Go Decaf!: implementing ‘decaf by default’ as a quality improvement intervention to reduce toileting-related falls and improve care environments in acute hospital settings

Abstract

Background

Caffeinated hot drinks are routinely offered as the default in hospitals despite guidance recommending caffeine reduction for urinary urgency. Caffeine may increase urinary frequency, disrupt sleep and contribute to toileting-related falls, particularly among older adults living with frailty.

Local problem

Routinely collected incident reporting data at an National Health Service (NHS) Foundation Trust identified that around one quarter of inpatient falls were toileting-related. Staff recognised caffeine as a potential contributor but reported limited ability to influence drink provision within existing ward routines.

Methods

A registered quality improvement initiative used two sequential Plan–Do–Study–Act cycles to implement a ‘decaf by default’ approach across acute and community wards, an emergency department and a care home, followed by evaluation after organisation-wide rollout. Routine incident reporting monitored toileting-related falls, while staff and manager experience was assessed using structured questionnaires.

Intervention

Decaffeinated hot drinks were offered as the default, with caffeinated drinks available on request. Implementation was supported by staff education, shared decision-making materials, face-to-face clinical support and a patient-facing ‘Taste the Difference’ activity.

Results

Across pilot sites, toileting-related falls as a proportion of all inpatient falls reduced from 19%–21% to 11%–15%. Toileting-related falls reduced by approximately 11% in Cycle 1 and 20% in Cycle 2, with seven wards achieving reductions greater than 50%. Following organisation-wide rollout, the proportion of inpatient falls that were toileting-related reduced from 33.2% to 25.9%. Managers reported calmer ward environments, improved sleep, reduced agitation and fewer urgency-driven toileting attempts. Patient and relative feedback was predominantly positive or neutral.

Conclusions

Changing default drink provision to decaffeinated options is a feasible, low-cost intervention associated with fewer toileting-related falls and improved ward environments while preserving patient choice. Sustained improvements following organisation-wide rollout suggest the approach is scalable using existing clinical infrastructure.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Toileting-related falls are common among older hospital inpatients, and although reducing caffeine intake is recommended for urinary urgency, there is little evidence on how to implement this successfully within routine hospital care.

WHAT THIS STUDY ADDS

• Introducing decaffeinated drinks as the default option was associated with reductions in toileting-related falls during both pilot testing and subsequent organisation-wide rollout, alongside perceived improvements in sleep, agitation and ward environment. The intervention achieved spread by changing the default rather than relying on education or individual behaviour change alone.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study suggests that changing clinical defaults, supported by local ownership and peer-to-peer diffusion, may be an effective strategy for translating existing evidence into routine practice while preserving patient choice. Similar implementation approaches could be applied to other quality improvement priorities where evidence exists but has proved difficult to embed consistently in everyday care.

Introduction

Problem description

Inpatient falls remain a major cause of harm, particularly among older adults. Toileting-related falls represent a substantial proportion, often occurring when patients mobilise urgently, frequently overnight, and in the context of frailty, cognitive impairment or delirium. These events are distressing for patients, carers and staff and are challenging to prevent using environmental or staffing interventions alone.

At an NHS Foundation Trust in the North-East of England, review of incident reporting data identified that more than one quarter of inpatient falls in a single month were related to toileting. Care of the elderly wards was disproportionately affected. Staff described recurrent patterns of urinary urgency, repeated toileting attempts and overnight wandering, often accompanied by agitation and sleep disruption.

Available knowledge

Caffeine is a central nervous system stimulant known to exacerbate lower urinary tract symptoms, including urgency and frequency.^1–4 It is also associated with sleep disruption and increased restlessness, particularly in older adults and people with cognitive impairment.⁵ National Institute for Health and Care Excellence (NICE) guidance on urinary incontinence and overactive bladder recommends reducing caffeine intake as part of conservative management, particularly where urinary urgency contributes to functional risk.⁶

Despite this guidance, caffeinated hot drinks remain the default option in many hospital wards. This reflects organisational routines rather than deliberate clinical decision-making. As a result, the contribution of caffeine to urinary urgency, sleep disturbance and fall risk may be under-recognised or viewed as difficult to address in practice.

