A structured programme to withdraw antipsychotics among adults with intellectual disabilities

J Appl Res Intellect Disabil. 2019;32:1389–1400. wileyonlinelibrary.com/journal/jar 
|
1389
Published for the British Institute of Learning Disabilities
© 2019 John Wiley & Sons Ltd
1 | INTRODUCTION
1.1 | The prevalence of antipsychotic use in people
with intellectual disabilities
Psychotropic medications such as antipsychotics, anti‐depressants,
mood stabilizers, including anti‐epileptic medications and lithium,
anti‐anxiety medications including benzodiazepines, psychostimu‐
lants, beta‐adrenergic blockers, opioid antagonists are used widely
among people with intellectual disabilities. The rate of psychotropic
use among people with intellectual disabilities varies between 32%
and 85%, the average being around 50%–63% (Deb, 2016; Doan,
Lennox, Taylor‐Gomez, & Ware, 2013; Sheehan et al., 2015). Most
widely used among psychotropics are the antipsychotics, which is
20%–45% of the psychotropics used (Bowring, Totsika, Hastings,
Toogood, & McMahon, 2017; Doan et al., 2013; de Kuijper et al.,
2010; Sheehan et al., 2015; Tsiouris, Kim, Brown, Pettinger, & Cohen,
2013). Antipsychotics are often used off‐licence in the absence of
mental illness as no mental disorder is recorded among 36%–71%
Received:12January2019
|
Revised:18April2019
|
Accepted:15May2019
DOI: 10.1111/jar.12635
ORIGINAL ARTICLE
A structured programme to withdraw antipsychotics among
adults with intellectual disabilities: The Cornwall experience
Rohit Shankar1,2 | Mike Wilcock3 | Shoumitro Deb4 | Rebecca Goodey1 |
Eve Corson1 | Charlotte Pretorius1 | Georgina Praed3 | Amanda Pell3 | Dee Vujkovic1 |
Ellen Wilkinson1 | Richard Laugharne1 | Sharon Axby1 | Rory Sheehan5 |
Regi Alexander6
1Cornwall Partnership NHS Foundation
Trust, Truro, UK
2University of Exeter Medical School, Exeter,
UK
3Kernow Clinical Commissioning Group,
Saint Austell, UK
4Imperial College London, London, UK
5University College London, London, UK
6University of Leicester, Leicester, UK
Correspondence
Rohit Shankar, Chygovenck Three Milestone
Industrial Estate, Truro TR4 9LD, Cornwall,
UK.
Email: [email protected]
Abstract
Background: Antipsychotic medications are used among 19%–58% of adults with
intellectual disabilities to manage challenging behaviour against the NICE guideline
recommendations. Studies show that it is possible to completely withdraw antipsy‐
chotics in about one third of adults with intellectual disabilities and a dose reduction
of 50% or more in another third.
Method: In Cornwall, over three years the present authors developed a structured
pathway to withdraw antipsychotics among adults with intellectual disabilities which
involved people with intellectual disabilities and their carers, GPs, community learn‐
ing disability team members and pharmacists.
Results: The present authors managed to withdraw antipsychotics totally among
46.5% (33/71) and reduced over 50% of dosage in another 11.3% (8/71) of adults
with intellectual disabilities. At three months follow‐up no one required hospital ad‐
mission or change in placement.
Conclusion: It is possible to withdraw/reduce antipsychotics in a high proportion of
adults with intellectual disabilities if a concerted effort is made involving all stake‐
holders from the outset.
