A Comparison of Patient Risk for Pressure Ulcer Development with Nursing Use of Preventive Interventions George C. Xakellis, MD,* Rita A. Franfz, PhD, RN,t Manuel Arteaga, MS,* Man Nguyen, BS, and Anne Lewis, MA, R N t Objective: (1) Determine if the Braden scale or Norton scale predicted the same patients to be at risk for pressure ulcer development as were receiving preventive nursing interventions. (2) Identify the items on the Braden and Norton risk assessment scales that the nurses used intuitively to determine a patient's need for a preventive intervention. Design: Cross-sectional study. Setting: Six hundred-bed, state-supported, long-term care facility. Patients: War veterans who were 82% male and 97% caucasian, mean age 73. Measurements: (1) Patients were categorized as at-risk or not-at-risk by the Norton and Braden scales. (2) The presence of a preventive nursing intervention was noted. Agreement in assignment of at-risk status among the two assessments and presence of a preventive intervention was analyzed using Cohen's Kappa. (3) The staff nurses' use of preventive inter-
ventions was modeled using stepwise logistic regression. The items from the Braden and Norton risk assessment scales were used as independent variables with staff nurse implementation of a preventive intervention as the dependant variable. ver 1 million people currently have pressure ulcers, and with the number of frail elderly persons increasing, the prevalence of pressure ulcers is likely to rise.' Prevention rather than treatment appears to be the key to reducing the number of pressure ulcers. As a consequence, numerous preventive strategies have been p r ~ p o s e d . ~Fundamental -~ to all of these strategies has been some method of determining which patients are at risk for developing pressure ulcers.6-'6 In an effort to standardize pressure ulcer prevention methods, the Agency for Health Care Policy and Research (AHCPR) of the US Public Health Service recently released the clinical practice guideline, Pressure Ulcers in Adults: Prediction and Pre~ention.'~ An essential component of this guideline is the recommendation that all bedbound or chairbound patients be assessed for pressure ulcer risk. The guideline recommends the use of one of two risk assessment scales for conducting
0
From the *Department of Family Practice, College of Medicine, and the tCollege of Nursing, The University of Iowa, Iowa City, Iowa. This project was supported in part by grants from the U S . Public Health Service (Grants for Predoctoral Training in Family Medicine, USPHS/National Institutes of Health, No. 5D15-PE87007-10) and The University of Iowa (Interdisciplinary Research Assistantships Program). Address correspondence to George C. Xakellis, MD, Associate Professor Dept. of Family Practice, The University of Iowa College of Medicine, 2134 Steindler Bldg., Iowa City, I A 52242.
IAGS 4O:I 250-1 254, 1992 0 1992 by the American Geriatrics Sociefy
Results: Nurse preventive interventions were found on 45% of patients. The Norton scale identified 38% and the Braden scale identified 27% of patients as at-risk. Agreement among the three methods was 0.53. Agreement between the Braden and Norton scales was 0.73. Agreement between use of a preventive intervention and a classification as at-risk by the Braden or Norton scale was 0.41 and 0.43, respectively.