Previous quality improvement work in another NHS organisation demonstrated that introducing decaffeinated drinks as the default, supported by staff engagement and patient education, was associated with reductions in toileting-related falls and improvements in continence-related symptoms.⁷ This suggested that default drink provision may represent an underused and modifiable lever within multifactorial falls prevention strategies.

Rationale

The clinical rationale for this project was that reducing caffeine exposure could decrease urinary urgency, reduce nocturnal waking and restless behaviours and improve sleep, thereby reducing toileting-related falls and agitation.^1–5 This aligns with NICE guidance recommending caffeine reduction as part of conservative management for urinary symptoms that increase functional risk.⁶

A default change was selected because defaults exert a powerful influence on behaviour in busy clinical environments, particularly where staff have limited time for repeated explanation. Changing the default was intended to reduce reliance on repeated persuasion while preserving patient autonomy and informed choice.

The implementation approach was informed by contemporary change-management theory, which challenges the assumption that large-scale change is achieved through information dissemination alone. Instead, change is understood to spread through social processes, including peer influence, local norms and visible early success.^{8 9} Staff attending an older people’s care workforce development programme frequently expressed awareness of potential harms of caffeine but described the issue as ‘outside their control’, reflecting a gap between knowledge and agency.

Rather than attempting a single organisation-wide launch, the project began in areas with local interest and leadership support. Early adopters were supported to test and adapt the approach, generating locally credible examples that could be shared across peer networks. This strategy aimed to reduce early resistance, normalise the practice change and create social proof that ‘people like us are already doing this’.⁹

Specific aims

The project aimed to:

1. Reduce toileting-related inpatient falls.

2. Improve staff understanding and advocacy of decaffeinated drinks while preserving patient autonomy and informed choice.

3. Improve aspects of patient experience related to sleep, agitation and ward environment.

4. Develop an intervention that could be sustained and spread across diverse care settings.

Methods

Context

The initiative took place within an NHS Foundation Trust in the North-East of England providing acute and community services across a large geographical area. The work emerged through an older people’s care workforce development programme, where frontline staff consistently raised concerns about caffeine but described limited ability to influence drink provision within existing ward routines.

The initiative was developed and led by registered healthcare practitioners working in quality improvement and workforce development roles, with backgrounds in patient-facing clinical practice across mental health and acute care pathways. The implementation team included senior nursing sponsors, an older people’s care improvement lead, a continence specialist clinician, ward leadership representatives and operational colleagues supporting procurement and communications. Ward-based hospitality and nutritional assistant staff supported delivery of the intervention in day-to-day practice and acted as local champions. Responsibilities were agreed locally: ward leaders led implementation and oversight; specialist teams provided education and troubleshooting; communications colleagues developed consistent resources for staff, patients and carers; and operational teams ensured reliable procurement and availability of decaffeinated products.

Cycle 1 was undertaken over 3 months (April–June 2024) across six pilot sites comprising four acute wards, an emergency department and one care home (approximately 180 beds). Following governance review and refinement of implementation materials, stakeholder engagement and preparation for wider testing, Cycle 2 was undertaken over 3 months (January–March 2025) across 15 sites, 10 acute wards, an emergency department and four community hospitals (approximately 400 beds).

3-month cycles were selected pragmatically to provide sufficient outcome observations while maintaining momentum and feasibility within routine clinical services. No major fall prevention initiatives or changes to incident-reporting processes were identified during the interval between cycles.

Senior nursing leadership supported the work and enabled approval through the organisation’s nutrition and hydration governance route. The project was designed with explicit recognition that awareness of NICE guidance alone was unlikely to produce sustained change. Implementation therefore focused on enabling staff to act differently within existing systems rather than increasing information or compliance messaging.⁷

As this was an iterative quality improvement initiative rather than a formal effectiveness study, outcome measurement was intentionally limited to routinely collected incident reporting data to minimise burden and support rapid cycles of learning. Falls were therefore reported using counts and proportions of toileting-related falls rather than standardised rates per 1000 bed-days, reflecting the measures routinely used by frontline teams to monitor improvement. We cannot exclude the possibility that unmeasured contextual factors influenced observed outcomes.

