The importance of adequate hydration for the maintenance of good health and wellbeing is widely recognised and is publicly advocated in the UK (Popkin et al, 2010; NHS website, 2018; 2019). Recent reminders in the published literature inform clinicians of the marked improvements in personal performance attained by adequate hydration (Parry et al, 2017). Patient malnutrition in the hospital setting, a known predisposing risk factor for morbidity and mortality, has received a significant amount of focus in recent literature, with both academic and political publications describing this as a clear issue within the UK health system (Brotherton et al, 2010; Parliamentary and Health Service Ombudsman (PHSO), 2011; Leach et al, 2013; Stratton et al, 2018).
Adequate hydration becomes even more significant in the inpatient hospital setting, where homeostatic mechanisms in the kidney may be impaired and where an illness state may alter fluid requirements. Alongside the admitting diagnosis, other vital factors to consider when assessing the importance and adequacy of hydration in the inpatient population include prescription medications and the use of laxatives or diuretics, as well as any concomitant acute or long-standing cognitive changes.
Dehydration in the inpatient setting is common, affecting over one third of elderly patients on admission, and is associated with subsequent acute kidney injury in about one quarter of cases (El-Sharkawy et al, 2015). The National Confidential Enquiry into Patient Outcome and Death (NCEPOD) report on acute kidney injury (AKI) demonstrated significant morbidity and mortality associated with avoidable causes of AKI, suggesting there were gaps in clinical care regarding providing hydration to patients (NCEPOD, 2009; MacLeod, 2009). Previous work has demonstrated difficulties in patient access to oral fluids that go beyond the availability of a jug at the bedside (Gaff et al, 2015; Johnstone et al, 2015). The Royal College of Nursing (RCN) also maintains a clinical framework that serves as an important reminder of the importance of ensuring patients are adequately hydrated (RCN, 2019).
The elderly patient population in hospital and institutional care is known to be at risk of dehydration (Lešnik et al, 2017). Among these, immobile patients are an especially vulnerable group; they may have impaired abilities to reach for items around the bedside. Further, evidence suggests that patients with dementia or those who have experienced stroke may be less sensitive to hypothalamic thirst centre signalling (Ishii and Iadecola, 2015), further stressing the important role of healthcare staff in ensuring adequate hydration. The RCN and National Patient Safety Agency's (NPSA) Water for Health toolkit (RCN and NPSA, 2007) highlights the need to make sure that drinking water is readily available at all times, that drinking while bedbound is actively encouraged and that patients should be enabled to attain a hydration standard of 2.5 litres of fluid per day (RCN and NPSA, 2007; Department of Health, 2010; RCN, 2019). There is inconsistency regarding measurement of dehydration, which may have hindered the benchmarking process to date within the patient care setting; clinical tools such as reduced skin turgor or dry mucous membranes are commonly used in everyday practice. However, these measures are subjective and generally have a low sensitivity (Fortes et al, 2015). Conversely, measurement of serum osmolality can provide an objective marker of hydration (El-Sharkawy et al, 2015), but is not widely adopted in clinical practice.
This study sought to assess the adequacy of provision of hydration at the bedside to patients who are unable to move from their bed without assistance.
Methods
Study design
This was a cross-sectional, point-prevalent audit of hospitalised medical patients in five enrolled NHS hospital sites in England. All sites were acute general hospitals.
Audit aims
To define the adequacy of provision of oral hydration to bedbound inpatients and define factors associated with a failure of provision. This was assessed by examining whether a water jug and cup were present at the bedside, and whether these were within reach of the patient.
Ethical approval
Each site received independent local ethics approval for the study as an audit of current practice against an identified gold-standard of care. An audit standard was defined as 100% of patients being able to access appropriate oral hydration where there were no contraindications to this, in keeping with the aforementioned RCN best practice toolkit (RCN and NPSA, 2007) and a Care Quality Commission (CQC) report on Dignity and Nutrition in NHS Hospitals (CQC, 2013).
