Abstract
We report incidence rates for pressure injuries seen in an acute hospital in Singapore that were classified as Stage 3 or Stage 4. The characteristics of patients and the factors that explain variation in the primary outcome of duration of hospital stay are summarized. Existing data were available from Singapore General Hospital for all admissions from January 2016 to December 2019. Univariable analysis was done and a multivariable Poisson regression model estimated. Incidence rates declined from 4.05 to 3.4 per 1000 admissions in the 48 months between 2016 and 2019. The vast majority were community acquired with 75% in admission from the patients' home. Factors that explain variation in length of stay were, ethnicity; site of injury; community versus healthcare associated; inter‐hospital transfer; fracture as reason for admission; and the number of days between admission and assessment of wound by specialist nurse. Stage 3 and 4 injuries arise in a home environment most often and are subsequently managed in acute hospital at high cost. These are novel epidemiological data from a hospital in the tropics where the potential to improve outcomes, implement screening and prevention, and thus increase the performance of health services is strong.
Keywords: hospitals, general; incidence; length of stay; pressure ulcer
1. INTRODUCTION
Pressure injuries are also known as pressure ulcers, decubitus ulcers, and bed sores. They arise from a failure of skin integrity because of unrelieved pressure such as from a bony area contacting external surfaces. The normal structure and function of the skin and soft tissue is inhibited and other causal factors are related to the patient and their living environment. 1 The elderly and those confined to bed are vulnerable and risks are augmented by skin fragility, poor blood supply, nutritional insufficiency, and moisture from incontinence. 2
Pressure injuries are classified between Stage 1 where the skin is intact but could be painful and might present as redness only and Stage 4 where the injury is open and deep, reaching into muscle and bone and causing extensive damage. 3 Serious complications such as infection of the bone or blood might occur if Stage 3 or 4 pressure injuries are not treated properly. Healing these wounds varies with frailty, infection, chronic diseases, deficiency in vitamin C, medications like steroids, and low perfusion of oxygen and blood flow to the wound in cases of hypoxia and cold temperature. Qualitative researchers have revealed contextual barriers and enablers among health care professionals that impact on the effectiveness of pressure injury management in clinical practice. 4
Pressure injuries often arise among residents of aged care facilities and people living at home. Some are “healthcare‐acquired” as they might arise during an admission to an acute hospital. 5 The trend is for these to be classified as “never events” under initiatives to improve the quality of care and in some jurisdictions hospitals are not re‐imbursed for the costs of treating them. 6
Published studies of the epidemiology of pressure injuries have been reviewed. 7 Point prevalence's are between 1.1% and 26.7% for the acute care setting, 6% and 29% for community settings, 7.6% to 53.2% for residential aged care facilities and between 13.1% and 28.7% for intensive care units. Published incidence rates of pressure injuries have been summarised in a recent review. 8 Among general medical patients in acute hospitals incidences varied from 6.5% to 16.6%. Among surgical patients, the incidence was between 11.1% and 11.8%. One study reported the incidence among intensive care unit (ICU) patients to be 30.4% over 12 months. 9 Incidences of pressure ulcer varied considerably by stage. A United States study reported an incidence of 21.5% for all stages of pressure injury and only 2% for stages 2 to 4 in the hospital setting. 10 An Australian study reported a much lower incidence of 6.5% for all stages of pressure ulcer with a 2% incidence for stage 2 to 4 in the hospital setting. 11
Pressure injuries impose substantial and often under recognised economic costs to individuals, providers of health care services and the wider economy. An Australian study reported the costs of pressure injuries to acute hospitals to be $USD1.64 billion annually. 12 This figure excludes the costs incurred in residential care settings and the individual's private costs. Comparable high costs have been reported in the United Kingdom (UK) health system 13 , 14 and it was estimated that 5.5% of total health expenditure in Wales is now spent on chronic wounds 15 with pressure injuries likely to make up a good proportion. Chronic wounds reduce the quality of life and working capacity and increase social isolation. 16 , 17 , 18 With aging populations and a greater burden from chronic diseases, it is reasonable to assume that pressure injuries and the associated negative health and economic outcomes will increase in the future unless interventions are successfully implemented.
Evidence exists for the prevention of pressure injury among hospital patients 19 , 20 , 21 and evidence is emerging for prevention programs in the community setting. 22 , 23 , 24 There seems to be insufficient attention paid to the sensible economic hypothesis that screening for early signs of pressure injury prior to a problem arising that warrants an expensive hospital admission would be a good policy. 25 Early identification and appropriate prevention could save large costs and deliver substantial health and economic benefits to high‐risk populations.
