|Home | About | Journals | Submit | Contact Us | Français|
Purpose:To estimate the frequency of use of pressure-redistributing support surfaces (PRSS) among hip fracture patients and to determine whether higher pressure ulcer risk is associated with greater PRSS use.Design and Methods:Patients (n = 658) aged ≥65 years who had surgery for hip fracture were examined by research nurses at baseline and on alternating days for 21 days. Information on PRSS use and pressure ulcer risk factors was recorded at each assessment visit. Other information was obtained by interview and chart review.Results:A PRSS was observed at 36.4% of the 5,940 study visits. The odds of PRSS use were lower in the rehabilitation setting (adjusted odds ratio [OR] 0.4, 95% confidence interval [CI] 0.3–0.6), in the nursing home (adjusted OR 0.2, 95% CI 0.1–0.3), and during readmission to the acute setting (adjusted OR 0.6, 95% CI 0.4–0.9) than in the initial acute setting. There was wide variation in frequency of PRSS use by admission hospital, even after adjusting for pressure ulcer risk factors. The relationships between PRSS use and pressure ulcer risk factors were not strong.Implications:In this study of hip fracture patients, adherence to guidelines for PRSS use was low and was based more on facility-related factors than on patient risk. There is an urgent need for health care providers to improve strategies for the prevention of pressure ulcers in high-risk patients.
Pressure ulcers are areas of localized injury to the skin or underlying tissue that arise when soft tissue is compressed between a bony prominence and an external surface for a prolonged period of time (Black et al., 2007). Immobility is the primary risk factor (Lindgren, Unosson, Fredrikson, & Ek, 2004); incontinence, poor nutritional status, and various disease states are also associated with higher risk (Thomas, 2001). Despite a number of important national prevention initiatives (Ayello & Lyder, 2007; Lyder & van Rijswijk, 2005; Rosenthal, 2007; US Department of Health & Human Services, 2000), pressure ulcers continue to have a negative impact on patient outcomes and health care costs in a variety of care settings (Allman, Goode, Burst, Bartolucci, & Thomas, 1999; Roghmann, Siddiqui, Plaisance, & Standiford, 2001; Russo & Elixhauser, 2006; Russo, Steiner, & Spector, 2008; Spilsbury et al., 2007). Focus on this problem has increased because the Centers for Medicare and Medicaid Services (CMS) designated pressure ulcers as preventable complications of medical care and no longer reimburse hospitals for the cost of treating hospital-acquired pressure ulcers (Rosenthal, 2007).
Clinical approaches to pressure ulcer prevention include risk assessment, active mobilization of patients who are able to walk, regular repositioning of bedbound patients, and use of preventive devices (Arnold, 2003). Devices used for pressure ulcer prevention include pressure-redistributing support surfaces (PRSS; i.e., mattresses and mattress overlays), chair and wheelchair cushions, and heel protectors. In general, these devices are intended to prevent tissue damage either by increasing the area of the support surface that is in contact with bony prominences or by reducing the intensity or duration of pressure on bony prominences (National Pressure Ulcer Advisory Panel, 2007).
Hip fracture patients are at high risk of pressure ulcers because they often experience long periods of immobility before, during, and after surgery (National Pressure Ulcer Advisory Panel, 2001). We have previously reported that approximately one third of hip fracture patients develop at least one new pressure ulcer Stage 2 or higher within 32 days following hospital admission (Baumgarten, 2009). Hip fracture patients tend to have a short hospital stay followed by transitions to one or more postacute settings within a short period of time (Gehlbach, Avrunin, & Puleo, 2007; Morris & Zuckerman, 2002); these transitions may threaten the quality and continuity of pressure ulcer prevention measures, including the use of PRSS. The aims of this study were to estimate the frequency of use of PRSS among hip fracture patients across the continuum of care and to determine whether PRSS use is higher among patients who are at higher risk of pressure ulcers. We hypothesized that frequency of PRSS use would be higher among patients with higher levels of pressure ulcer risk factors because current guidelines recommend that devices and other preventive interventions be targeted at patients at high risk of pressure ulcers (Ratliff, 2005). Furthermore, we hypothesized that frequency of PRSS use would be highest in the acute care setting because that is where pressure ulcer risk (in terms of immobility, incontinence, illness severity, and other risk factors) is likely to be highest and because hospitals may have more resources that can be devoted to pressure ulcer prevention than other types of health care facilities.