KEYWORDS
adults, antipsychotics withdrawal, challenging behaviour, intellectual disabilities, structured
pathway

1390
|
Published for the British Institute of Learning Disabilities
SHANKAR et Al.
of those who are prescribed antipsychotics (Clarke, Kelley, Thinn,
& Corbett, 1990; Sheehan et al., 2015). The most common reason
for the off‐licence use of antipsychotics is the management of prob‐
lem (challenging) behaviour in people with intellectual disabilities
as among those who receive antipsychotics, in 19%–58% of cases
they are used for the management of problem behaviours (Clarke
et al., 1990; de Kuijper et al., 2010; Tsiouris et al., 2013). However,
National (NICE, ) and International Guidelines (Deb et al., 2009) rec‐
ommend that antipsychotics should not be used to treat problem
behaviour unless other non‐pharmacological approaches have been
tried and failed and the person with intellectual disabilities or others
are at serious risk of harm.
1.2 | Concerns relating to antipsychotic use and the
STOMP programme
The off‐licence use of antipsychotics in people with intellectual dis‐
abilities is a cause of major public health concern because (a) these
medications are used in addition to existing high use of medication
for physical problems (Deb & Fraser, 1994), (b) this may lead to ad‐
verse effects and people with intellectual disabilities are prone to
develop adverse effects of antipsychotics more frequently than their
non‐ intellectual disabilities counterpart (Sheehan et al., 2017), (c)
also adverse effects in this population may be difficult to assess, and
inappropriate concomitant use of medicine to counteract adverse
effects is a common practice, (Deb, Unwin, & Deb, 2015), (d) once
started it is difficult to withdraw these medications (Deb, Bertelli,
& Rossi, 2019; Sheehan & Hassiotis, 2017), (e) lack of evidence base
demonstrating effectiveness (Deb, 2013, 2016), (f) use of antipsy‐
chotic drugs at a higher than recommended dose as well as polyp‐
harmacy of antipsychotic use (Deb et al., 2015), (g) long‐term use
without reviews (Deb, 2018), (h) difficult ethical issues involved in
the use of these medicines and difficulty in securing explicit informed
consent in many cases (Unwin & Deb, 2008) and (i) difficulty in car‐
rying out necessary investigations (Unwin & Deb, 2008). Because of
these concerns, the NHS England in the UK has embarked on a major
campaign called “STopping Over Medication of People with intellec‐
tual disability, autism or both (STOMP)” (NHS England, 2016 cited in
Branford, Gerrard, Saleem, Shaw, & Webster, 2019).
1.3 | Withdrawal studies
One practical way to reduce overmedication in this population is
to withdraw antipsychotic medication. A recent systematic review
showed that withdrawal is possible in a proportion of people with in‐
tellectual disabilities (4%–74%) (Sheehan & Hassiotis, 2017). However,
most of the studies included in this systematic review are from the
United States that included patients from long‐term institutions.
More relevant to our practice in the UK are the withdrawal studies in
Europe that included people with intellectual disabilities from com‐
munity settings. Deb et al.’s (2019) review showed that there were
two studies in the UK (Ahmed et al., 2000; Branford, 1996) and one
recent one from the Netherlands (de Kuijper, Evenhuis, Minderaa,
& Hoekstra, 2014) where a concerted structured effort was made
to withdraw antipsychotic medication. In Branford's (1996) study,
25% (31/123) achieved a total withdrawal, and in 42% (52/123) of
cases, an attempt to withdraw or reduce dose precipitated problem
behaviour leading to reinstatement of antipsychotics. In Ahmed et
al.’s study, 33% (12/36) achieved complete withdrawal, and another
19% (7/36) achieved at least a 50% reduction in dose. In de Kuijper et
al.’s (2014) study, 44% (43/98) achieved a complete withdrawal, but
in 16% (7/43) of cases antipsychotics were reinstated at 12 weeks
follow‐up, so 37% (36/98) remained off antipsychotics at 12 weeks.