The aims of this paper are to report the incidence rates of both healthcare‐acquired and community‐acquired pressure injuries seen in an acute care hospital in Singapore that were classified as Stage 3 or Stage 4, to summarise the characteristics of the patients affected and to quantify patient and treatment factors that contribute to the variation in the lengths of hospital stay incurred by the acute hospital sector. Prior to the design of good primary prevention programmes, the scale and scope of the problem in the acute care setting should be understood. This finding from this analysis will show the potential value or cost savings from prevention programmes that reduce the numbers of pressure injuries that are managed in the acute sector.
2. METHODS
The setting is Singapore General Hospital (SGH) a 1785 bed tertiary care teaching hospital. It provides 19% of all inpatient public beds in Singapore 26 and the staff deliver a comprehensive range of medical services including cardiothoracic surgery, bone marrow transplantation and solid‐organ transplantation, and it is the regional referral centre for management of major burns. The hospital employs over 9800 clinical, research and support staff. In 2019, the hospital discharged a total of 80 817 patients, had an Emergency Department attendance of 128 660 patients, and a Specialist Outpatient Clinic attendance of 724 480 patients.
The information used for this analysis is from a risk management system at SGH for pressure injuries that occurred from hospital or community sources in the 48 months between Jan 2016 to Dec 2019. Values were available for variables that describe the demographics of the patients, whether the injury was community or healthcare‐acquired, admitting clinical specialty, dates of important events such as hospital admission and discharge and staging of the wound, features of the wound such as stage and dimensions, concurrent patient risk factors for pressure injury, and interventions during the admission to treat the wound. Braden scores were available for each patient, which is a summary of the risks of pressure injury. 27 Statistical analysis was performed using open‐source software R, version 3.4.4.
Monthly incidence rates were calculated over 4 years for all pressure injuries seen in the hospital and then for those designated “community acquired” and “healthcare acquired.” The method was to divide the number of pressure injuries staged during a month by the total admissions to the corresponding hospital wards for that month. Healthcare‐acquired pressure injuries were defined as those not present on admission from either the patients home, a nursing home, or an inter‐hospital transfer. Yearly incidence rates were estimated by dividing the number of pressure injuries staged during a year by the total admissions to the corresponding hospital wards for that year.
The outcome of interest was the length of hospital stay, which was counted as the number of days between the admission date and the discharge date. Patients who were admitted from an aged care facility who had stays greater than 30 days were censored as it was thought likely they were waiting for placement in a community facility and any stay >30 days was not primarily because of a pressure injury.
All continuous variables were summarised by their respective mean and standard deviation and categorical variables were presented by the counts and percentages. Because the length of hospital stay for an individual is a count outcome and the distribution is positively skewed, we employed a generalised linear model with Poisson regression to assess the effect of baseline variables. All variables used in this analysis are summarised in Appendix A.
A multivariable Poisson regression model was specified from variable selection process of fitting individual simple linear regression model on each baseline covariate and reported P‐value. The unknown regression parameters in Poisson regression were estimated using maximum likelihood estimation readily available in glm() function in R software. Rather than interpreting the Poisson regression results through the regression coefficients, we reported adjusted rate ratios (RRs) for each individual covariate with their respective 95% confidence intervals (CIs). The definition of RR for a binary covariate X is defined as the ratio of the expectation of Y given Xi = 1 and the expectation of Y given Xi = 0, given that the other covariates in the model are fixed; whereas for a numerical covariate X it is the ratio of the expectation of Y given Xi = x + 1 and the expectation of Y given Xi = x for any arbitrary x. This is a similar idea to reporting odds ratio in logistic regression. For variable selection, we used a 25% alpha level. Otherwise, we used the standard 5% alpha level as statistical significance throughout reported 95% CI of various effects.
3. RESULTS
Figure 1 is used to summarise incidence rates by source for Stage 3 and 4 pressure ulcers per 1000 admissions to hospital for the 48 months between January 2016 and December 2019. The monthly rates are the sharply fluctuating points, the relatively steady blue line is the trend over time and the 95% confidence interval is shown with grey shading. The mean incidence rates for each year by the source of pressure injury are shown in Table 1.
FIGURE 1.