Details of the study design and procedures have been reported previously (Baumgarten, 2009). Briefly, this prospective cohort study was carried out between 2004 and 2007 in nine hospitals that are part of the Baltimore Hip Studies network (Magaziner et al., 2000) and in 105 postacute facilities to which patients from these hospitals were discharged. Eligible patients were aged 65 years or older, had surgery for hip fracture (International Classification of Diseases, Ninth Revision code 820), and provided written consent while in the acute care hospital. Verbal assent was obtained from patients, and written consent from proxies for patients whose score on the Mini-Mental State Exam (Folstein, Folstein, & McHugh, 1975) was less than 20. Proxy consent was obtained for patients who were unconscious or noncommunicative. The study was approved by the institutional review boards of the University of Maryland, Baltimore, and of each of the participating hospitals. Study participants were examined by specially trained research nurses at baseline and on alternating days for 21 days for a total of 11 assessments. The follow-up assessments were carried out in the care setting where the patient was at the time of the scheduled examination.
At each assessment visit, the research nurse recorded whether the patient had a PRSS in use and, if so, the type. A PRSS was considered to be in use if the device was observed to be on the patient's bed, even if the patient was not in bed at the time of the assessment visit. Care setting (initial acute hospital, nursing home, rehabilitation unit in hospital or other location, readmission to acute hospital, and home) and presence of a pressure ulcer were observed and recorded by the research nurses at each assessment visit. For days on which there was no scheduled assessment visit, care setting was ascertained retrospectively at the next visit by consulting the patient, the caregiver, and the facility chart.
The research nurses evaluated pressure ulcer risk at each visit using the Braden Scale (Kring, 2007), a widely used tool that assesses sensory perception, moisture, activity, mobility, nutrition, friction, and shear. The maximum Braden Scale score is 23, with lower scores indicating higher risk. For the analysis, we dichotomized the Braden scores into two categories: low risk (score >16) and high risk (score ≤16; Braden & Bergstrom, 1994).
Nutritional status was assessed at the baseline study visit using the Subjective Global Assessment of Nutritional Status (Detsky, Smalley, & Chang, 1994), which classifies individuals as being at low, moderate, or high risk of nutrition-associated complications. Severity of illness was measured using the Rand Sickness at Admission Scale (hip fracture version; Keeler et al., 1990) and comorbidity by the Charlson comorbidity index (Charlson, Pompei, Ales, & MacKenzie, 1987); both of these were completed using information abstracted from the medical record. The patient's activity level (walks, chairbound, and bedbound) was assessed at each study visit using the activity item from the Braden Scale (Kring, 2007). Urinary or fecal incontinence status was recorded at each visit based on observation and discussion with clinical staff. Finally, the patient's orientation to person, place, and time was assessed at each study visit by the research nurse, and the patient was classified according to the number of dimensions (0–3) to which the patient was oriented; a larger number of dimensions indicates better mental status.
Descriptive information was summarized using means and standard deviations, or percentages, as appropriate. The relationship between the use of a PRSS at a given visit and the predictor variables (pressure ulcer risk factors and care setting at the same study visit) was analyzed using generalized estimating equations (GEE; Liang & Zeger, 1986) in SAS version 9.1 (SAS Institute Inc., Cary, NC). A binomial distribution with logit link and compound symmetry covariance structure was specified in PROC GENMOD. For each pressure ulcer risk factor and for care setting (initial acute hospital, nursing home, rehabilitation unit in hospital or other location, and readmission to acute hospital), we fit a GEE model accounting for within-subject correlation and time since initial hospital admission, with the risk factor or care setting as the predictor variable and use of a PRSS as the binary outcome variable. In addition, we fit a fully adjusted model, which included care setting and all the pressure ulcer risk factors (except Braden score) simultaneously, while accounting for within-subject correlation and time since initial hospital admission. Time-dependent covariates (presence of a pressure ulcer, activity, orientation, incontinence, and care setting) measured at each study visit were included in the GEE models. The other covariates (nutritional risk, age, Rand severity of illness score, and Charlson comorbidity score) were measured only at baseline. To avoid overfitting, we did not include the Braden Scale score in this model because the Braden Scale is a measure that includes items that overlap conceptually and are highly correlated with the risk factors in the model. Statistical significance was defined as a p value <.05.
Given the large number of acute and postacute facilities (more than 100) represented in this study, it was not possible to explore the association between the individual facility in which the patient resided and the use of a PRSS at a given visit. However, because there were only nine acute hospitals, it was possible to examine the effect of facility on PRSS use in the acute setting. To do this, we performed a set of analyses using only data from study visits conducted while the patient was in the initial acute setting. These were similar to the analyses described previously except that, instead of care setting, admission hospital was included as a predictor variable.
