- Email: [email protected]
Predictive validity of the braden scale and nurse perception in identifying pressure ulcer risk
Predictive Validity of the Braden Scale and Nurse Perception in Identifying Pressure Ulcer Risk Terry VandenBosch, Cecelia Montoye, Martha Satwicz, Karen Durkee-Leonard, and Barbara Boylan-Lewis
Before the implementation of the Braden Scale, an institutional study was conducted to determine the cut-off point at which patients would be at risk for developing a pressure ulcer. Nurses' perception of patient risk for pressure ulcer development also was compared with the Braden Scale. One hundred and three subjects hospitalized for a minimum of 7 days and randomly selected from routine hospital admissions were rated with the Braden Scale, and skin assessments were made three times per week for up to 2 weeks starting 24 to 48 hours after admission. In addition, bedside nurses were asked to use clinical judgment to identify subjects at risk for pressure ulcer development. Twenty-nine subjects developed pressure ulcers. Nurse's judgements of pressure ulcer risk were not significant in predicting pressure ulcer positive or negative status. Results of the t test demonstrated the Braden Scale score is the most highly significant finding (p = .0038) to predict pressure ulcer positive and pressure ulcer negative groups. For this study, the Braden Scale cut-off point was set at 17 with a sensitivity of .59 and a specificity of .59. Clinical implementation of the Braden Scale must be combined with frequent and thorough skin assessment practices because some patients will develop pressure ulcers even though the tool does not predict the patients to be at risk.
Copyright © 1996 by W.B. Saunders Company
A
SYSTEMATIC method to assist the caregiver in predicting patients at risk for developing pressure ulcers is a useful clinical tool. In the process of developing institutional standards for pressure ulcer development, the authors reviewed several published tools that predict pressure ulcer risk and selected the Braden Scale 1 for this study. The Braden Scale was used because it is highly sensitive and has a lower rate of overprediction than other scales (Bergstrom, Braden, Laguzza, & From Mission Health, St. Joseph Mercy Hospital, Ann Arbor, MI; Michigan Peer Review Organ&ation, Plymouth, MI; and the Visiting Nurses'Association, Port Huron, MI. Terry VandenBosch, MS, RN, CS: Research Specialist, Mission Health, St. Joseph Mercy Hospital, Ann Arbor, MI; Cecelia Montoye, MSN, RN: Senior Staff Associate for Education and Feedback, Michigan Peer Review Organization, Plymouth, M1; Martha Satwicz, MSN, RN: Senior Staff Associate for Education and Feedback, Michigan Peer Review Organization, Plymouth, M1; Karen Durkee-Leonard, MSN, ICN: Program Coordinator, Rehabilitation Services, Visiting Nurses' Association, Port Huron, MI; Barbara Boylan-Lewis, MA, RN, CETN: Enterostomal Therapy Nurse, Mission Health, St. Joseph Mercy Hospital, Ann Arbor, ML Address reprint requests to Terry VandenBosch MS, RN, CS, Nursing Development Services, 5301 East Huron River Dr, P.O. Box 992, Ann Arbor, M148106. Copyright © 1996 by W.B. Saunders Company 0897-1897/96/0902-000755.00/0 80
Holman, 1987). The registered nurse (RN) caregiver uses the Braden Scale to assess the patient on admission and periodically thereafter. Each patient is scored according to criteria on the scale and then given an overall numerical score that represents risk of pressure ulcer development. The lower the overall numerical score, the more risk to the patient. Institutions that use the Braden Scale must decide on a cut-off point (numerical value) at which the patient is determined to be at risk. The cut-off point signifies the need for implementation of the institution's standard of care for pressure ulcer prevention. Preventive care is targeted to those patients who are at risk and can reduce the time and cost of implementing preventive care for patients not at risk. The developers of the Braden Scale recommend conducting a study to calculate the institutional critical cut-off point for risk and to determine predictive validity of the tool. Therefore, a replication study was conducted at the institution. BACKGROUND
Pressure ulcers are a serious problem in the United States. Incidence of pressure ulcers in hospitals has been reported from 2.7% to 29.5% (Panel for Treatment of Pressure Ulcers, 1994). Usual estimates of hospital pressure ulcer inciApplied Nursing Research,Vol. 9, No. 2 (May), 1996: pp 80-86
PREDICTIVE VALIDITY OF THE BRADEN SCALE
dence rates are reported at 1% to 5% (Allman et al., 1986; Anderson, 1982; National Pressure Ulcer Advisory Panel, 1989) with many more patients identified as being at risk. The in-hospital mortality rate for patients with pressure ulcers is estimated at 23% to 37% (Allman, 1989; Allman et al., 1986). Pressure ulcers develop quickly and heal slowly. Deep lesions require weeks or months to heal and sometimes require surgical closure. Complications include bacteremia, sepsis, and osteomyelitis (Allman, 1989). Pressure ulcers can also be painful and are an emotional burden for patients and families. The cost of treating pressure ulcers is high. Studies report that a single pressure ulcer costs from $1,300 to $35,000 (Alterescu, 1989; Van Ness, 1989). Estimations of cost depend on the severity and stage of the pressure ulcer and on the method used to analyze costs. Consistent methods for determining costs have not been determined (Alterescu, 1989). In 1986, 3.5 to 7 billion dollars was estimated as the annual cost of pressure ulcer treatment (Maklebust, Mondous, & Sieggreen, 1986). The causes of pressure ulcers are complex. Critical determinates of pressure ulcers are factors that affect the intensity and duration of pressure and factors that affect the tolerance of the skin and supporting structures for pressure. Mobility status, activity, and sensory perception are factors related to intensity and duration of pressure. Tissue tolerance is affected by extrinsic factors of skin moisture, friction, shear, and by intrinsic factors such as nutritional status, age, and arteriolar pressure (Braden & Bergstrom, 1987; Copeland-Fields & Hosiko, 1989). Prevention and care of pressure ulcers are important. However, the first aim is to target measures that will prevent pressure ulcers. When preventive programs are instituted, a reduction in the incidence of pressure ulcers of up to 50% has been reported (Anderson, 1982; Seiler & Stahelin, 1985). The Braden Scale identifies critical factors in pressure ulcer formation (Braden & Bergstrom, 1989). The scale is made up of six subscales: mobility, activity, sensory perception, moisture, nutrition, and friction and shear. Each subscale contains a title, a key concept descriptor, and a brief sentence description of qualifying attributes. Subscales are scored by level of mobility, activity, etc. The highest possible score is 23 points. The lower the score the greater the risk of pressure ulcer
81
development. Cut-off points are used to predict risk and institute preventive protocols. They provide a balance between the cost of prevention and the cost of treatment. A cut-off score of 16 or less is suggested by Braden and Bergstrom (1989). Darby (1990) compared nurse perception of patient risk for pressure ulcers with the Braden Scale. Nurse perception alone was not as accurate as the Braden Scale in predicting patients at risk. Salvadalena, Synder, and Brodgon (1992) found the Braden Scale was less accurate than nurse judgment in a clinical trial on an acute care medical unit. Jones, Burger, Piraino, and Utley (1993) and Van Ness (1989) demonstrated a decrease in prevalence and incidence or pressure ulcers following implementation of pressure ulcer prevention programs that used risk assessment tools. The results of these studies, along with the Agency for Health Care Policy and Research Clinical Practice Guideline (Panel for the Prediction and Prevention of Pressure Ulcers in Adults, 1992), suggest that the use of a systematic research-based tool improves the clinical judgment of nurses in predicting pressure ulcer development. Thus, the purpose of this study was twofold: to determine the cut-off point, sensitivity, and specificity for use of the Braden Scale in a specific inpatient hospital setting; and to compare nurses' perception of risk of pressure ulcer development with and without the Braden Scale. METHODS Subjects The study took place in a 550-bed tertiary care community teaching hospital. Subjects were drawn from general care, intensive care, and a 40-bed inpatient rehabilitation unit. Hospitalized inpatients over 18 years of age with intact skin and no reddened areas over any bony prominence were asked to participate. Using a random numbers table and the daily list of admissions, subjects with an expected length of stay of at least 1 week were selected. Subjects were selected three times a week and were approached for informed consent within 48 hours of admission to the hospital. Instruments Three instruments, the Braden Scale, a skin assessment form, and a tool to record the subject's clinical status, were used to collect data. All of the tools were developed and used by Bergstrom and
82
Braden in previous studies (Braden & Bergstrom, 1987). Nurse perception of pressure ulcer risk also was used.
