The National Health Service continues to spend at a rate that will be difficult to maintain. Ninety percent of health care contact in the UK occurs in general practice, and consultations for children make up one tenth of the general practice workload. Diagnostic tests, such as blood tests or x-rays, can be used to help general practitioners (GPs) diagnose conditions in children. A GP may miss a serious diagnosis if a child does not receive the test they need. On the other hand, if a child receives a test they do not need, they may experience unnecessary pain and/or anxiety and it may lead to a cascade of further unnecessary testing and treatments.
We will use data from the Clinical Practice Research Datalink which contains patient and practice-level anonymised data for 60 million patients in the UK, to examine how testing in children has changed over time and explore differences in test use between general practices. Possible reasons for these differences will be explored, including the child’s level of social and financial disadvantage, whether they have other medical conditions, and how many children have the disease related to the test.
If there are tests that have large differences in their use with no clear explanation, then these may represent wasteful or unnecessary testing. These results will support the need to update testing guidance for GPs in order to prevent harms to children. Reducing unnecessary testing will also cut down spending and improve the efficiency of the health system.
Due to mounting financial pressure on the National Health Service (NHS), interventions are required to improve the sustainability of the health system. Children constitute one-tenth of the general practice workload. Substantial variation has been reported in the use of diagnostic tests in adults in UK primary care, however comparatively little is known about the rates of diagnostic test utilisation and variation in children in the same setting. This highlights an evidence gap that is important to address so as to prevent harms to children from under-testing and over-testing. This study aims to use data from CPRD GOLD, CPRD Aurum linked to the practice-level index of multiple deprivation and HES Admitted Patient Care datasets. We will analyse test records for children in English primary care from January 31 2006 to December 31 2021, to: determine total change in test use during the time period; quantify rates of test use by age and sex; rank the most frequently requested tests; assess temporal change in specific tests using joinpoint regression; calculate between-practice variation in overall and specific test use between 2018 and 2019 using the co-efficient of variation (CoV). We also aim to explore factors associated with test-use (including deprivation level, presence of co-morbidities, disease prevalence, ethnicity) using multi-level mixed effects logistic regression. Quantifying and understanding the drivers of variation in diagnostic test use is essential to improve the quality of care delivered to children seen in general practice by preventing harms from inappropriate testing. The results of this study will fill an important research gap to better understand diagnostic testing practices for children, identify opportunities to improve service provision for children in primary care and contribute to a more efficient and sustainable health system.
Primary outcomes:
1. Proportion of children who receive a diagnostic test in primary care overall, and by age and sex
2. Crude and adjusted rates of diagnostic test utilisation (tests per 1000 child-years) ordered for children in each year from 2006 – 2021, stratified by sex, age by year and developmental age groups - 0 to 1 month (neonates), 1 month to 1 year (infants), 1 to 4 years (early childhood), 5 to 12 years (middle childhood) and 13 to 16 years (adolescence)
3. Rank order of the top 25 most commonly requested diagnostic tests for children in primary care
4. Degree of variation in test utilisation between general practices for the most commonly requested tests (measured by co-efficient of variation).
Secondary outcomes:
1. Number of tests ordered per patient in primary care per year and overall.
2. The difference between disease prevalence and test utilisation for disease-specific tests e.g., prevalence of asthma and spirometry rates, to determine whether variation in testing is related to prevalence of disease.
3. Association between presence of co-morbidity and rates of test utilisation.
4. Association between deprivation levels, ethnicity and variation in test use.
5: Determine how many presentations/admissions to hospital occur within 28 days of a test being performed, stratified by test name.
Rafael Perera - Chief Investigator - University of Oxford
Elizabeth Thomas - Corresponding Applicant - University of Oxford
Carl Heneghan - Collaborator - University of Oxford
Clare Bankhead - Collaborator - University of Oxford
Peter Gill - Collaborator - University of Toronto
HES Admitted Patient Care;Practice Level Index of Multiple Deprivation