There is another recent open‐label discontinuation study from
the Netherlands (de Kuipjer & Hoekstra, 2018). In this study, of
129 participants, 61% had completely discontinued antipsychotics
at 16 weeks, 46% at 28 weeks, and 40% at 40 weeks. In 49% of
participants, behaviour deteriorated at 16 weeks follow‐up leading
to reinstatement of medication. It is worth mentioning here that
recruitment to these studies could be a major problem, particularly
when a placebo‐controlled, randomized, double‐blind discontinu‐
ation study design is used where the people with intellectual dis‐
abilities were blindly randomly allocated to either an antipsychotic
discontinuation group (antipsychotics replaced by a placebo) or an‐
tipsychotic continuation group (McNamara, et al., 2017; Ramerman
et al., 2019). Similar recruitment problems were encountered in pla‐
cebo‐controlled RCT assessing efficacy of antipsychotics in people
with intellectual disabilities (Oliver‐Africano et al., 2010; Tyrer et
al., 2008). Open‐label studies also faced similar problems with re‐
cruitment (Ahmed et al., 2000). However, a recent study from the
Netherlands (de Kuijper et al., 2014) and our current study are en‐
couraging in that respect. These show that assessment of risk factors
affecting withdrawal using appropriate instrument and involving all
appropriate stakeholders such as people with intellectual disabili‐
ties and their carers, and also CLDT team members at the outset are
likely to help with the recruitment.
1.4 | Factors affecting withdrawal
A number of factors affect withdrawal. For example, Branford (1996)
found that a lower dose of antipsychotics, minimal psychopathology,
lack of aggression, stereotype and hyperactivity at baseline helped
with the withdrawal. Ahmed et al. (2000) on the other hand high‐
lighted the environmental and organisational factors. For example,
they suggested that experienced and full‐time staff in regular em‐
ployment, low staff turnover, staff training, courses for managing
problem behaviour, less reliance on environmental restrictions are
likely to facilitate the withdrawal process. deKuijper and Hoekstra
(2018) found that female gender, a lower rate of baseline prob‐
lem behaviours, and lower baseline dosage are in favour, and the
presence of severe behaviour, and autonomic and extrapyramidal
symptoms at baseline are factors against a successful withdrawal.
deKuijper and Hoekstra (2018) also found that the presence of co‐
morbid autism, a higher dose of antipsychotic drug, higher behaviour
rating and akathisia scores, and more‐frequent worsening of health
during discontinuation were associated with a lower incidence of

|
1391
Published for the British Institute of Learning Disabilities
SHANKAR et Al.
complete discontinuation. These factors also affect reinstatement
rate. For example, Janowsky, Barnhill, Khalid, and Davis (2006)
found within a large population in an American residential institu‐
tion, 66.3% (55/83) of individuals remained antipsychotic‐free al‐
most 10 years after withdrawal. However, in a subsequent study,
Janowsky, Barnhill, Khalid, and Davis (2008) demonstrated that,
unfortunately, it becomes difficult to withdraw antipsychotic medi‐
cations altogether in a very high proportion of those who showed
worsening of behaviour after one or two attempts of withdrawal.
1.5 | Withdrawal symptoms
Although most studies have shown that behaviour improves after
withdrawal of antipsychotics but in a proportion of cases an attempt
of withdrawal fails because of worsening of behaviour. Part of this
could be due to withdrawal symptoms which might occur after the
long‐term use of antipsychotics. These include akathisia, dyskinesia,
anxiety, sleep problems and agitation (Cerovecki et al., 2013; Correll,
2010 cited in Beumer, 2017). These symptoms might be misinterpreted
as recurrence of the original problem behaviour (Beumer, 2017).
Rebound akathisia might appear within the first few days, whereas
rebound parkinsonism usually emerges after a week and rebound
dyskinesia might only become apparent within a month (Cerovecki et
al., 2013). However, most studies show that emergent extrapyramidal
symptoms seem to improve after a few weeks (Christian, Snycerski,
Singh, & Poling, 1999; deKuijper et al., 2014). This is an indicator for
the clinicians who are considering withdrawal of antipsychotics that
instead of reinstating antipsychotics straight away, they should wait
(if necessary with the help of as necessary PRN prescription) until the
behaviour improves (Deb et al., 2009) (see Figure 1).
