Patient expectations do matter - experimental evidence on antibiotic prescribing decisions

Abstract   Inappropriate prescription of antibiotics remains a major contributor to the global antimicrobial resistance crisis despite clear linkages between antibiotic utilization and resistance spread. This study aims to better understand the simultaneous and independent effect of previous prescription behavior, patient expectation, and clinical uncertainty on antibiotic prescribing. This discrete choice experiment was embedded within a routine organizational climate survey administered to all physicians working in the Tuscany healthcare system administered between Nov 11 and Nov 20, 2019 (Qualtrics). Participants were provided with a patient encounter vignette and subsequently asked to in which of two alternatives they were more likely to prescribe antibiotics. The two alternatives varied in levels of clinical uncertainty, patient expectations, and the physician’s past behavior. We fitted a conditional logistic regression model. Respondents included 1,436 hospital-based physicians, of which 52% were female, 78% practiced in a general hospital setting, and 33% were between the ages of 50 and 59. Results show that the odds of prescribing antibiotics decrease when a patient requests it (OR = 0.80, 95%CI [0.72,0.89]) and increase when the physician has prescribed antibiotics to a patient under similar circumstances previously (OR = 1.15, 95%CI [1.03,1.27]). We found no significant effect of clinical uncertainty on the odds of prescribing antibiotics (OR = 0.96, 95%CI [0.87, 1.07]). We show that patient expectation has a significant negative association with antibiotic prescribing among hospital-based physicians. Our findings inform the design of antibiotic stewardship programs in Tuscany and highlight the importance of cultural context in shaping the physician’s disposition when confronted with patient expectations. We suggest shared decision-making to improve prudent prescribing without compromising on patient satisfaction. Key messages • Health administrators should address patient expectations when designing hospital antibiotic stewardship programs. • Physicians’ past prescribing behaviour influences antibiotic prescribing decisions and should be considered during intervention design.


Background:
The COVID-19 pandemic led to an 'infodemic', as defined by the WHO, which made it difficult to be accurately informed on public health topics. For this purpose, many people use social media as a source of information, mainly YouTube. Given the great resonance of this platform, our study aims at assessing quality and reliability of its content regarding the COVID-19 vaccination. Methods: During March 2022, six searches were performed on the Italian YouTube platform using the following terms: ''Covid vaccination '',''Covid vaccine'',''Coronavirus vaccination'',''Coronavirus vaccine'', vaccine''. A total of 329 videos were analysed, after removing 271 duplicated videos, and classified in seven types of channel. The reliability of the content was evaluated through the HoNCode score, while quality was tested using the validated DISCERN tool.

Results:
The most frequent category was 'Internet Media' (33%), while the less frequent one was 'Educational Medical' (7%). The content reliability (i.e. HoNCode score) resulted higher for videos produced by medical healthcare workers than nonmedical ones. Concerning the quality, the DISCERN score resulted significantly higher for the Educational channels (median 46.0 for medical and 41.3 non-medical ones) as compared to Internet Media (26.5) and New Agencies (24.3).

Conclusions:
Although YouTube has implemented a policy against misinformation related to the COVID-19 vaccination, the study highlights that there is extreme heterogeneity in reliability and quality of videos. Content produced by non-medical users, especially ''Internet Media'' and ''News Agencies'' categories should be evaluated with attention by users, as their quality is not appropriate to the importance of the topic.

Key messages:
Because of to the heterogeneity of its content, YouTube should be evaluated carefully when used as a source of information for Covid-19 vaccination. Content produced by non-medical users, is generally of poor quality, not appropriate to the importance of the topic.