Braden Scale The Braden Scale is a sensitive (83% to 100%) and specific (64% to 90%) instrument (Bergstrom, Braden, et al., 1987; Bergstrom, et al., 1987). Sensitivity and specificity are indicators of predictive validity. Reliability studies indicate the instrument is also highly reliable. Reliability for nurses aides and licensed practical nurses ranged from r = .83 to r = .94 (Bergstrom, Braden, et al., 1987). Reliability for RNs has been reported at r = .99 (Bergstrom, Braden, et al., 1987). Bergstrom, Braden, et al. (1987) assessed validity and reliability of the Braden Scale in acute care, rehabilitation, and skilled nursing facility settings.
Skin Assessment Form A skin assessment form was used to document the status of the hospitalized patient's skin over every bony prominence, to document the presence and location of a pressure ulcer, and tO record the staging of any ulcers that were present. The form contained a human body diagram with a number at every potential pressure ulcer site. The form has construct validity and captures 99% of the possible pressure ulcer sites (Bergstrom, 1994, personal communication). A classification schema of four pressure ulcer stages was used: stage I was defined as "nonblanchable erythema of intact skin that does not disappear for 24 hours after pressure is relieved"; stage II was defined as a break in skin such as blisters or abrasions; stage III was defined as a break in skin exposing subcutaneous tissue; and stage IV was defined as a break in skin extending through tissue and subcutaneous layers exposing muscle or bone. Clinical Status
Information about subject's clinical status was collected. Data collected were age, gender, history of smoking, diagnosis, and systolic and diastolic blood pressure.
Nurse Perception of Pressure Ulcer Risk Nurse perception of pressure ulcer risk was measured with a question developed by Darby (1990) and used by Salvadalena et al. (1992). At the admission rating/observation period, the staff nurse
VANDENBOSCH ET AL
caring for the study patient was asked to answer "yes" or "no" to the question: "Do you think this patient will develop a pressure ulcer while they are in the hospital?" Procedure
Selected patients who had been admitted 24 to 48 hours before a data collection day were approached and asked for informed consent. Only those patients whose length of stay was estimated as 7 days or longer were approached. Data were collected on Mondays, Wednesdays, and Fridays. All subjects had a minimum of three assessments and a maximum of six. The research data collectors were nurses on the staff of the hospital. Interrater reliability was conducted 1 week before implementation of the study and all ratings were at a 90% or higher level of agreement. Eleven nurses were trained as consistent data collectors and five additional nurses were trained as relief data collectors. To avoid bias on the part of the data collectors, two data collection roles were used. The first, Nurse I, was trained to use the Braden Scale. The second, Nurse II, was trained to assess subjects' skin and pressure ulcer status. Eight nurses were trained for the Nurse I procedures and eight were trained for the Nurse II procedures. Subjects were assessed by Nurse I and Nurse II on each data collection day.
Nurse I 1. Obtained informed verbal consent and completed demographic data within 48 hours of admission to the hospital. 2. Used the Braden Scale to rate the subject at the onset of the study and every Monday, Wednesday, and Friday for 2 weeks or until the subject was discharged. Nurses were instructed not to give data about the subject's risk for pressure ulcer development to the nurse caregiver.
Nurse II 1. Completed the skin assessment form on the day of entry into the study and every Monday, Wednesday, and Friday until discharge. If a nurse caregiver asked the nurse collecting data for information, the nurse could inform the caregiver if a pressure ulcer was present. 2. Asked the staff nurse assigned to the subject for the day about the subject's risk for
PREDICTIVE VALIDITY OF THE BRADEN SCALE
83
Table 2. Demographic Data
pressure ulcer development at the time of the first skin assessment. 3. Collected information on the patients' clinical status.