Stepwise logistic regression revealed that low Braden mobility scores (Odds Ratio: 2.74) and low Braden friction/shear scores (Odds Ratio: 3.29) were associated with an increased likelihood of a patient receiving a preventive nursing intervention. Conclusions: The overall level of agreement among the two scales predicting risk and the presence of a preventive intervention was not high. Agreement, however, between the two risk assessment scales was close. The staff nurses apparently relied on a patients' mobility, their exposure to friction/shear, and additional unidentified factors to guide implementation of a preventive intervention. Further study is needed to define the cost, efficacy, and related cost effectiveness of routine pressure ulcer risk assessment. J Am Geriatr SOC 40:12501254,1992 this systematic assessment: the Norton Scale or the Braden Scale for Predicting Pressure Sore Risk. Both scales have been tested in a variety of settings and patient populations. Although the performance of these tools in predicting pressure ulcer development has been variable, incorporating them into clinical practice ensures a systematic approach to assessment of individual risk factors. This recommendation represents a significant departure from what has been routine practice in most long-term care facilities. Historically, the responsibility for pressure ulcer prevention has rested with nursing staff, who have used their intuitive clinical judgement to determine patients' level of risk and their need for prevention measures. It is unknown what effect implementation of a standard risk assessment tool will have on numbers of patients judged to be at risk. The AHCPR clinical practice guideline further recommends that preventive measures be instituted for all patients assessed to be at risk for pressure ulcer development. Since such preventive measures create additional expenditures for health-care facilities, determining the relative impact of using the scales versus allowing the nursing staff to use their clinical judgement is essential. To date, the relative performance of the Norton Scale in predicting patient risk has not been compared with that of the Braden scale on the same population of patients. Furthermore, the staff nurses' intuitive judgement of risk, as opera-
0002-861 4/92/$3.50
PRESSURE ULCER RISK VS PREVENTIVE INTERVENTION
/AGS-DECEMBER 7 992-VOL. 40. NO.I 2
tionalized by the use of preventive interventions such as pressure reducing mattresses and turning schedules, has not been compared with the at-risk predictions of these scales. This study was undertaken to determine if the Norton scale and the Braden scale predicted the same patients to be at risk for pressure ulcer development as were receiving preventive nursing interventions. The questions to be answered were: (1)is there a high level of agreement among these three methods of categorizing patients’ risk for pressure ulcer development, and ( 2 ) do staff nurses who have not been trained in the use of formal risk assessment scales intuitively use components of the Braden Scale or Norton Scale to determine if a patient should receive a preventive intervention.
METHODS Setting and Subjects The study was conducted in a 600-bed, state-supported, long-term care facility that serves primarily veterans. This facility was chosen because of its large size, its low prevalence of pressure ukers, and its lack of experience with use of a pressure ulcer risk assessment scale. Further, at this facility the staff nurses make the decisions to implement preventive interventions such as a patient repositioning schedule or a pressure-reducing mattress. A physician’s order is not necessary. Internal audits by the nursing department during the preceding year revealed a 476-576 prevalence of Stage 11-IV pressure ulcers in the study facility. Sixty-seven patients on two nursing units were used to pilot-test the study methods, and all remaining patients in this long-term care facility were assessed for risk of developing pressure ulcers according to the protocol. Patients were excluded from the study if they had a pressure ulcer at the time of initial risk assessment. Risk Assessment Instruments The Norton risk assessment scale is composed of five clinical items, each rated on a 1-4 scale with 1 representing a poor clinical condition and 4 representing a good clinical condition.”, The five items are: physical condition, mental condition, activity, mobility, and incontinence. The Norton risk assessment scale has a total possible score of 20 and defines patients as at-risk if their score is 14 or less (Figure 1).
Physical condition
Mental condition
1251
The Braden scale for predicting pressure sore risk is composed of six clinical item^.'^ Five of the items are rated on a 1-4 scale, with 1 representing a poor clinical condition and 4 representing a good clinical condition. These five items are: (1)sensory perception, defined as the patient’s ”ability to respond meaningfully to pressure-related discomfort”; (2) moisture, defined as the “degree to which skin is exposed to either incontinence, perspiration, or wound drainage”; (3) activity, which quantifies the patient’s ”degree of physical activity” from bedfast to walks frequently; (4) mobility, which rates the patient’s ”ability to change and control body position”; and (5) nutrition, which quantifies the patient’s ”usual food intake pattern.” The sixth Braden item is friction/shear and is rated on a 1-3 scale, with 1 indicating inability to lift freely without sliding across linen and 3 meaning the patient has sufficient muscle strength to lift completely off the supporting surface when repositioning. The scores from each of the items are totalled. The maximum possible score is 23, and patients are at risk if their total score is 16 or less (Figure 2).l2-I4 The sensitivity and specificity of these risk assessment tools has been reported elsewhere.lO,12, 13, 15. 18.19
Study Protocol A research nurse was trained in the use of the Braden and Norton risk assessment and performed all risk assessments using these scales. Staff nurses were consulted when a rating for a particular scale item (such as amount of incontinence or level of nutritional intake) could not be determined from a single patient assessment and a review of the record by the research nurse. The Braden risk assessment was performed on all patients on a particular nursing ward first. After an interval of at least 1 day and without reference to the Braden score, risk assessment was repeated using the Norton scale. Given the large number of risk assessments that were performed on a given day (30-60 subjects reside on any given nursing ward), it was felt that contamination of the second risk score by the first could be minimized by allowing at least 1 day between assessments. All Norton scale risk assessments were performed within 2 weeks of Braden scale assessments. During risk assessment, the research nurse also noted whether a pressure-reducing mattress was present on the subject’s bed and noted if the subject had a repositioning schedule documented in his/her treatment cardex. The patient was classified as
Incontinent
Mobility
Activity
Total Good
Fair
4.
