PROGRESS: Point-of-caRe diagnOstics to Guide appRopriate antimicrobial thErapy of urinary tract infectionS in nursing homeS

Antimicrobial resistance (AMR) is on the rise worldwide and is mainly driven by inappropriate antibiotic use in humans and animals. Nursing homes are increasingly regarded as an important reservoir for the emergence of AMR. Indeed, surveillance studies in Dutch nursing homes have shown high rates of AMR against commonly used antibiotics with substantial differences between homes, even in the same region. Suspected urinary tract infections (UTI) rank among the most common reasons for antibiotic use in nursing homes. However, diagnosing UTI in this setting is challenging because of frequent non-specific symptomatology combined with a high prevalence of asymptomatic bacteriuria, which complicates attribution of causality to detection of bacteria in urine. The difficulty of distinguishing true UTI from bacterial colonization of the urinary tract results in frequent inappropriate antibiotic use. Given the diagnostic challenge and the high and variable AMR prevalence, unmet needs exists for (1) point-of-care (POC) diagnostic tests to support clinical rules for diagnosing UTIs, and (2) timely availability of AMR prevalence data at the level of individual nursing homes to guide empirical treatment choices. Both needs are addressed by the PROGRESS project. Procalcitonin (PCT) and C-reactive protein (CRP) are inflammatory blood markers that have been proven useful to support diagnosis and monitoring of (bacterial) respiratory tract infections and sepsis. While limited studies suggest their usefulness in supporting UTI diagnosis, their utility has not been studied in elderly populations for this purpose. In two consecutive prospective studies, PROGRESS will assess and compare the utility of rapid POC measurements of blood CRP and PCT levels to support clinical rules for diagnosing UTI in nursing home residents (study A), followed by assessment of the impact of the best performing POC test on antibiotic exposure (study B). Study A is a 1-year matched diagnostic accuracy study comparing the sensitivities of both POC tests for UTI diagnosis in 600 nursing home residents suspected of UTI by the attending physician. For the purpose of this study, a ‘true UTI’ is defined post-hoc using stringent criteria, including microbiology results and clinical response to adequate antibiotic therapy. The best performing test will be investigated further in study B, which is an 18-month cluster randomized study in 8 nursing homes (100 subjects per home) to assess the impact of POC testing on antibiotic prescription in nursing home residents with suspected UTI. Testing will be performed on-site and results are communicated immediately to the attending physician accompanied by an interpretation and recommendation (e.g. “UTI likely: consider antibiotic treatment” and “UTI not likely: consider withholding antibiotic treatment”). The primary outcome measure is the total duration of antibiotic exposure during 30 days after study enrolment. Secondary outcomes include the duration of initial antibiotic therapy, and the rates of clinical resolution, recurrences and complications (e.g. hospitalization). Since effectiveness of a new test does not just depend on its proven efficacy in research studies, but also on successful implementation after the study, qualitative implementation studies will be conducted in parallel to both clinical studies. These studies are aimed at identifying facilitators and key barriers for implementation, followed by development and evaluation of implementation strategies. Finally, results from bacterial urine cultures performed in the course of studies A and B will be used to assess the usefulness of Lot Quality Assurance Sampling (LQAS)–based surveillance in providing relevant AMR prevalence data to guide local empirical treatment choices. LQAS is a classification tool originating from industry in which pre-defined prevalence thresholds of a relevant outcome (e.g. faulty products) leads to responsive action (e.g. rejection of production batches). In our setting, thresholds refer to the prevalence of antimicrobial resistance (AMR) and batches to nursing homes. A classification tool requires markedly smaller sample sizes (tens of bacterial isolates) than conventional surveillance to assess precise AMR prevalence (hundreds of bacterial isolates), while such classification is highly compatible with clinical practice since guidelines for empirical treatment choices are typically based on threshold prevalences of AMR. LQAS thus provides opportunities to generate relevant local AMR data in a timely fashion. LQAS classifications will be determined for relevant antibiotics in each participating nursing home which will be evaluated in the light of current local empirical treatment. Taken together, the combined approaches addressed in PROGRESS will have the important potential to reduce AMR in nursing homes by better informed decisions about who to treat and how to treat.