One hundred three subjects participated. Twentynine subjects developed pressure ulcers. First occurrence of pressure ulcers ranged from 3 to 13 days after hospital admission with a mean of 4.8 days and a SD of 1.9 days. The coccyx/sacrum area was the most common site. Four pressure ulcers on the ear were reported. Sites and pressure ulcer staging are listed in Table 1. The number of pressure ulcer sites per subject ranged from one to eight. Demographic data by pressure ulcer positive and pressure ulcer negative status is included in Table 2. The sample included mostly White patients. Females and former smokers had a higher percent of pressure ulcers. However, there was no significant increase in the odds of developing pressure ulcers associated with gender or smokers. Nurse prediction about risk for developing pressure ulcers is described in Table 2. Nurses correctly predicted 51.7% of the subjects would develop a pressure ulcer. They incorrectly predicted that 41.1% of the patients who did not develop pressure ulcers would develop them. These predictions were not significant when analyzed using chi-square analysis. Results of t-test of selected variables by pressure ulcer positive and pressure ulcer negative status are presented in Table 3. There was no significant Table 1. Pressure Ulcer Sites
Coccyx/sacrum Elbow Heel Ear Vertebrae Scapula Toe Ankle Trochanter Knee Lower leg Back of head Ischial tuber Other
Stage 1
Stage 2
Stage 3
Frequency %
Frequency %
Frequency %
13 8 4 3 3 2
18,8 11.5 5.7 4.3 4.3 2.9
1
1.5
1 1 1
1.5 1.5 1.5
Frequency
Pressure Ulcer Positive (n = 29)
Percent Frequency Percent
Gender
RESULTS
Site
Pressure Ulcer Negative (n = 74)
5 5 7 1 1
7.2 10.0 1.5 1.5
2 1 1 1
2.9 1,5 1.5 1.5
1
1.5
10
14.4
There were no stage 4 pressure ulcers.
1
1.5
I
1.5
Male
39
52.7
10
34.5
Female
35
47.3
19
65.5
White Black
63 9
85.1 12.2
25 3
89.3 10.7
Other
2
2,7
--
--
33
44.6
8
27.6
10 31
13.5 41.9
4 17
13.8 58.6
43 30
58.9 41.1
14 15
48.3 51.7
Race
Smoke Never Current Former Nurse prediction of pressure ulcer No Yes
A total n less than 103 indicates missing data,
difference between the ages of patients who were pressure ulcer positive and those who were pressure ulcer negative. There was a significant difference between the pressure ulcer positive and pressure ulcer negative subjects on lowest systolic blood pressure with pressure ulcer positive patients having higher systolic readings. The Braden Scale score was the most highly significant finding (p = .0038) between the pressure ulcer positive and pressure ulcer negative groups. The Braden Scale score used for t-test analysis was the second observation period for pressure ulcer negative subjects and the observation before the first pressure ulcer occurrence for pressure ulcer positive subjects. Sensitivity and specificity are common predictive tools in research. Sensitivity is the percentage of subjects who develop pressure ulcers and whom the scale predicted would develop them. Specificity is the percentage of those who do not develop pressure ulcers and whom were not predicted to develop them. A standard formula for determining sensitivity and specificity, as reported in Lilienfeld and Lilienfeld (1980) and as described in Bergstrom, Braden, et al. (1987), was used in this study. The cut-off point for assessing risk was determined using the guidelines in Bergstrom, Demuth, and Braden (1987). At a score of 17, 59% of the subjects predicted to develop pressure ulcers did develop a pressure ulcer (sensitivity), whereas 41% of the subjects predicted to develop a pressure ulcer did not develop one (false-positive rate). At a score
84
VANDENBOSCH ET AL
Table 3. t-Test Results of Selected Variables by Pressure Ulcer Positive and Negative Pressure Ulcer PressureUlcer Negative Positive Variables Age
M
SD
M
67.0
SD
t score
1 3 . 8 -1.3411
df
62.4
16.4
BPsystolic
126.9
20.8
1 3 4 . 5 24.9
-1.5665
100
BP lowest systolic
110.4
17.1
1 2 0 . 4 21.2
-2.5048*
101
BY lowest diastolic
59.5
10.8
63.8
Braden score
18.2
2.4
16.6
1 0 . 4 -1.8418 3,0
2.9652**
101
100 101
*p = .01; * * p = .0038.
of 18 or greater, 41% of the patients who were not predicted to develop a pressure ulcer developed one (false-negative rate). Fifty-nine percent who were not predicted to develop a pressure ulcer did not do so (specificity). A score of 17 or below was determined as the cut-off point to implement standards of care for pressure ulcer prevention. DISCUSSION
The incidence of pressure ulcers reported in this study is higher than in previous studies in acute care settings (Allman et ai., 1986; Anderson, 1982; Langemo, et al., 1991; National Pressure Ulcer Advisory Panel, 1989). The higher incidence rate may be related to the exclusion of large numbers of patients with a length of stay of less than 7 days, the tertiary care setting, the type of patients entered into the study, and the inclusion of acute rehabilitation patients. The higher incidence rate in this study may also be related to lack of consistency regarding classification and use of stage I pressure ulcers in published incidence and prevalence rates. At times, stage I pressure ulcers are not included or different descriptors for stage I pressure ulcers are used. Stage I pressure ulcers were used in the calculation of the incidence rate in this study. Pressure ulcer development occurred between 3 and 13 days of hospitalization with a mean of 4.8 days. The span of days is consistent with reports by Langemo et al. (1991) of 3 to 16 days. Increased age is reported as a significant factor predicting pressure ulcer positive status (Bergstrom, Braden, Norvell, Lenaghan, & Boynton, 1988). Age was not a significant risk factor in the development of pressure ulcers in this study; how-
ever, the mean age for pressure ulcer positive subjects was 5 years higher than for pressure ulcer negative subjects. Bedside nurse ability to predict pressure ulcers was not confirmed in this study. Salvadalena et al. (1992) reported nurse judgement as more accurate than the Braden Scale. The current study used different pressure ulcer staging criteria than the Salvadalena et al (1992) study. Salvadalena et al. used continued redness after 10 minutes to determine stage I pressure ulcers, whereas the current study used nonblanchable erythema after 30 minutes to determine stage I pressure ulcers. Differences in pressure ulcer staging criteria is a possible explanation for the contradictory findings (Bergstrom, 1993). The most significant t-test finding that distinguished between pressure ulcer positive and pressure ulcer negative status was the Braden score. This finding is consistent with Bergstrom's work (Bergstrom, Braden, et al., 1987; Bergstrom, 1991, personal communication). The Braden Scale provides the tools to objectify and improve nurse judgment concerning risk of pressure ulcers. A score of 17 or less was determined as the Braden Scale cut-off point in this study. The decision to use a score of 17 represents a balance of the need to adequately predict pressure ulcer risk without overprediction. Underprediction can result in pressure ulcer development, while overprediction leads to misuse of nursing time and increased costs. The results of this study confirm that a cut-off point of 16 for the Braden Scale may not be a universal standard. Institutions considering the use of the Braden Scale may need to replicate this study to determine their own cut-off point. LIMITATIONS
