A healthcare failure mode and effect analysis to optimise the process of blood culture performance.

BACKGROUND: Blood cultures are essential diagnostic tools to identify pathogens in systemic infections. However, logistics of blood culture performance is often suboptimal. This study analyses the pre-analytic phase of blood culture processing through different types of risk assessments. METHODS: We performed direct observations to gain in-depth knowledge of the root causes of suboptimal blood culture performance. These findings were summarised in a Bow-Tie chart. We then utilised a healthcare failure mode and effect analysis to prioritise failures per step in the process and to organise improvement activities. Finally, improvement actions were planned. RESULTS: Not obtaining a second set of blood cultures in the logistics of blood culture performance had the highest priority for action. Several failure modes, including human and system factors, were identified. Improvement actions included training and clinical lessons for nurses in the emergency department, updating hospital search engines to ease identification of relevant protocols, and an evaluation of the workload at the emergency department. Failure modes caused by human factors appear easy to address, however changing human behaviour is challenging. CONCLUSIONS: The analysis provided useful insight into the different steps in the logistics of blood culture performance and facilitated the organisation of actions focused on addressing the most urgent root causes.

Clinical condition and comorbidity as determinants for blood culture positivity in patients with skin and soft-tissue infections.

The utility of performing blood cultures in patients with a suspected skin infection is debated. We investigated the association between blood culture positivity rates and patients' clinical condition, including acute disease severity and comorbidity. We performed a retrospective study, including patients with cellulitis and wound infection who had been enrolled in three Dutch multicenter studies between 2011 and 2015. Patients' acute clinical condition was assessed using the Modified Early Warning Score (MEWS; severe: MEWS ≥2) and comorbidity with the Charlson Comorbidity Index (CCI; severe: CCI ≥2). A total of 334 patients with a suspected skin infection were included. Blood cultures were performed in 175 patients (52%), 28 of whom (16%) had a positive blood culture. Data on the clinical condition were collected in 275 patients. Blood cultures were performed in 76% of the patients with a severe acute condition, compared with 48% with a non-severe acute condition (OR 3.5; 95% confidence interval: 2.0-6.2; p < 0.001). Blood cultures were positive in 18% and 12% respectively (OR 1.7 (0.7-4.1); p = 0.3). Blood cultures were performed in 53% of patients with severe comorbidity, compared with 61% without severe comorbidity (OR 0.7; 0.4-1.2; p = 0.2). Blood cultures were positive in 25% and 10% respectively (OR = 3.1; 1.2-7.5; p = 0.02). The blood culture positivity rate among hospitalized patients diagnosed with skin infections was higher than the rates reported by the Infectious Diseases Society of America guidelines, particularly in patients with severe comorbidity. Therefore, the recommendations concerning blood culture performance in patients with a skin infection should be reconsidered.

Effect of an antibiotic checklist on length of hospital stay and appropriate antibiotic use in adult patients treated with intravenous antibiotics: a stepped wedge cluster randomized trial.

OBJECTIVES: Quality indicators (QIs) have been developed to define appropriate antibiotic use in hospitalized patients. We evaluated whether a checklist based on these QIs affects appropriate antibiotic use and length of hospital stay. METHODS: An antibiotic checklist for patients treated with intravenous antibiotics was introduced in nine Dutch hospitals in a stepped wedge cluster randomized trial. Prophylaxis was excluded. We included a random sample before (baseline), and all eligible patients after (intervention) checklist introduction. Baseline and intervention outcomes were compared. Primary endpoint was length of stay (LOS), analysed by intention to treat. Secondary endpoints, including QI performances, QI sum score (performance on all QIs per patient), and quality of checklist use, were analysed per protocol. RESULTS: Between 1 November 2014 and 1 October 2015 we included 853 baseline and 5354 intervention patients, of whom 993 (19%) had a completed checklist. The LOS did not change (baseline geometric mean 10.0 days (95% CI 8.6-11.5) versus intervention 10.1 days (95% CI 8.9-11.5), p 0.8). QI performances increased between +3.0% and +23.9% per QI, and the percentage of patients with a QI sum score above 50% increased significantly (OR 2.4 (95% CI 2.0-3.0), p<0.001). Higher QI sum scores were significantly associated with shorter LOS. Discordance existed between checklist-answers and actual performance. CONCLUSIONS: Use of an antibiotic checklist resulted in a significant increase in appropriateness of antibiotic use, but not in a reduction of LOS. Low overall checklist completion rates and discordance between checklist-answers and actual provided care might have attenuated the impact of the checklist.

A survey to identify barriers of implementing an antibiotic checklist.

A checklist is an effective implementation tool, but addressing barriers that might impact on the effectiveness of its use is crucial. In this paper, we explore barriers to the uptake of an antibiotic checklist that aims to improve antibiotic use in daily hospital care. We performed an online questionnaire survey among medical specialists and residents with various professional backgrounds from nine Dutch hospitals. The questionnaire consisted of 23 statements on anticipated barriers hindering the uptake of the checklist. Furthermore, it gave the possibility to add comments. We included 219 completed questionnaires (122 medical specialists and 97 residents) in our descriptive analysis. The top six anticipated barriers included: (1) lack of expectation of improvement of antibiotic use, (2) lack of expected patients' satisfaction by checklist use, (3) lack of feasibility of the checklist, (4) negative previous experiences with other checklists, (5) the complexity of the antibiotic checklist and (6) lack of nurses' expectation of checklist use. Remarkably, 553 comments were made, mostly (436) about the content of the checklist. These insights can be used to improve the specific content of the checklist and to develop an implementation strategy that addresses the identified barriers.