3 Poor 2 VeryBad 1
Alert
4
Apathetic 3
confused 2
Stupor
1
Ambulant
4
Wak/help 3 Chairbound 2 Bcdbound 1
Full
4
SLlimited 3 Veryhited 2 Immobile
1
None
4
Scnre
3 usuauy/urinc 2 Doubly 1
Occasional
Name/Date
FIGURE 1. Norton Scale. From Norton D, McLaren R, Exton-Smith AN. An Investigation of Geriatric Nursing Problems in Hospitals. London: National Corporation for the Care of Old People, 1962 (reprinted by Churchill Livingston, Edinburgh, 1975). Reproduced by permission of NCCOP (now known as the Centre for Policy on Ageing).
1252
XAKELLIS ET A L
SEWWRY tERCEPTlOW
1. Ud#: Unresponsive (does not moan. flinch. or grasp) to ability to respond mem- painful stimuli. due to ingfully to pressure-re- diminished level of conlated discomfort sciousness or sedation. OR limited ability to feel pain over most of body Surlace.
IAGS-DECEMBER 1992-VOL. 40, NO. 1 2
2. V a q UuIW Responds only to painful stimuli. Cannot communicate discomfort except by moaning or restlessness.
3. tll#MIlUUltJ:
4. hlapahent
Responds to verbal commands. but cannot always communicate discomfort or need to be turned.
Responds to verbal commands. Has no sensory deficit which would limit ability to feel or voice pain or discomfort.
OR
OR has a sensory impairment which limits the ability to feel pain or discomfort over 112 of body.
has some sensory impairment which limits ability to feel pain or discomfort in 1 or 2 extremities.
2. Vtryldrt Skin is often. but not always moist. Linen must be changed at least once a shift.
3. ossulrrlly ud8t Skin is occasionally moist. requiring an extra linen change approximately once a day.
4. lhnll Uabt
1. Wat Confined to W
2. rnldwt Ability to walk severely limited or non-existent. Cannot bear own weight andlor must be assisted into chair or wheelchair.
3. W l l b accaldly: Walks occasionally during day, but for very short distances. with or without assistance. Spends majority of each shift in bed or chair.
4. w m Fmmaar Walks outside the room at least twice a day a d inside room at least O m every 2 hours during waking hours.
mourn
1. Cwphmly larllk:
2. bq UmltJ:
ability to change and control body position
Doas not make even slight changes in body or extremity position without assistance.
Makes occasional slight changes in body or extremity position but unable to make frequent or signficiant changes independently.
3. r l l ~ U vm M 4. lhUalb(lr+ Makes frequent though Makes major and frequent slight changes in body or changes in position withextremity position indeout assistance. pendently.
UOISTURE
degree to which skin is exposed to moisture
AcTlvm
1.
cmaaul Ud8t
Skin is kept moist almost constantly by perspiration. urine. etc. Dampmss is detected every time patient is moved or turned.
degree of physical activity
I. vwhw: Never eats a complete usuulfood intake pattern meal. Rarely eats more than 113 of any food offered. Eats 2 servings or leSs of protein (meat or dairy products) per day. Takes fluids poorly. Does not take a liquid dietary supplement.
MUTRnlOM
FRICTIOW AN0 SHEAR
2. PmbIbly Inrdqultt Rarely eats a complete meal and generally eats only about 112 of any food offered. Protein intake includes only 3 servings of meat or dairy products per day. Occasionally will take a dietary supplement. OR OR is NPO andlor maintained receives less than option clear liquids or IV’s for mum amount of liquid diet more than 5 days. or tube feeding.
