Antimicrobial Resistance Surveillance in Low- and Middle-Income Countries: Progress and Challenges in Eight South Asian and Southeast Asian Countries.

Antimicrobial resistance (AMR) is a serious global health threat and is predicted to cause significant health and economic impacts, particularly in low- and middle-income countries (LMICs). AMR surveillance is critical in LMICs due to high burden of bacterial infections; however, conducting AMR surveillance in resource-limited settings is constrained by poorly functioning health systems, scarce financial resources, and lack of skilled personnel. In 2015, the United Nations World Health Assembly endorsed the World Health Organization's Global Action Plan to tackle AMR; thus, several countries are striving to improve their AMR surveillance capacity, including making significant investments and establishing and expanding surveillance networks. Initial data generated from AMR surveillance networks in LMICs suggest the high prevalence of resistance, but these data exhibit several shortcomings, such as a lack of representativeness, lack of standardized laboratory practices, and underutilization of microbiology services. Despite significant progress, AMR surveillance networks in LMICs face several challenges in expansion and sustainability due to limited financial resources and technical capacity. This review summarizes the existing health infrastructure affecting the establishment of AMR surveillance programs, the burden of bacterial infections demonstrating the need for AMR surveillance, and current progress and challenges in AMR surveillance efforts in eight South and Southeast Asian countries.

2021 update to HIV-TRePS: a highly flexible and accurate system for the prediction of treatment response from incomplete baseline information in different healthcare settings.

OBJECTIVES: With the goal of facilitating the use of HIV-TRePS to optimize therapy in settings with limited healthcare resources, we aimed to develop computational models to predict treatment responses accurately in the absence of commonly used baseline data. METHODS: Twelve sets of random forest models were trained using very large, global datasets to predict either the probability of virological response (classifier models) or the absolute change in viral load in response to a new regimen (absolute models) following virological failure. Two 'standard' models were developed with all baseline variables present and 10 others developed without HIV genotype, time on therapy, CD4 count or any combination of the above. RESULTS: The standard classifier models achieved an AUC of 0.89 in cross-validation and independent testing. Models with missing variables achieved AUC values of 0.78-0.90. The standard absolute models made predictions that correlated significantly with observed changes in viral load with a mean absolute error of 0.65 log10 copies HIV RNA/mL in cross-validation and 0.69 log10 copies HIV RNA/mL in independent testing. Models with missing variables achieved values of 0.65-0.75 log10 copies HIV RNA/mL. All models identified alternative regimens that were predicted to be effective for the vast majority of cases where the new regimen prescribed in the clinic failed. All models were significantly better predictors of treatment response than genotyping with rules-based interpretation. CONCLUSIONS: These latest models that predict treatment responses accurately, even when a number of baseline variables are not available, are a major advance with greatly enhanced potential benefit, particularly in resource-limited settings. The only obstacle to realizing this potential is the willingness of healthcare professions to use the system.

Association of Intrapartum Risk Factors and Infant Clinical Indicators with Culture Confirmed Early Onset Neonatal Sepsis in a Secondary Care Rural Hospital in India.

OBJECTIVE: The aim of this study is to determine the association of intrapartum risk factors and infant clinical indicators using the National Institute for Health and Care Excellence (NICE) criteria with culture-positive early-onset neonatal sepsis (EONS) from a rural secondary healthcare facility where intrapartum prophylactic antibiotics are routinely administered to high-risk mothers. METHODS: A single-center prospective observational study was conducted between July 2017 and September 2018. All intramural neonates with at least one NICE criteria at less than 72 h of life, were included. Univariate logistic regression and multivariable logistic backward elimination analyses were conducted to investigate individual risk factors and predictive models for culture proven EONS. RESULTS: Of 236 newborns who were at risk for EONS by NICE criteria, 32 (13.8%) had positive blood cultures. Klebsiella species (n = 13, 39.4%) and Acinetobacter species (n = 11, 33.3%) were the most common isolated bacteria. In univariate analysis, the number of infant clinical indicators were associated with culture positive EONS (OR 1.36; 95% CI 1.01-1.81), but not the number of intrapartum risk factors (OR 0.76; 95% CI 0.4-1.29). The multivariate logistic regression with backward elimination procedure suggested that a model including absolute neutrophil count [adjusted OR (aOR) 0.81; 95% CI 0.72-0.92], C-reactive protein (aOR 1.24; 95% CI 1.08-1.43) and the number of clinical indicators (aOR 1.29; 95% CI 0.93-1.80) could be useful to predict culture positive EONS in our setting. CONCLUSION: In this maternal and neonatal cohort, infant clinical indicators rather than intrapartum risk factors were associated with culture confirmed EONS.