Monthly incidence rates for all pressure injuries, then by community acquired and healthcare acquired
TABLE 1.
Means incidence rates for Stage 3 and 4 pressure injuries by source per 1000 admissions with 95% intervals
| 2016 | 2017 | 2018 | 2019 | |
|---|---|---|---|---|
| All PI | 4.05 (3.61, 4.49) | 3.75 (3.33, 4.17) | 3.32 (2.92, 3.71) | 3.40 (3.00, 3.80) |
| Community | 3.97 (3.53, 4.40) | 3.61 (3.20, 4.02) | 3.16 (2.77, 3.54) | 3.15 (2.77, 3.53) |
| Healthcare | 0.087 (0.022, 0.15) | 0.14 (0.056, 0.22) | 0.16 (0.073, 0.25) | 0.25 (0.15, 0.36) |
Out of the 1143 patients in the dataset with Stage 3 or 4 Pressure injury, 1130 or 98.86% were community acquired and only 13 or 1.13% were healthcare acquired. The source of admission for these patients is shown in Table 2.
TABLE 2.
Sources of Pressure Injury for those that were community acquired, number of cases, and percentage
| Community acquired (%) | Healthcare acquired (%) | |
|---|---|---|
| Admitted from home | 867 (77) | 11 (85) |
| From inter hospital transfer | 89 (8) | 1 (7.7) |
| Nursing home | 174 (15) | 1 (7.7) |
| Total | 1130 (100) | 13 (100) |
Very few injuries were healthcare acquired suggesting Stage 1 and 2 ulcers are managed appropriately in the acute care setting preventing progression to more serious Stage 3 and 4 injuries. From those that arose in the community, the majority started in the home environment. Inter‐hospital transfers represent patients who need specialist care in a tertiary centre and we do not know whether these were originally community acquired or health care acquired. Nursing homes—at 15%—represent a modest but important source of Stage 3 and 4 injuries. The characteristics of the in‐patients with Stage 3 and Stage 4 are summarised in Tables 3 and 4 respectively.
TABLE 3.
Descriptive statistics for Stage 3 patients (n = 661)
| All PI (661) | Community acquired (619) | Healthcare acquired (42) | |
|---|---|---|---|
| Age (mean ± SD) | 73.30 ± 14.64 | 73.42 ± 14.64 | 71.48 ± 14.78 |
| Gender (n (%)) | |||
| Male | 320 (46.99) | 296 (46.32) | 24 (57.14) |
| Female | 361 (53.01) | 343 (53.68) | 18 (42.86) |
| Race (n (%)) | |||
| Chinese | 542 (79.01) | 507 (78.73) | 35 (83.33) |
| Indian | 48 (7.00) | 47 (7.30) | 1 (2.38) |
| Malay | 66 (9.62) | 61 (9.47) | 5 (11.90) |
| Others | 30 (4.37) | 29 (4.50) | 1 (2.38) |
| Braden Score | |||
| No risk | 11 (1.60) | 11 (1.71) | 0 (0) |
| At risk | 155 (22.59) | 147 (22.83) | 8 (19.05) |
| Moderate risk | 143 (20.85) | 132 (20.50) | 11 (26.19) |
| High risk | 287 (41.84) | 268 (41.61) | 19 (45.24) |
| Very high risk | 90 (13.12) | 86 (13.35) | 4 (9.52) |
| Source of injury (n (%)) | |||
| From home | 508 (78.15) | 503 (78.11) | 5 (83.33) |
| Inter‐Hospital transfer | 49 (7.54) | 48 (7.45) | 1 (16.67) |
| Nursing home | 93 (14.31) | 93 (14.44) | 0 (0) |
| Length of stay, censored (mean [SD]) | 15.39 (32.10) | 13.72 (30.78) | 41.45 (40.67) |
TABLE 4.
