2026, Vol. 6, Issue 6, Part A

Keywords: Deprescribing, Psychotropic polypharmacy, Autism spectrum disorder, Intellectual disability, Clinician decision-making, Prescribing barriers, Learning disability services

Introduction

Deprescribing refers to the systematic process of tapering, reducing, or discontinuing medications that are no longer indicated, where potential harms outweigh benefits, or where continued use lacks an evidence base (Tyrer et al., 2008; Sheehan et al., 2015) ^{[1, 2]}. In the context of autistic adults with co-occurring intellectual disability, psychotropic polypharmacy presents a particularly resistant target for deprescribing efforts, as the original prescribing rationale is frequently unclear and behavioural responses to medication changes are difficult to predict (Branford et al., 2019; Branford et al., 2023) ^{[3, 4]}.

The scale of psychotropic prescribing in populations with intellectual disability has been documented across multiple healthcare systems. Sheehan and colleagues analysed primary care records of 33,016 adults with intellectual disability in England and found that 49% received at least one psychotropic medication, with antipsychotics being the most commonly prescribed class at 21% (Sheehan et al., 2015) ^[2]. Henderson and colleagues extended this analysis through data linkage and demonstrated that adults with both autism and intellectual disability received psychotropic prescriptions at rates substantially higher than those with either condition alone (Henderson et al., 2022) ^[5]. The STOMP (Stopping the Over-Medication of People with a learning disability, autism, or both) initiative, described by Branford and colleagues, was launched in England in 2016 as a national campaign to address this overprescribing, yet progress has been slow and uneven across services (Branford et al., 2019) ^[3].

The evidence base for long-term psychotropic use in this population is particularly weak. Tyrer and colleagues, in a landmark randomised controlled trial involving 86 adults with intellectual disability and challenging behaviour, found that both haloperidol and risperidone were no more effective than placebo in reducing aggressive behaviour over four weeks (Tyrer et al., 2008) ^[1]. This finding, subsequently replicated and extended by other research groups, raised fundamental questions about the continued prescription of antipsychotics for behaviour management in this population (NICE, 2015; Branford et al., 2023) ^{[6, 4]}. Yet discontinuation rates remain low, suggesting that factors beyond efficacy evidence govern prescribing decisions.

Several qualitative investigations have begun to identify the barriers that impede deprescribing. Branford and colleagues conducted a systematic review of 23 investigations examining psychotropic discontinuation in people with intellectual disability and found that clinician confidence, carer anxiety, and institutional inertia were recurring themes across healthcare systems (Branford et al., 2023) ^[4]. Bailey and colleagues explored stakeholder perspectives through qualitative interviews and identified a pervasive fear of behavioural deterioration as the dominant barrier, even among clinicians who acknowledged the lack of evidence supporting continued prescribing (Bailey et al., 2023) ^[7]. O'Dwyer and colleagues, examining prescribing patterns in residential care settings in Ireland, documented the role of organisational culture in perpetuating polypharmacy, with risk-averse management structures discouraging medication changes (O'Dwyer et al., 2016) ^[8].

Despite this growing body of qualitative research, quantitative measurement of deprescribing barriers across professional groups remains limited. Existing instruments measure general deprescribing attitudes in elderly care populations but do not capture the specific clinical, interpersonal, and systemic factors that distinguish psychotropic deprescribing in autism and intellectual disability services (de Kuijper et al., 2017; Bowring et al., 2019) ^{[9, 10]}. Adams and colleagues explored stakeholder experiences of the deprescribing process and identified systemic barriers that have not been examined across the broader multidisciplinary team (Adams et al., 2024) ^[11]. At the same time, research to date has been concentrated in high-income Western healthcare systems, with no published data from the Middle East and North Africa region, where service structures, family dynamics, and regulatory frameworks differ substantially from those in the United Kingdom or Australia.

This research addresses these gaps through a mixed-methods investigation of deprescribing barriers among four professional groups working in community intellectual disability services in Jordan. The quantitative arm measures barrier severity using a purpose-built inventory validated through Delphi consensus, while the qualitative arm examines how barriers interact and reinforce one another in clinical practice.

Materials and methods

This mixed-methods research employed a convergent parallel design, with quantitative survey data and qualitative interview data collected concurrently between January and September 2023 across three health directorates in Amman, Irbid, and Zarqa, Jordan. The research protocol received ethical approval from the Amman Academy of Clinical Psychiatry Research Ethics Committee (reference AACP-2022-064) and the Jordanian Ministry of Health Institutional Review Board.