3. Adr*lmc Eats over half of most meals. Eats a total of 4 servings of protein (meat. dairy products) each day. Occasionally will refuse a meal, but will usually take a supplement if offered.
2. PSwamI m1.c Moves lecbly or requires minimum assistance. During a move skin probably slides to some extent against sheets, chair, restraints. or other dovices. Maintains relatively good position in chair or bed most of the time but occasionally slides down.
3. M. &mlt P M r n Moves in bed and in chair independently and has sufficient muscle strength to lift up completely during move. Maintains good position in bed or chair at all times.
1. frabhc Requires moderate to maximum assistance in moving. Complete lifting without sliding against sheets is impossible. Frepwntly slides down in brd or chair. requiring frequent repositioning with maximum assistance. Spasticity. conbactures or agitation leads to almost constant friction.
Copyright Barbara Bndrn and Nancy Lrgstrom. 1988
Skin is usually dry. linen only requiries changing at routine ifltervalS.
4. wkrt
Eats most of every meal. Never refuses a meal. Usually eats a total of 4 or more servings of meat and dairy products. Ocasionally eats between meals. Does not require supplementation.
OR is on a tube feeding or
TPN regimen which probably meets most of nutritional needs.
Total Scorc
FIGURE 2. Braden Scale for Predicting Pressure Sore Risk (reprinted by permission).
receiving a nursing preventive intervention if either a repositioning schedule was documented in the treatment cardex or a pressure reducing mattress was found on the bed. Demographic information was obtained from the subject’s medical record at the time of risk assessment. This information included the subject’s age, gender, and race. The presence or absence of a pressure ulcer was recorded.
Agreement in assignment of at-risk status between the two assessments and receipt of a preventive intervention was analyzed using Cohen’s Kappa. Cohen’s Kappa reports the percentage agreement between different methods of assessment after correcting for chance agreement. An additional analysis was performed in order to determine if any of the items of the Norton or Braden Scales for Predicting Pressure Sore Risk were used
PRESSURE ULCER RISK VS PREVENTIVE INTERVENTION
IAGS-DECEMBER 1992-VOL. 40, NO.12
TABLE 1. BASELINE CHARACTERISTICS OF STUDY RESIDENTS (n = 478) Age, yrs. Male Caucasian Underlying Medical Conditions Diabetes Malignancy Organic Heart Disease Stroke Hemiparesis/hemiplegia Quadriplegia/paraplegia Dementia Arthritis
COPD Activity level Bed/chair bound Ambulant withlwithout help Incontinence none occasionally of urine usually of urine Both urine and fecal
73 k 12 82% 97% 79 48 187 140 82 12 156 135 121 160 318 234 88 32 124
TABLE 2. CLASSIFICATION OF SUBJECTS’ PRESSURE ULCER RISK STATUS BY THE BRADEN SCALE AND THE NORTON SCALE COMPARED WITH THE PRESENCE OF PREVENTIVE NURSING INTERVENTION IN A STUDY OF 478 SUBJECTS Nurse At Risk At Risk Preventive by Braden by Norton Number of Subiects Intervention Scale Scale Yes Yes Yes Yes No No No No
Yes Yes No No Yes Yes No No
Yes No Yes No Yes No Yes No
106 1 26 80 23 0 28 214
intuitively by the staff nurses to assign risk. The item scores on the Norton and Braden scales were considered as independent variables in a stepwise logistic regression with the presence or absence of a preventive intervention as the dependent variable.
