We report trends in the recording of mental illness, challenging behaviour, and psychotropic medication prescription in people with intellectual disability (ID) in a large primary care dataset between 1999 and 2013. Rates of new recording of severe mental illness declined significantly over the observation period, whereas rates of new recording of depression and anxiety disorders did not change. Rates of new prescription of antipsychotic and mood stabilising drugs decreased significantly between 1999 and 2013, and rates of new prescription of anxiolytics/hypnotics and anti-depressants remained stable.
We do not report rates of recording of ID during the observation period, however, these data have previously been published by our group (1). Although there is a spike in recording of certain ID Read codes in 2006 (coinciding with the introduction of the learning disability Quality Outcome Framework, which incentivised general practitioners to keep a register of people with ID), there is no similar spike in the recording of chromosomal or other causes of ID, suggesting that the introduction of the QOF was associated with increased formal recording of ID, but not necessarily enhanced recognition of the disorder. Similarly, our analysis does not show any corresponding change in mental illness recording or psychotropic prescribing at the introduction of the QOF, suggesting that this incentive was not associated with identification of a new group of previously undiagnosed individuals who would then be recognised as a high-risk group for mental illness and therefore more likely to receive a diagnosis of mental illness or prescription of psychotropic drugs. Moreover, we considered a diagnosis of ID at any point during an individual’s electronic health record to indicate the presence of ID, since the condition is fixed and lifelong. Any changes in GP recognition of ID at the introduction of QOF would not have a knock-on effect of dramatically changing our cohort size at that time, since the individual was considered ‘exposed’ from the point at which they started contributing data (cohort entry was that latest date of GP registration, quality filter dates, age 18, or January 1999).
The health check scheme (Directed Enhanced Scheme, DES) for people with ID covers only England, is targeted to people with more severe disability, and is not standardised in terms of content. Under half of the eligible population receives a health check, as shown by PHE reports (2), therefore the impact of DES on our cohort is unlikely to influence the overall results. Furthermore, in our previous assessment of the health check scheme using the same dataset we showed that practices adopting DES did not differ from practices that did not in rates of newly-diagnosed depression (3).
The Health Improvement Network (THIN) is a dynamic dataset, with individuals entering and leaving at different times. The denominator used in our calculation of incidence is person-years exposed to ID. The calculations therefore take into account changes in the size of the cohort over time. Given this, we believe our time trend results to represent a real phenomenon rather than an artefact of the data collection. Falling rates of antipsychotic prescription might be expected given the negative publicity surrounding their use and as awareness of adverse side-effects (particularly of the second generation agents) has grown, and mirrors falls in antipsychotic use in other groups (4,5).
The major finding of this work is the discrepancy between psychotropic prescribing and mental illness recording in people with ID. The regression analysis demonstrates that challenging behaviour, advancing age, and a diagnosis of autism or dementia, are all independently associated with antipsychotic prescription. We welcome the report by Public Health England, the findings of which are broadly in agreement with our own. There is now good evidence that people with ID are prescribed psychotropic medication at a rate higher than indicated by the rates of mental illness they experience.
References
(1) Osborn DPJ, Horsfall L, Hassiotis A, Petersen I, Walters K, Nazareth I. Access to cancer screening in people with learning disabilities in the UK: cohort study in the Health Improvement Network, a primary care research database. PLoS ONE 7(8):e43841
(2) Buszewicz M, Welch C, Horsfall L, et al. Assessment of an incentivised scheme to provide annual health checks in primary care for adults with intellectual disability: a longitudinal cohort study. Lancet Psychiatry 2014;1:522-30
(3) Glover G. The uptake of learning disability health checks. 2014. Available online at https://www.improvinghealthandlives.org.uk/gsf.php5?f=313365&fv=20829
(4) Olfson M, King M, Schoenbaum M. Treatment of young people with antipsychotic medications in the United States. JAMA Psychiatry 2015;72:867-74
(5) Martinez C, Jones RW, Rietbrock S. Trends in the prevalence of antipsychotic drug use among patients with Alzheimer's disease and other dementias including those treated with antidementia drugs in the community in the UK: a cohort study. BMJ Open 2013;3:e002080
Rapid Response:
Author reply to Glover
We report trends in the recording of mental illness, challenging behaviour, and psychotropic medication prescription in people with intellectual disability (ID) in a large primary care dataset between 1999 and 2013. Rates of new recording of severe mental illness declined significantly over the observation period, whereas rates of new recording of depression and anxiety disorders did not change. Rates of new prescription of antipsychotic and mood stabilising drugs decreased significantly between 1999 and 2013, and rates of new prescription of anxiolytics/hypnotics and anti-depressants remained stable.