Intervention

The intervention introduced ‘decaf by default’ for hot drinks in participating areas. Decaffeinated drinks were offered routinely, while caffeinated alternatives remained available on request.

Implementation was supported by education materials for staff, patients and carers outlining the benefits of reducing caffeine, potential withdrawal effects and mitigation strategies such as gradual reduction. Face-to-face support was provided by the older people’s care team and continence specialists. A ‘Taste the Difference’ challenge enabled patients to sample drinks and make informed choices. Existing suppliers and stock were used, avoiding additional cost or infrastructure. No changes were made to patients’ access to caffeinated drinks, which remained available at any time on request.

Staff were encouraged to use shared decision-making language and to personalise discussions based on patient preference and clinical context.

Study of the intervention

Impact was assessed using routinely collected incident reporting data on toileting-related falls. Staff were not asked to collect additional clinical data to minimise burden and avoid disrupting routine care. Following organisation-wide rollout, routinely collected incident reporting data were also compared for the same 3-month period before (January–March 2025) and after (January–March 2026) trust-wide adoption, using the same outcome definitions and reporting processes. This formed the basis of the organisation-wide evaluation presented in figure 1.

Figure 1. Proportion of inpatient falls that were toileting-related before and after organisation-wide rollout of Decaf by Default.

To strengthen attribution within a pragmatic design, several approaches were used. These included the use of consistent outcome definitions across cycles; comparison of pilot-site patterns with contemporaneous organisation-wide trends; examination of variation by ward type and time of day and triangulation of quantitative trends with qualitative reports from ward managers describing contemporaneous changes in urgency-driven toileting, sleep and agitation. These approaches do not eliminate confounding but were intended to increase confidence that observed patterns were plausibly related to the intervention rather than background variation.

Additional outcome measures, including delirium screening using validated tools, were explored during project planning. However, their introduction was not feasible without increasing staff workload and risking disengagement. Instead, structured questionnaires were used to capture manager and staff perceptions of patient behaviour, sleep, agitation and ward environment. This reflected a deliberate trade-off between measurement precision and implementation feasibility in pressured clinical environments.⁸

Measures and analysis

Primary outcome measures were the number and proportion of toileting-related inpatient falls. Secondary measures included care home overnight pad use as a proxy for nocturnal continence burden, ward and care home manager questionnaires on patient safety and care environment and staff knowledge and experience assessments.

Quantitative data were analysed descriptively, comparing pilot sites with overall organisational patterns during the same period. Qualitative responses were analysed thematically to identify recurring patterns related to sleep, agitation, continence and ward atmosphere. No concurrent organisational changes to fall reporting processes occurred during the study period.

A toileting-related fall was defined as an inpatient fall recorded in the incident reporting system where the narrative or category indicated toileting, transfer to toilet or urgency-related mobilisation. Data completeness and accuracy depended on routine reporting and categorisation practices. To enhance consistency, the project team used a shared operational definition and reviewed samples of incident narratives to confirm coding where ambiguity existed. Contextual factors influencing implementation were monitored through regular site check-ins, including staffing pressures, ward engagement, availability of decaffeinated stock and local championing.

Quantitative data were analysed descriptively across cycles and sites. Given the pragmatic design and small numbers at ward level, formal statistical process control was not undertaken. Instead, we explored variation across time, ward type and care setting and triangulated quantitative patterns with qualitative findings.

Ethical considerations

The project was registered as a quality improvement initiative within the organisation’s governance framework and did not require formal research ethics review. Patient autonomy was preserved by maintaining access to caffeinated drinks on request. Education materials addressed potential withdrawal symptoms and mitigation strategies. No reduction in overall fluid intake or increase in hydration-related concerns was reported during routine ward monitoring.

Results

Implementation and evolution

Cycle 1 involved six sites across four acute wards, an emergency department and one care home. Cycle 2 expanded to 15 sites, including all care of the elderly wards.

Between cycles, implementation materials were refined to improve clarity on shared decision-making and caffeine withdrawal mitigation, and communication resources were standardised to support scale-up. Operational processes were strengthened to reduce stock variability and support wards newly adopting the approach. These refinements reflected learning from early sites and were incorporated prior to wider rollout.

