Hospital-Acquired Pressure Injuries in Critical and Progressive Care: Avoidable Versus Unavoidable

This study used a descriptive, retrospective, comparative design to examine rates of avoidable and unavoidable HAPIs in adult critical and progressive care patients in 6 acute care hospitals within a large academic health care system in the midwestern United States. The critical care areas included surgical, trauma, cardiovascular surgical, cardiac, neurologic, and medical intensive care and corresponding progressive care units. The study was approved by the Indiana University institutional review board before the start of data collection.

An investigator-developed conceptual framework for differentiating unavoidable from avoidable pressure ulcers (Figure 1) was used to guide this study. The aspects of this model that are new and unique incorporate (1) new epidemiological evidence on pressure injury risk factors,^9,10  (2) risk-based prevention strategies consistent with the 2014 Pressure Ulcer International Guideline,¹¹  and (3) guidance for determining whether the pressure injury was avoidable or unavoidable based on the implementation of appropriate risk-based interventions.⁷ 

The sample consisted of patients who (1) had a HAPI develop while in critical and progressive care (NPUAP stage 2, 3, or 4, unstageable, or deep tissue pressure injury,¹²  including pressure injuries from medical devices and on mucous membranes), (2) were hospitalized between 2012 and 2015, and (3) were aged 18 years or older. Patients were excluded if the HAPI developed outside the critical or progressive care unit. We used the National Database of Nursing Quality Indicators methods for identifying unit-acquired pressure injuries.¹³  On the basis of a previous pilot study,⁷  we determined potential sample size by estimating the number of HAPIs occurring during a 12-month period and the proportion of those estimated as unavoidable. Approximately 160 patients with a HAPI were estimated to provide an adequate sample to detect meaningful information.

To achieve the first study aim, eligible patients were identified by using monthly pressure injury prevalence surveys, medical record reporting mechanisms, medical coding procedures, and/or quality reporting processes. Patients were divided into 2 groups (avoidable and unavoidable) using the PUPI tool.

The PUPI contains 13 items (Figure 2). If all items in the PUPI are answered “yes” (all interventions were appropriately performed and documented), the HAPI was identified as unavoidable.⁷ 

The Braden Inventory Worksheet, an investigator-developed data collection tool specific to our electronic medical record (EMR) documentation, was used to collect Braden Scale subscale and total scores¹⁴  for the 3 days before the first documentation of the HAPI and the interventions documented that correspond to each Braden Scale subscale (Figure 3). These data were then used to complete the PUPI.

Data were collected from the EMR upon admission and 3 days before first documentation of the HAPI, including (1) demographic and clinical information, (2) Braden Scale subscale and total scores, (3) additional risk factors identified as significant in epidemiological studies (eg, poor perfusion, body temperature), and (4) pressure injury prevention interventions. The types and number of comorbidities were determined. For a variable to be marked yes, it had to be documented on any of the 4 days. For example, if “chemically sedated (IV drip)” was marked no on the HAPI date, yes on 1 day before, yes on 2 days before, and yes on 3 days before, then it was counted as yes (Figure 4).

Descriptive statistics were used to summarize avoidable and unavoidable HAPIs. Risk factors were compared between the 2 groups (avoidable and unavoidable). Continuous variables were reported by using mean and SD and were compared between patient groups by using the Student t test. Categorical variables were reported as number (percentage) and compared between patient groups using the χ² test. Logistic regression analyses were used to describe factors associated with unavoidable HAPIs. A P value of .05 was considered to represent statistical significance.

A total of 165 participants were included in this study. Participants’ mean (SD) age was 59.9 (16.4) years. The mortality rate was 27% (n = 45), and the mean (SD) length of stay (LOS) in the hospital before development of the HAPI was 15.1 (13.7) days (Table 1).

Almost 60% (n = 98) of the HAPIs were determined to be avoidable and 41% (n = 67) were determined to be unavoidable (Table 1). Most HAPIs were deep tissue pressure injuries (n = 102, 63%), followed by stage 2 (n = 34, 21%) and unstageable (n = 25, 15%). Approximately 36% (n = 60) of the HAPIs were related to medical devices. Most of the HAPIs were on the sacrum (n = 70, 42%) or heel (n = 23, 14%). Almost 79% (n = 130) of the participants with HAPIs were receiving mechanical ventilation, 56% (n = 92) were chemically sedated, 64% (n = 106) had systolic blood pressure (SBP) less than 90 mm Hg, 61% (n = 100) had mean arterial pressure less than 60 mm Hg, 41% (n = 68) were receiving 1 or more vasopressors, 55% (n = 91) were incontinent, and 21% (n = 34) had a bowel management system used at least once in the 3 days before and the day when the HAPI was first documented (Table 1).

In this study, the PUPI demonstrated fair inter-rater reliability (κ = 0.40), and raters were in total agreement 92.5% of the time (310 of 335) for the 13 PUPI items.