RESULTS A total of 504 subjects were assessed for risk of pressure ulcer formation. Of these, 22 (4%) currently had pressure ulcers and were excluded. Four additional subjects either left the facility or died before all risk assessments were completed. This resulted in a total sample of 478 subjects. The mean age of the subjects was 73 f 12 yrs (range 35-101). The subject group was 82% male and 97% Caucasian. Functional and physical characteristics of the study subjects are presented in Table 1. Staff nurse preventive interventions were found on 45% of all patients, while the Norton scale identified
1253
38% and the Braden scale identified 27% as at-risk for pressure ulcer development. The number of subjects defined as at-risk by each scale and the number of subjects receiving preventive intervention is presented in Table 2. Cohen‘s Kappa showed that overall agreement above chance among all three was 0.53. Cohen’s Kappa for agreement in patient classification between the Braden and Norton scales was 0.73. Agreement between the Braden scale classification and receipt of a nursing preventive intervention was 0.41. Agreement between the Norton scale and the nursing intervention was 0.43. Stepwise logistic regression analysis revealed that two items from the Braden scale significantly contributed toward the predictability of the staff nurse implementing a preventive intervention. These were the mobility item (BMOB) (x2 = 27.3, P < 0.01) and the friction and shear item (BFS) (x2= 17.2, P < 0.01). The nurses were 21 times more likely to have implemented a preventive intervention if a patient’s Braden mobility score was 1 (the minimum score) than if it was 4 (the maximum score) (odds ratio = 2.74 times more likely for each unit decrease in score). The staff nurses were 11 times more likely to implement a preventive intervention if a patient’s Braden friction/shear score was 1 (the minimum score) than if it were 3 (the maximum score) (odds ratio = 3.29 times more likely for each unit decrease in score). The model using these two items to predict the nurse assessment of patient risk is expressed by the following equation: In{prob(x)/l - prob(x)j = 6.03 - 1.01 (BMOB) - 1.19 (BFS)
where: prob(x) = the probability that the staff nurses implemented a preventive intervention BMOB = the patient score on the Braden mobility item BFS = the patient score on the Braden friction and shear item prob(x)/l - prob(x) = the odds of a patient having received a preventive intervention The probability that an individual subject had received a preventive intervention from the staff nurses was calculated for each subject in the study using the regression model above. Probabilities ranged from .OO to 1.00. Using 0.50 to divide “no intervention present” and “intervention present,” the model correctly predicted staff nurse action for 75% of the subjects.
DISCUSSION The overall level of agreement among the three measures of patient pressure ulcer risk was not high (approximately 50%). However, agreement between the Norton and Braden risk assessment scales was close, and the discrepancy between the two risk scales was due almost entirely to the Norton scale classifying patients as at-risk while the Braden scale classified them as not-at-risk. The poor agreement between the
1254
XAKELLIS ET AL
patients receiving a nursing preventive intervention and their being classified as at-risk by either risk assessment scale was due to the fact that nearly 50% of patients were receiving a preventive intervention, although many of these patients were classified as notat-risk by the risk assessment tools. At the same time, however, the study facility maintained a relatively low prevalence (496-576) of patients with pressure ulcers. Whether this low prevalence was due to the high frequency of preventive interventions or some other inherent patient characteristic is not known. Regular use of the Braden scale in place of the nurses' intuitive clinical assessment as a guide to the use of preventive interventions would most likely reduce the number of patients receiving these interventions and their associated costs. These study findings suggest that for facilities just initiating routine risk assessment, the Braden scale offers a method for systematic assessment of risk factors, while minimizing the number of patients at risk and requiring preventive measures. The success of such an approach could be evaluated by following serial point-prevalence rates. This study attempted to make explicit the intuitive expertise of the nurses in assessing risk for pressure ulcer formation. Since this facility has a low pressure ulcer prevalence (4%-5%), these nurses could be considered adept in pressure ulcer prevention. If their assessment process could be explicitly described, it might serve as a model for other facilities. Stepwise logistic regression created a model using two of the Braden items (Mobility and Friction/shear) which correctly predicted the presence or absence of an intervention in 75% of patients. While this was a reasonably good result for such a model, 25% of patients were misclassified. Other unidentified factors must have contributed to the staff nurses' intuitive judgement as to when and when not to implement an intervention. These additional factors will need to be identified before the model of staff nurse clinical judgement can be further refined. This study had several limitations. Since it was conducted in a single long-term care facility with a predominately male population, findings may not be generalizable to other long-term care facilities. In addition, the study included only two types of interventions in the classification of "preventive intervention present." There are other types of preventive interventions, such as pressure-reducing wheelchair cushions, which were not included. Expanding the definition of a preventive intervention would have increased the discrepancy between the number of patients receiving a preventive intervention and the number classified as being at-risk by the scales. Also, some subjects with a pressurereducing mattress might have received this intervention for a reason other than the staff nurses' concern about pressure ulcer formation, although this was probably unlikely. Additionally, the study did not allow for the possibility that the patient's condition had improved between the time the mattress was placed
IAGS-DECEMBER 1992-VOL. 40, NO. 12
on their bed and the time of the study assessment. It is possible that some of the patients with pressurereducing mattresses experienced an improvement in their clinical status, eliminating the need for preventive interventions, although this would not be a common occurrence in long-term care. Finally, this study did not attempt to determine which method of risk assessment would prevent the largest number of pressure ulcers. Rather, it compared the number of patients who would be defined as at-risk by tools which are being recommended for widespread implementation. Consequently, this study needs to be seen as providing a look at the impact of implementing a governmentsponsored clinical guideline. The time has come when pressure ulcer risk assessment is viewed as an essential part of patient assessment in the long-term care facility. The implication of routine risk assessment is that a preventive intervention will be used for at-risk patients. These interventions will not be without cost, and the magnitude of the cost incurred will be directly related to the number of patients defined as at-risk. Further study is needed to define the cost, efficacy and related cost effectiveness of routine pressure ulcer risk assessment.
