The goals of medicine include alleviating pain and suffering, promoting health, preventing disease and death, achieving a peaceful death, curing disease when possible, and caring for patients for whom there is no cure. Reference Callahan1 The International Nosocomial Infection Control Consortium (INICC) is aligned with these objectives through the surveillance and prevention of healthcare-associated infections (HAIs). Reference Rosenthal2 The INICC is the first and largest multinational HAI surveillance, prevention, and investigation network, and it has operated internationally since 2002. Reference Rosenthal, Maki and Graves3 The INICC has been publishing international reports that provide data on HAI rates and related mortality. Its first report was in 2006, Reference Rosenthal, Maki and Salomao4 and subsequent reports were published in 2008, Reference Rosenthal, Maki and Mehta5 2010, Reference Rosenthal, Maki and Jamulitrat6 2012, Reference Rosenthal, Bijie and Maki7 2014, Reference Rosenthal, Maki and Mehta8 2016, Reference Rosenthal, Al-Abdely and El-Kholy9 2019, Reference Rosenthal, Bat-Erdene and Gupta10 and 2021. Reference Rosenthal, Duszynska and Ider11
In addition, an INICC investigation showed that patients without an HAI have a mortality rate of 17%, whereas those with 1 HAI have a mortality rate of 30%–48% and patients with 3 HAIs have a mortality rate of 63%. Reference Rosenthal, Duszynska and Ider11
In multivariate logistic regression analysis, ICU-acquired infection remained an independent risk factor for hospital mortality after adjustment for APACHE II score and age. Reference Ylipalosaari, Ala-Kokko, Laurila, Ohtonen and Syrjala12
In a study analyzing clinical outcomes and risk factors for mortality from ventilator-associated events of patients hospitalized in ICU, older age, higher APACHE II score on ICU admission, blood transfusion, immunosuppressive drugs, and presence of a central line were associated with higher risks for all-cause mortality in an ICU. Reference Zhu, Wang and Kang13 In a study of risk factors for long-term mortality of critically ill elderly patients admitted to an ICU, in-hospital mortality risk was independently associated with mechanical ventilation, use of vasopressors, and neurological disease. Reference Mukhopadhyay, Tai and See14
These studies have analyzed the impact of severity of illness scores, underlying disease, age, sex, presence of HAI, and a few other variables. However, no study has included multiple nations on multiple continents simultaneously, and no study has been conducted prospectively over 18 years. Furthermore, no other study has examined the following factors and their association with mortality: income per country (ie, low, lower-middle, upper-middle, and high income), hospital ownership (ie, private, public, university, and nonprofit), type of hospitalization (medical or surgical), ICU type (ie, cardiothoracic, coronary, medical, medical-surgical, neurosurgical, neurologic, adult oncology, pediatric oncology, pediatric, respiratory, surgical, trauma, and burn), device days (ie, central-line days, mechanical ventilation days, and urinary catheter days), device utilization ratio (ie, central-line usage, mechanical ventilation, and urinary catheter usage). We analyzed the impact of these variables and others on risk of mortality.
Methods
Study population and design
This multicenter, multinational, cohort, prospective study was carried out with patients admitted to 317 ICUs from 96 hospitals in 44 cities in 9 Asian countries over 18 years between March 27, 2004, and February 11, 2022.