FIGURE 1 Flow chart for considering
withdrawal of medication (adapted from
Deb et al., 2009; http://www.ld‐medic
ation.bham.ac.uk)
Follow-up,
Assess
Rate of withdrawal will depend on:
-type (e.g. depot vs. oral), dosage,
duration, adverseeffects of
medications
-individual circumstances.
Monitor
Medication not
withdrawn
Factors to consider:
•review of medications and
reasons for prescription,
including medications for
problem behaviours that were
prescribed by someone else
•type, frequency, severity and
duration of problem
behaviours
•previous response to
withdrawal
•individual circumstances
•whether alternatives are
available
•relapse plan.
Develop a relapse plan:
•wait, see and monitor
behaviour
•specify a timescale
•consider non-medication
intervention
•consider ‘prn’ medication
•reconsider withdrawal
symptoms which may
improve at follow up
•consider re-prescribing
medication
Deterioration in behaviour
Consider relapse plan
Continue with regular reviews if
necessary
No deterioration in behaviour
Medication
withdrawn
CONSIDER WITHDRAWAL

1392
|
Published for the British Institute of Learning Disabilities
SHANKAR et Al.
1.6 | Need for a structured pathway
It is clear that in order to succeed with the withdrawal of antipsychot‐
ics, it is imperative to follow a structured pathway, which includes a
withdrawal plan presented in Figure 1 and proper consideration of
initiation of medicinal treatment as per the International Guideline
(Deb et al., 2009) (see Table 1) recommendations. Furthermore, the
clinicians need to consider complex factors involved in prescribing
which will affect the withdrawal process such as variables related to
patients, treatment modalities, type and dosage of medication, envi‐
ronment which may not be appropriate for the person, and other re‐
sources, including appropriate psychosocial support structure around
the person as well as their carers. In the process of withdrawal, clini‐
cians also need to consider factors mentioned earlier in the text that
may influence successful withdrawal, including carer training and or‐
ganisational support structure. No study so far described a structured
pathway for withdrawal. Therefore, in Cornwall, the present authors
have decided to develop a structure around the withdrawal pathway
before embarking on antipsychotic withdrawal in adults with intellec‐
tual disabilities to support the NHS England STOMP initiative, which
is supported by many organisations including the Royal College of
Psychiatrists in the UK (Royal College of Psychiatrists, 2016).
The project was set out as a quality improvement (QI) initiative
using the “Plan, Do, Study, Act (PDSA)” model (NHS Improvements,
2018). The “Plan” was to identify all people with intellectual disabili‐
ties on APT in Cornwall, having no major mental disorder. The aim of
the first QI cycle was to reduce the antipsychotic burden by 20% of
the identified target population in one year and then realign expec‐
tations for next cycles subsequently. “Do” was to undertake a struc‐
tured reduction plan as identified in the methods section. “Study” was
to collect data post‐attempt and compare to see if 20% reduction was
achieved and if not why not. It would also allow an opportunity to
reflect on the impact, barriers and what went well of the change and
what was learned including looking into what worked and what did
not and why. “Act” was to find solutions to overcome the challenges
and to plan the next cycle. Four cycles of PDSA were carried out with
the final one leading to using the principals suggested by theoretical
evidence established (Shankar, Wilcock, Oak, McGowan, & Sheehan,
2019). It was recognized during the QI cycles that new approaches
and tools need to be developed to overcome the challenges as avail‐
able methods could not lend itself directly on occasions to the QI
improvement cycles. However, QI methodology has been the back‐
ground for the current project to guide and evidence improvements.
In this paper, the present authors have described the methods
used for this structured programme and the overall outcome.
2 | METHOD
In Cornwall, UK, the present authors proceeded with the withdrawal
programme in several steps that are described here. At the heart
of our programme was the wider stakeholder involvement including
people with intellectual disabilities and their carers (both family and
paid carers), GPs, community pharmacists and community learning
disability team (CLDT) members.