REFERENCES 1. National Pressure Ulcer Advisory Panel. Pressure ulcer prevalence, cost,
and risk assessment: Consensus development conference statement. Decubitus 1989;2:24-28. 2 Allman RM. Pressure ulcers among the elderly. N Engl J Med 1989;320: 850-853. 3. Andersen KE, Jensen 0, Kvorning SA, Bach E. Prevention of pressure sores by identifying patients at risk. Br Med J 1982;284:1370-1371. 4. Maklebust J, Sieggreen M. Pressure Ulcers: Guidelines for Prevention and Nursing Management. West Dundee, ILS-N Publications, 1991. 5. Braun JL, Silvetti AN, Xakellis GC. Decubitus ulcers: What works best. Patient Care 1988;22(16):22-23. 6. Gosnell DJ. An assessment tool to identify pressure sores. Nurs Res 1973; 22:35-59. 7. Gosnell D. Pressure sore risk assessment: A critique. Part 1. The Gosnell Scale. Decubitus 1989;2:32-38. 8. Waterloo J. A risk assessment card. Nurs Times 1985;81:49-55. 9. Towney AP, Erland SM. Validity and reliability of an assessment tool for pressure ulcer risk. Decubitus 1988;1:40-48. 10. Norton D, McLaren R, Exton-Smith NA. An Investigation of Geriatric Nursing Problems in Hospitals, London: National Corporation for the Care of Old People, 1962. 11. Norton D. Calculating the risk Reflections on the Norton Scale. Decubitus 1989;2:24-31. 12. Bergstrom N, Braden BJ, Laguzza A, Holman V. The Braden Scale for Predicting Pressure Sore Risk. N u n Res 1987;36:205-210. 13. Bergstrom N, Demuth PJ, Braden B). A clinical trial of the Braden Scale for predicting pressure sore risk. Nurs Clin North Am 1987;22(2): 41 7-428. 14. Braden BJ. Clinical utility of the Braden Scale for Predicting Pressure Sore Risk. Decubitus 1989;2:44-51. 15. Goldstone LA, Goldstone J. The Norton score: An early warning of pressure sores? J Adv Nurs 1982;7:419-426. 16. Ek AC. Prediction of pressure sore development. Scand J Caring Sci 1987; 1 :77-84. 17. Panel for the Prediction and Prevention of Pressure Ulcers in Adults. 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, US Department of Health and Human Services, May 1992. 18. Lincoln R, Roberts R, Maddox A, et al. Use of the Norton pressure sore risk assessment scoring system with elderly patients in acute care. J Enterostom Ther 1986;13:132-138. 19. Roberts BV, Goldstone LA. A survey of pressure sores in the over-sixties on two orthopaedic wards. Int J Nurs Studies 1979;16(4):355-364.