Acute undifferentiated fever in India: a multicentre study of aetiology and diagnostic accuracy.

BACKGROUND: The objectives of this study were to determine the proportion of malaria, bacteraemia, scrub typhus, leptospirosis, chikungunya and dengue among hospitalized patients with acute undifferentiated fever in India, and to describe the performance of standard diagnostic methods. METHODS: During April 2011-November 2012, 1564 patients aged ≥5 years with febrile illness for 2-14 days were consecutively included in an observational study at seven community hospitals in six states in India. Malaria microscopy, blood culture, Dengue rapid NS1 antigen and IgM Combo test, Leptospira IgM ELISA, Scrub typhus IgM ELISA and Chikungunya IgM ELISA were routinely performed at the hospitals. Second line testing, Dengue IgM capture ELISA (MAC-ELISA), Scrub typhus immunofluorescence (IFA), Leptospira Microscopic Agglutination Test (MAT), malaria PCR and malaria immunochromatographic rapid diagnostic test (RDT) Parahit Total™ were performed at the coordinating centre. Convalescence samples were not available. Case definitions were as follows: Leptospirosis: Positive ELISA and positive MAT. Scrub typhus: Positive ELISA and positive IFA. Dengue: Positive RDT and/or positive MAC-ELISA. Chikungunya: Positive ELISA. Bacteraemia: Growth in blood culture excluding those defined as contaminants. Malaria: Positive genus-specific PCR. RESULTS: Malaria was diagnosed in 17% (268/1564) and among these 54% had P. falciparum. Dengue was diagnosed in 16% (244/1564). Bacteraemia was found in 8% (124/1564), and among these Salmonella typhi or S. paratyphi constituted 35%. Scrub typhus was diagnosed in 10%, leptospirosis in 7% and chikungunya in 6%. Fulfilling more than one case definition was common, most frequent in chikungunya where 26% (25/98) also had positive dengue test. CONCLUSIONS: Malaria and dengue were the most common causes of fever in this study. A high overlap between case definitions probably reflects high prevalence of prior infections, cross reactivity and subclinical infections, rather than high prevalence of coinfections. Low accuracy of routine diagnostic tests should be taken into consideration when approaching the patient with acute undifferentiated fever in India.

An update to the HIV-TRePS system: the development of new computational models that do not require a genotype to predict HIV treatment outcomes.

OBJECTIVES: The optimal individualized selection of antiretroviral drugs in resource-limited settings is challenging because of the limited availability of drugs and genotyping. Here we describe the development of the latest computational models to predict the response to combination antiretroviral therapy without a genotype, for potential use in such settings. METHODS: Random forest models were trained to predict the probability of a virological response to therapy (<50 copies HIV RNA/mL) following virological failure using the following data from 22,567 treatment-change episodes including 1090 from southern Africa: baseline viral load and CD4 cell count, treatment history, drugs in the new regimen, time to follow-up and follow-up viral load. The models were assessed during cross-validation and with an independent global test set of 1000 cases including 100 from southern Africa. The models' accuracy [area under the receiver-operating characteristic curve (AUC)] was evaluated and compared with genotyping using rules-based interpretation systems for those cases with genotypes available. RESULTS: The models achieved AUCs of 0.79-0.84 (mean 0.82) during cross-validation, 0.80 with the global test set and 0.78 with the southern African subset. The AUCs were significantly lower (0.56-0.57) for genotyping. CONCLUSIONS: The models predicted virological response to HIV therapy without a genotype as accurately as previous models that included a genotype. They were accurate for cases from southern Africa and significantly more accurate than genotyping. These models will be accessible via the online treatment support tool HIV-TRePS and have the potential to help optimize antiretroviral therapy in resource-limited settings where genotyping is not generally available.