Descriptive statistics for Stage 4 patients (n = 474)
| All PI (474) | Community Acquired (464) | Healthcare Acquired (10) | |
|---|---|---|---|
| Age (mean ± SD) | 71.28 ± 15.66 | 71.21 + 15.66 | 75.10 ± 15.66 |
| Gender (n (%)) | |||
| Male | 235 (48.35) | 231 (48.53) | 4 (40) |
| Female | 251 (51.65) | 245 (51.47) | 6 (60) |
| Race (n (%)) | |||
| Chinese | 396 (79.84) | 388 (79.84) | 8 (80) |
| Indian | 28 (5.65) | 28 (5.76) | 0 (0) |
| Malay | 53 (10.69) | 52 (10.70) | 1 (10) |
| Others | 19 (3.83) | 18 (3.70) | 1 (10) |
| Braden score | |||
| No risk | 3 (0.60) | 3 (0.62) | 0 (0) |
| At risk | 61 (12.30) | 59 (12.14) | 2 (20) |
| Moderate risk | 90 (18.15) | 89 (18.31) | 1 (10) |
| High risk | 240 (48.39) | 236 (48.56) | 4 (40) |
| Very high risk | 102 (20.56) | 99 (20.37) | 3 (30) |
| Source of injury (n (%)) | |||
| From home | 370 (75.05) | 364 (74.90) | 6 (85.71) |
| Inter‐hospital transfer | 41 (8.32) | 41 (8.44) | 0 (0) |
| Nursing home | 82 (16.63) | 81 (16.67) | 1 (14.29) |
| Length of stay, censored (mean [SD]) | 12.82 (21.52) | 12.81 (21.65) | 12.22 (12.49) |
These findings emerged from censored data on 1006 patients after 137 cases were removed because of stays >30 days in the hospital. Age, gender, race, and source of injury are similar for patients with Stage 3 and 4 injuries. As we might expect, “Braden scores” are elevated for patients with Stage 4 as compared with Stage 3. Mean lengths of stay were longest for healthcare‐acquired Stage 3 pressure injuries, but this statistic arises from very few data points. All patients with Stage 3 had longer stays compared with Stage 4.
The findings from univariable analyses of the relationship between candidate variables and the length of stay outcome are shown in Appendix A. A good number of these were found to be individually associated with variation in length of stay. Following variable selection, we implemented a multivariable Poisson regression model. Only 307 observations out of 1006 observations had no missing values and so we employed the “mice” package in R to impute observations. The statistical inference in terms of 95% CI of relative risks were modified using an appropriate method. 28 , 29 The findings from the multivariable statistical model based on imputed data are shown in Table 5.
TABLE 5.
Regression results based on Poisson regression with missing under consideration
| Variable | Regression parameter estimate (SE) | Rate ratio (95% CI) | P‐value |
|---|---|---|---|
| Race (Chinese is reference category) | |||
| Indian | −0.016 (0.046) | 0.984 (0.894, 1.073) | .726 |
| Malay | 0.054 (0.038) | 1.056 (0.978, 1.133) | .150 |
| Others | 0.220 (0.053) | 1.246 (1.117, 1.374) | <.05* |
| PI site (Buttock is reference category) | |||
| Ear | 0.092 (0.129) | 1.096 (0.820, 1.373) | .478 |
| Elbow | 0.472 (0.094) | 1.603 (1.309, 1.898) | <.05* |
| Heel | 0.035 (0.063) | 1.035 (0.907, 1.163) | .583 |
| Iliac crest | −0.076 (0.185) | 0.927 (0.591, 1.263) | .683 |
| Knee | −0.332 (0.144) | 0.717 (0.515, 0.919) | .022* |
| Malleolus | 0.181 (0.060) | 1.199 (1.058, 1.340) | .003* |
| Others | 0.078 (0.050) | 1.082 (0.975, 1.188) | .121 |
| Sacrum/Coccyx | 0.031 (0.035) | 1.032 (0.961, 1.102) | .369 |
| Scapula | −0.037 (0.136) | 0.963 (0.706, 1.221) | .785 |
| Spine | 0.120 (0.110) | 1.128 (0.884, 1.372) | .275 |
| Toe | −0.240 (0.156) | 0.787 (0.546, 1.028) | .125 |
| Trochanter | 0.011 (0.052) | 1.011 (0.907, 1.114) | .837 |
| Community acquired injury (yes) | −0.247 (0.087) | 0.781 (0.648, 0.914) | .008* |
| Braden score (6‐24) | 0.004 (0.005) | 1.004 (0.995, 1.013) | .355 |
| Admission source (home is reference category) | |||
| Inter‐hospital transfer | 0.296 (0.045) | 1.345 (1.227, 1.462) | <.05* |
| Nursing home | −0.029 (0.035) | 0.972 (0.905, 1.038) | .409 |
| Admitted for fracture (yes) | 0.325 (0.355) | 1.384 (0.421, 2.346) | .375 |
| Admitted for fracture with co‐morbidities (yes) | 0.339 (0.185) | 1.404 (0.895, 1.914) | .072 |
| Surgery performed during hospitalisation (admit for fracture) | 0.867 (0.233) | 2.381 (1.293, 3.468) | <.05* |
| Surgery performed during hospitalisation + Co‐morbidities | 0.583 (0.233) | 1.792 (0.974, 2.610) | .022* |
| Days between admission and air mattress | −0.001 (0.014) | 0.999 (0.973, 1.026) | .962 |
| Days between admission and wound nurse assessment | 0.106 (0.032) | 1.112 (1.042, 1.182) | .008* |
P‐Value < = .05.
