Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Name | Class |
|---|---|
| Wellcome Trust | OTHER |
Not provided
Not provided
Not provided
Not provided
The main goal of this retrospective observational study is to understand how stepping down antibiotic treatment (called antibiotic de-escalation) affects patients who receive it compared to those who don't after received a short-course (≤7 days) of parenteral antibiotics. The investigators will use past medical records from four public referral hospitals in Thailand from the year 2019 to 2024. The investigators will firstly evaluate which types of patients are more likely to receive antibiotic de-escalation. Then, the investigators will estimate the impact of antibiotic de-escalation, while taking those differences into account. This way, it will help us understand the impact of antibiotic de-escalation in real-world clinical practice. The investigators also aim to assess how accurate automated outbreak detection systems are at detecting outbreaks, evaluate patterns of antimicrobial use and antimicrobial-resistant infections, and develop new indicators for antimicrobial stewardship that are applicable for local and national actions in low and middle-income countries.
Rationale:
Despite the availability of guidance documents for implementing antimicrobial stewardship (AMS) programs at both hospital and national levels, there is a lack of robust data to monitor and evaluate the current antimicrobial use practice in low and middle-income countries (LMICs). Evidence on the impact of antibiotic de-escalation on antimicrobial-resistant (AMR) infection is limited. In addition, practical guidelines on how to utilize and analyze routine data for LMICs have not been established.
Objective:
In this study, our primary objective is (a) to evaluate impact of antibiotic de-escalation. Our secondary objectives are (b) to estimate accuracy of automated outbreak detection systems, (c) to evaluate epidemiology of antimicrobial use (AMU) and AMR infections and associations between AMU and AMR infections, and (d) to develop AMS outcome indicators that are applicable for local and national actions in low and middle-income countries (LMICs)
Methodology:
The investigators will conduct a retrospective data analysis using individual-level electronic databases of hospital admission, ward transfer, microbiology laboratory and drug dispensing from four hospitals (including Chiangrai Prachanukroh Hospital, Chiangrai; Sunpasitthiprasong Hospital, Ubon Ratchathani; Phrachomklao Hospital, Phetchaburi; and Chaoprayayommarat Hospital, Suphanburi) from January 2019 to December 2024 in Thailand.
The hospital admission data collected will include hospital number (HN) and admission number (AN), sex, age, admission wards, admission dates, discharge dates, discharge outcomes (discharge type and discharge status) and ICD-10. The ward transfer data collected will include HN, AN, ward, data transfer in and date transfer out. The microbiology laboratory data collected will include HN, AN, ward, specimen type, specimen collection date, report dates, culture results and antimicrobial susceptibility results. The drug dispensing data collected will include, HN, AN, drug name, drug code, route of drug administration, the dosage regimen, drug start dates, drug stop dates, wards and departments of prescribing physicians. The HN and AN will be used to link the three databases together. The ward data will also be used to evaluate the cluster of AMR infection. The department of prescribing physicians will be used to understand AMS and antimicrobial use (AMU) by department because multiple wards are mixed wards (i.e. having patients from multiple departments in the same wards). Infection and Prevention Control (IPC) team's records of cluster will be used to evaluate the accuracy of automated outbreak detection systems. All data will be protected at the highest security.
Inverse Probability of Treatment Weighting (IPTW) will be used to evaluate impact of antibiotic de-escalation on new AMR BSI. Descriptive analysis and regression models will also be used.
Outcomes:
In this study, our primary outcomes are (a1) the relative risk of new AMR BSI in patients who receive antibiotic de-escalation compared to those who do not after a short-course (≤7 days) of parenteral antibiotics, and (a2) the relative risk of in-hospital 30-day mortality. Our secondary outcomes include (b) the accuracy of automated outbreak detection systems, (c) factors associated with AMU and AMR infections, and (d) factors associated with new AMS outcome indicators.
Findings of this study, particularly impact of antibiotic de-escalation on new AMR BSI and new AMS outcome indicators will be used to expand AutoMated tool for Antimicrobial resistance Surveillance System (AMASS) with AMS modules (i.e. AMASS version 4.0), under collaboration with the Ministry of Public Health Thailand. The investigators anticipate that findings of this study will support the monitoring and evaluation of AMS practice at both hospital and national levels in Thailand and other LMICs in the future.