We also performed all the analyses described previously with use of any type of preventive device (including chair and wheelchair cushions, heel protectors, and positioning pillows) as the outcome variable. Because the results were very similar, and because PRSS are likely to be of most interest because they are the subject of recent national initiatives (National Pressure Ulcer Advisory Panel, 2007; Posthauer, Jordan, & Sylvia, 2006), only the results with PRSS as the outcome variable are presented in this article.
The mean age of study participants was 83.2 years (SD 6.6); 23.1% were male and 98.0% were White (Table 1). More than two thirds had resided in the community before admission for the hip fracture. The mean length of hospital stay was 5.9 days (SD 3.2), and the mean time between the initial hospital admission and the baseline study assessment was 2.9 days (SD 2.0); 11.7% had the baseline assessment before the day of surgery.
A PRSS was observed to be in use at 36.4% of the 5,940 study visits (Table 2). PRSS were observed somewhat more frequently at visits where the patient was at higher risk according to the Braden Scale than at visits where pressure ulcer risk was lower, with an odds ratio (OR) of 1.2 (95% confidence interval [CI] 1.0–1.4) comparing the high risk to the lower risk group. In the adjusted analysis, patients with a pressure ulcer at a given visit had slightly higher odds of PRSS use than those without a pressure ulcer (adjusted OR 1.2, 95% CI 1.0–1.5). None of the other pressure ulcer risk factors was significantly associated with PRSS use in the adjusted analysis. The odds of PRSS use was lower in the rehabilitation setting (adjusted OR 0.4, 95% CI 0.3–0.6), in the nursing home (adjusted OR 0.2, 95% CI 0.1–0.3), and during readmission to the acute setting (adjusted OR 0.6, 95% CI 0.4–0.9) than in the initial acute setting.
The analysis limited to study visits in the initial acute setting revealed that a PRSS was in use at 56.8% of the 1,406 study visits in that setting (Table 3). Neither the Braden Scale score nor any of the pressure ulcer risk factors was significantly associated with PRSS use. However, there was considerable variation by admission hospital with, for example, hospital H having odds of PRSS use that were nearly 80 times as high as in hospital A, after adjustment for pressure ulcer risk factors.
Frequency of PRSS use was only 57% in the initial acute care setting, despite the fact that hip fracture patients are at high risk of pressure ulcers, especially in the acute hospital (Baumgarten, 2009; National Pressure Ulcer Advisory Panel, 2001), and in spite of clinical guidelines recommending the use of PRSS for high-risk patients (Ratliff, 2005). Care setting was an important determinant of PRSS use in this study, with odds of PRSS use in the rehabilitation setting less than half those in the initial acute setting and odds of PRSS use in the nursing home setting less than one quarter those in the initial acute setting. Patients in the initial acute setting may be more likely to receive a PRSS because they are at higher risk of pressure ulcers during the acute care stay than later in the recovery period when they have moved to a different setting. However, these differences among care settings persisted almost unchanged after adjustment for multiple pressure ulcer risk factors and for time since admission, suggesting that the type and quality of preventive care in the different settings may also be important factors. We also found tremendous variability among individual hospitals in the frequency of use of PRSS during the initial acute stay (range 23.0%–93.6%). As with the results for care setting, large differences between hospitals persisted even after adjusting for pressure ulcer risk factors and time since initial hospital admission, suggesting that differences in the frequency of PRSS use are not attributable only to differences in case mix across hospitals.
There are several possible explanations for the lower odds of PRSS use during readmission to the acute hospital than during the initial acute stay (OR 0.6, 95% CI 0.4–0.9). It may be that, when patients are readmitted, they go to a hospital unit whose pressure ulcer prevention practices and resources differ from those of the hospital unit the patient was on during the initial acute stay. Also, the need for readmission indicates the presence of a serious acute condition, and hospital staff may consider the acute condition to be a higher priority than the pressure ulcer prevention.
A surprising finding was the lack of association between pressure ulcer risk factors and use of PRSS, both in the analysis that included all settings and in the analysis limited to the initial acute hospital stay. This finding, together with the strong associations between PRSS use and both type of setting and individual hospital, suggests that PRSS use among hip fracture patients is based more on facility-related factors than on patient risk, despite guideline recommendations that prevention efforts be based on each patient's pressure ulcer risk status (Ratliff, 2005).