Following the completion of the pilot and governance review, ‘decaf by default’ was adopted as standard practice and rolled out across acute wards in the organisation. Wards introduced to the intervention at scale, without prior involvement in the pilot, reported similarly positive early experiences. This suggested that the intervention was acceptable and implementable even in areas ‘coming to it cold’, without the need for prolonged preparatory engagement.

Fall outcomes

Figure 2 shows the temporal pattern of toileting-related falls across the baseline period and both improvement cycles. The run chart demonstrates an initial reduction following implementation in the pilot sites that was sustained during subsequent testing and expansion, despite the iterative nature of the project and changing participating wards.

Figure 2. Run chart showing the proportion of inpatient falls that were toileting-related across baseline and two Plan-Do-Study-Act (PDSA) cycles.

During the preintervention period, toileting-related falls accounted for 30.6% of all falls across the baseline wards. Following implementation, this reduced to 19.1% during Cycle 1 and remained lower than baseline at 17.5% during Cycle 2 despite expansion to additional wards. Seven wards achieved reductions greater than 50%.

Trust-wide implementation

Following governance approval, the intervention was adopted across acute inpatient wards within the organisation. Analysis of routine incident reporting before and after trust-wide implementation demonstrated a reduction in the proportion of all inpatient falls that were toileting-related from 33.2% (326/982) to 25.9% (530/2050), representing an absolute reduction of 7 percentage points and a relative reduction of approximately 22% (figure 1). Overall inpatient falls also reduced during this period (618 to 502), although the project was not designed to attribute this change to the intervention. No obvious difference was observed between day and night reductions, and no clear correlation was found between ward specialty and magnitude of change.

Qualitative findings

Managers reported improvements in patient behaviour and ward environment, particularly among patients with cognitive impairment. One manager described reduced nocturnal disturbance and a shift in default practice:

The patients we care for often have dementia, delirium, cognitive impairment. Over the last 12 months we have noticed a huge difference. Less waking and wandering, especially overnight, and better sleep. We now offer decaf drinks as the norm, only offering caffeinated drinks if requested.

Several wards described improvements following time-based caffeine reduction:

Since decaf was introduced patients sleep better, urinary frequency is reduced, less agitated, the ward seems calmer.

Others reported smaller but still positive changes:

Slight improvement in agitation of patients.

Toileting behaviour was frequently referenced:

Patients are calmer. Less trips to the toilet for a wee.

Less amount of falls overnight, patients going to the toilet. Patients state that they have a more restful night.

Patient acceptability was high, with most patients not reporting concerns about taste. No negative feedback from relatives was reported. Managers planned to continue offering decaf by default.

No unintended harms were identified during the pilot or subsequent rollout. In particular, there were no reports of reduced overall fluid intake, increased dehydration concerns or patient dissatisfaction related to drink provision. Missing data were limited to under-reporting inherent in routine incident reporting and variable completeness of free-text narratives. No additional data collection was mandated.

Discussion

Interpretation of findings

This quality improvement initiative found that introducing decaffeinated drinks as the default was associated with reductions in toileting-related falls during both pilot testing and subsequent organisation-wide rollout. Qualitative findings suggest reduced urgency-driven toileting, improved sleep and calmer behaviour, particularly overnight. These observations are consistent with the proposed mechanism of caffeine reduction.^1–5

As an observational quality improvement initiative, these findings demonstrate association rather than causation. The pragmatic design means that unmeasured confounding, including seasonal variation, staffing changes and other contextual influences, cannot be excluded. In addition, no direct process measures (such as toileting frequency, overnight toileting episodes or call-bell use) were collected, limiting our ability to examine the proposed mechanism through which reduced caffeine exposure may have influenced fall outcomes. Qualitative reports provide indirect support for this mechanism but may be subject to perception bias. Future evaluations should incorporate these measures alongside patient-level outcomes to better understand both how and for whom the intervention is most effective. Nevertheless, the consistency of findings across diverse settings, including wards adopting the intervention for the first time and subsequent organisation-wide rollout, supports its feasibility and practical value.