Not provided
Not provided
Not provided
Not provided
Not provided
| Label | Type | Description | Intervention Names |
|---|---|---|---|
| No antibiotic de-escalation | No antibiotic de-escalation group is defined as patients who are still hospitalized and are receiving a parenteral antibiotic classified as Medium Watch (e.g. Piperacillin-tazobactam), High Watch (e.g. carbapenem) or Reserve (e.g. Tigecycline, Fosfomycin, and Colistin) on day 8 after starting a parenteral antibiotic therapy. Only patients who receive parenteral antibiotics for at least four consecutive days are included as a proxy for presumed severe infection. Day 8 is used as a proxy to represent the day immediately following the end of a short-course (≤7 days) of parenteral antibiotics. | ||
| De-escalation by using Access or Low Watch parenteral antibiotics | De-escalation by using Access or Low Watch parenteral antibiotics group is defined as patients who are still hospitalized and are receiving a parenteral antibiotic classified as Access (e.g. ampicillin, gentamicin and amoxicillin-clavulanic acid) or Low Watch (e.g. ceftriaxone) on day 8 after starting a parenteral antibiotic therapy. Only patients who receive parenteral antibiotics for at least four consecutive days are included as a proxy for presumed severe infection. Day 8 is used as a proxy to represent the day immediately following the end of a short-course (≤7 days) of parenteral antibiotics. | ||
| De-escalation by cessation of parenteral antibiotics | De-escalation by cessation of parenteral antibiotics group is defined as patients who are still hospitalized and are not receiving any parenteral antibiotics on day 8 after starting a parenteral antibiotic therapy. Only patients who receive parenteral antibiotics for at least four consecutive days are included as a proxy for presumed severe infection. Day 8 is used as a proxy to represent the day immediately following the end of a short-course (≤7 days) of parenteral antibiotics. |
Not provided
| Measure | Description | Time Frame |
|---|---|---|
| New antimicrobial-resistant bloodstream infection | Antimicrobial-resistant (AMR) bloodstream infection (BSI) is defined as having blood culture positive for methicillin-resistant S. aureus (MRSA), vancomycin-resistant Enterococcus spp., 3rd-generation cephalosporin-resistant Gram-negative bacterium and carbapenem-resistant Gram-negative bacterium from day 8 to day 30 after starting a parenteral antibiotic therapy. | 6 years |
| In-hospital 30-day mortality | In-hospital mortality is defined as having discharge outcomes recorded as 'death without autopsy', 'death with autopsy' or 'died'. In hospital 30-day mortality will be defined as having in-hospital mortality from day 8 to day 30 after starting a parenteral antibiotic therapy from day 8 to day 30 after starting a parenteral antibiotic therapy. | 6 years |
| Measure | Description | Time Frame |
|---|---|---|
| Sensitivity of automated outbreak detection systems | Sensitivity is defined as the proportion of positive cluster signals identified by automated outbreak systems among all clusters defined by expert opinion. Data from the year 2019 is used as the initial baseline for automated outbreak detection systems. | 5 years, from 2020 to 2024 |
Not provided
For primary objectives Inclusion Criteria
Exclusion Criteria
Antimicrobial-resistant (AMR) organism is defined as an organism that is resistant to Access and Low Watch antibiotics, and if the organism is the cause of infection, the recommended antimicrobial therapy involves the use of Medium Watch, High Watch or Reserve antibiotics. The common organisms include methicillin-resistant S. aureus, methicillin-resistant coagulase-negative Staphylococcus spp., ampicillin-resistant Enterococcus spp., vancomycin-resistant Enterococcus spp., 3rd-generation cephalosporin-resistant Gram-negative bacterium and carbapenem-resistant Gram-negative bacterium. The definition of organism includes organisms frequently associated with contamination including coagulase-negative staphylococci, viridans group streptococci, Corynebacterium spp., Bacillus spp., Diptheroid spp., Micrococcus spp. and Propionibacterium spp.. All types of specimens are included (e.g. sputum and tracheal suction). We excluded such patients because the study has no clinical data to differentiate whether the isolated AMR organisms are causing infections or represent colonization.