Many of the explanatory factors cannot be modified. The race classification of “others” shows as compared with “Chinese” the mean length of stay is associated with a 24.6% increase in length of stay. The site of the injury at the “Elbow” is associated with increasing stay by 60.3% as compared with “Buttock”, while the associated decreased stay is 21.3% for “toe” and 28.3% for “knee.” It is plausible that for patients who are unable to lift their elbow then even 2‐hourly turning does not relieve pressure. Whether the injury was community acquired versus healthcare acquired is associated with 21.9% reduction in length of stay. Surgery for fracture when combined with comorbidities is associated with increased lengths of stay. A modifiable factor is time to nurse wound assessment. For every one‐day increase in the days between admission and nurse wound assessment, length of stay is 11.2% longer.
4. DISCUSSION
Little is known about the epidemiology, economic costs, and effectiveness of prevention programs in Singapore and other countries in the region. A survey of pressure ulcers was been conducted in a tertiary hospital in Singapore in 2005 and the prevalence of pressure ulcers was 18.1% and the incidence was 8.1%. 30 A realist case study conducted in Singapore 31 found pressure injury to be a nurse‐sensitive quality indicator in hospitals. The prevention and management of pressure injury was suboptimal, despite the availability of interventions. While nurses strived to improve patient safety, important deficits were identified relating to contextual factors and the mechanisms of the prevention. On a positive note, an intervention for an orthopaedic ward in Singapore to prevent heel pressure ulcers among adult patients increased compliance with heel off‐loading techniques to 93.3% and resulted in more than 50% reduction in stage one heel pressure ulcers. 25
This analysis has revealed that serious pressure inquiries at Stage 3 and 4 are frequently occurring among all admissions to a large tertiary hospital in Singapore. They arise among 3 to 4 patients per 1000 admissions and shows hospital stays between 12 to 15 days. We do not know whether the pressure injury was the main reason for the admission or other factors played a role. However, given the average length of stay in Singapore hospitals for a cohort of frequent attenders was 4.64 days 32 there is some evidence that pressure injury does indeed use up valuable bed days. Most start at home and are present on admission suggesting an elderly and possibly vulnerable population of individuals are unable to manage the risks of this preventable event. Around 15% of community‐acquired injuries come from nursing homes where strategies are in place to reduce risks. It is encouraging to see that very few are healthcare acquired, around 1 per 10 000 in our data. Stage one and two injuries are common in hospital but it seems progression to the more serious stage 3 and 4 is very rare. This is likely because of the emphasis placed on this as a quality of care indicator and the effective prevention programmes now embedded in acute services.
The analyses of factors that explain variation in length of stay outcomes was unremarkable with many factors unmodifiable. These findings might be used to highlight patients for whom more stringent nursing care or mandatory nurse wound assessment is warranted. The time interval between admission and nurse wound assessment is modifiable, and for each extra day's delay, the length of stay was associated with a 11.2% increase. One interpretation of this finding is that wound assessment on the first day of the admission could change the care plan such that length of stay is shortened. This simple practice change might well pay for itself many times over in saved costs.
There are limitations to this study. The data were retrospective and not collected specifically to address the questions presented here. Detailed information was only available once the patient entered into the acute care setting. It will be important to understand the incidences, risk factors, and outcomes for patients in care homes and other places who may well bounce around primary care and community care and never be admitted to hospital or a hospital admission might only represent a small part of their experience. Good policy for prevention will likely require information for the whole system and not just the acute hospital setting. The Braden score is established by imperfect measure of risk and will likely be surpassed by more sensitive and specific tools in the future. 33 We only reported in Stage 3 and 4 ulcers as data were not complete for the entire time series for Stage 1 and 2 injuries. Regardless of this, the impact on health outcomes and length of stay are likely to be far less for the less serious injuries. The decision to censor the length of stay variable was based on tacit knowledge and theory. The uncensored outcome variable was very long tailed with a maximum length of stay of >600 days, see Appendix B. In this case, the mean is not always a useful summary measure as it is pulled upwards by a small number of observations. 34 The effect of censoring on the distribution of length of stay outcomes is shown in Appendix C.
