Search

Abstract Background Infective endocarditis is an uncommon but serious infection, where evidence for giving antibiotic prophylaxis before invasive dental procedures is inconclusive. In England, antibiotic prophylaxis was offered routinely to patients at risk of infective endocarditis until March 2008, when new guidelines aimed at reducing unnecessary antibiotic use were issued. We investigated whether changes in infective endocarditis incidence could be detected using electronic health records, assessing the impact of inclusion criteria/statistical model choice on inferences about the timing/type of any change. Methods Using national data from Hospital Episode Statistics covering 1998–2017, we modelled trends in infective endocarditis incidence using three different sets of inclusion criteria plus a range of regression models, identifying the most likely date for a change in trends if evidence for one existed. We also modelled trends in the proportions of different organism groups identified during infection episodes, using secondary diagnosis codes and data from national laboratory records. Lastly, we applied non-parametric local smoothing to visually inspect any changes in trend around the guideline change date. Results Infective endocarditis incidence increased markedly over the study (22.2–41.3 per million population in 1998 to 42.0–67.7 in 2017 depending on inclusion criteria). The most likely dates for a change in incidence trends ranged from September 2001 (uncertainty interval August 2000–May 2003) to May 2015 (March 1999–January 2016), depending on inclusion criteria and statistical model used. For the proportion of infective endocarditis cases associated with streptococci, the most likely change points ranged from October 2008 (March 2006–April 2010) to August 2015 (September 2013–November 2015), with those associated with oral streptococci decreasing in proportion after the change point. Smoothed trends showed no notable changes in trend around the guideline date. Conclusions Infective endocarditis incidence has increased rapidly in England, though we did not detect any change in trends directly following the updated guidelines for antibiotic prophylaxis, either overall or in cases associated with oral streptococci. Estimates of when changes occurred were sensitive to inclusion criteria and statistical model choice, demonstrating the need for caution in interpreting single models when using large datasets. More research is needed to explore the factors behind this increase.

ABSTRACT Objectives This study shows how to improve PRL when using a limited number of personal identifiable information (PIIs) available in HES such as NHS number, DOB, Hospital Number, Gender and Postcode. The focus is on new approaches in categorisation of m-u probabilities for fields related to Postcode. In addition, a brief outline of our procedure to generate an appropriate training dataset will be presented as well as reasons for splitting HES dataset into single and multi-admissions, and proposed techniques to avoid computational cost. Approach PIIs such as DOB and Postcode can be categorised in order to increase the number of potential comparison vectors which is necessary in order to deal with the multi-million size of the HES dataset. We propose a multinomial approach for comparison between postcodes both for agreement and disagreement case scenarios, based on geographical information and a similarity score resulting from a Probit model. In the absence of a ‘gold-standard’ dataset, generating an appropriate training set is possible by combining two-way Deterministic Linkage (DL) and simple random sampling. Splitting HES into single and multi-admissions is recommended for reasons related to computational performance and demographic differences of the underlying population. Selection of the lower and upper thresholds regarding linkage weights can be done in a practical way by using conditional Normal mixture modelling, or more formally by using sampling techniques based on Fellegi-Sunter decision rule. We show how the latter can be further improved avoiding the computational cost by introducing appropriate mapping functions. Results We show the advantages of PRL over DL between Surveillance Laboratory Flu Data and HES. Preliminary results indicate an additional 10% of matches for < 1% expected false positives. EM convergence problems when treating missing values as separate outcome category will be discussed comparing results with the suggested re-normalisation approach in the bibliography. Conclusion PRL between Laboratory Surveillance Data and Hospital Episode Statistics (HES) is a common practice and performed in a number of epidemiological studies conducted at Public Health England (PHE). Therefore, improving linkage will subsequently benefit those studies. Furthermore, this study proposes new approaches for Postcode comparison, and the utilisation of mapping functions to avoid the computational cost of the Fellegi-Sunter decision rule. These approaches could be used in a wide range of applications. Finally, we suggest a structural way of handling the HES dataset in order to achieve high efficiency and robustness of linkage.

ABSTRACT Objectives This study aims to describe the burden of respiratory syncytial virus (RSV) in children