In comparison of clinical risk factors between groups, participants who had a comorbid disease of congestive heart failure (CHF) were less likely to have an unavoidable HAPI (odds ratio [OR], 0.22; 95% CI, 0.06-0.76; P = .02). Similar results were found for those who were chemically sedated (OR, 0.38; 95% CI, 0.20-0.72; P = .003), had SBP less than 90 mm Hg (OR, 0.52; 95% CI, 0.27-0.99; P = .047), and received at least 1 vasopressor (OR, 0.44; 95% CI, 0.23-0.86; P = .01). However, those who had a bowel management system were more likely to have an unavoidable HAPI than those who did not (OR, 2.19; 95% CI, 1.02-4.71; P = .04). Similarly, those who had a previous pressure injury were more likely to have an unavoidable HAPI, although this difference was not statistically significant (OR, 2.32; 95% CI, 0.84-6.44; P = .10).

The LOS before pressure injury identification was longer in the unavoidable HAPI group (17.6 vs 13.4 days). A 1-day increase in LOS before pressure injury identification was associated with an almost 2% increase in the odds of being an unavoidable HAPI. Participants in the unavoidable group consistently had higher daily Braden Scale total scores as well as individual subscale scores (mobility, activity, sensory perception, and nutrition), although the difference was statistically significant only for nutrition (P = .01). A 1-unit increase in nutrition score was associated with a 182% increase in the odds of being an unavoidable HAPI, a statistically significant difference (P = .01). When the number of preventive interventions was analyzed, patients with avoidable HAPIs had fewer preventive interventions implemented than did patients with unavoidable HAPIs. For all of the Braden Scale subscales except moisture, 1 more intervention was associated with a significant increase in the odds of being an unavoidable HAPI (mobility: OR, 4.02; 95% CI, 2.17-7.43; activity: OR, 4.12; 95% CI, 2.24-7.58; sensory perception: OR, 2.96; 95% CI, 1.55-5.64; nutrition: OR, 3.09; 95% CI, 1.75-5.47; friction and shear: OR, 1.70; 95% CI, 1.22-2.36; all P values < .01; Table 2).

Using multivariate logistic regression analysis controlling for clinical risk factors (LOS before pressure injury identification, comorbidities, daily Braden Scale total score, history of pressure injury, chemically sedated, CHF, SBP < 90 mm Hg, vasopressors, bowel management system, and smoking), participants who had CHF (OR, 0.028; 95% CI, 0.002-0.36; P = .006) were less likely to have an unavoidable HAPI, whereas those who had longer LOS before pressure injury development (OR, 1.04; 95% CI, 1.002-1.08; P = .04) or a history of pressure injury (OR, 5.27; 95% CI, 1.20-23.15; P = .03) were more likely to have an unavoidable HAPI. Specifically, for every additional day before pressure injury identification, there was a 4% increase in the likelihood of having an unavoidable HAPI, and participants with a history of pressure injury were 5 times more likely to have an unavoidable HAPI (Table 3).

An important finding of this study was the identification of 41% (n = 67) of the HAPIs as being unavoidable. Using a valid and reliable instrument (PUPI) provided an objective measure to identify unavoidable HAPIs. No other similar studies using such a tool were found in the literature.

In addition, this study is unique in being the first of its kind with the aim of objectively quantifying pressure injury prevention interventions in use before pressure injury development. The only other study we found that examined appropriate pressure injury preventive care was that of Beeckman and colleagues,¹⁵  a randomized controlled trial of 464 nursing home residents in 4 nursing homes in Belgium. In that study, the authors examined adherence to guideline-based pressure injury preventive care recommendations. Pressure injury preventive protocol was tailored to the resident and was described as skin observation, use of support surface, repositioning, and heel elevation. Preventive interventions were defined as either fully adequate, meaning that all were performed, or not. The results of the study were limited, as the researchers found that fully adequate preventive care was provided to the intervention group only when patients were seated; no improvements to preventive care were found while patients were in bed.¹⁵  Preventive interventions were not described in as much detail as in the present study. The main factor distinguishing avoidable from unavoidable HAPIs in the present study was the number of interventions documented. The unavoidable HAPI group had more interventions documented than the avoidable HAPI group; this finding makes sense given that the documented nursing care was deemed appropriate to the patient condition in the unavoidable HAPI group.

Another interesting finding of the present study was the high proportion (36%) of medical device–related HAPIs. This finding is most likely due to the nature of the study population and the large number of devices used in critical care areas. The finding is consistent with the results of Black and colleagues,¹⁶  who reported a HAPI rate of 5.4% (113 of 2079), with 34.5% (39 of 113) being related to medical devices.

In this study, unavoidable HAPIs were defined as those that developed in spite of consistent documentation of evidence-based preventive interventions. Thus, the question arises: Why did a HAPI develop in these individuals? The literature includes discussion of the complexity of pressure injury etiology and the potential for acute skin failure in critically ill patients^4-6,17-20 ; however, with our current level of evidence, the distinction between acute skin failure and unavoidable pressure injury remains obscure.