Maybe the most important point to discuss from these findings is the potential to intervene to reduce risks of pressure injury arising in the home. There is a well‐developed literature about the effectiveness of programs delivered in residential care settings, including care bundles 35 that aimed to change the behaviour of healthcare workers. And there are applications of these interventions in real world settings. 36 Yet there is very little research about interventions that aim to reduce risks among individuals living in their own home. The presence of a serious pressure injury for someone living at home might be an indicator they can no longer care for themselves and so need to consider a move to residential care. Alternatively, the caregiver might not be knowledgeable or able to prevent a pressure injury at home. Screening tools might be relevant for primary care professionals to identify those at high risk of pressure injury. The information shown in Table 5 suggest these individuals are not Chinese, Indian or Malay, rather they will be Caucasian, Eurasian, and Arabic, they are inter‐hospital transfers and were admitted for fracture related issues and surgery.
5. CONCLUSION
Pressure injuries are a common and challenging clinical problem. Most serious injuries arise in the home and contribute to long stays in the acute care setting. If reducing these events is a priority then processes and interventions to evaluate their effectiveness, cost‐effectiveness, and fidelity are required. These must be however designed within the wider context of the health systems to counter the issue of fragmentation of services between hospitals, care homes, and home care.
ACKNOWLEDGEMENTS
We acknowledge the work of the nursing team who collect and manage the data reported here: Ivy Goh Hui Qi; Nur Liyana Agus; Ng Xin Ping; Lee Jia Hui; Raden Nurheryany Sunari. This research was supported by the Agency for Science, Technology and Research (A*STAR) under its Industry Alignment Fund—Pre‐Positioning Programme (IAF‐PP) grant number H19/01/a0/CC9 and H19/01/a0/0Y9 as part of Wound Care Innovation for the Tropics (WCIT) Programme.
Appendix A. INDIVIDUAL VARIABLES AND THEIR RESPECTIVE P‐VALUES
| Variables | Min P‐value (<.25) |
|---|---|
| Demographic | |
| Age | .870 |
| Gender | .793 |
| Race | .052* |
| Pressure Injury | |
| Pressure Injury site | .007* |
| Ulcer present on admission(y/n) | <.05* |
| Braden score | .239* |
| Source of ulcer, if present on admission | <.05* |
| Clinical co‐information | |
| Stroke | .690 |
| Diabetes | .690 |
| Vascular | .690 |
| Incontinence | .303 |
| Admitted for fracture | .184* |
| Others (other predisposing factors) | .693 |
| Admitted for fracture (only) | .184* |
| Admitted for fracture with incontinence | .850 |
| Admitted for fracture with co‐morbidities | .094* |
| Admitted for fracture with incontinence and co‐morbidities | .915 |
| Admitted for fracture with Incontinence and others | .830 |
| Incontinence (only) | .594 |
| Incontinence, others | .398 |
| Co‐morbidities | .463 |
| Co‐morbidities, Incontinence | .303 |
| Co‐morbidities, Incontinence, others | .607 |
| Co‐morbidities, others | .521 |
| Surgery performed during hospitalisation | .683 |
| Surgery performed during hospitalisation, admission for fracture | .021* |
| Surgery performed during hospitalisation, co‐morbidities | .980 |
| Surgery performed during hospitalisation, co‐morbidities, incontinence | .004* |
| Surgery performed during hospitalisation, co‐morbidities, others | .362* |
| Interventions | |
| Duration between admission date and Date of 1st documentation | .770 |
| Duration between admission date and Date of air mattress | <.05* |
| Duration between admission date and Date referred to wound nurse | .001* |
P‐Value < .25.
Appendix B. UNCENSORED LENGTH OF STAY OUTCOME VARIABLE
Appendix C. CENSORED LENGTH OF STAY OUTCOME VARIABLE AT 30 DAYS FOR ADMISSIONS FROM NON‐HOME SETTING
Graves N, Maiti R, Aloweni FAB, Yuh AS, Lo ZJ, Harding K. Pressure injuries among admissions to a hospital in the tropics. Int Wound J. 2020;17:1659–1668. 10.1111/iwj.13448
Funding information Agency for Science, Technology and Research, Grant/Award Numbers: H19/01/a0/0Y9, H19/01/a0/CC9
REFERENCES
- 1. Al Aboud AM, Manna B. Wound Pressure Injury Management. Wound Pressure Injury Management [Updated 2020 Feb 14] In: StatPearls [Internet]. 2020. January.