2.1 | Step1: Primary care and
identification of the cohort
As a first step, the present authors invited local GPs to attend a STOMP‐
related one‐hour tutorial in which the present authors described the
STOMP initiative and our proposal to withdraw antipsychotic medication.
A GP prescribing lead from each primary care practice was invited to at‐
tend these meetings and disseminate the learning within their own prac‐
tices. There was a follow‐up meeting after 12 months of the initial tutorial
TABLE 1 Main recommendation in the International Guideline
(Deb et al., 2009; http://ld‐medic ation.bham.ac.uk)
• Ensure that an assessment has been conducted and recorded
prior to initiating treatment.
• Ensure that an appropriate formulation is carried out and a treat‐
ment plan drawn up, prior to instigating any interventions.
• Ensure that appropriate physical examinations and investigations
have been carried out.
• Assess the person’s capacity to consent to treatment.
• Allow the person and/ or their family or carers to influence the
decisions that are made and included in the treatment plan.
• Clarify to the person and/ or their family or carers if the medica‐
tion is prescribed outside their licensed indication. If this is the
case, they should be told about the type and quality of evidence
that is available to demonstrate its effectiveness.
• Where possible, and when necessary, discuss the formulation and
treatment plan with other relevant professionals.
• The treatment plan should be part of a broader person‐centred
care plan.
• The treatment plan must comply with the country’s legal
framework.
• Identify a key person who will ensure that medication is adminis‐
tered appropriately and communicate all changes to the relevant
parties.
• Medication might be used at the same time as non‐medication
managements.
• Provide the person and/ or their family or carers with a written
treatment plan at the time of prescribing.
• Discuss with the person and/ or their family, carer, or key person,
common and serious adverse events related to the treatment
(where possible, they should provide accessible information). Also
advise what action is needed in case of a serious adverse event.
• Consider the communication needs of the person.
• When “prn”/ “as required” medication is prescribed, provide as
much information as possible about why and when the medica‐
tion may be used and monitor this regularly.
• The method and timing of the assessment of treatment outcome
should be set at the beginning of the treatment.
• As far as possible, there should be an objective way to assess
outcomes (the use of standardized measures is recommended).
• Ensure that follow‐up assessments of treatment outcome and
adverse effects have taken place.
• As far as possible, one medication should be prescribed at a time.
• In general, the medication should be used within the national
recommended dose range.
• Consideration for withdrawing medication and exploring non‐
medication management options should be ongoing.
• Document all appropriate information and share it with appropri‐
ate individuals when necessary.

|
1393
Published for the British Institute of Learning Disabilities
SHANKAR et Al.
in which the present authors have administered a 21‐item questionnaire
to assess GPs knowledge on the subject (Shankar & Wilcock, 2018).
Involving GPs at the outset made it easy for us to access GP reg‐
ister‐based database. The present authors have used Read codes (a
coded thesaurus of clinical terms) to identify adults with intellectual
disabilities who are on antipsychotic medication in the area. The pres‐
ent authors also used a software program called ECLIPSE (Electronic
Checking Leading to Improved Prescribing Safety and Efficiency)
to identify all people on the Learning Disability Quality Outcome
Framework (QoF) register on antipsychotics who did not have a re‐
corded mental disorder in 44 practices in Cornwall.
The present authors also carried out an audit of people with intel‐
lectual disabilities on antipsychotics who were discharged back to pri‐
mary care from the secondary care between 2010 and 2015 to assess
in what proportion of cases an annual health check (AHC) involved a
satisfactory antipsychotic medication review (Shankar et al., 2016).
2.2 | Step 2: Involving all stakeholders
The present authors organized a meeting where the people with intel‐
lectual disabilities and their carers were invited, and the present authors
frankly discussed all the issues involved in the withdrawal process. The
benefits and risks were highlighted. It was expressed that over time
diagnostic systems have changed (Autism used to be mistaken as child‐
hood schizophrenia), medication withdrawal might expose a hitherto
undiagnosed mental illness or uncover an “unmet need” such as an
unsuitable placement which will need addressing. The present authors
discussed that in some people withdrawal may lead to hospital admis‐
sion in the short term to keep patients and others safe. However, pa‐
tients and their carers and other stakeholders remained totally involved
in the withdrawal and relapse prevention strategy (see Figure 1).