Sixteen nurse data collectors completed interrater reliability training. Although data collectors worked in teams and usually followed up on their own individual subject assignments, the large number of data collectors may have influenced consistency in judgments concerning skin assessment and Braden Scale scores. Limitations of the study are nurse behaviors, ethical concerns, and pressure ulcer prevention strategies already in place. Bedside nurses may have been influenced to provide more vigilant nursing care by the frequent visits of data collectors to identified subjects. In addition, bedside nurses were frequently asked to use their
PREDICTIVE VALIDITY OF THE BRADEN SCALE
85
knowledge of a subject to assist in the Braden Scale assessment, which may have increased their awareness of a potential problem. Preventive measures were already in place for some patients and ethically it would not be appropriate to discontinue or change a practice already in place. Effective preventative measures could reduce a subject's risk of pressure ulcers as well as reduce sensitivity and increase specificity. CLINICAL IMPLICATIONS
This study d e t e r m i n e d the cut-off point, sensitivity, and specificity o f the B r a d e n Scale, and identified the B r a d e n S c a l e as m o r e useful than nurse j u d g m e n t alone. T h e B r a d e n Scale should be implem e n t e d with a c o m p l e t e understanding of what it does and does not do for patient care. It assists nurses in m a k i n g j u d g m e n t s about factors that predict risk o f pressure ulcer d e v e l o p m e n t . As this study d e m o n s t r a t e d the tool is not 100% accurate in prediction as s o m e subjects d e v e l o p e d pressure ulcers w h o w e r e not predicted to do so. W h e n planning care for individual patients, clinicians must r e m e m b e r the B r a d e n Scale can underpredict risk. The B r a d e n Scale m u s t be c o m b i n e d with frequent, thorough skin assessment practices and p r e v e n t i v e strategies. A s s u m i n g nurses c o m p l e t e physical assessment o f patients, the B r a d e n S c a l e does not require
additional assessment; rather, it requires use of assessment information in a prescribed manner. Hospital-wide implementation of the Braden Scale requires mobilization of effort and is not without expense. This cost is balanced against the expense of treating pressure ulcers, thepain and suffering of patients, the customer image resulting from "bedsores," and the satisfaction of nursing staff. In the past, more resources have been allocated towards treatment than towards the evaluation of the effectiveness and cost benefits of pressure ulcer prevention (Petro, 1990). The results of this study, along with the research agenda suggested by the Panel for the Prediction and Prevention of Pressure Ulcers in Adults (1992), support the need for continued research and refinement of risk detection tools in order to increase sensitivity and specificity. FOOTNOTE
IThe Braden Scale can be ordered by writing to Nancy Bergstrom, PhD, RN, FAAN, Interim Association, Dean Graduate Nursing Programs, University of Nebraska, Omaha, NE 68198-5330; or to Barbara J. Braden, PhD, RN, Dean of Graduate School, Creighton University Graduate School, 2500 California Plaza, Omaha, NE 68178. ACKNOWLEDGMENT
The authors wish to thank Dr. Nancy Bergstrom for her assistance with the study.
REFERENCES
Allman, R.M. (1989). Pressure ulcers among the elderly. The New England Journal of Medicine, 320, 850-853. Allman, R.M., Laprade, C.A., Noel, L.B., Walker, J.A., Moorer, C.A., Dear, M.R., & Smith, C.R. (1986). Pressure sores among hospitalized patients. Annals of Internal Medicine, 105, 337-342. Alterescu, V. (1989). The financial costs of inpatient pressure ulcers to an acute care facility. Decubitus, 2, 14-23. Anderson, K. (1982). Prevention of pressure sores by pressure ulcers identifying patients at risk. British Medical Journal, 284, 1370-1371. Bergstrom, N. (1993). Braden scale and clinical judgment. Journal of ET Nursing, 20, 133-134. Bergslrom, N., Braden, B.J., Laguzza, A., & Holman, V. (1987). The Braden Scale© for predicting pressure sore risk. Nursing Research, 36, 205-210. Bergstrom, N., Braden, B., Norvell, K., Lenaghan, R, & Boynton, R (1988). Diminished tissue tolerance: Influence on pressure sore development in the institutionalized elderly. Applied Nursing Research, l, 96. Bergstrom, N., Demuth, RJ., & Braden, B. (1987). A clinical trial of the Braden Scalee for predicting pressure sore risk. Nursing Clinics of North America, 22, 417-418.
Braden, B.J., & Bergstrom, N. (1987). Conceptual schema for the study of the etiology of pressure sores. Rehabilitation Nursing, 12, 8-12. Brad©n, B.J., & Bergstrom, N. (1989). Clinical utility of the Braden Scale © for predicting pressure sore risk. Decubitus, 2, 44-46, 50-51. Copeland-Fields, L.D., & Hosiko, B.R. (1989). Clinical validation of Braden and Bergstrom's conceptual schema of pressure sore risk factors. Journal of Enterostomal Therapy, 16, 257-260. Darby, S. (1990). A comparison of predictive validity between nurse perception and Braden Scale© score in identifying pressure sore risk. Unpublished manuscript, Mercy Hospital, Detroit, MI. Jones, S., Burger, C., Piraino, J., & Utley, S. (1993). A pressure ulcer prevention program. Ostomy/Wound Management, 39, 33-39. Langemo, D., Olson, B., Hunter, S., Hanson, D., Burd, C., & Cathcart-Silberberg, T. (1991). Incidence and prediction of pressure ulcers in five patient care settings. Decubitus, 4, 25-26, 28, 30. Lilienfeld, A.M., & Lilienfeld, D.E. (1980). Foundations of epidemiology (2nd ed.). New York: Oxford University Press.
86
Maklebust, J., Mondous, L., & Sieggreen, M. (1986). Pressure relief characteristics of various support surfaces used in prevention and treatment of pressure ulcers. Journal of Enterostomal Therapy, 13, 85-89. National Pressure Ulcer Advisory Panel. (1989). Consensus Development Conference Statement. Buffalo, NY: State University of New York. Panel for the Prediction and Prevention of Pressure Ulcers in Adults. (1992). Pressure ulcers in adults: Prediction and prevention. Clinical practice guideline, number 3. (AHCPR Publication No. 92-0047). Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services. Panel for the Prediction and Prevention of Pressure Ulcers. (1994). Treatment of pressure ulcers. Clinical practice guideline, number 15. (AHCRP Publication No. 95-0652. Rockville,
VANDENBOSCH ET AL
MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services. Petro, J. (1990). Ethical dilemmas of pressure ulcers. Decubitus, 3, 28-31. Salvadalena, G., Synder, M., & Brodgon, K. (1992). Clinical trial of the Braden Scale on an acute care medical unit. Journal of ETNursing, 19, 160-165. Seller, W.O., & Stahelin, H.B. (1985). Decubitus ulcers: preventive techniques for the elderly patient. Geriatrics, 40, 53-60. Van Ness, C. (1989). The implementation of a quality assurance study and program to reduce the incidence of hospital acquired pressure ulcers. Journal of Enterostomal Therapy, 16, 61-64.