- 2. Coleman S, Gorecki C, Nelson EA, et al. Patient risk factors for pressure ulcer development: systematic review. Int J Nurs Stud. 2013;50(7):974‐1003. [DOI] [PubMed] [Google Scholar]
- 3. Australian Wound Management Association . Clinical Practice Guidelines for the Prediction and Prevention of Pressure Ulcers. West Leederville, Perth, Australia: Cambridge Publishing; 2001. [Google Scholar]
- 4. Coyer F, Cook JL, Doubrovsky A, Campbell J, Vann A, McNamara G. Understanding contextual barriers and enablers to pressure injury prevention practice in an Australian intensive care unit: an exploratory study. Aust Crit Care. 2019;32(2):122‐130. [DOI] [PubMed] [Google Scholar]
- 5. Graves N, Birrell F, Whitby M. Effect of pressure ulcers on length of hospital stay. Infect Control Hosp Epidemiol. 2005;26(3):293‐297. [DOI] [PubMed] [Google Scholar]
- 6. Centers for Medicare and Medicaid Services (CMS), HHS . Medicare program: changes to the hospital inpatient prospective payment systems and fiscal year 2009 rates; payments for graduate medical education in certain emergency situations; changes to disclosure of physician ownership in hospitals and physician self‐referral rules; updates to the long‐term care prospective payment system; updates to certain IPPS‐excluded hospitals; and collection of information regarding financial relationships between hospitals. Final rules. Fed Regist. 2008;73(161):48433‐49084. [PubMed] [Google Scholar]
- 7. Graves N, Zheng H. The prevalence and incidence of chronic wounds: a literature review. Wound Pract Res. 2014;22(1):4. [Google Scholar]
- 8. McCosker L, Tulleners R, Cheng Q, et al. Chronic wounds in Australia: a systematic review of key epidemiological and clinical parameters. Int Wound J. 2019;16(1):84‐95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Coyer F, Gardner A, Doubrovsky A, et al. Reducing pressure injuries in critically ill patients by using a patient skin integrity care bundle (InSPiRE). Am J Crit Care. 2015;24(3):199‐209. [DOI] [PubMed] [Google Scholar]
- 10. Schultz A, Bien M, Dumond K, Brown K, Myers A. Etiology and incidence of pressure ulcers in surgical patients. AORN J. 1999;70(3):434, 437‐440, 443‐444, 446‐449. [DOI] [PubMed] [Google Scholar]
- 11. Charlier C. Prevalence, incidence and risk: a study of pressure ulcers at a rural base hospital. Prim Intent. 2001;9(1):12‐21. [Google Scholar]
- 12. Graves N, Zheng H. Modelling the direct health care costs of chronic wounds in Australia. Wound Pract Res. 2014;22(1):20. [Google Scholar]
- 13. Posnett J, Franks PJ. The burden of chronic wounds in the UK. Nurs Times. 2008;104:44‐45. [PubMed] [Google Scholar]
- 14. Posnett J, Franks PJ. Skin Breakdown: the Silent Epidemic. Hull, UK: The Smith and Nephew Foundation; 2007. [Google Scholar]
- 15. Phillips CJ, Humphreys I, Fletcher J, Harding K, Chamberlain G, Macey S. Estimating the costs associated with the management of patients with chronic wounds using linked routine data. Int Wound J. 2016;13(6):1193‐1197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Phillips T, Stanton B, Provan A, Lew R. A study of the impact of leg ulcers on quality of life: financial, social, and psychologic implications. J Am Acad Dermatol. 1994;31(1):49‐53. [DOI] [PubMed] [Google Scholar]
- 17. Cheng Q, Kularatna S, Lee XJ, Graves N, Pacella RE. Comparison of EQ‐5D‐5L and SPVU‐5D for measuring quality of life in patients with venous leg ulcers in an Australian setting. Qual Life Res. 2019;28(7):1903‐1911. [DOI] [PubMed] [Google Scholar]
- 18. Kapp S, Santamaria N. The financial and quality‐of‐life cost to patients living with a chronic wound in the community. Int Wound J. 2017;14(6):1108‐1119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Gaspar S, Peralta M, Marques A, Budri A, Gaspar de Matos M. Effectiveness on hospital‐acquired pressure ulcers prevention: a systematic review. Int Wound J. 2019;16(5):1087‐1102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Holte HH, Underland V, Hafstad E. Systematic reviews on preventing pressure ulcers 2016, Norwegian Institute of Public Health. https://www.fhi.no/en/publ/2016/pressure-ulcers/. Accessed March 16, 2020. [PubMed]
