that the increase may be explained by the high proportion of HAIs in the COVID-19 population specifically. Our findings suggest that, despite the unquestionable stressors that hospitals experienced during the COVID-19 pandemic, key safety processes appear to have been maintained, and HAI increases may be associated with the unique risks and intensive needs of patients with COVID-19. This analysis suggests that the greatest opportunity to improve outcomes may involve targeting additional resources to provide even greater attention to the hospitalized COVID-19 population.

Supplement.

Data Sharing Statement