Predictors of delayed entry into medical care of children diagnosed with HIV infection: data from an HIV cohort study in India.

Data about the attrition before entry into care of children diagnosed with HIV in low- or middle-income countries are scarce. The aim of this study is to describe the attrition before engagement in HIV medical care in 523 children who were diagnosed with HIV from 2007 to 2012 in a cohort study in India. The cumulative incidence of children who entered into care was 87.2% at one year, but most children who did not enter into care within one year were lost to followup. The mortality before entry into care was low (1.3% at one year) and concentrated during the first three months after HIV diagnosis. Factors associated with delayed entry into care were being diagnosed after mother's HIV diagnosis, belonging to scheduled castes, age<18 months, female gender, and living >90 minutes from the HIV centre. Children whose parents were alive and were living in a rented house were at a higher risk of delayed entry into care than those who were living in an owned house. The results of this study can be used to improve the linkage between HIV testing and HIV care of children diagnosed with HIV in India.

Clinical features and risk factors for death in acute undifferentiated fever: A prospective observational study in rural community hospitals in six states of India.

BACKGROUND: Acute undifferentiated fever (AUF) ranges from self-limiting illness to life-threatening infections, such as sepsis, malaria, dengue, leptospirosis and rickettsioses. Similar clinical presentation challenges the clinical management. This study describes risk factors for death in patients hospitalized with AUF in India. METHODS: Patients aged ≥5 y admitted with fever for 2-14 d without localizing signs were included in a prospective observational study at seven hospitals in India during 2011-2012. Predictors identified by univariate analysis were analyzed by multivariate logistic regression for survival analysis. RESULTS: Mortality was 2.4% (37/1521) and 46.9% (15/32) died within 2 d. History of heart disease (p=0.013), steroid use (p=0.011), altered consciousness (p<0.0001), bleeding (p<0.0001), oliguria (p=0.020) and breathlessness (p=0.015) were predictors of death, as were reduced Glasgow coma score (p=0.005), low urinary output (p=0.004), abnormal breathing (p=0.006), abdominal tenderness (p=0.023), leucocytosis (p<0.0001) and thrombocytopenia (p=0.001) at admission. Etiology was identified in 48.6% (18/37) of fatal cases. CONCLUSIONS: Bleeding, cerebral dysfunction, respiratory failure and oliguria at admission, suggestive of severe organ failure secondary to systemic infection, were predictors of death. Almost half of the patients who died, died shortly after admission, which, together with organ failure, suggests that delay in hospitalization and, consequently, delayed treatment, contribute to death from AUF.

Microbiology Clinical Culture Diagnostic Yields and Antimicrobial Resistance Proportions before and during the COVID-19 Pandemic in an Indian Community Hospital and Two US Community Hospitals.

Studies comparing the impact of the COVID-19 pandemic on diagnostic microbiology culture yields and antimicrobial resistance proportions in low-to-middle-income and high-income countries are lacking. A retrospective study using blood, respiratory, and urine microbiology data from a community hospital in India and two community hospitals (Hospitals A and B) in St. Louis, MO, USA was performed. We compared the proportion of cultures positive for selected multi-drug-resistant organisms (MDROs) listed on the WHO's priority pathogen list both before the COVID-19 pandemic (January 2017-December 2019) and early in the COVID-19 pandemic (April 2020-October 2020). The proportion of blood cultures contaminated with coagulase-negative Staphylococcus (CONS) was significantly higher during the pandemic in all three hospitals. In the Indian hospital, the proportion of carbapenem-resistant (CR) Klebsiella pneumoniae in respiratory cultures was significantly higher during the pandemic period, as was the proportion of CR Escherichia coli in urine cultures. In the US hospitals, the proportion of methicillin-resistant Staphylococcus aureus in blood cultures was significantly higher during the pandemic period in Hospital A, while no significant increase in the proportion of Gram-negative MDROs was observed. Continuity of antimicrobial stewardship activities and better infection prevention measures are critical to optimize outcomes and minimize the burden of antimicrobial resistance among COVID-19 patients.