Surveillance of healthcare-associated infections
The data were collected on each patient at the time of their ICU admission. From admission to discharge, infection prevention professionals (IPPs) went to the bedside of each patient daily. All adult and pediatric patients with or without HAIs admitted to an ICU were prospectively included in this investigation, and their data were collected from admission to discharge using the INICC Surveillance Online System (ISOS). Reference Rosenthal2, Reference Rosenthal, Maki and Graves3 IPPs visit the bedside of each hospitalized patient in the ICU with a computer tablet; they log into the ISOS and upload the patient’s data in real time. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3
At the time the patient is admitted, details about the setting (ie, country, city, hospital name, and type of ICU) are uploaded. Patient details are also input, including type of hospitalization (surgical or medical), use of invasive devices (central line, mechanical ventilator, urinary catheter, and/or peripheral catheter), and presence of infection. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3 IPPs further upload information about invasive devices (ie, central line, mechanical ventilator, urinary catheter, and/or peripheral catheter), positive cultures (ie, blood cultures, urine cultures, and respiratory samples), and medications every day until the patient is released. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3
If the patient has signs or symptoms of infection, an infectious disease specialist approaches the patient to determine the presence of an HAI (CLABSI, VAE, CAUTI, or other). According to the Centers for Disease Control and Prevention/National Healthcare Safety Network (CDC/NHSN), IPPs look at a patient’s signs and symptoms, cultures, radiographs, and other described criteria to fulfill definitions of HAI. Reference Emori, Culver and Horan15, Reference Horan, Andrus and Dudeck16 When IPPs upload the result of a culture to the ISOS, the ISOS immediately shows a message to the IPP and leads the IPP to an online module of the ISOS to check all the criteria of CDC NHSN to confirm the presence of an HAI and kind of HAI (CLABSI, VAE, CAUTI, or other). Reference Rosenthal2, Reference Rosenthal, Maki and Graves3
The ISOS automatically checks device utilization daily. From admission to discharge, the ISOS sends messages to the IPPs when a bias on bed days or use of devices is detected. If ISOS detects lack of use of any kind of device on any given day, it will send a message to IPP to remind him or her to upload missing devices or upload the discharge of the patient. If the patient is hospitalized in an ICU without any device in place, it is most probably because the IPP forgot to upload the use of the devices to the ISOS or forgot to upload the patient’s discharge. Thus, the ISOS requests that IPPs investigate why a patient in an ICU has no device in place. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3 This approach significantly reduces biases associated with device utilization, bed days, and discharge conditions. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3
Patients with missing data for their age or sex were excluded from this analysis. This study was approved by the institutional review boards of the hospitals involved. The identities of patients and hospitals are confidential.
Recorded data
Income group by World Bank (low, lower-middle, upper-middle, or high income), country name, city name, hospital name, hospital ownership (public, university, private, or nonprofit), ICU type, age, sex, device utilization (central-line days, mechanical ventilation days, urinary catheter days, central-line utilization ratio, mechanical ventilation utilization ratio, and/or urinary catheter utilization ratio), bed days, cultures, presence of HAI (CLABSI, VAE, or CAUTI), and clinical outcome at discharge (dead or alive) were recorded. The treatment outcome of HAIs was evaluated as discharge all-cause mortality.
INICC surveillance online system
According to standard CDC/NSHN methods, HAI numerators and denominators are device days collected from all patients as pooled data without specifying patient characteristics or the number of device days related to each patient. Reference Emori, Culver and Horan15, Reference Horan, Andrus and Dudeck16
INICC HAI surveillance is carried out using an online platform, the ISOS, which includes CDC/NHSN criteria and methods. Reference Emori, Culver and Horan15, Reference Horan, Andrus and Dudeck16 ISOS also adds the collection of patient-specific data on all patients, with and without HAI, including their risk factors and other variables per patient. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3
Data from all patients admitted to the ICU, infected or uninfected, allowed matching by various characteristics, which facilitated the estimation of risk factors for HAIs and mortality. Data were collected through the INICC ISOS, and the latest CDC/NHSN criteria and methodology were applied to diagnose HAIs, calculate HAI rates, and calculate device utilization ratio, among other variables. Reference Emori, Culver and Horan15, Reference Horan, Andrus and Dudeck16 Definitions of HAI used during surveillance were those published by CDC in 1991 Reference Emori, Culver and Horan15 and their subsequent updates through 2017. Reference Horan, Andrus and Dudeck16
Statistical analysis
Surviving and deceased patients were compared using multiple logistic regression. Independent variables are included in Table 2. Statistically significant variables were associated with an increased mortality risk. The test statistic used was the Wald test, and the statistical significance level was set at 0.05. Calculated using the outputs of logistic regression, adjusted odds ratios (aORs), and the corresponding 95% confidence intervals (CIs) of statistically significant variables were also reported. We estimated variables independently associated with the outcome adjusted to confounders. All statistical analyses were performed using R version 4.1.3 software (R Foundation for Statistical Computing, Vienna, Austria).
Results
From March 27, 2004, through February 11, 2022, a course of 18 years, a cohort, prospective, multicenter surveillance study of HAIs was conducted in 317 ICUs of 96 hospitals in 44 cities in 9 countries from Asia (China, India, Malaysia, Mongolia, Nepal, Pakistan, Philippines, Sri Lanka, Thailand, and Vietnam) currently participating in the INICC. Of all 96 hospitals, 20 (20.8%) were academic, 23 (23.9%) were public, 49 (51.1%) were private, and the remaining 4 (4.2%) were philanthropic. Among all 317 ICUs, 78 (24.6%) were medical-surgical, 59 (18.6%) were medical, 28 (8.8%) were pediatric, 34 (11.7%) were surgical, 29 (9.2%) were coronary, 16 (5.1%) were neurologic, 27 (8.5%) were neurosurgical. In addition, 7 ICUs (2.2%) were trauma units, and 10 ICUs (2.5%) were respiratory units. Furthermore, 5 ICUs (1.6%) were oncology pediatric ICUs, 13 (4.1%) were cardiothoracic ICUs, 5 (1.6%) were oncology adult ICUs, and 9 (9.2%) were burn units.