The results of prior studies that have examined care setting as a predictor of pressure ulcer preventive device use are inconsistent with each other and with the findings of the current study. A Swiss study found that use of preventive devices was more frequent in rehabilitation and subacute care units than in medical and surgical units (Perneger, Heliot, Rae, Borst, & Gaspoz, 1998). A German study found that, among patients at high pressure ulcer risk according to the Braden Scale, the prevalence of PRSS use was higher in nursing homes than in hospitals (Lahmann, Halfens, & Dassen, 2005). However, a study in the Netherlands found that the prevalence of PRSS use did not differ between nursing homes and hospitals (Tannen, Dassen, & Halfens, 2008). Given the large differences in health care organization and financing between the United States and Europe, the results of these studies are not directly comparable to those of the current study. Other studies of PRSS use in the United States have focused on a single type of facility (Lyder, Shannon, Empleo-Frazier, McGeHee, & White, 2002), usually acute hospitals (O’Dea, 1999), or have examined the use of devices without regard to the setting or facility (Bergstrom, Braden, Kemp, Champagne, & Ruby, 1996; Goodridge et al., 1998; Gunningberg, 2004; Pieper, Sugrue, Weiland, Sprague, & Heimann, 1997; Pieper, Sugrue, Weiland, Sprague, & Heiman, 1998).
The Braden Scale is widely used in a variety of health care settings in the United States to assess pressure ulcer risk (Bergstrom, 2005), and we hypothesized that care providers use the results of the Braden assessment to make decisions about the need for pressure ulcer preventive devices. Therefore, we expected that use of PRSS in our study would be associated with lower Braden scores. In the analysis that included all care settings, we did find a significantly higher odds of PRSS use at study visits where patients had lower scores, but the OR was only 1.2 and, in the analysis restricted to the acute hospital setting, Braden scores were not significantly associated with PRSS use. We also expected that use of PRSS would be associated with the individual pressure ulcer risk factors. In the analysis that included all care settings having a pressure ulcer at a given study visit was associated with slightly higher odds of using a PRSS at the same visit (OR 1.2, 95% CI 1.0–1.5). It is surprising that this association is not stronger because PRSS are often used for treatment of pressure ulcers as well as prevention (Thompson, Anderson, Langemo, Hanson, & Hunter, 2008). None of the other pressure ulcer risk factors was significantly associated with PRSS use in either analysis.
Our finding of little association between pressure ulcer risk factors and PRSS use is in contrast with the results of older studies (Bergstrom et al., 1996; Pieper et al., 1997, 1998). A recent study limited to long-term care facilities did not find a significant relationship (Lyder et al., 2002), and European studies have been inconsistent in their findings (Goodridge et al., 1998; Gunningberg, 2004, 2005; O’Dea, 1999; Perneger et al., 1998). A possible reason for the difference between the current findings and the results of prior U.S. studies (Bergstrom et al.; Pieper & Weiland, 1997; Pieper et al., 1998) is that the current study was limited to hip fracture patients, whereas all prior studies were conducted in a broader patient population with no eligibility criteria based on diagnosis. Assuming that hip fracture patients are recognized by care providers as being a high-risk group, allocation of PRSS to patients in the current study may have been determined primarily based on their hip fracture diagnosis, reducing the impact of the individual pressure ulcer risk factors examined in the study.
However, the finding that pressure ulcer risk factors had little impact on the use of PRSS may also indicate a problem with the quality of care, as pressure ulcer prevention guidelines recommend the allocation of devices based on presence of the risk factors that were examined in this study (e.g., incontinence, immobility, and poor nutritional status; Ratliff, 2005). In this study, we did not have information on policies and clinical practices within study hospitals and postacute facilities that would allow us to explain why pressure ulcer risk factors were not associated with use of PRSS. One explanation may be health care providers’ lack of knowledge about recommended methods of pressure ulcer prevention. Also, providers may have a global perception of pressure ulcer risk for all patients in a given setting or facility. For example, in the rehabilitation setting, care providers may assume that patients are fully mobilized and thus not at risk for pressure ulcers. Although it is true that hip fracture patients who have moved to the rehabilitation setting are no longer restricted to bed, frail elderly patients with multiple comorbid conditions may spend much of their time sitting in a chair where pressure relief is more difficult than on an appropriate mattress (Bliss & Simini, 1999). In a prior study, only 67.5% of inpatient rehabilitation units reported assessing pressure ulcer risk for all patients daily, and risk assessment was commonly based on facility-developed tools and clinical judgment rather than on established tools (Sae-Sia & Wipke-Tevis, 2002).
Given that all the hospitals in this study are in the same geographic region and operate within the same reimbursement environment, differences between the hospitals in PRSS use should depend only on patient characteristics. However, the large differences observed between hospitals in the frequency of PRSS use, even after adjusting for case mix, suggest that facility-level policies and resources, rather than patient characteristics, play an important role in determining whether a patient receives a PRSS. This may be an example of supply-sensitive variation in care, which occurs when the availability of resources, not patient need or scientific evidence, drives the frequency of resource use (Fisher & Wennberg, 2003; Fisher et al., 2000). Anecdotally, we know that the policy in some of the hospitals in our study is to provide PRSS to all patients on medical and surgical units regardless of pressure ulcer risk.