Early HIV viral load determination after initiating first-line antiretroviral therapy for indentifying patients with high risk of developing virological failure: data from a cohort study in a resource-limited setting.

OBJECTIVES: To evaluate the performance of a single determination of HIV viral load (VL) 6-12 months after starting antiretroviral therapy (ART) for identifying patients who will subsequently develop virological failure. METHODS: We selected HIV-infected patients with at least two VL determinations after 6 months of ART from an HIV cohort study in India. Patients were divided in two groups depending on whether the first VL was below (Group 1) or above (Group 2) 1000 copies/ml. Cut-off for virological failure was defined according to World Health Organization recommendation (>5000 copies/ml). RESULTS: The study included 584 patients and 560.1 person-years of follow-up. Of all virological failures, 83% were diagnosed at the first VL determination. The cumulative incidence of virological failure after 1 and 2 years since the first VL was 0.9% [95% confidence interval (CI), 0.3-2.7] and 1.7% (95% CI, 0.6-5), respectively, for Group 1, and 58.2% (95% CI, 47-69.7) and 63.1% (95% CI, 49.8-76.4), respectively, for Group 2. Compared with Group 1, patients from Group 2 had a hazard ratio for virological failure of 78.3 (95% CI, 27.8-220.2). CONCLUSIONS: A single VL determination after 6 months of ART was able to identify patients with high risk of virological failure.

Comparison of essential drug list in a rural secondary care hospital in south India with Indian & World Health Organization list 2011.

OBJECTIVE: Fixed drug combinations are a major marketing strategy in India but it can compromise the rational use of medicines. In this study we compared the fixed drug combinations and dosage forms in the hospital pharmacy before and after introducing the essential drug list. We also compared the Hospital Essential Drug List (HEDL) 2011 with the World Health Organization (WHO) Essential Drug List (EDL) 2011 and the National Essential Drug List of India (NEDL) 2011. METHODS: The study was done in a secondary level care charity hospital at Anantapur, AP with a bed size of 315 and an average OP per day of 1200-1700 visits. We compared the three essential drug lists (HEDL, WHOEDL and NEDL) and the hospital drug list before introducing EDL. Drugs which were present in NEDL and not present in the HEDL were also screened. Microsoft excel was used to tabulate the results and for graphs. RESULTS: The number of medicines used in the hospital before and after the introduction of the HEDL was 1627 and 424 respectively. On comparison, WHOEDL 2011 have 350 and NEDL of India have 348 medicines. While preparing the HEDL, 46 double drug combinations decreased to 15 and 9 triple drug combinations decreased to 1. In the case of injections, 20 double drug combinations decreased to 6 and 1 triple drug combination increased to 2. The number of tablets, capsules, injections, syrups, powders and inhalers was reduced to almost half. The great reductions were in 51 ointments to 9, 69 drops to 5, 11 paste to 0, 21 solutions to 3 and 14 creams to 1. The dosage forms removed included elixirs, insulin pens, gums, paste, paints, gargles and mouthwashes. CONCLUSIONS: There was drastic reduction in the number of medicines and dosage forms when the HEDL was implemented. Many of the fixed drug combinations were also removed for improving the rational use of medicines. The WHO essential drug list 2011, national essential drug list of India 2011 and the hospital essential drug list 2011 were comparable with few exceptions.

Development and Validation of the RCOS Prognostic Index: A Bedside Multivariable Logistic Regression Model to Predict Hypoxaemia or Death in Patients with SARS-CoV-2 Infection.