Participants were 126 clinicians actively involved in the care of autistic adults with intellectual disability: 34 psychiatrists, 41 general practitioners, 29 learning disability nurses, and 22 pharmacists. Inclusion criteria required at least two years of clinical experience with this population and current involvement in medication-related decision-making. Recruitment used purposive sampling through professional networks and service directories, supplemented by snowball sampling. Informed consent was obtained from all participants prior to data collection.

The Deprescribing Barriers Inventory (DBI) was developed through a three-stage process. Stage one involved systematic review of existing deprescribing barriers literature and extraction of candidate barrier items. Stage two refined these items through a two-round Delphi process involving 12 subject-matter experts (four psychiatrists, three pharmacists, three learning disability nurses, and two clinical psychologists), with items retained if they achieved 75% or greater consensus on relevance and clarity. Stage three piloted the 42-item instrument with 20 clinicians not included in the main sample, assessing comprehension, completion time, and internal consistency. The final DBI comprises 42 items rated on a 10-point Likert scale across six subscales, with Cronbach alpha values ranging from 0.79 to 0.91.

Quantitative data were analysed using SPSS version 28. Descriptive statistics characterised barrier scores by professional group. One-way ANOVA compared composite and subscale barrier scores across the four groups, with Tukey HSD post-hoc tests for pairwise comparisons. Exploratory factor analysis with varimax rotation confirmed the six-factor structure of the DBI. Effect sizes were reported as partial eta-squared.

The qualitative component involved semi-structured interviews with a purposive subsample of 38 participants (10 psychiatrists, 12 general practitioners, 9 learning disability nurses, and 7 pharmacists), selected to maximise variation in barrier profiles as identified from preliminary DBI scores. Interviews were conducted in Arabic, audio-recorded, transcribed verbatim, and translated into English by a bilingual research assistant with clinical background. Member checking was performed with eight participants to verify translation accuracy and interpretive validity.

Qualitative data were analysed using Braun and Clarke's six-phase reflexive thematic analysis. Two researchers independently coded the first 12 transcripts, achieving a pooled Cohen kappa of 0.83 before proceeding to code the remaining transcripts. Themes were developed through iterative discussion and mapped against the quantitative barrier domains to identify areas of convergence and divergence. Integration of quantitative and qualitative findings followed the joint display approach recommended for convergent designs.

Results

Participant demographics and response characteristics

Of 154 clinicians invited, 126 completed the DBI (response rate = 81.8%). Participants had a mean of 8.7 years of clinical experience (SD = 4.3) with this population. Sixty-seven percent reported managing at least one medication review per week for autistic adults with intellectual disability, and 84.1% reported that at least half of their patients with intellectual disability were receiving two or more concurrent psychotropic medications.

Barrier domain scores

The six-factor structure of the DBI was confirmed by exploratory factor analysis, accounting for 68.4% of total variance. Table 1 presents mean subscale scores by professional group. Clinical uncertainty regarding behavioural outcomes emerged as the highest-scoring domain overall (mean = 8.2, SD = 1.3), followed by fear of behavioural deterioration (mean = 7.8, SD = 1.5) and perceived lack of alternatives (mean = 7.4, SD = 1.6). Organisational risk aversion scored lowest overall (mean = 6.3, SD = 2.1) but showed the greatest between-group variation.

Between-group comparisons

Composite barrier scores differed significantly across professional groups (F(3, 122) = 4.18, p = 0.003, partial eta-squared = 0.09; Figure 1). General practitioners reported the highest composite barrier score (mean = 7.9, SD = 1.2), significantly exceeding pharmacists (mean = 6.1, SD = 1.8; Tukey HSD p = 0.001). Psychiatrists (mean = 6.8, SD = 1.4) and learning disability nurses (mean = 7.2, SD = 1.6) occupied intermediate positions. Subscale analysis revealed that general practitioners scored significantly higher than psychiatrists on clinical uncertainty (8.7 vs 7.4, p = 0.006) and lack of alternatives (8.3 vs 6.9, p = 0.011), while learning disability nurses scored highest on carer resistance (8.6 vs group mean 7.1, p = 0.004), reflecting their greater direct contact with families and residential care staff.