Inclusion and exclusion criteria
The number of wards contributing from each of the five sites was limited by the number of members of the research team at each site. The study included general medical and specialist elderly care wards only. Wards in which common treatments would have confounding influences on fluid balance and provision were excluded, such as surgical, specialist cardiology and renal.
Definitions
Patients were defined as bedbound if they were unable to mobilise from their bed without the assistance of a member of staff; we did not include patients who were independent with an apparatus (such as a frame or walking stick). Patients who were kept nil by mouth for clinical reasons were separated in the analysis—and these indications were noted. Patients who were on continuous enteral feeding, as well as patients who were nil by mouth pending investigations or procedures, were excluded.
Data collection
Data were collected by off-shift junior doctors on audit proformas approved by all local audit boards. Data collected included patient demographics and admission-related information, including diagnosis and whether the patient was bedbound, was unable to swallow safely or was receiving tube-assisted feeding, whether or not a water jug or alternative was present at the bedside, and whether this was within reach of the patient.
Heterogeneity and bias
In an attempt to reduce expectation bias ward staff were unaware of the audit taking place before the arrival of the study personnel on the ward. The study was conducted across all sites at mid-afternoon (between 14:30 and 17:00), aiming not to coincide with meal times or medical or nursing drug-dispensing rounds, which would be anticipated confounders.
Analysis
Data were collated and analysed with appropriate statistical modelling, using a combination of Microsoft Excel and SPSSv23. Quantitative data were compared using unpaired Student's t-test and ANOVA, proportional data using Fisher's exact test with an associated odds ratio and 95% confidence interval (OR + CI). Multivariate logistic regression modelling was performed to identify factors associated with inadequate supply of water at the bedside. The statistical significance level for P values was taken as 0.05.
For representation within the results section, data are displayed as a median (range) within tables. In performing uni- and multivariate modelling, age and duration of stay were converted to categorical data using NHS England definitions. Patients were defined as over 75 years of age in keeping with categorisation of Elderly Medicine and Care (NHS Interim Management and Support (now part of NHS Improvement), 2012; NHS England, 2014). Prolonged admission was defined as 25 days, in keeping with the 95th centile of hospital stay duration, according to NHS Hospital Episode Statistics 2014–2015 (Health and Social Care Information Centre (HSCIC) (now NHS Digital), HSCIC, 2015).
Results
Patient demographics and length of stay
Across a total surveyed patient population of 559, we identified 138 patients who were bedbound. Full data for study entry were available for 100% of bedbound patients. Table 1 demonstrates the number of wards studied per site, the number of patients per site, and the age and admission duration of all patients in the study and specifically the bedbound patient population. The median age of the bedbound population was 82 years with a range of 22–99 years. Within all centres but one (Centre E), the bedbound population was older than the mobile population. However, this was statistically significant in one centre only (Centre A). In pooled analysis, the mean age of the bedbound population was 2 years older than the mobile population (78 versus 76). Bedbound patients' duration of stay was significantly longer (21 days versus 14 days, P=0.009). There was no difference in gender between populations across and between all centres.
| Centre (number of wards surveyed) | Number of patients (male) | Age (years) | Duration of admission (days) | Cases of inadequate water provision (%) |
|---|---|---|---|---|
| A (8) | 190 | 76 (17–97) | 10 (1–165) | – |
| Bedbound | 68 (30) | 81 (22–96)* | 14 (1–165) | 14 (20.1) |
| B (1) | 19 | 84 (64–94) | 7 (1–40) | – |
| Bedbound | 10 (2) | 87 (72–94) | 13 (1–40) | 2 (20) |
| C (8) | 280 | 77 (38–99) | 12(1–140) | – |
| Bedbound | 50 (14) | 82 (38–99) | 11 (1–140) | 0 (-) |
| D (2) | 52 | 85 (51–96) | 11 (1–65) | – |
| Bedbound | 6 (3) | 85 (72–91) | 14 (4–48) | 1 (16.7) |
| E (2) | 18 | 84 (66–97) | 15 (1–77) | – |
| Bedbound | 4 (3) | 74 (66–83)* | 16 (2–33) | 1 (25) |
We examined further the admission diagnoses and comorbidities of the bedbound patient group (n=138). Admitting diagnoses varied widely. Over half (72 patients; 52%), were admitted with an infection as the primary diagnosis. A further 18 patients (13%) were admitted with a diagnosis where fluid provision would be the primary treatment (AKI, hypercalcaemia or dehydration with reduced oral intake).