Introduction: Previous COVID-19 prognostic models have been developed in hospital settings and are not applicable to COVID-19 cases in the general population. There is an urgent need for prognostic scores aimed to identify patients at high risk of complications at the time of COVID-19 diagnosis. Methods: The RDT COVID-19 Observational Study (RCOS) collected clinical data from patients with COVID-19 admitted regardless of the severity of their symptoms in a general hospital in India. We aimed to develop and validate a simple bedside prognostic score to predict the risk of hypoxaemia or death. Results: 4035 patients were included in the development cohort and 2046 in the validation cohort. The primary outcome occurred in 961 (23.8%) and 548 (26.8%) patients in the development and validation cohorts, respectively. The final model included 12 variables: age, systolic blood pressure, heart rate, respiratory rate, aspartate transaminase, lactate dehydrogenase, urea, C-reactive protein, sodium, lymphocyte count, neutrophil count, and neutrophil/lymphocyte ratio. In the validation cohort, the area under the receiver operating characteristic curve (AUROCC) was 0.907 (95% CI, 0.892-0.922), and the Brier Score was 0.098. The decision curve analysis showed good clinical utility in hypothetical scenarios where the admission of patients was decided according to the prognostic index. When the prognostic index was used to predict mortality in the validation cohort, the AUROCC was 0.947 (95% CI, 0.925-0.97) and the Brier score was 0.0188. Conclusions: The RCOS prognostic index could help improve the decision making in the current COVID-19 pandemic, especially in resource-limited settings with poor healthcare infrastructure such as India. However, implementation in other settings is needed to cross-validate and verify our findings.

Association of Empiric Antibiotic Regimen Discordance With 30-Day Mortality in Neonatal and Pediatric Bloodstream Infection-A Global Retrospective Cohort Study.

BACKGROUND: While there have been studies in adults reporting discordant empiric antibiotic treatment associated with poor outcomes, this area is relatively unexplored in children and neonates despite evidence of increasing resistance to recommended first-line treatment regimens. METHODS: Patient characteristics, antibiotic treatment, microbiology, and 30-day all-cause outcome from children <18 years with blood-culture-confirmed bacterial bloodstream infections (BSI) were collected anonymously using REDCap™ through the Global Antibiotic Prescribing and Resistance in Neonates and Children network from February 2016 to February 2017. Concordance of early empiric antibiotic treatment was determined using European Committee on Antimicrobial Susceptibility Testing interpretive guidelines. The relationship between concordance of empiric regimen and 30-day mortality was investigated using multivariable regression. RESULTS: Four hundred fifty-two children with blood-culture-positive BSI receiving early empiric antibiotics were reported by 25 hospitals in 19 countries. Sixty percent (273/452) were under the age of 2 years. S. aureus, E. coli, and Klebsiella spp. were the most common isolates, and there were 158 unique empiric regimens prescribed. Fifteen percent (69/452) of patients received a discordant regimen, and 7.7% (35/452) died. Six percent (23/383) of patients with concordant regimen died compared with 17.4% (12/69) of patients with discordant regimen. Adjusting for age, sex, presence of comorbidity, unit type, hospital-acquired infections, and Gram stain, the odds of 30-day mortality were 2.9 (95% confidence interval: 1.2-7.0; P = 0.015) for patients receiving discordant early empiric antibiotics. CONCLUSIONS: Odds of mortality in confirmed pediatric BSI are nearly 3-fold higher for patients receiving a discordant early empiric antibiotic regimen. The impact of improved concordance of early empiric treatment on mortality, particularly in critically ill patients, needs further evaluation.

Lung disease caused by nontuberculous mycobacteria.

PURPOSE OF REVIEW: Although the incidence of tuberculosis has reduced in developed countries, there is a growing interest in nontuberculous mycobacteria (NTM) as a cause of lung disease. However, NTM are a heterogeneous group and most of the data come from only three species: Mycobacterium avium complex, Mycobacterium kansasii and Mycobacterium abscessus. Still, information about these three species is confusing because it is based mainly on retrospective studies and series of clinical cases performed in developed countries. In recent years, new information has appeared about other species and the pathogenesis of NTM. RECENT FINDINGS: Epidemiological studies show that NTM infection is a worldwide phenomenon with an increasing presence in developing countries perhaps because of the implementation of tap water. Women with characteristic phenotype are at higher risk of acquiring NTM infection along with patients with defects on cystic fibrosis transmembrane conductance regulators. New studies on Mycobacterium fortuitum, Mycobacterium xenopi, Mycobacterium szulgai and Mycobacterium simiae indicate that the American Thoracic Society criteria for diagnosing NTM disease may not be useful for all species of NTM. SUMMARY: New multicentric and prospective studies are needed to clarify the pathogenesis and treatment of NTM. These organisms form a numerous and heterogeneous group and each species should be studied separately.