Qualitative findings: The self-reinforcing barrier cycle

Thematic analysis of 38 interview transcripts identified six themes that corresponded directly to the DBI subscales, together with a superordinate theme describing the cyclical reinforcement between barriers. Figure 2 illustrates this self-reinforcing cycle. Participants described how clinical uncertainty about post-reduction behavioural outcomes created a default preference for maintaining current prescriptions. One psychiatrist stated: "Even when I know the evidence does not support continuing, the question is always what happens next. And nobody can answer that with certainty" (Participant P09). This uncertainty reduced exposure to successful deprescribing experiences, which in turn limited familiarity with the process and heightened fear of adverse behavioural outcomes.

Fear of behavioural deterioration was described as both a clinical concern and a professional risk. General practitioners, who typically had less specialist training in intellectual disability, expressed particular anxiety about managing behavioural crises that might follow dose reduction: "If the behaviour gets worse, the family calls emergency services, and then everyone asks why you changed the medication" (Participant G14). This professional exposure concern linked directly to organisational risk aversion, with several participants noting that service-level incident reporting systems implicitly penalised medication changes that preceded behavioural incidents, regardless of the clinical appropriateness of the change.

Carer and family resistance operated as both a direct barrier and an amplifier of clinician reluctance. Learning disability nurses, who served as the primary interface between prescribers and families, described repeated experiences of families opposing dose reduction on the basis that current medications were "working" even when objective functional outcomes were poor: "The family says he is calm, but calm is not the same as well. He sleeps sixteen hours a day and they call that calm" (Participant N22). Pharmacists, despite reporting the lowest composite barrier scores, identified systemic gaps in their involvement, noting that dispensing data and medication review expertise were rarely sought during deprescribing discussions.

Discussion

The international movement toward reducing psychotropic overprescribing in people with intellectual disability has generated policy commitments, awareness campaigns, and a growing evidence base documenting the scope of the problem (Sheehan et al., 2015; Branford et al., 2019; Branford et al., 2023) ^{[2, 3, 4]}. What has received less attention is why progress remains slow despite this accumulated knowledge. These findings provide a structured account of the barriers that sustain prescribing inertia and, critically, reveal the cyclical interdependence among them.

The identification of general practitioners as the professional group reporting the highest barrier scores aligns with their structural position in the care pathway. GPs in Jordan, as in many other healthcare systems, serve as the primary prescribing and repeat-prescribing interface for patients with intellectual disability but typically receive limited specialist training in this area (O'Dwyer et al., 2016; Henderson et al., 2022) ^{[8, 5]}. The combination of high prescribing responsibility and low specialist confidence creates conditions favourable to defensive continuation of established regimens. Bailey and colleagues reached similar conclusions in their qualitative investigation of stakeholder experiences in England, where GPs described feeling "caught between the recommendation to reduce and the lack of support to do so safely" (Bailey et al., 2023) ^[7].

The self-reinforcing barrier cycle identified through qualitative analysis represents a contribution beyond the enumeration of individual barriers. Prior literature has documented high rates of psychotropic medication use across intellectual disability populations (Song et al., 2023) ^[12] and catalogued deprescribing barriers as discrete items—knowledge gaps, time pressures, family concerns—without examining how they interact dynamically (de Kuijper et al., 2017; Young-Southward et al., 2017) ^{[9, 13]}. The cyclical model proposed here suggests that interventions targeting a single barrier in isolation are likely to encounter resistance from the remaining interconnected barriers. De Kuijper and colleagues found that structured medication review programmes produced modest deprescribing rates even with substantial resource investment, which is consistent with the cyclical resistance mechanism observed in these data (de Kuijper et al., 2017) ^[9].

These findings carry several limitations. The cross-sectional design captures barrier perceptions at a single time point and cannot track how they shift in response to deprescribing attempts. Social desirability bias may have inflated reported willingness to deprescribe, particularly among interviewees aware of the research aims. The Jordanian healthcare context, while sharing structural features with other middle-income settings, differs from Western European systems in family involvement norms, regulatory infrastructure, and specialist service availability, limiting direct generalisability. The DBI, though validated through Delphi consensus and pilot testing, requires further psychometric evaluation across larger and more diverse samples.

Clinical implications

The cyclical nature of deprescribing barriers implies that effective interventions must target multiple points in the cycle simultaneously rather than addressing barriers in sequence. Table 2 presents a proposed stepped deprescribing protocol designed to disrupt the barrier cycle at each phase by assigning clear responsibilities, providing structured monitoring, and embedding family engagement as a procedural step rather than leaving it to ad hoc clinical judgement (NICE, 2015; Branford et al., 2023) ^{[6, 4]}.