2.3 | Step 3: Secondary care
Within the secondary care, the present authors involved CLDT mem‐
bers regularly to discuss the strategy for withdrawal. The multi‐disci‐
plinary team (MDT) was also involved in the discussion of individual
cases when withdrawal was considered. However, the present authors
integrated primary and secondary care within a whole system ap‐
proach by establishing a STOMP oversight committee (project team).
This committee was led by the clinical director of the local special‐
ist Learning Disability service with multiagency representation from
service users, primary care liaison nurses, community pharmacists,
commissioners, social care and private providers with regular meet‐
ings to discuss ideas and review progress. After multiagency consulta‐
tion, and receiving patient/family/carers feedback, the project team
developed a framework to stratify, support and assess withdrawals.
In secondary care, the present authors also carried out an audit,
specifically looking at antipsychotic prescribing patterns in the local
CLDT during the whole month of October 2015 (Niven, Goodey,
Webb, & Shankar, 2018). The present authors were particularly in‐
terested to identify patterns and associations between prescribing
and MDT working and the use of the bio‐psycho‐social formulation.
2.4 | Step 4: Development of tools
The present authors developed a number of tools to facilitate the
process of withdrawal. These included ECLIPSE (already mentioned
in the earlier section), Risk Assessment Pyramid (RAP), E‐Connect,
the Purple book and the STOMP STAMP.
The RAP describes the factors (personal, behaviour related,
drug‐related, psychosocial‐environment, carer/staff training etc.)
associated with a high‐risk versus low‐risk withdrawal process. This
allows triaging and stratification for withdrawal and allows an appro‐
priate risk assessment and communication with people with intellec‐
tual disabilities and their carers.
E‐connect was developed to identify nine commonly identified
dimensions of treatment response that are clinically and holistically
important and which are commonly discussed in care plan meetings. E‐
connect summarizes the outcome score in a visual format (see Figure 2).
Table 2 describes the information gathering tool that allows us to use an
algorithm to summarize the findings in a visual format. In principle, it is a
care planning tool bringing together information collected from current
good practice policies and established tools as a single infographic.
STOMP STAMP describes three possibilities after the antipsy‐
chotic withdrawal is considered, which may indicate a successful
withdrawal (Green) or inability to withdraw (Red) (see Figure 3).
No formal training was provided on the tools. However, case stud‐
ies were used to implement them. A power point presentation was
used to help outline to new carers/patients the reasons why the
Behaviour that concerns (BtC) Connect is helpful.
2.5 | Step 5: Assessment of patients for withdrawal
At this stage, all referrals for withdrawing antipsychotic medication
were assessed using the risk assessment pyramid (RAP). This allowed
us to prioritize patients for withdrawal and have a full discussion about
risk assessment and management with people with intellectual disabil‐
ities and their carers. This allowed the clinicians to consider all factors
that may influence the withdrawal process. Usually dose changes were
10%–25% of the baseline dose reduced every 6–8 weeks.
2.6 | Step 6: Follow‐up and contingency plan
An appropriate arrangement was made to follow up patients during
and after the gradual withdrawal. This involved a contingency plan (see
Figure 1) which was discussed fully with the people with intellectual
disabilities and their carers. This was devised to alleviate patients’ and
their carers’ anxiety about the withdrawal as they were worried about
losing their placement or worsening of behaviour or hospital admission.
3 | RESULTS
3.1 | Step 1: Primary care: findings
Of the 60 expected GP participants, the tutorial were attended by 44
GPs (73%) and the follow‐up meeting by 42 (70%). Ninety per cent GPs