ventions was modeled using stepwise logistic regression. The items from the Braden and Norton risk assessment scales were used as independent variables with staff nurse implementation of a preventive intervention as the dependant variable. ver 1 million people currently have pressure ulcers, and with the number of frail elderly persons increasing, the prevalence of pressure ulcers is likely to rise.' Prevention rather than treatment appears to be the key to reducing the number of pressure ulcers. As a consequence, numerous preventive strategies have been p r ~ p o s e d . ~Fundamental -~ to all of these strategies has been some method of determining which patients are at risk for developing pressure ulcers.6-'6 In an effort to standardize pressure ulcer prevention methods, the Agency for Health Care Policy and Research (AHCPR) of the US Public Health Service recently released the clinical practice guideline, Pressure Ulcers in Adults: Prediction and Pre~ention.'~ An essential component of this guideline is the recommendation that all bedbound or chairbound patients be assessed for pressure ulcer risk. The guideline recommends the use of one of two risk assessment scales for conducting
0
From the *Department of Family Practice, College of Medicine, and the tCollege of Nursing, The University of Iowa, Iowa City, Iowa. This project was supported in part by grants from the U S . Public Health Service (Grants for Predoctoral Training in Family Medicine, USPHS/National Institutes of Health, No. 5D15-PE87007-10) and The University of Iowa (Interdisciplinary Research Assistantships Program). Address correspondence to George C. Xakellis, MD, Associate Professor Dept. of Family Practice, The University of Iowa College of Medicine, 2134 Steindler Bldg., Iowa City, I A 52242.
IAGS 4O:I 250-1 254, 1992 0 1992 by the American Geriatrics Sociefy
Results: Nurse preventive interventions were found on 45% of patients. The Norton scale identified 38% and the Braden scale identified 27% of patients as at-risk. Agreement among the three methods was 0.53. Agreement between the Braden and Norton scales was 0.73. Agreement between use of a preventive intervention and a classification as at-risk by the Braden or Norton scale was 0.41 and 0.43, respectively.
Stepwise logistic regression revealed that low Braden mobility scores (Odds Ratio: 2.74) and low Braden friction/shear scores (Odds Ratio: 3.29) were associated with an increased likelihood of a patient receiving a preventive nursing intervention. Conclusions: The overall level of agreement among the two scales predicting risk and the presence of a preventive intervention was not high. Agreement, however, between the two risk assessment scales was close. The staff nurses apparently relied on a patients' mobility, their exposure to friction/shear, and additional unidentified factors to guide implementation of a preventive intervention. Further study is needed to define the cost, efficacy, and related cost effectiveness of routine pressure ulcer risk assessment. J Am Geriatr SOC 40:12501254,1992 this systematic assessment: the Norton Scale or the Braden Scale for Predicting Pressure Sore Risk. Both scales have been tested in a variety of settings and patient populations. Although the performance of these tools in predicting pressure ulcer development has been variable, incorporating them into clinical practice ensures a systematic approach to assessment of individual risk factors. This recommendation represents a significant departure from what has been routine practice in most long-term care facilities. Historically, the responsibility for pressure ulcer prevention has rested with nursing staff, who have used their intuitive clinical judgement to determine patients' level of risk and their need for prevention measures. It is unknown what effect implementation of a standard risk assessment tool will have on numbers of patients judged to be at risk. The AHCPR clinical practice guideline further recommends that preventive measures be instituted for all patients assessed to be at risk for pressure ulcer development. Since such preventive measures create additional expenditures for health-care facilities, determining the relative impact of using the scales versus allowing the nursing staff to use their clinical judgement is essential. To date, the relative performance of the Norton Scale in predicting patient risk has not been compared with that of the Braden scale on the same population of patients. Furthermore, the staff nurses' intuitive judgement of risk, as opera-
0002-861 4/92/$3.50
PRESSURE ULCER RISK VS PREVENTIVE INTERVENTION
/AGS-DECEMBER 7 992-VOL. 40. NO.I 2
tionalized by the use of preventive interventions such as pressure reducing mattresses and turning schedules, has not been compared with the at-risk predictions of these scales. This study was undertaken to determine if the Norton scale and the Braden scale predicted the same patients to be at risk for pressure ulcer development as were receiving preventive nursing interventions. The questions to be answered were: (1)is there a high level of agreement among these three methods of categorizing patients’ risk for pressure ulcer development, and ( 2 ) do staff nurses who have not been trained in the use of formal risk assessment scales intuitively use components of the Braden Scale or Norton Scale to determine if a patient should receive a preventive intervention.