The length of participation of hospitals in INICC ranged from 1.23 to 201.93 months (mean, 40.50; SD, 41.31). More participating hospitals and patient characteristics are shown in Table 1. Data on 157,667 critical patients were gathered, and these patients were followed from admission to discharge from ICU during 957,517 bed days, and they acquired 8,157 HAIs. The mortality rate for patients at adult oncology ICUs was 29%. At medical ICUs, the mortality rate was 13%; at surgical ICUs, the mortality rate was 7%; at pediatric ICUs, the mortality rate was 7%, and at coronary ICUs, the mortality rate was 6%.
Table 1. Setting and Patient Characteristics
Note. ICU, intensive care unit; LOS, length of stay; CLABSI, central-line–associated bloodstream infection; VAE, ventilator-associated events; CAUTI, catheter associated urinary tract infection; SD, standard deviation.
Table 2. Multiple Logistic Regression Analysis of Risk Factors Associated With Death
Note. aOR, adjusted odds ratio; CI, confidence interval; ICU, intensive care unit; LOS, length of stay; CLABSI, central-line–associated bloodstream infection; VAE, ventilator-associated event; CAUTI, catheter-associated urinary tract infection.
Using multiple logistic regression, we identified the following variables as statistically significantly independently associated with a higher risk of mortality (Table 2): Acquisition of a CLABSI (aOR: 2.36; P < .0001), VAE (aOR,1.51; P < .0001), or CAUTI (aOR, 1.04; P < .0001) was associate with higher mortality risk. Female sex had a higher mortality risk than male sex (aOR, 1.06; P < .0001). Older age (aOR, 1.01; P < .0001) increased mortality risk by 1% per year. More bed days (aOR, 1.01; P < .0001) increased mortality risk by 1% per bed day. More central-line days (aOR, 1.02; P < .0001) increased mortality risk by 2% per central-line day. More urinary catheter days (aOR, 1.04; P < .0001) increased mortality risk by 4% per urinary catheter day. Higher mechanical ventilation utilization ratio (aOR, 12.48; P < .0001) increased mortality risk. Upper-middle income country was associated with higher mortality risk (aOR, 1.09; P = .033). The variables with the highest mortality risk were surgical hospitalization (aOR, 2.17; P < .0001), pediatric oncology ICU stay (aOR, 9.90; P < .0001), and adult oncology ICU stay (aOR, 4.52; P < .0001). Meanwhile, surgical ICU stay (aOR, 0.43; P = .01), cardiothoracic ICU stay (aOR, 0.41; P = .01), and surgical ICU stay (aOR, 0.24; P < .0001) showed the lowest mortality risks. Compared with public, private, and nonprofit hospitals, patients at university hospitals had the lowest mortality risk (aOR, 0.61; P < .0001).
Discussion
In this study, CLABSI, VAE, CAUTI, age, sex, length of stay, mechanical ventilation utilization ratio, central-line days, urinary catheter days, upper-middle income countries, surgical hospitalization, and oncology ICU were identified associated with an increased mortality risk. The associations between CLABSI, VAE, and CAUTI and increased mortality risk in our study are consistent with findings of Toufen et al, Reference Toufen, Franca, Okamoto, Salge and Carvalho17 in which mortality risk was independently associated with nosocomial infection in patients in the surgical ICU.
Older age was also associated with increased mortality risk in our study. We observed a 1% increase in mortality risk per year of age. Similarly, in a study at an ICU in northern Thailand, age was a risk factor for mortality. Reference Chittawatanarat, Jaipakdee, Chotirosniramit, Chandacham and Microbiology18 Additionally, we identified an association between female sex and mortality risk in our study. Conversely, in an ICU in southern India, male sex had a higher mortality risk. Reference Garg, Tellapragada and Shaw19
Moreover, we detected an association between prolonged length of stay and mortality, which is consistent with a previous study that analyzed risk factors for mortality in VAE in ICU patients in which prolonged hospital stay was an independent risk factor for mortality in multivariate analysis. Reference But, Yetkin and Kanyilmaz20
Furthermore, we identified an association between invasive mechanical ventilation utilization ratio and increased mortality risk. This finding is consistent with the findings of Toufen et al, Reference Toufen, Franca, Okamoto, Salge and Carvalho17 in which mortality was independently associated with the need for mechanical ventilation among patients in the surgical ICU.