Training programmes should equip general practitioners with the specialist knowledge and supervised experience needed to initiate deprescribing with confidence, rather than relying on their willingness to seek specialist referral for each medication change (Adams et al., 2024; Adams et al., 2025) ^{[11, 14]}. Pharmacists, who reported the lowest barrier scores and possess medication review expertise, should be formally integrated into deprescribing decision-making teams rather than confined to dispensing roles. Learning disability nurses, given their unique position at the interface between prescribers and families, require specific training in managing carer anxiety and facilitating shared decision-making around medication reduction (Bailey et al., 2023) ^[7].

Safety considerations

Psychotropic deprescribing in autistic adults with intellectual disability carries specific safety risks that distinguish it from medication discontinuation in other populations. Abrupt withdrawal from antipsychotics can precipitate withdrawal dyskinesia, rebound psychosis, and autonomic instability, while benzodiazepine discontinuation risks seizure activity in individuals with pre-existing epilepsy (Tyrer et al., 2008; de Kuijper et al., 2017) ^{[1, 9]}. Any deprescribing protocol must therefore incorporate graduated tapering schedules, specify minimum inter-reduction intervals, and define clear criteria for pausing or reversing the taper if clinically indicated.

Behavioural monitoring during and after dose reduction requires instruments validated for use with individuals who have limited verbal communication. Standard psychiatric rating scales, designed for self-report or verbal interview, may not capture behavioural changes in this population. The Aberrant Behavior Checklist and the Behavior Problems Inventory, both validated for use in intellectual disability populations, offer alternative monitoring approaches that rely on caregiver observation rather than patient self-report (Matson et al., 2009; Bowring et al., 2019) ^{[15, 10]}.

Legal and regulatory framework

The regulatory context for psychotropic prescribing in people with intellectual disability varies substantially across jurisdictions (Table 3). In England, the NICE guideline NG11 recommends that antipsychotics prescribed for challenging behaviour should be reviewed after three months and discontinued if no benefit is evident (NICE, 2015) ^[6]. The STOMP campaign, endorsed by NHS England and supported by professional bodies, has increased awareness of overprescribing but lacks binding regulatory force (Branford et al., 2019) ^[3]. In Australia, the Comprehensive Health Assessment Program (CHAP) includes a medication review component for adults with intellectual disability but does not specify deprescribing protocols.

In Jordan, the Ministry of Health regulates psychiatric prescribing through general pharmaceutical legislation, with no specific provisions for intellectual disability populations. The absence of specialised guidance means that deprescribing decisions rest entirely on individual clinical judgement, without institutional frameworks to support or standardise the process (Sheehan et al., 2015; Henderson et al., 2022) ^{[2, 5]}. Development of region-specific deprescribing guidance, informed by local family dynamics and service structures, should be a priority for Jordanian mental health policy.

Conclusion

This research has identified and quantified the barriers that impede psychotropic deprescribing in autistic adults with intellectual disability within Jordanian community services, demonstrating that these barriers are distributed across clinical, interpersonal, and institutional domains and that they operate through a self-reinforcing cycle that resists isolated intervention. The finding that general practitioners report the highest barrier scores while pharmacists report the lowest points to a structural mismatch between prescribing responsibility and specialist expertise that characterises many healthcare systems beyond Jordan.

At the field level, these data add to the growing international evidence that overprescribing in intellectual disability populations is maintained not by ignorance of the evidence but by systemic conditions that make continuation safer and easier than change. The parallel findings from the United Kingdom, Ireland, and now Jordan suggest that the barrier cycle is not specific to any single healthcare system but rather reflects a more general pattern in the organisation of psychiatric care for people with intellectual disability (Sheehan et al., 2015; O'Dwyer et al., 2016; Branford et al., 2019; Branford et al., 2023) ^{[2, 8, 3, 4]}.

The cyclical barrier model proposed here offers a theoretical framework for understanding why deprescribing initiatives that address individual barriers—such as prescriber education alone or carer information campaigns alone—have produced limited results. Each barrier in the cycle draws strength from adjacent barriers: clinical uncertainty feeds prescribing inertia, which limits exposure to successful deprescribing, which maintains uncertainty. Breaking this cycle requires coordinated, simultaneous intervention at multiple points, as proposed in the stepped protocol presented in this research.