Importantly, reductions observed during the pilot were also evident following organisation-wide adoption, with the proportion of inpatient falls that were toileting-related reducing from 33.2% to 25.9% across the Trust (figure 1). During the same period, the overall number of inpatient falls also reduced from 618 to 502, although this observational study was not designed to attribute this change to the intervention alone.

While these observational data cannot establish causality, they suggest that the intervention remained feasible and acceptable when implemented at scale, including in wards with no previous involvement in the improvement work. This reduces the likelihood that improvements were confined to early-adopter sites and normalisation of the intervention and diffusion through routine clinical systems.

These findings extend previous work demonstrating reductions in continence-related symptoms following decaffeinated default provision by showing that the intervention can be implemented successfully across multiple acute and community settings and sustained following organisation-wide adoption.

The findings are consistent with contemporary theories of change in complex adaptive systems, including diffusion of innovation and social influence models, which emphasise the role of defaults, peer norms and local ownership in enabling behaviour change.^8–10 Although NICE guidance on caffeine reduction has been available for several years, staff reported limited ability to operationalise this guidance within routine practice. By reframing caffeine reduction as a system-level default rather than an individual behavioural request, the intervention enabled staff to enact guidance more consistently. The intervention required minimal staff time beyond existing workflows and did not displace or compete with other fall prevention activities.

More broadly, this work illustrates how redesigning routine defaults may provide a practical mechanism for implementing evidence-based care in complex healthcare systems. Rather than relying on repeated education or persuasion, changing the default altered the environment in which decisions were made while preserving patient choice.

Reflexive statement on the use of PDSA

PDSA methodology was selected to support iterative testing in real-world clinical settings while minimising disruption to staff. In practice, the cycles functioned less as hypothesis-testing experiments and more as protected spaces for early adopters to trial, adapt and normalise a new default. This supported learning, local ownership and subsequent scale-up.

This approach inevitably limited causal attribution. However, in the context of a complex healthcare system, the team prioritised feasibility, engagement and sustainability over experimental control, consistent with contemporary critiques of PDSA use in improvement science.¹¹

Sustainability, spread and system support

Sustainability was supported by several reinforcing mechanisms. First, manager endorsement created a strong mandate for continuation. Second, the intervention required no additional staffing, equipment or recurrent funding, reducing dependency on short-term resources. Third, the default change reduced cognitive and workload burden for staff, embedding the practice into routine workflows rather than relying on ongoing vigilance.

Rollout was supported by coordinated communications work to develop consistent, accessible information resources for staff, patients and carers. These materials reinforced key messages about choice, benefits and withdrawal mitigation, supporting implementation at scale. System-level support, including back-office and supply-chain alignment, ensured reliable availability of decaffeinated options and reduced operational friction.

Ward-based hospitality and nutritional assistant staff played a practical role as local champions. They reinforced the default in everyday practice and modelled conversations with patients and relatives. Their involvement helped normalise the change and support consistency across shifts.

Sustainability was further supported by factors known to enable hospital-wide improvement, including leadership endorsement, integration into routine workflows, minimal additional resource requirements and reinforcement through local ownership, all of which have been identified as key enablers of sustained change in previous hospital-wide improvement initiatives.¹²

The successful transition from pilot testing to routine organisation-wide practice, together with sustained reductions in toileting-related falls following trust-wide implementation (figure 1), demonstrates that the intervention can be embedded within routine care. The simplicity and adaptability of the model also suggest transferability to community hospitals, care homes and intermediate care environments, where urinary urgency, sleep disruption and fall risk remain important drivers of harm.

Conclusions

Offering decaffeinated drinks as the default is a feasible, acceptable and low-cost intervention associated with reduced toileting-related falls and improvements in ward environment. Preserving patient choice, reducing staff burden and enabling local ownership were central to success. This approach represents a pragmatic addition to multifactorial fall prevention strategies aligned with existing clinical guidance.

Next steps include evaluation in intermediate care and community settings and exploration of more robust patient-level measures of sleep and urinary symptoms where feasible without increasing staff burden.

Data availability statement

Data are available upon reasonable request.