Strengths of this study include its large sample size and inclusion of a variety of care settings and facilities. The availability of a large amount of detailed information on pressure ulcer risk factors meant that we could effectively adjust for case mix when comparing care settings and individual facilities. Because we collected data at multiple time points, we were able to use analytic methods that assessed the association between PRSS at a given study visit and the predictor variables (care setting and pressure ulcer risk factors) at the same study visit.
However, the study has several limitations. First, the use of PRSS was based on observations by research nurses at brief study visits performed every other day during the patient's participation in the study. It is not known to what extent the observations at the study visits are an accurate representation of PRSS use between study visits. Because it is unlikely that misclassification of this type would depend on care setting, however, any bias would be toward the null, suggesting that the associations between PRSS and care setting may be even greater than estimated in this study. Second, given the large number of facilities in this study (more than 100), we could not examine the differences in PRSS use among individual facilities, although we did estimate differences between individual hospitals in an analysis limited to visits in the initial acute setting. Future studies should examine whether the high degree of variability we observed between hospitals extends to rehabilitation and nursing home facilities as well. Third, given the large degree of variability among individual hospitals in the frequency of PRSS use, it would have been informative to examine the role of facility characteristics and policies. However, because we did not gather facility-specific information, such an analysis was not possible. Finally, although clinical guidelines recommend that high-risk patients should be cared for on a PRSS, the guidelines do not provide an operationalized definition of “high risk.” Therefore, we cannot compare the observed frequency of PRSS use to what the frequency should have been, based on the guidelines. However, even in the highest risk group (patients in the initial acute setting with a Braden score <16), PRSS were only in use at 54.8% of study visits.
Although use of PRSS for high-risk patients in all care settings is advocated by all existing clinical practice guidelines (American Medical Directors Association, 2008; Ayello, 2003; Benbow, 2006; National Collaborating Centre for Nursing and Supportive Care, 2003; National Guideline Clearinghouse [Registered Nurses Association of Ontario], 2005; Panel for the Prediction and Prevention of Pressure Ulcers in Adults, 1992; Ratliff, 2005), the appropriateness of this recommendation is unknown. The evidence base for the effectiveness of PRSS is weak and only a few types of support surface have been evaluated using rigorous methods (Cullum, McInnes, Bell-Syer, & Legood, 2004). Similarly, there is conflicting evidence regarding the validity of risk assessment scales such as the Braden Scale (Bolton, 2007; Olshansky, 2008), despite the fact that these scales are widely used to assess pressure ulcer risk and, therefore, to predict need for preventive measures. A recent Cochrane review (Moore & Cowman, 2008) concluded that there has not been a single randomized study evaluating the effect of using a pressure ulcer risk assessment tool on pressure ulcer incidence. Studies are needed to define clearly which patients should be cared for on a PRSS and to provide conclusive information on the effectiveness of PRSS and other preventive measures. However, until such studies become available, current clinical guidelines based on existing research and expert opinion represent the best alternative and adherence to these guidelines should be encouraged.
Pressure ulcers in elderly patients can have a negative impact on quality of life and can result in longer hospital stays, higher health care costs, poor rehabilitation outcomes, and serious complications (Allman et al., 1999; Kumar et al., 2004; Roghmann et al., 2001; Spilsbury et al., 2007). Pressure ulcer frequency has not declined in recent years (Whittington & Briones, 2004), despite growing recognition of the problem. Although there are questions about the strength of the evidence underlying existing clinical guidelines, these guidelines represent the best recommendations currently available. The results of this study suggest that, at least in one high-risk patient population, adherence to the guidelines for preventive device use is low and is not guided by rational decisions about patient risk. Whether CMS’ new reimbursement policies will have an impact on the incidence of acquired pressure ulcers in hospitals remains to be seen. In the mean time, there is an urgent need for health care providers to improve care for the prevention of pressure ulcers in high-risk patients.
Supported by grants from National Institute of Arthritis and Musculoskeletal and Skin Diseases (5R01 AR 47711); University of Maryland General Clinical Research Center Grant, General Clinical Research Centers Program, National Center for Research Resources (M01 RR 16500); and National Institute on Aging Claude D. Pepper Older Americans Independence Center (P30 AG028747).
Sincere thanks to the staff, patients, and families at all the study hospitals and postacute facilities that participated in this study.