A new study has found clear evidence of inappropriate prescribing of antipsychotic treatments to children and young people (CYP) and low levels of aftercare monitoring by GPs but has also raised concerns about the clarity of best practice guidelines across different care settings and the completeness of data available to researchers.
The research team who carried out the study, suspect that this is partly explained by a lack of consistency in NICE best practice advice across different treatment settings and gaps in the anonymised general practice patient information available to researchers.
The study, which is published in BMJ Mental Health, was led by Dr Yana Vinogradova in the School of Medicine at the University of Nottingham, and was funded by the National Institute for Health and Care Research (NIHR) Nottingham Biomedical Research Centre.
Antipsychotics have serious physical side-effects that particularly affect CYP, so NICE has developed clinical guidelines for prescribing and aftercare for these patients. NICE guidelines for the two conditions for which antipsychotics can be core treatments (psychosis and bipolar disorder) are comprehensive. They recommend that secondary care specialists should prescribe any antipsychotics to CYP and also undertake/manage initial regular health monitoring for at least the first 12 months before patients may be transferred to primary care (general practitioners) for continuing aftercare under 'shared care' arrangements.
Although antipsychotics are not recommended for core treatment of autism or Tourette's, NICE guidelines for CYP with these conditions both include information about short off-label use for specific severe symptoms. The advice on antipsychotics prescribing and aftercare is, however, rudimentary and makes no reference to the more detailed best practice guidelines.
Inappropriate use
Antipsychotics may also be used more widely, and concern about inappropriate use of psychotropic drugs to manage 'challenging behaviour' has led to an NHS-supported project (started in 2018) to reduce their use among CYP and to some recent more comprehensive NICE recommendations for CYP with learning difficulties.
This new study focused on the four mental health conditions explicitly linked to antipsychotics treatment in NICE guidelines, using encoded English general practice data linked to hospital admission and outpatient data. The sample of CYP, aged 5 to 17 years, was representative of the English general population and the study had three aims – describe recorded trends in in antipsychotics prescribing; describe recoded reasons for prescribing and treatment lengths; and compare recorded levels of health monitoring in general practices between CYP with antipsychotics prescriptions and those with mental health condition diagnoses but no antipsychotics prescription.
Dr Vinogradova said: "Our findings showed rises in levels of antipsychotics prescribing to CYP between 2006 and 2021 – greater in adolescence than in childhood and greatest among adolescent females. Encoded indications for CYP prescribed antipsychotics appeared concerning – only just under half were for one of our focus mental health conditions and 62% of these were for autism. Just over a third had only mental health symptoms or associated illnesses recorded, and just under one fifth had no records of mental health problems. Some antipsychotics treatment lengths also appeared concerning, particularly those linked to autism, Tourette's, ADHD (attention-deficit/hyperactivity disorder), learning difficulties and self-harm.
"Compared to NICE recommendations for CYP with psychosis or bipolar disorder, recorded health monitoring levels were generally low, but lower for autism or Tourette's than for psychosis or bipolar disorder. More patients had records for basic health monitoring (heart or other body-related measurements) than for laboratory tests."
Further study needed with more complete data
She added: "Taken at face value, these findings appear to be disturbing evidence of lack of adherence to NICE guidelines, both with respect to the prescribing of antipsychotics to CYP and to the levels of aftercare given by general practitioners. Our interpretation, however, is more nuanced. We suspect that the encoded general practitioner records for these complex and multifaceted conditions fail to reflect the full history of individual treatments, and that studies with access to more complete data will be needed to audit accurately adherence to best practice.
"Our study also identified the need for streamlined, centralised and condition independent best practice recommendations for antipsychotics treatments for CYP, covering core and off-label treatments and long and short exposures, to which all condition-related guidelines could point. Together with efforts to clarify key data flows and encoding responsibilities, these could optimise adherence across treatment environments and establish more reliable records for national audits."
Possible reasons for the incompleteness of encoded general practitioner records are many, and impossible to rank clearly. The root cause is the separation of primary and secondary care patient records, with secondary care information transferred in text form when patients are sent back to primary care for ongoing treatment. General practices then need to re-encode the information, and incompleteness at any point in the process will lead to loss of data from specialists and hospitals. Although this is true for any illness involving primary and secondary care, it is particularly an issue here where the conditions are complex and often inter-related, and much of initial treatment and aftercare is allocated to secondary care.
Similar issues may also arise with respect to records of health monitoring in general practices. How this is done and by whom is not defined, but there are almost certainly a range of different data flows, and subsequent encoding and access solutions, all reflecting prevailing local resources. Auditing studies need certainty about the completeness of available data, so we conclude that a different research approach will be needed to establish reliable estimates of adherence to accepted best practice in the short to medium term.
The full study can be found here.