- 21. Reddy M, Gill SS, Rochon PA. Preventing pressure ulcers: a systematic review. JAMA. 2006;296(8):974‐984. [DOI] [PubMed] [Google Scholar]
- 22. Lavallee JF, Gray TA, Dumville J, Cullum N. Preventing pressure ulcers in nursing homes using a care bundle: a feasibility study. Health Soc Care Community. 2019;27(4):e417‐e427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Lavallee JF, Gray TA, Dumville J, Cullum N. Barriers and facilitators to preventing pressure ulcers in nursing home residents: a qualitative analysis informed by the theoretical domains framework. Int J Nurs Stud. 2018;82:79‐89. [DOI] [PubMed] [Google Scholar]
- 24. Tappen RM, Newman D, Huckfeldt P, et al. Evaluation of nursing facility resident safety during implementation of the INTERACT quality improvement program. J Am Med Dir Assoc. 2018;19(10):907‐913. e1. [DOI] [PubMed] [Google Scholar]
- 25. Norman RE, Gibb M, Dyer A, et al. Improved wound management at lower cost: a sensible goal for Australia. Int Wound J. 2016;13(3):303‐316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Ministry of Health . Beds in inpatient facilities and places in non‐residential long‐term care facilities. https://www.moh.gov.sg/resources‐statistics/singapore‐health‐facts/beds‐in‐inpatient‐facilities‐and‐places‐in‐non‐residential‐long‐term‐care‐facilities2020. Accessed March 16, 2020.
- 27. Braden BJ. The Braden scale for predicting pressure sore risk: reflections after 25 years. Adv Skin Wound Care. 2012;25(2):61. [DOI] [PubMed] [Google Scholar]
- 28. Schafer JL, Olsen MK. Multiple imputation for multivariate missing‐data problems: a data analyst's perspective. Multivariate Behav Res. 1998;33(4):545‐571. [DOI] [PubMed] [Google Scholar]
- 29. Rubin DB, Schenker N. Multiple imputation for interval estimation from simple random samples with ignorable nonresponse. J Am Stat Assoc. 1986;81:366‐374. [Google Scholar]
- 30. Chan EY, Tan SL, Lee CK, Lee JY. Prevalence, incidence and predictors of pressure ulcers in a tertiary hospital in Singapore. J Wound Care. 2005;14(8):383‐384, 386–388. [DOI] [PubMed] [Google Scholar]
- 31. Teo CSM, Claire CA, Lopez V, Shorey S. Pressure injury prevention and management practices among nurses: a realist case study. Int Wound J. 2019;16(1):153‐163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Low LL, Tay WY, Ng MJ, Tan SY, Liu N, Lee KH. Frequent hospital admissions in Singapore: clinical risk factors and impact of socioeconomic status. Singapore Med J. 2018;59(1):39‐43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Nixon J, Nelson EA, Rutherford C. Pressure UlceR Programme Of reSEarch (PURPOSE): using mixed methods (systematic reviews, prospective cohort, case study, consensus and psychometrics) to identify patient and organisational risk, develop a risk assessment tool and patient‐reported outcome Quality of Life and Health Utility measures. Southampton (UK):. NIHR Journals Library; 2015 September (Programme Grants for Applied Research, No 36) 2015. [PubMed]
- 34. Malehi AS, Pourmotahari F, Angali KA. Statistical models for the analysis of skewed healthcare cost data: a simulation study. Health Econ Rev. 2015;5:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Resar R, Griffin F, Haraden C. Using Care Bundles to Improve Health Care Quality. Cambridge, MA: Institute for Healthcare Improvement; 2012. [Google Scholar]
- 36. Lavallée JF, Gray TA, Dumville JC, Cullum N. Preventing pressure injury in nursing homes: developing a care bundle using the behaviour change wheel. BMJ Open. 2019;9(6):e026639. [DOI] [PMC free article] [PubMed] [Google Scholar]