For secondary objectives
Inclusion Criteria:
Exclusion Criteria:
• Admitted as day admissions to four collaborating hospitals from 1 Jan 2019 to 31 Dec 2024
Not provided
Not provided
Not provided
Not provided
All patients admitted to four collaborating hospitals from 1 Jan 2019 to 31 Dec 2024. We expect that the total sample size would be about 10,800,000 inpatients.
| Name | Role | Phone | Extension | |
|---|---|---|---|---|
| Direk Limmathurotsakul, MD, PhD | Contact | +66 85 998 7679 | direk@tropmedres.ac | |
| Preeyarach Klaytong | Contact | +66 89 896 3419 | preeyarach@tropmedres.ac |
| Name | Affiliation | Role |
|---|---|---|
| Direk Limmathurotsakul, MD, PhD | Mahidol Oxford Tropical Medicine Research Unit | Principal Investigator |
| Facility | Status | City | State | ZIP | Country | Contacts |
|---|---|---|---|---|---|---|
| Chiangrai Prachanukroh Hospital | Chiang Rai | Chiangrai | Thailand |
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| ID | Term |
|---|---|
| D016470 | Bacteremia |
| ID | Term |
|---|---|
| D001424 | Bacterial Infections |
| D001423 | Bacterial Infections and Mycoses |
| D007239 | Infections |
| D018805 | Sepsis |
Not provided
Not provided
Not provided
Not provided
Not provided
| Specificity of automated outbreak detection systems |
Specificity is defined as the proportion of ward-weeks without positive cluster signals identified by automated outbreak systems among all ward-weeks without clusters defined by expert opinion. Data from the year 2019 is used as the initial baseline for automated outbreak detection systems. |
| 5 years, from 2020 to 2024 |
| Positive predictive value (PPV) of automated outbreak detection systems | PPV is defined as the proportion of clusters defined by expert opinion among all positive cluster signals. Data from the year 2019 is used as the initial baseline for automated outbreak detection systems. | 5 years, from 2020 to 2024 |
| Negative predictive value (NPV) of automated outbreak detection systems | NPV is defined as the proportion of ward-weeks without positive cluster signals defined by expert opinion among all ward-weeks without positive cluster signals. Data from the year 2019 is used as the initial baseline for automated outbreak detection systems. | 5 years, from 2020 to 2024 |
| Day of therapy (DOT) per 1,000 bed-days | DOT is defined as the number of days that a patient receives an antimicrobial agent (regardless of dose). Any dose of an antimicrobial that is received during a calendar day represents 1 DOT. | 6 years |
| Length of therapy (LOT) per 1,000 bed-days | LOT is defined as the number of days a patient receives any antimicrobial agents (irrespective of the number of different antimicrobials). Any day of at least one antimicrobial that is received represent 1 LOT. | 6 years |
| Frequency of antimicrobial resistance bloodstream infection | Frequency of AMR BSI per 100,000 tested patients, per 100,000 admissions and per 100,000 bed-days will be used to assess the level of antimicrobial resistance infection. Community-origin BSI is defined as a confirmed BSI occurring in an individual who has been admitted to a hospital for two or less calendar days, with calendar day one equal to the day of admission. Hospital-origin BSI is defined as a confirmed BSI occurring in an individual who has been admitted to a hospital for more than two calendar days, with calendar day one equal to the day of admission. | 6 years |
| Utility of new AMS indicators | 'Proportion of patients who had a blood culture taken within ±1 day when a parenteral antibiotic was started' among 'patients who received parenteral antibiotics for at least four consecutive days (i.e. a proxy of patients with severe infection)' will be evaluated as a potential new indicator for diagnostic stewardship (as part of AMS). 'Proportion of patients who received antibiotic de-escalation after a short-course (≤7 days) antibiotic therapy' among 'patients who received parenteral antibiotics for at least four consecutive days (i.e. a proxy of patients with severe infection) and who were still hospitalized on day 8' will be evaluated as a potential new indicator for AMS. Factors associated these outcomes will be evaluated using regression models. Difference in these proportion between wards, departments and hospitals will be explored to assess the potential utility of the indicator. | 6 years |
| Phrachomklao Hospital | Phetchaburi | Thailand |
|
| Chaoprayayommarat Hospital | Suphan Buri | Thailand |
|
| Sunpasitthiprasong Hospital | Ubon Ratchathani | Thailand |
|
| D018746 |
| Systemic Inflammatory Response Syndrome |
| D007249 | Inflammation |
| D010335 | Pathologic Processes |
| D013568 | Pathological Conditions, Signs and Symptoms |