Some 45/138 bedbound patients (33%) had a prior diagnosis of dementia. Outside of these, 25/93 patients (27%) presented with a reduced Abbreviated Mental Test Score (AMTS), with 18 patients presenting with a marked reduction in AMTS (less than 6/10) (Hodkinson, 1972).
A flowchart of the patient population displaying the provision of water at the bedside is shown in Figure 1. Eight (6%) of the bedbound patients were not provided with any means of hydration on their table, although the table was within reach. Seven of these patients were classed as being unable to swallow safely and required nasogastric (NG) or percutaneous endoscopic gastrostomy (PEG) feeding; the remaining patient was nil by mouth awaiting radiology investigations.
Among the 138 bedbound patients there was adequate water provided and within reach in 86 cases (62%). Some 34 patients (25%) had a contraindication to water at the bedside (8 were not provided with water, and in 26 cases the water jug was out of reach of the patient).
The remaining 18 patients would have been able to drink for themselves, with no contraindication for free oral fluid intake; however, the bedside table was not within reach of the patient. The reasons for this were uniformly that it had been moved by a member of staff (doctors, nurses and physiotherapists) when performing their role. Two patients within this group had visual impairment and were not aware of where their water had been center—on further questioning neither patient had known where it had been center all day or the day before, and were only drinking when prompted. However, both were able to reach and drink for themselves when the table was moved closer to the bedside.
There was no significant difference in the mean age or length of stay of the group for whom water was out of reach. We performed univariate analysis using factors associated with an inadequate provision of water at the bedside using Fisher's exact test and an associated odds ratio and confidence interval (Table 2): age (over 75 years), gender and prolonged stay (more than 25 days) were not significantly associated with inadequate water provision. Severe cognitive impairment (AMTS<6) was significantly more common in patients who had inadequate water provision (OR5.5[1.7-17.8], P=0.008).
| Variable | OR | 95% CI | P (Fisher's exact test) |
|---|---|---|---|
| Age ≥75 years | 1.32 | 0.40-4.32 | 0.78 |
| Male gender | 0.72 | 0.27-1.97 | 0.605 |
| Length of stay ≥25 days | 1.09 | 0.33 – 3.59 | 1.00 |
| AMTS <6 | 5.5 | 1.70-17.80 | 0.008 |
OR=odds ratio; CI=confidence interval; AMTS=Abbreviated Mental Test Score
When assessing the effect of admitting centre (numbers shown in Table 1), owing to small numbers in a number of categories the authors were unable to perform appropriate statistical modelling using the chi-square test; because one centre had 0 cases reported, we were unable to include influence of centre on multivariate modelling.
The authors performed a multivariate regression analysis for factors associated with inadequate water provision, demonstrating that age, gender and length of stay were non-significant variables. Severe cognitive impairment (AMTS<6) was found to continue to be significantly associated with a lack of provision of water at the bedside (OR 4.49 [1.38-14.63], P=0.013).
Discussion
To the authors' knowledge, this study represents the first of its kind to demonstrate provision of access to water at the bedside across several UK NHS centres. The study shows that 13% of bedbound patients were unable to reach a drink at the side of their bed. Reasons for this included the bedside table having been moved away from the bedside, and the water jug/cup being placed too far from the reach of the patient one it had been refilled. This demonstrates a shortfall in essential care provision to some of the most at-risk hospital inpatients in one in eight cases.