Before the implementation of the Braden Scale, an institutional study was conducted to determine the cut-off point at which patients would be at risk for developing a pressure ulcer. Nurses' perception of patient risk for pressure ulcer development also was compared with the Braden Scale. One hundred and three subjects hospitalized for a minimum of 7 days and randomly selected from routine hospital admissions were rated with the Braden Scale, and skin assessments were made three times per week for up to 2 weeks starting 24 to 48 hours after admission. In addition, bedside nurses were asked to use clinical judgment to identify subjects at risk for pressure ulcer development. Twenty-nine subjects developed pressure ulcers. Nurse's judgements of pressure ulcer risk were not significant in predicting pressure ulcer positive or negative status. Results of the t test demonstrated the Braden Scale score is the most highly significant finding (p = .0038) to predict pressure ulcer positive and pressure ulcer negative groups. For this study, the Braden Scale cut-off point was set at 17 with a sensitivity of .59 and a specificity of .59. Clinical implementation of the Braden Scale must be combined with frequent and thorough skin assessment practices because some patients will develop pressure ulcers even though the tool does not predict the patients to be at risk.
Copyright © 1996 by W.B. Saunders Company
A
SYSTEMATIC method to assist the caregiver in predicting patients at risk for developing pressure ulcers is a useful clinical tool. In the process of developing institutional standards for pressure ulcer development, the authors reviewed several published tools that predict pressure ulcer risk and selected the Braden Scale 1 for this study. The Braden Scale was used because it is highly sensitive and has a lower rate of overprediction than other scales (Bergstrom, Braden, Laguzza, & From Mission Health, St. Joseph Mercy Hospital, Ann Arbor, MI; Michigan Peer Review Organ&ation, Plymouth, MI; and the Visiting Nurses'Association, Port Huron, MI. Terry VandenBosch, MS, RN, CS: Research Specialist, Mission Health, St. Joseph Mercy Hospital, Ann Arbor, MI; Cecelia Montoye, MSN, RN: Senior Staff Associate for Education and Feedback, Michigan Peer Review Organization, Plymouth, M1; Martha Satwicz, MSN, RN: Senior Staff Associate for Education and Feedback, Michigan Peer Review Organization, Plymouth, M1; Karen Durkee-Leonard, MSN, ICN: Program Coordinator, Rehabilitation Services, Visiting Nurses' Association, Port Huron, MI; Barbara Boylan-Lewis, MA, RN, CETN: Enterostomal Therapy Nurse, Mission Health, St. Joseph Mercy Hospital, Ann Arbor, ML Address reprint requests to Terry VandenBosch MS, RN, CS, Nursing Development Services, 5301 East Huron River Dr, P.O. Box 992, Ann Arbor, M148106. Copyright © 1996 by W.B. Saunders Company 0897-1897/96/0902-000755.00/0 80
Holman, 1987). The registered nurse (RN) caregiver uses the Braden Scale to assess the patient on admission and periodically thereafter. Each patient is scored according to criteria on the scale and then given an overall numerical score that represents risk of pressure ulcer development. The lower the overall numerical score, the more risk to the patient. Institutions that use the Braden Scale must decide on a cut-off point (numerical value) at which the patient is determined to be at risk. The cut-off point signifies the need for implementation of the institution's standard of care for pressure ulcer prevention. Preventive care is targeted to those patients who are at risk and can reduce the time and cost of implementing preventive care for patients not at risk. The developers of the Braden Scale recommend conducting a study to calculate the institutional critical cut-off point for risk and to determine predictive validity of the tool. Therefore, a replication study was conducted at the institution. BACKGROUND
Pressure ulcers are a serious problem in the United States. Incidence of pressure ulcers in hospitals has been reported from 2.7% to 29.5% (Panel for Treatment of Pressure Ulcers, 1994). Usual estimates of hospital pressure ulcer inciApplied Nursing Research,Vol. 9, No. 2 (May), 1996: pp 80-86
PREDICTIVE VALIDITY OF THE BRADEN SCALE
dence rates are reported at 1% to 5% (Allman et al., 1986; Anderson, 1982; National Pressure Ulcer Advisory Panel, 1989) with many more patients identified as being at risk. The in-hospital mortality rate for patients with pressure ulcers is estimated at 23% to 37% (Allman, 1989; Allman et al., 1986). Pressure ulcers develop quickly and heal slowly. Deep lesions require weeks or months to heal and sometimes require surgical closure. Complications include bacteremia, sepsis, and osteomyelitis (Allman, 1989). Pressure ulcers can also be painful and are an emotional burden for patients and families. The cost of treating pressure ulcers is high. Studies report that a single pressure ulcer costs from $1,300 to $35,000 (Alterescu, 1989; Van Ness, 1989). Estimations of cost depend on the severity and stage of the pressure ulcer and on the method used to analyze costs. Consistent methods for determining costs have not been determined (Alterescu, 1989). In 1986, 3.5 to 7 billion dollars was estimated as the annual cost of pressure ulcer treatment (Maklebust, Mondous, & Sieggreen, 1986). The causes of pressure ulcers are complex. Critical determinates of pressure ulcers are factors that affect the intensity and duration of pressure and factors that affect the tolerance of the skin and supporting structures for pressure. Mobility status, activity, and sensory perception are factors related to intensity and duration of pressure. Tissue tolerance is affected by extrinsic factors of skin moisture, friction, shear, and by intrinsic factors such as nutritional status, age, and arteriolar pressure (Braden & Bergstrom, 1987; Copeland-Fields & Hosiko, 1989). Prevention and care of pressure ulcers are important. However, the first aim is to target measures that will prevent pressure ulcers. When preventive programs are instituted, a reduction in the incidence of pressure ulcers of up to 50% has been reported (Anderson, 1982; Seiler & Stahelin, 1985). The Braden Scale identifies critical factors in pressure ulcer formation (Braden & Bergstrom, 1989). The scale is made up of six subscales: mobility, activity, sensory perception, moisture, nutrition, and friction and shear. Each subscale contains a title, a key concept descriptor, and a brief sentence description of qualifying attributes. Subscales are scored by level of mobility, activity, etc. The highest possible score is 23 points. The lower the score the greater the risk of pressure ulcer
81
development. Cut-off points are used to predict risk and institute preventive protocols. They provide a balance between the cost of prevention and the cost of treatment. A cut-off score of 16 or less is suggested by Braden and Bergstrom (1989). Darby (1990) compared nurse perception of patient risk for pressure ulcers with the Braden Scale. Nurse perception alone was not as accurate as the Braden Scale in predicting patients at risk. Salvadalena, Synder, and Brodgon (1992) found the Braden Scale was less accurate than nurse judgment in a clinical trial on an acute care medical unit. Jones, Burger, Piraino, and Utley (1993) and Van Ness (1989) demonstrated a decrease in prevalence and incidence or pressure ulcers following implementation of pressure ulcer prevention programs that used risk assessment tools. The results of these studies, along with the Agency for Health Care Policy and Research Clinical Practice Guideline (Panel for the Prediction and Prevention of Pressure Ulcers in Adults, 1992), suggest that the use of a systematic research-based tool improves the clinical judgment of nurses in predicting pressure ulcer development. Thus, the purpose of this study was twofold: to determine the cut-off point, sensitivity, and specificity for use of the Braden Scale in a specific inpatient hospital setting; and to compare nurses' perception of risk of pressure ulcer development with and without the Braden Scale. METHODS Subjects The study took place in a 550-bed tertiary care community teaching hospital. Subjects were drawn from general care, intensive care, and a 40-bed inpatient rehabilitation unit. Hospitalized inpatients over 18 years of age with intact skin and no reddened areas over any bony prominence were asked to participate. Using a random numbers table and the daily list of admissions, subjects with an expected length of stay of at least 1 week were selected. Subjects were selected three times a week and were approached for informed consent within 48 hours of admission to the hospital. Instruments Three instruments, the Braden Scale, a skin assessment form, and a tool to record the subject's clinical status, were used to collect data. All of the tools were developed and used by Bergstrom and
82
Braden in previous studies (Braden & Bergstrom, 1987). Nurse perception of pressure ulcer risk also was used.