Furthermore, we observed an association between central-line days and mortality. We detected a 2% increase in mortality risk per day of use of a central line. A similar association was shown in a study at a surgical ICU that analyzed risk factors and mortality of patients with intravenous catheter-related bloodstream infections. In that study, in multivariable analyses, catheterization of patients in the general wards was the sole independent risk factor of CRBSI occurrence (OR, = 8.67; P < .01) and males (OR, 7.20; P = .03) had the highest mortality risk. Reference Cheewinmethasiri, Chittawatanarat, Chandacham, Jirapongchareonlap and Chotirosniramit21
In this study, we further detected an association between urinary catheter days and increased mortality risk, with a 4% increased mortality per day of urinary catheter use. No other study has reported this association.
In the current study, patients in university hospitals had a significantly lower mortality risk than patients in private or public hospitals. Regarding this association, Eggleston et al Reference Eggleston, Shen, Lau, Schmid and Chan22 conducted a systematic review to analyze factors that explain the diversity of findings associated with the ownership and quality of hospitals. Government-controlled, for-profit hospitals had higher mortality rates than nonprofit hospitals. Reference Eggleston, Shen, Lau, Schmid and Chan22 Devereaux et al Reference Devereaux, Choi and Lacchetti23 conducted a systematic review and meta-analysis of studies to compare the mortality rates of private for-profit hospitals with those of and private nonprofit hospitals. They calculated the relative mortality risk of patients in private for-profit hospitals and compared it with the mortality risk for patients in private nonprofit hospitals. In this adult population, private for-profit hospitals were associated with an increased mortality risk compared to privately owned nonprofit hospitals. Reference Rosenthal2, Reference Rosenthal, Maki and Graves3
Additionally, patients in upper middle-income countries had a higher mortality risk than patients in lower middle-income countries. This finding can probably be explained by the inadequate quality of healthcare programs in the upper middle-income countries that participated in this study. Surgical hospitalization was associated with a significantly increased mortality risk than medical hospitalization in our study. To our knowledge, no other studies have analyzed this variable. Finally, patients admitted to oncology ICUs had the highest mortality risk, whereas patients admitted to coronary ICU had a significantly lower mortality risk. To our knowledge, this variable has not been analyzed in other research.
Some of the mortality risk factors identified in our study are unlikely to change, such as age, sex, the national economy, medical or surgical hospitalization, and type of ICU. However, some risk factors for death can be modified, such as length of stay, use of a central line, use of a urinary catheter, use of a mechanical ventilator, and acquisition of CLABSI, VAE, or CAUTI. As INICC data have already shown, HAI rates in low- and middle-income countries are 3–5 times higher than in the United States; therefore, there is room for improvement. Our recommendation for reducing mortality risk in ICUs is to focus on strategies to reduce length of stay, strategies to reduce utilization of mechanical ventilator, central lines, and urinary catheters, and to apply an evidence-based set of recommendation to prevent HAIs, such as those recently published by the Society for Healthcare Epidemiologists of America (SHEA), the Infections Diseases Society of America (IDSA), and the Association for Professionals in Infection Control and Epidemiology (APIC). Reference Klompas, Branson and Cawcutt24, Reference Buetti, Marschall and Drees25
This study had several limitations. This study is not representative of all hospitals of Asia because it relied on a surveillance system that included hospitals participating voluntarily. The hospitals that participate in this system are likely those with a better-quality program and surveillance and prevention of HAIs. Therefore, the rates of HAI shown in this study may be lower than the actual rates found worldwide. Lastly, data regarding the severity of illness score and underlying diseases were not collected, which may have contributed to identifying them as another mortality risk factor.
Acknowledgments
The authors thank the many healthcare professionals who assisted in conducting surveillance in their hospitals, the INICC advisory board, country directors, and secretaries who have so generously supported this unique international infection control network. The institutional review board of each hospital agreed to the study protocol, and patient confidentiality was protected by codifying the recorded information, making it only identifiable to the infection control team.
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors report no conflicts of interest related to this article.