Practically, three priorities emerge from these findings. First, structured deprescribing protocols specific to psychotropic medications in autistic adults with intellectual disability need to be developed, piloted, and disseminated within and beyond the Jordanian context. The stepped protocol proposed here provides a starting framework that can be adapted to different service configurations and regulatory environments. Second, professional training in deprescribing must extend beyond psychiatrists to include general practitioners, learning disability nurses, and pharmacists, each of whom encounters different barrier profiles and plays a different role in the deprescribing process. Third, family engagement should be formalised as a procedural step within deprescribing protocols rather than left to individual clinician discretion, with specific communication tools and decision aids developed for families and carers of autistic adults with intellectual disability.

Future research should test the cyclical barrier model using longitudinal designs that track how barrier perceptions shift as clinicians gain deprescribing experience. Intervention trials should evaluate whether multi-component programmes targeting several barriers simultaneously produce greater deprescribing rates than single-component approaches. Cross-cultural replication of the DBI in other Middle Eastern healthcare systems would clarify whether the barrier profiles observed in Jordan generalise across the region.

References

1. 1.Tyrer P, Oliver-Africano PC, Ahmed Z, et al. Risperidone, haloperidol, and placebo in the treatment of aggressive challenging behaviour in patients with intellectual disability: a randomised controlled trial. Lancet. 2008;371(9606):57-63.
2. 2.Sheehan R, Hassiotis A, Walters K, Osborn D, Strydom A, Horsfall L. Mental illness, challenging behaviour, and psychotropic drug prescribing in people with intellectual disability: UK population based cohort study. BMJ. 2015;351:h4326.
3. 3.Branford D, Gerrard D, Saleem N, Shaw C, Webster A. Stopping over-medication of people with intellectual disability, autism or both (STOMP) in England part 1 - history and background of STOMP. Adv Ment Health Intellect Disabil. 2019;13(1):31-40.
4. 4.Branford D, Gerrard D, Saleem N, Shaw C, Taylor L. Stopping or reducing psychotropic drug treatment in people with learning disabilities: systematic review of the evidence informing clinical practice. BMC Psychiatry. 2023;23(1):234.
5. 5.Henderson A, Kinnear D, Fleming M, et al. Antipsychotic prescribing for people with intellectual disabilities, autism or both: a data linkage cohort study in Scotland. Br J Psychiatry. 2022;220(5):275-282.
6. 6.National Institute for Health and Care Excellence. Challenging behaviour and learning disabilities: prevention and interventions for people with learning disabilities whose behaviour challenges. NICE guideline NG11. London: NICE; 2015.
7. 7.Bailey T, Kiani R, Patel N, et al. The lived experience of deprescribing psychotropic medication for behaviours that challenge in people with intellectual disabilities: a qualitative study. Br J Learn Disabil. 2023;51(4):532-544.
8. 8.O'Dwyer M, Peklar J, McCallion P, McCarron M, Henman MC. Factors associated with polypharmacy and excessive polypharmacy in older people with intellectual disability differ from the general population: a cross-sectional observational nationwide study. BMJ Open. 2016;6(4):e010505.
9. 9.de Kuijper G, Hoekstra P. Physicians' reasons not to discontinue long-term used off-label antipsychotic drugs in people with intellectual disability. J Intellect Disabil Res. 2017;61(10):899-908.
10. 10.Bowring DL, Painter J, Hastings RP. Prevalence of challenging behaviour in adults with intellectual disabilities, correlates, and association with mental health. Curr Dev Disord Rep. 2019;6(4):173-181.
11. 11.Adams D, Hastings RP, Maidment I, Shah C, Langdon PE. Stakeholder experiences of deprescribing psychotropic medicines for challenging behaviour in people with intellectual disabilities. Tizard Learning Disability Review. 2024;29(2):99-106.
12. 12.Song M, Rubin BS, Ha JWT, Ware RS, Doan TN, Harley D. Use of psychotropic medications in adults with intellectual disability: a systematic review and meta-analysis. Aust N Z J Psychiatry. 2023;57(5):661-674.
13. 13.Young-Southward G, Rydzewska E, Philo C, Cooper SA. Physical and mental health of young people with and without intellectual disabilities: cross-sectional analysis of a whole country population. J Intellect Disabil Res. 2017;61(10):984-993.
14. 14.Adams D, Hastings RP, Maidment I, Langdon PE. Factors supporting optimisation of psychotropic deprescribing in people with intellectual disabilities within the UK: a modified Delphi study. Front Psychiatry. 2025;16:1652988.
15. 15.Matson JL, Neal D. Psychotropic medication use for challenging behaviors in persons with intellectual disabilities: an overview. Res Dev Disabil. 2009;30(3):572-586.