Braden Scale The Braden Scale is a sensitive (83% to 100%) and specific (64% to 90%) instrument (Bergstrom, Braden, et al., 1987; Bergstrom, et al., 1987). Sensitivity and specificity are indicators of predictive validity. Reliability studies indicate the instrument is also highly reliable. Reliability for nurses aides and licensed practical nurses ranged from r = .83 to r = .94 (Bergstrom, Braden, et al., 1987). Reliability for RNs has been reported at r = .99 (Bergstrom, Braden, et al., 1987). Bergstrom, Braden, et al. (1987) assessed validity and reliability of the Braden Scale in acute care, rehabilitation, and skilled nursing facility settings.
Skin Assessment Form A skin assessment form was used to document the status of the hospitalized patient's skin over every bony prominence, to document the presence and location of a pressure ulcer, and tO record the staging of any ulcers that were present. The form contained a human body diagram with a number at every potential pressure ulcer site. The form has construct validity and captures 99% of the possible pressure ulcer sites (Bergstrom, 1994, personal communication). A classification schema of four pressure ulcer stages was used: stage I was defined as "nonblanchable erythema of intact skin that does not disappear for 24 hours after pressure is relieved"; stage II was defined as a break in skin such as blisters or abrasions; stage III was defined as a break in skin exposing subcutaneous tissue; and stage IV was defined as a break in skin extending through tissue and subcutaneous layers exposing muscle or bone. Clinical Status
Information about subject's clinical status was collected. Data collected were age, gender, history of smoking, diagnosis, and systolic and diastolic blood pressure.
Nurse Perception of Pressure Ulcer Risk Nurse perception of pressure ulcer risk was measured with a question developed by Darby (1990) and used by Salvadalena et al. (1992). At the admission rating/observation period, the staff nurse
VANDENBOSCH ET AL
caring for the study patient was asked to answer "yes" or "no" to the question: "Do you think this patient will develop a pressure ulcer while they are in the hospital?" Procedure
Selected patients who had been admitted 24 to 48 hours before a data collection day were approached and asked for informed consent. Only those patients whose length of stay was estimated as 7 days or longer were approached. Data were collected on Mondays, Wednesdays, and Fridays. All subjects had a minimum of three assessments and a maximum of six. The research data collectors were nurses on the staff of the hospital. Interrater reliability was conducted 1 week before implementation of the study and all ratings were at a 90% or higher level of agreement. Eleven nurses were trained as consistent data collectors and five additional nurses were trained as relief data collectors. To avoid bias on the part of the data collectors, two data collection roles were used. The first, Nurse I, was trained to use the Braden Scale. The second, Nurse II, was trained to assess subjects' skin and pressure ulcer status. Eight nurses were trained for the Nurse I procedures and eight were trained for the Nurse II procedures. Subjects were assessed by Nurse I and Nurse II on each data collection day.
Nurse I 1. Obtained informed verbal consent and completed demographic data within 48 hours of admission to the hospital. 2. Used the Braden Scale to rate the subject at the onset of the study and every Monday, Wednesday, and Friday for 2 weeks or until the subject was discharged. Nurses were instructed not to give data about the subject's risk for pressure ulcer development to the nurse caregiver.
Nurse II 1. Completed the skin assessment form on the day of entry into the study and every Monday, Wednesday, and Friday until discharge. If a nurse caregiver asked the nurse collecting data for information, the nurse could inform the caregiver if a pressure ulcer was present. 2. Asked the staff nurse assigned to the subject for the day about the subject's risk for
PREDICTIVE VALIDITY OF THE BRADEN SCALE
83
Table 2. Demographic Data
pressure ulcer development at the time of the first skin assessment. 3. Collected information on the patients' clinical status.
One hundred three subjects participated. Twentynine subjects developed pressure ulcers. First occurrence of pressure ulcers ranged from 3 to 13 days after hospital admission with a mean of 4.8 days and a SD of 1.9 days. The coccyx/sacrum area was the most common site. Four pressure ulcers on the ear were reported. Sites and pressure ulcer staging are listed in Table 1. The number of pressure ulcer sites per subject ranged from one to eight. Demographic data by pressure ulcer positive and pressure ulcer negative status is included in Table 2. The sample included mostly White patients. Females and former smokers had a higher percent of pressure ulcers. However, there was no significant increase in the odds of developing pressure ulcers associated with gender or smokers. Nurse prediction about risk for developing pressure ulcers is described in Table 2. Nurses correctly predicted 51.7% of the subjects would develop a pressure ulcer. They incorrectly predicted that 41.1% of the patients who did not develop pressure ulcers would develop them. These predictions were not significant when analyzed using chi-square analysis. Results of t-test of selected variables by pressure ulcer positive and pressure ulcer negative status are presented in Table 3. There was no significant Table 1. Pressure Ulcer Sites
Coccyx/sacrum Elbow Heel Ear Vertebrae Scapula Toe Ankle Trochanter Knee Lower leg Back of head Ischial tuber Other
Stage 1
Stage 2
Stage 3
Frequency %
Frequency %
Frequency %
13 8 4 3 3 2
18,8 11.5 5.7 4.3 4.3 2.9
1
1.5
1 1 1
1.5 1.5 1.5
Frequency
Pressure Ulcer Positive (n = 29)
Percent Frequency Percent
Gender
RESULTS
Site
Pressure Ulcer Negative (n = 74)
5 5 7 1 1
7.2 10.0 1.5 1.5
2 1 1 1
2.9 1,5 1.5 1.5
1
1.5
10
14.4
There were no stage 4 pressure ulcers.