The NPUAP hosted a multidisciplinary conference in 2014 to explore the issue of pressure injury unavoidability using an organ system framework.⁵  Participants achieved consensus that unavoidable pressure injuries do indeed occur and that risk factors such as multiorgan dysfunction syndrome, shock or sepsis, hemodynamic instability and impaired tissue oxygenation, cardiac dysfunction, CHF, and skin failure are associated with pressure injury development and increase the likelihood of unavoidable pressure injury development.⁵  These factors have also been described as part of the concept of acute skin failure.^18-21 

Langemo and Brown conducted a systematic review and defined skin failure as “an event in which skin and underlying tissues die due to hypoperfusion concurrent with severe dysfunction or failure of other organ systems.”^18(p208) They reported that the term skin failure appeared in the literature as early as 1993 and was described as the damage that occurs in skin and underlying tissue at the end of life and in the intensive care setting.¹⁸  Langemo and Brown stated that the distinguishing factor between skin failure and a pressure injury is the coexistence of a significant disease process or organ failure. Their definition of acute skin failure is consistent with the overall characteristics of the patients in our study.

In a study of 552 critical care patients, Delmore and colleagues¹⁹  developed a predictive risk model for development of acute skin failure versus pressure injury in patients admitted to critical care units. Specifically, they identified 5 predictors of acute skin failure: peripheral artery disease, mechanical ventilation lasting more than 72 hours, respiratory failure, liver failure, and severe sepsis or septic shock.¹⁹  In a retrospective correlational study of 347 critical care patients, Cox²⁰  identified norepinephrine as a predictive risk factor for pressure injuries. Our study results are consistent with these findings, as almost 79% (n = 130) of the participants with HAPIs in our study were receiving mechanical ventilation, 56% (n = 92) were chemically sedated, 64% (n = 106) had SBP below 90 mm Hg, 61% (n = 100) had mean arterial pressure below 60 mm Hg, and 41% (n = 68) were receiving 1 or more vasopressors.

Advances in medicine have enabled critically ill patients to survive situations that in the past led to death. Now, increasing numbers of patients with multiorgan failure are surviving but remain highly susceptible to adverse events involving skin integrity, specifically skin failure. In a prospective descriptive study of 29 critical care patients with acute skin failure, Curry and colleagues²¹  reported that 5 patients had 2 organ failures at the time that skin failure was noted, 15 patients had either 3 organ failures or 2 organ failures plus sepsis, and 9 patients experienced 4 or more organ system failures and/or sepsis. All patients in that study had a diagnosis of nonskin organ failure and low albumin levels. Although the study of Curry et al²¹  provides additional information, it lacks an objective definition of acute skin failure and does not differentiate between pressure injuries and acute skin failure. The present study provides an objective, although retrospective, means to identify pressure injuries that are unavoidable. The results may suggest a process more aligned with acute skin failure than with pressure injury.

The present study’s findings regarding risk factors and avoidable HAPIs are difficult to explain, but they may be due to nurses’ perceptions of patients’ hemodynamic instability and thus their inability to implement preventive interventions, primarily repositioning. Patients with CHF, chemical sedation, SBP less than 90 mm Hg, or vasopressor use were more likely to lack sufficient preventive measures. These findings support current discussions in the literature on the challenges of repositioning critical care patients. Brindle and colleagues²²  reported on the development of consensus recommendations by a group of experts related to safe repositioning of patients. They noted that the critical care unit’s culture and clinicians’ perceptions about hemodynamic instability may lead to staff members’ not repositioning patients. Krapfl and colleagues,²³  in a review of the literature, noted that the complexity and instability of the patient’s condition often limit repositioning by the nurse. However, best practice continues to suggest that slow, gradual turning allows sufficient time for stabilization of blood pressure and oxygen saturation and should be considered.¹¹ 

Several limitations of this study are evident. The study had a retrospective design and relied on the accuracy of documentation, which did not allow confirmation of findings through observation. All participants in the study had pressure injuries, and no case-control group was used for comparison, thus limiting the ability to compare risk factors across groups. The study was conducted at a large health care system with high patient acuity, and the findings may not be generalizable to other populations and settings. Data retrieval to complete the PUPI relied on the accuracy of the documentation in the EMR. The complexity of the EMR was evident during data collection, even though the EMR software is commonly used and provided by an international EMR software company. Another limitation was the inclusion of progressive care patients, which may have diluted the acuity of the sample. Finally, the PUPI is based on the Braden Scale subscales and was not originally intended for use with medical device–related pressure injuries, thus limiting its application to this type of HAPI.

The findings of this study provide important information and new knowledge for critical care nurses and other health care providers and highlight the importance of pressure injury prevention documentation. Attention to standard nursing care (preventive interventions) and accuracy of documentation is essential in the complex critical care setting. In addition, this study objectively identified the occurrence of unavoidable pressure injuries, suggesting a possible etiology of acute skin failure rather than a lack of preventive nursing care. Further research—particularly rigorous, controlled studies—is needed to investigate the occurrence of unavoidable HAPIs and acute skin failure in critically ill patients.

This study was performed at Indiana University Health, Indianapolis, Indiana.