It was noted that there was ‘appropriate non-provision’ of water in patients who were at risk of aspiration, although it could be argued that cognitively normal individuals ought to make their own decisions about this and might be provided with water. The question of ‘at-risk’ feeding is controversial and a balanced commentary published earlier this year scrutinises much of the common dogma adopted by many health professionals around the area (Murray et al, 2019); the authors advocate an individualised decision-making process regarding enteral feeding in those deemed at-risk of aspiration, including the use of thickened fluids; and they conclude that branding a patient absolutely ‘nil by mouth’ is something that ought to be avoided where possible.
Some of the results within this study are not altogether surprising, with the bedbound inpatient population overall being older and with a statistically longer hospital stay being consistent with national data (HSCIC, 2015). The adequacy of bedside water provision was not affected by age, gender or length of stay, but seemed to be significantly worse in those patients with cognitive impairment. Specific to hospitalised patients with reduced cognition, inadequate hydration has been seen to be a significant predisposing factor to delirium episodes (Collier, 2012; Schofield et al, 2012), defining this patient group as especially vulnerable to the consequences of dehydration. The findings of this study mirror those identified in similar assessments in an institutionalised population of patients within the UK, demonstrating links between cognitive impairment and inadequate hydration (Hooper et al, 2016).
Investigating the impact of the lack of water at the bedside in these specific patients was not within the remit of this study. However, approximately two-thirds (90/138) of the bedbound population are affected by conditions such as infection and kidney injury, where adequate fluid intake is a mainstay of treatment. The authors were also unable to assess fluid intake per patient by way of studying fluid balance charts within the limitations of this study. It is an important point to bear in mind that the provision of fluids and nutrition to patients is not only a cornerstone of care but a legal duty of the profession (Care Quality Commission (CQC), 2013).
Publications from the CQC and the PHSO have stated that there is room to improve the basic provision of hydration to hospital patients (CQC, 2013; PHSO, 2011). Clearly, adequate enteral nutrition and hydration may also reduce the need for intravenous fluid therapy (Brotherton et al, 2010), which is known to be a challenge within the frail patient populations we have described here. Shells and Morrell-Scott (2018) discussed the use of bedside water devices, which were not used in any centre in this study but would have been a potential solution to the issues seen here. It is worth noting that simply having a jug and glass within reach may underestimate the ability of a frail patient to access drinking water, as noted in other studies where many frail or immobile patients lacked the strength or sufficient coordination to drink without some assistance (Gaff et al, 2015; Johnstone et al, 2015).
Limitations
As a single episode, point-prevalent study this audit provides a snapshot of day-to-day care within the NHS. However, wider conclusions should be drawn with caution owing to the small patient cohort and limited number of hospitals included. However, the finding of inadequate provision across multiple centres points to a system-wide failure to meet a gold standard, as opposed to an institutional one. Patient data were gathered by observation of the patient environment and contemporaneous notes reviewed by the study team. There is a possibility for uncontrolled inter-observer error in the inclusion of diagnoses within the study data because patient notes may not have been explored to the same extent and historic and acute diagnoses may not have been combined. There was heterogeneity among sample sizes contributed by each centre; this may mask potential differences in performance between centres because comparative statistics were not appropriate given small group numbers. Expectation bias is possible, but in all centres the audit was approved but not publicised to ward staff to accurately depict current practice.
Conclusion
This study demonstrates that there is a potential lack of attention to ensuring that patients can reach a means of hydration across several NHS trusts. Previous authors have discussed the practical reasons for a ‘less-than-perfect’ delivery of care in this respect (Bowman and Meyer, 2014). However the authors consider that something as simple as moving a bedside table back with reach of the patient should be in the forefront of the mind of every health professional. Responsibility for this demonstrated shortfall, and the necessary changes in culture and attitude to correct it, needs to be shouldered by all members of the multidisciplinary team.