1
1.5
I
1.5
Male
39
52.7
10
34.5
Female
35
47.3
19
65.5
White Black
63 9
85.1 12.2
25 3
89.3 10.7
Other
2
2,7
--
--
33
44.6
8
27.6
10 31
13.5 41.9
4 17
13.8 58.6
43 30
58.9 41.1
14 15
48.3 51.7
Race
Smoke Never Current Former Nurse prediction of pressure ulcer No Yes
A total n less than 103 indicates missing data,
difference between the ages of patients who were pressure ulcer positive and those who were pressure ulcer negative. There was a significant difference between the pressure ulcer positive and pressure ulcer negative subjects on lowest systolic blood pressure with pressure ulcer positive patients having higher systolic readings. The Braden Scale score was the most highly significant finding (p = .0038) between the pressure ulcer positive and pressure ulcer negative groups. The Braden Scale score used for t-test analysis was the second observation period for pressure ulcer negative subjects and the observation before the first pressure ulcer occurrence for pressure ulcer positive subjects. Sensitivity and specificity are common predictive tools in research. Sensitivity is the percentage of subjects who develop pressure ulcers and whom the scale predicted would develop them. Specificity is the percentage of those who do not develop pressure ulcers and whom were not predicted to develop them. A standard formula for determining sensitivity and specificity, as reported in Lilienfeld and Lilienfeld (1980) and as described in Bergstrom, Braden, et al. (1987), was used in this study. The cut-off point for assessing risk was determined using the guidelines in Bergstrom, Demuth, and Braden (1987). At a score of 17, 59% of the subjects predicted to develop pressure ulcers did develop a pressure ulcer (sensitivity), whereas 41% of the subjects predicted to develop a pressure ulcer did not develop one (false-positive rate). At a score
84
VANDENBOSCH ET AL
Table 3. t-Test Results of Selected Variables by Pressure Ulcer Positive and Negative Pressure Ulcer PressureUlcer Negative Positive Variables Age
M
SD
M
67.0
SD
t score
1 3 . 8 -1.3411
df
62.4
16.4
BPsystolic
126.9
20.8
1 3 4 . 5 24.9
-1.5665
100
BP lowest systolic
110.4
17.1
1 2 0 . 4 21.2
-2.5048*
101
BY lowest diastolic
59.5
10.8
63.8
Braden score
18.2
2.4
16.6
1 0 . 4 -1.8418 3,0
2.9652**
101
100 101
*p = .01; * * p = .0038.
of 18 or greater, 41% of the patients who were not predicted to develop a pressure ulcer developed one (false-negative rate). Fifty-nine percent who were not predicted to develop a pressure ulcer did not do so (specificity). A score of 17 or below was determined as the cut-off point to implement standards of care for pressure ulcer prevention. DISCUSSION
The incidence of pressure ulcers reported in this study is higher than in previous studies in acute care settings (Allman et ai., 1986; Anderson, 1982; Langemo, et al., 1991; National Pressure Ulcer Advisory Panel, 1989). The higher incidence rate may be related to the exclusion of large numbers of patients with a length of stay of less than 7 days, the tertiary care setting, the type of patients entered into the study, and the inclusion of acute rehabilitation patients. The higher incidence rate in this study may also be related to lack of consistency regarding classification and use of stage I pressure ulcers in published incidence and prevalence rates. At times, stage I pressure ulcers are not included or different descriptors for stage I pressure ulcers are used. Stage I pressure ulcers were used in the calculation of the incidence rate in this study. Pressure ulcer development occurred between 3 and 13 days of hospitalization with a mean of 4.8 days. The span of days is consistent with reports by Langemo et al. (1991) of 3 to 16 days. Increased age is reported as a significant factor predicting pressure ulcer positive status (Bergstrom, Braden, Norvell, Lenaghan, & Boynton, 1988). Age was not a significant risk factor in the development of pressure ulcers in this study; how-
ever, the mean age for pressure ulcer positive subjects was 5 years higher than for pressure ulcer negative subjects. Bedside nurse ability to predict pressure ulcers was not confirmed in this study. Salvadalena et al. (1992) reported nurse judgement as more accurate than the Braden Scale. The current study used different pressure ulcer staging criteria than the Salvadalena et al (1992) study. Salvadalena et al. used continued redness after 10 minutes to determine stage I pressure ulcers, whereas the current study used nonblanchable erythema after 30 minutes to determine stage I pressure ulcers. Differences in pressure ulcer staging criteria is a possible explanation for the contradictory findings (Bergstrom, 1993). The most significant t-test finding that distinguished between pressure ulcer positive and pressure ulcer negative status was the Braden score. This finding is consistent with Bergstrom's work (Bergstrom, Braden, et al., 1987; Bergstrom, 1991, personal communication). The Braden Scale provides the tools to objectify and improve nurse judgment concerning risk of pressure ulcers. A score of 17 or less was determined as the Braden Scale cut-off point in this study. The decision to use a score of 17 represents a balance of the need to adequately predict pressure ulcer risk without overprediction. Underprediction can result in pressure ulcer development, while overprediction leads to misuse of nursing time and increased costs. The results of this study confirm that a cut-off point of 16 for the Braden Scale may not be a universal standard. Institutions considering the use of the Braden Scale may need to replicate this study to determine their own cut-off point. LIMITATIONS
Sixteen nurse data collectors completed interrater reliability training. Although data collectors worked in teams and usually followed up on their own individual subject assignments, the large number of data collectors may have influenced consistency in judgments concerning skin assessment and Braden Scale scores. Limitations of the study are nurse behaviors, ethical concerns, and pressure ulcer prevention strategies already in place. Bedside nurses may have been influenced to provide more vigilant nursing care by the frequent visits of data collectors to identified subjects. In addition, bedside nurses were frequently asked to use their
PREDICTIVE VALIDITY OF THE BRADEN SCALE
85