Reduction in neutrophil count during hepatitis C treatment: drug toxicity or predictor of good response?

BACKGROUND: Bone marrow suppression is a well-recognized toxicity of the treatment of hepatitis C virus (HCV). Reduction of the peginterferon dose because of neutropenia is common in clinical practice. However, reduction of peginterferon dose during the first weeks of HCV treatment is associated with failure to achieve sustained virological response. AIMS: The objective of this study is to investigate whether the fall of neutrophil count during hepatitis C treatment is associated with achieving sustained virological response. METHODS: We performed an observational study of patients who completed peginterferon and ribavirin treatment in an Infectious Diseases Department in Manchester, UK. RESULTS: Of the 74 patients included in the analysis, 78% had genotype 2 or 3 hepatitis C and 15% had liver cirrhosis. Sustained virological response was achieved in 78% of patients. On univariate analysis, factors related to achieving sustained virological response were younger age, genotype 2 or 3, baseline neutrophil count, and fall of neutrophil count during treatment. Multivariate analysis showed baseline neutrophil count >3.5 x 10(3) cells/mm(3) [odds ratio (OR) 5.7; 95% confidence interval (CI) 1.24-26.3] and a reduction of neutrophil count >60% (OR 4.5; 95% CI 1.03-19.9) to be independently associated with achieving sustained virological response. Neutropenia was not associated with an increased risk of infections. CONCLUSIONS: In this observational study, higher baseline neutrophil count and fall of neutrophil count during the treatment of hepatitis C was associated with achieving sustained virological response. These findings could have important implications for the monitoring and management of HCV treatment with peginterferon if they are confirmed in other studies.

Factors associated with treatment failure of patients with psychiatric diseases and injecting drug users in the treatment of genotype 2 or 3 hepatitis C chronic infection.

BACKGROUND: Genotype 2/3 hepatitis C virus (HCV) has a good response to treatment with peginterferon and ribavirin. Patients with psychiatric disorders and injecting drug users (IDUs) are considered 'difficult to treat' and are often excluded from treatment despite the lack of evidence supporting this decision. AIMS: To investigate the outcome and factors associated with treatment failure in these groups. METHODS: This is an observational study of a cohort of patients infected by genotype 2/3 HCV. IDUs and patients with psychiatric diseases were not excluded from treatment. We performed an intention-to-treat analysis to evaluate factors related to treatment failure. RESULTS: A sustained virological response (SVR) was achieved in 91 of the 125 patients treated (72.8%). Patients with chronic psychotic disorders or former IDUs had SVR rates similar to other groups. After multivariate analysis, independent factors associated with treatment failure were liver cirrhosis [odds ratio (OR) 3.4, 95% confidence interval (CI) 1.1-10.4], a history of depression and not being on antidepressants at the commencement of HCV treatment (OR 4.4, 95% CI 1.2-16) and active IDUs (OR 7.3, 95% CI 1.77-30.4). CONCLUSIONS: Patients with a history of depression who were not receiving antidepressants and active IDUs are more likely to fail treatment for genotype 2/3 HCV and will need additional support.

Is the efficacy of antibiotic prophylaxis for surgical procedures decreasing? Systematic review and meta-analysis of randomized control trials.