DH Epidemiology
Abstract Inappropriate prescription of antibiotics remains a major contributor to the global antimicrobial resistance crisis despite clear linkages between antibiotic utilization and resistance spread. This study aims to better understand the simultaneous and independent effect of previous prescription behavior, patient expectation, and clinical uncertainty on antibiotic prescribing. This discrete choice experiment was embedded within a routine organizational climate survey administered to all physicians working in the Tuscany healthcare system administered between Nov 11 and Nov 20, 2019 (Qualtrics).
Participants were provided with a patient encounter vignette and subsequently asked to in which of two alternatives they were more likely to prescribe antibiotics. The two alternatives varied in levels of clinical uncertainty, patient expectations, and the physician's past behavior. We fitted a conditional logistic regression model. . We show that patient expectation has a significant negative association with antibiotic prescribing among hospital-based physicians. Our findings inform the design of antibiotic stewardship programs in Tuscany and highlight the importance of cultural context in shaping the physician's disposition when confronted with patient expectations. We suggest shared decision-making to

Background:
The COVID-19 pandemic has changed the way infectious diseases are perceived. Global healthcare systems have faced challenges since the start of the COVID-19 epidemic, particularly in developing countries. Some individuals with an acute COVID-19 diagnosis have developed symptoms persisting beyond 90 days. Long-Covid is the new term for this syndrome (LC). LC, on the other hand, is poorly known and appears to cause a wide range of symptoms, particularly among Brazilian patients. As a result, utilizing retrospective data from patients in Petrolina, Brazil's largest city in the northeast, we conducted an exploratory epidemiology study.

Methods:
A retrospective, cohort study design was used with a real-world dataset. The primary aim was to evaluate the prevalence of LC within Petrolina. The sample size was 1,164 LC patients. A comparative and subgroup analysis was conducted to evaluate demographics, comorbidities, clinical symptoms, and mortality. A k means model was used to assess disease severity using a clustering analysis based on the presence of comorbidities.

Results:
The prevalence of physical symptoms identified was 69Á5%. The strongest physical symptom was fever with resultant of 64Á09% followed by pain, 43Á64%. The prevalence of autonomic and neurological symptomatology was 8Á59% and 8Á16% respectively. A higher prevalence of autonomic symptoms were reported among older men of Black and Caucasian in comparison to Pardo. Disease severity within the sample could be associated with the presence of comorbidities which were identified based on medication history. Pregnant women have high rate of comorbidities. 529 patients have at least one comorbidity and 28Á73% of them are pregnant.

Conclusions:
It is useful to evaluate symptoms although a definitive diagnosis of LC is essential. This study provides insightful information around LC within a Brazilian population to develop better infection control protocols, as well as future management of similar pandemics.

Key messages:
This study could potentially improve the prognosis and mortality among LC patients with comorbidities. Our findings could be combined with other regional datasets to predict pattern inferences of LC spread, prognosis and morbidity, including for multimorbidity and pregnant patients.

Background:
In May 2020, considering gradual restoration of all economical activities the Government of Pakistan updated containment strategy from locking down the whole country to locking down high-risk areas to mitigate COVID-19 spread. All districts having !300 cases/100,000 population. COVID-19 case incidence and test positivity rates by real-time RT-PCR before and after zonal lockdown were compared to assess whether the locality-based lockdowns can be used as an alternative to country lockdown to contain COVID-19 spread.

Methods:
Smart lockdowns were implemented in ten localities in the Islamabad Capital Territory (ICT), having a population of 60,000 from 12 May 2020 to June 3, 2020. Movements were restricted. Entry and exit points were guarded by police. Any person with symptoms of fever, cough, or sore throat tested by real-time RT-PCR methods and reported within 24 hours of collection. To compare the rate of active cases and positivity rate by weeks, we performed a z-test for two proportions and set p < 0.05 as the level of significance.

Results:
The red zone had 60,000 persons in 2.00 square kilometers. The rate of active COVID-19 cases significantly decreased (p < 0.0001) during intervention from 300/100,000 population pre-containment time to 22/100,000 population after the first three weeks of lockdown. The COVID-19 positivity rate also decreased significantly (p < 0.0001) from 24% (24/78) precontainment to 5.3% during containment. A total of 3800 people were tested in the following three weeks of intervention and 26 cases were detected.

Conclusions:
The smart lockdowns approach reduced COVID-19 transmission in the ICT district. This type of intervention was