knowledge of a subject to assist in the Braden Scale assessment, which may have increased their awareness of a potential problem. Preventive measures were already in place for some patients and ethically it would not be appropriate to discontinue or change a practice already in place. Effective preventative measures could reduce a subject's risk of pressure ulcers as well as reduce sensitivity and increase specificity. CLINICAL IMPLICATIONS
This study d e t e r m i n e d the cut-off point, sensitivity, and specificity o f the B r a d e n Scale, and identified the B r a d e n S c a l e as m o r e useful than nurse j u d g m e n t alone. T h e B r a d e n Scale should be implem e n t e d with a c o m p l e t e understanding of what it does and does not do for patient care. It assists nurses in m a k i n g j u d g m e n t s about factors that predict risk o f pressure ulcer d e v e l o p m e n t . As this study d e m o n s t r a t e d the tool is not 100% accurate in prediction as s o m e subjects d e v e l o p e d pressure ulcers w h o w e r e not predicted to do so. W h e n planning care for individual patients, clinicians must r e m e m b e r the B r a d e n Scale can underpredict risk. The B r a d e n Scale m u s t be c o m b i n e d with frequent, thorough skin assessment practices and p r e v e n t i v e strategies. A s s u m i n g nurses c o m p l e t e physical assessment o f patients, the B r a d e n S c a l e does not require
additional assessment; rather, it requires use of assessment information in a prescribed manner. Hospital-wide implementation of the Braden Scale requires mobilization of effort and is not without expense. This cost is balanced against the expense of treating pressure ulcers, thepain and suffering of patients, the customer image resulting from "bedsores," and the satisfaction of nursing staff. In the past, more resources have been allocated towards treatment than towards the evaluation of the effectiveness and cost benefits of pressure ulcer prevention (Petro, 1990). The results of this study, along with the research agenda suggested by the Panel for the Prediction and Prevention of Pressure Ulcers in Adults (1992), support the need for continued research and refinement of risk detection tools in order to increase sensitivity and specificity. FOOTNOTE
IThe Braden Scale can be ordered by writing to Nancy Bergstrom, PhD, RN, FAAN, Interim Association, Dean Graduate Nursing Programs, University of Nebraska, Omaha, NE 68198-5330; or to Barbara J. Braden, PhD, RN, Dean of Graduate School, Creighton University Graduate School, 2500 California Plaza, Omaha, NE 68178. ACKNOWLEDGMENT
The authors wish to thank Dr. Nancy Bergstrom for her assistance with the study.
REFERENCES
Allman, R.M. (1989). Pressure ulcers among the elderly. The New England Journal of Medicine, 320, 850-853. Allman, R.M., Laprade, C.A., Noel, L.B., Walker, J.A., Moorer, C.A., Dear, M.R., & Smith, C.R. (1986). Pressure sores among hospitalized patients. Annals of Internal Medicine, 105, 337-342. Alterescu, V. (1989). The financial costs of inpatient pressure ulcers to an acute care facility. Decubitus, 2, 14-23. Anderson, K. (1982). Prevention of pressure sores by pressure ulcers identifying patients at risk. British Medical Journal, 284, 1370-1371. Bergstrom, N. (1993). Braden scale and clinical judgment. Journal of ET Nursing, 20, 133-134. Bergslrom, N., Braden, B.J., Laguzza, A., & Holman, V. (1987). The Braden Scale© for predicting pressure sore risk. Nursing Research, 36, 205-210. Bergstrom, N., Braden, B., Norvell, K., Lenaghan, R, & Boynton, R (1988). Diminished tissue tolerance: Influence on pressure sore development in the institutionalized elderly. Applied Nursing Research, l, 96. Bergstrom, N., Demuth, RJ., & Braden, B. (1987). A clinical trial of the Braden Scalee for predicting pressure sore risk. Nursing Clinics of North America, 22, 417-418.
Braden, B.J., & Bergstrom, N. (1987). Conceptual schema for the study of the etiology of pressure sores. Rehabilitation Nursing, 12, 8-12. Brad©n, B.J., & Bergstrom, N. (1989). Clinical utility of the Braden Scale © for predicting pressure sore risk. Decubitus, 2, 44-46, 50-51. Copeland-Fields, L.D., & Hosiko, B.R. (1989). Clinical validation of Braden and Bergstrom's conceptual schema of pressure sore risk factors. Journal of Enterostomal Therapy, 16, 257-260. Darby, S. (1990). A comparison of predictive validity between nurse perception and Braden Scale© score in identifying pressure sore risk. Unpublished manuscript, Mercy Hospital, Detroit, MI. Jones, S., Burger, C., Piraino, J., & Utley, S. (1993). A pressure ulcer prevention program. Ostomy/Wound Management, 39, 33-39. Langemo, D., Olson, B., Hunter, S., Hanson, D., Burd, C., & Cathcart-Silberberg, T. (1991). Incidence and prediction of pressure ulcers in five patient care settings. Decubitus, 4, 25-26, 28, 30. Lilienfeld, A.M., & Lilienfeld, D.E. (1980). Foundations of epidemiology (2nd ed.). New York: Oxford University Press.
86
Maklebust, J., Mondous, L., & Sieggreen, M. (1986). Pressure relief characteristics of various support surfaces used in prevention and treatment of pressure ulcers. Journal of Enterostomal Therapy, 13, 85-89. National Pressure Ulcer Advisory Panel. (1989). Consensus Development Conference Statement. Buffalo, NY: State University of New York. Panel for the Prediction and Prevention of Pressure Ulcers in Adults. (1992). Pressure ulcers in adults: Prediction and prevention. Clinical practice guideline, number 3. (AHCPR Publication No. 92-0047). Rockville, MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services. Panel for the Prediction and Prevention of Pressure Ulcers. (1994). Treatment of pressure ulcers. Clinical practice guideline, number 15. (AHCRP Publication No. 95-0652. Rockville,
VANDENBOSCH ET AL
MD: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services. Petro, J. (1990). Ethical dilemmas of pressure ulcers. Decubitus, 3, 28-31. Salvadalena, G., Synder, M., & Brodgon, K. (1992). Clinical trial of the Braden Scale on an acute care medical unit. Journal of ETNursing, 19, 160-165. Seller, W.O., & Stahelin, H.B. (1985). Decubitus ulcers: preventive techniques for the elderly patient. Geriatrics, 40, 53-60. Van Ness, C. (1989). The implementation of a quality assurance study and program to reduce the incidence of hospital acquired pressure ulcers. Journal of Enterostomal Therapy, 16, 61-64.