OBJECTIVE: Rising antibiotic resistance could reduce the effectiveness of antibiotics in preventing postoperative infections. We investigated trends in the efficacy of antibiotic prophylaxis regimens for 3 commonly performed surgical procedures-appendectomy, cesarean section, and colorectal surgery-and 1 invasive diagnostic procedure, transrectal prostate biopsy (TRPB). DESIGN: Systematic review and meta-analysis. METHODS: We searched PubMed and Cochrane databases (through October 31, 2017) for randomized control trials (RCTs) that measured the efficacy of antibiotic prophylaxis for 4 index procedures in preventing postoperative infections (surgical site infections [SSIs] following the 3 surgical procedures and a combination of urinary tract infections [UTIs] and sepsis following TRPB). RESULTS: Of 399 RCTs, 74 studies (9 appendectomy, 11 cesarean section, 39 colorectal surgery, and 15 TRPB) were included. Multilevel logistic regression models with random intercepts for each study showed no statistically significant increase in SSIs over time for appendectomy (adjusted odds ratio [aOR] per year, 1.03; 95% confidence interval [CI], 0.92-1.16; P=.57), cesarean section (aOR per year, 1.01; 95% CI, 0.96-1.05; P=.80), and TRPB (aOR per year, 0.95; 95% CI, 0.77-1.18; P=.67). However, there was a significant increase in SSIs proportion following colorectal surgery (aOR per year, 1.049; 95% CI, 1.03-1.07; P<.001). CONCLUSION: The efficacy of antibiotic prophylaxis agents in preventing SSIs following colorectal surgery has declined. Small number of RCTs and low infections rates limited our ability to assess true effect for simple appendectomy, cesarean section, or TRPB.

Quantifying uncertainty about future antimicrobial resistance: Comparing structured expert judgment and statistical forecasting methods.

The increase of multidrug resistance and resistance to last-line antibiotics is a major global public health threat. Although surveillance programs provide useful current and historical information on the scale of the problem, the future emergence and spread of antibiotic resistance is uncertain, and quantifying this uncertainty is crucial for guiding decisions about investment in antibiotics and resistance control strategies. Mathematical and statistical models capable of projecting future rates are challenged by the paucity of data and the complexity of the emergence and spread of resistance, but experts have relevant knowledge. We use the Classical Model of structured expert judgment to elicit projections with uncertainty bounds of resistance rates through 2026 for nine pathogen-antibiotic pairs in four European countries and empirically validate the assessments against data on a set of calibration questions. The performance-weighted combination of experts in France, Spain, and the United Kingdom projected that resistance for five pairs on the World Health Organization's priority pathogens list (E. coli and K. pneumoniae resistant to third-generation cephalosporins and carbapenems and MRSA) would remain below 50% in 2026. In Italy, although upper bounds of 90% credible ranges exceed 50% resistance for some pairs, the medians suggest Italy will sustain or improve its current rates. We compare these expert projections to statistical forecasts based on historical data from the European Antimicrobial Resistance Surveillance Network (EARS-Net). Results from the statistical models differ from each other and from the judgmental forecasts in many cases. The judgmental forecasts include information from the experts about the impact of current and future shifts in infection control, antibiotic usage, and other factors that cannot be easily captured in statistical forecasts, demonstrating the potential of structured expert judgment as a tool for better understanding the uncertainty about future antibiotic resistance.

Predicting Virological Response to HIV Treatment Over Time: A Tool for Settings With Different Definitions of Virological Response.

OBJECTIVE: Definitions of virological response vary from <50 up to 1000 copies of HIV-RNA/mL. Our previous models estimate the probability of HIV drug combinations reducing the viral load to <50 copies/mL, with no indication of whether higher thresholds of response may be achieved. Here, we describe the development of models that predict absolute viral load over time. METHODS: Two sets of random forest models were developed using 50,270 treatment change episodes from more than 20 countries. The models estimated viral load at different time points following the introduction of a new regimen from variables including baseline viral load, CD4 count, and treatment history. One set also used genotypes in their predictions. Independent data sets were used for evaluation. RESULTS: Both models achieved highly significant correlations between predicted and actual viral load changes (r = 0.67-0.68, mean absolute error of 0.73-0.74 log10 copies/mL). The models produced curves of virological response over time. Using failure definitions of <100, 400, or 1000 copies/mL, but not 50 copies/mL, both models were able to identify alternative regimens they predicted to be effective for the majority of cases where the new regimen prescribed in the clinic failed. CONCLUSIONS: These models could be useful for selecting the optimum combination therapy for patients requiring a change in therapy in settings using any definition of virological response. They also give an idea of the likely response curve over time. Given that genotypes are not required, these models could be a useful addition to the HIV-TRePS system for those in resource-limited settings.