We do not report rates of recording of ID during the observation period, however, these data have previously been published by our group (1). Although there is a spike in recording of certain ID Read codes in 2006 (coinciding with the introduction of the learning disability Quality Outcome Framework, which incentivised general practitioners to keep a register of people with ID), there is no similar spike in the recording of chromosomal or other causes of ID, suggesting that the introduction of the QOF was associated with increased formal recording of ID, but not necessarily enhanced recognition of the disorder. Similarly, our analysis does not show any corresponding change in mental illness recording or psychotropic prescribing at the introduction of the QOF, suggesting that this incentive was not associated with identification of a new group of previously undiagnosed individuals who would then be recognised as a high-risk group for mental illness and therefore more likely to receive a diagnosis of mental illness or prescription of psychotropic drugs. Moreover, we considered a diagnosis of ID at any point during an individual’s electronic health record to indicate the presence of ID, since the condition is fixed and lifelong. Any changes in GP recognition of ID at the introduction of QOF would not have a knock-on effect of dramatically changing our cohort size at that time, since the individual was considered ‘exposed’ from the point at which they started contributing data (cohort entry was that latest date of GP registration, quality filter dates, age 18, or January 1999).
The health check scheme (Directed Enhanced Scheme, DES) for people with ID covers only England, is targeted to people with more severe disability, and is not standardised in terms of content. Under half of the eligible population receives a health check, as shown by PHE reports (2), therefore the impact of DES on our cohort is unlikely to influence the overall results. Furthermore, in our previous assessment of the health check scheme using the same dataset we showed that practices adopting DES did not differ from practices that did not in rates of newly-diagnosed depression (3).
The Health Improvement Network (THIN) is a dynamic dataset, with individuals entering and leaving at different times. The denominator used in our calculation of incidence is person-years exposed to ID. The calculations therefore take into account changes in the size of the cohort over time. Given this, we believe our time trend results to represent a real phenomenon rather than an artefact of the data collection. Falling rates of antipsychotic prescription might be expected given the negative publicity surrounding their use and as awareness of adverse side-effects (particularly of the second generation agents) has grown, and mirrors falls in antipsychotic use in other groups (4,5).
The major finding of this work is the discrepancy between psychotropic prescribing and mental illness recording in people with ID. The regression analysis demonstrates that challenging behaviour, advancing age, and a diagnosis of autism or dementia, are all independently associated with antipsychotic prescription. We welcome the report by Public Health England, the findings of which are broadly in agreement with our own. There is now good evidence that people with ID are prescribed psychotropic medication at a rate higher than indicated by the rates of mental illness they experience.
References
(1) Osborn DPJ, Horsfall L, Hassiotis A, Petersen I, Walters K, Nazareth I. Access to cancer screening in people with learning disabilities in the UK: cohort study in the Health Improvement Network, a primary care research database. PLoS ONE 7(8):e43841
(2) Buszewicz M, Welch C, Horsfall L, et al. Assessment of an incentivised scheme to provide annual health checks in primary care for adults with intellectual disability: a longitudinal cohort study. Lancet Psychiatry 2014;1:522-30
(3) Glover G. The uptake of learning disability health checks. 2014. Available online at
https://www.improvinghealthandlives.org.uk/gsf.php5?f=313365&fv=20829
(4) Olfson M, King M, Schoenbaum M. Treatment of young people with antipsychotic medications in the United States. JAMA Psychiatry 2015;72:867-74
(5) Martinez C, Jones RW, Rietbrock S. Trends in the prevalence of antipsychotic drug use among patients with Alzheimer's disease and other dementias including those treated with antidementia drugs in the community in the UK: a cohort study. BMJ Open 2013;3:e002080
Competing interests: No competing interests