Hospital Treatment Insights is a novel database that combines information on diagnoses in English hospitals and the type and amount of medications that were used to treat them. For the first time, this data allows to investigate relationships between individual diagnoses and hospital dispensing in England on a large scale. What is more, its link to the Clinical Practice Research Datalink enables investigations into associations with primary care prescribing. Before they can be used for research, however, the properties of both datasets have to be assessed separately. We propose to partially do so using the case of antibiotics. We will first summarise patient’s characteristics, hospital diagnoses for infections and overall use of antibiotics present in each dataset and compare them to what we would expect. In a further step, we will then only look at patients specifically admitted for infections. As we should see at least one record of dispensed antibiotics for each of those, this will help us to further describe and understand uncaptured dispensing. The combined findings of our study are expected to highlight the strengths and limitations of Hospital Treatment Insights and its link to the Clinical Practice Research Datalink for research on antibiotic usage in hospitals.
Due to the way in which data for Hospital Treatment Insights is collected, an absence of a dispensing record is not equal to there not having been a drug administered and there is need to quantify the extent to which this affects recording for different drugs. Thus, the usefulness of the full Hospital Treatment Insights database and the subset linked to the Clinical Practice Research Datalink for research on antibiotic dispensing in hospitals in England will be explored using descriptive statistics. We will calculate prevalence rates and 95% confidence intervals for demographic variables, diagnoses for clinical infectious syndromes and antibiotic dispensing to provide insight into the representativeness of the separate populations included in the two datasets. Variations will be explored by stratifying by calendar year, age, sex and type of antibiotic. Similar statistics will be used to further explore dispensing for community-acquired pneumonia and under-recorded individual-level dispensing will be quantified where possible. If time and sample sizes allow for it, the results will be compared to the same estimates for urinary tract infection, cellulitis and sepsis.
The primary outcome is the number of dispensed antibiotics of a list of clinically relevant antibiotics within the same Hospital Episode Statistics episode.
Joseph Kim - Chief Investigator - IQVIA Ltd ( UK )
Eleanor King - Collaborator - IQVIA ( IMS Health ) France
Elizabeth Williamson - Collaborator - London School of Hygiene & Tropical Medicine ( LSHTM )
Laura Shallcross - Collaborator - University College London ( UCL )
Matthew Hankins - Collaborator - IQVIA - UK
Patrick Rockenschaub - Collaborator - University College London ( UCL )
Rachel Tham - Collaborator - IQVIA Ltd ( UK )
Other