METHODS Setting and Subjects The study was conducted in a 600-bed, state-supported, long-term care facility that serves primarily veterans. This facility was chosen because of its large size, its low prevalence of pressure ukers, and its lack of experience with use of a pressure ulcer risk assessment scale. Further, at this facility the staff nurses make the decisions to implement preventive interventions such as a patient repositioning schedule or a pressure-reducing mattress. A physician’s order is not necessary. Internal audits by the nursing department during the preceding year revealed a 476-576 prevalence of Stage 11-IV pressure ulcers in the study facility. Sixty-seven patients on two nursing units were used to pilot-test the study methods, and all remaining patients in this long-term care facility were assessed for risk of developing pressure ulcers according to the protocol. Patients were excluded from the study if they had a pressure ulcer at the time of initial risk assessment. Risk Assessment Instruments The Norton risk assessment scale is composed of five clinical items, each rated on a 1-4 scale with 1 representing a poor clinical condition and 4 representing a good clinical condition.”, The five items are: physical condition, mental condition, activity, mobility, and incontinence. The Norton risk assessment scale has a total possible score of 20 and defines patients as at-risk if their score is 14 or less (Figure 1).
Physical condition
Mental condition
1251
The Braden scale for predicting pressure sore risk is composed of six clinical item^.'^ Five of the items are rated on a 1-4 scale, with 1 representing a poor clinical condition and 4 representing a good clinical condition. These five items are: (1)sensory perception, defined as the patient’s ”ability to respond meaningfully to pressure-related discomfort”; (2) moisture, defined as the “degree to which skin is exposed to either incontinence, perspiration, or wound drainage”; (3) activity, which quantifies the patient’s ”degree of physical activity” from bedfast to walks frequently; (4) mobility, which rates the patient’s ”ability to change and control body position”; and (5) nutrition, which quantifies the patient’s ”usual food intake pattern.” The sixth Braden item is friction/shear and is rated on a 1-3 scale, with 1 indicating inability to lift freely without sliding across linen and 3 meaning the patient has sufficient muscle strength to lift completely off the supporting surface when repositioning. The scores from each of the items are totalled. The maximum possible score is 23, and patients are at risk if their total score is 16 or less (Figure 2).l2-I4 The sensitivity and specificity of these risk assessment tools has been reported elsewhere.lO,12, 13, 15. 18.19
Study Protocol A research nurse was trained in the use of the Braden and Norton risk assessment and performed all risk assessments using these scales. Staff nurses were consulted when a rating for a particular scale item (such as amount of incontinence or level of nutritional intake) could not be determined from a single patient assessment and a review of the record by the research nurse. The Braden risk assessment was performed on all patients on a particular nursing ward first. After an interval of at least 1 day and without reference to the Braden score, risk assessment was repeated using the Norton scale. Given the large number of risk assessments that were performed on a given day (30-60 subjects reside on any given nursing ward), it was felt that contamination of the second risk score by the first could be minimized by allowing at least 1 day between assessments. All Norton scale risk assessments were performed within 2 weeks of Braden scale assessments. During risk assessment, the research nurse also noted whether a pressure-reducing mattress was present on the subject’s bed and noted if the subject had a repositioning schedule documented in his/her treatment cardex. The patient was classified as
Incontinent
Mobility
Activity
Total Good
Fair
4.
3 Poor 2 VeryBad 1
Alert
4
Apathetic 3
confused 2
Stupor
1
Ambulant
4
Wak/help 3 Chairbound 2 Bcdbound 1
Full
4
SLlimited 3 Veryhited 2 Immobile
1
None
4
Scnre
3 usuauy/urinc 2 Doubly 1
Occasional
Name/Date
FIGURE 1. Norton Scale. From Norton D, McLaren R, Exton-Smith AN. An Investigation of Geriatric Nursing Problems in Hospitals. London: National Corporation for the Care of Old People, 1962 (reprinted by Churchill Livingston, Edinburgh, 1975). Reproduced by permission of NCCOP (now known as the Centre for Policy on Ageing).
1252
XAKELLIS ET A L
SEWWRY tERCEPTlOW
1. Ud#: Unresponsive (does not moan. flinch. or grasp) to ability to respond mem- painful stimuli. due to ingfully to pressure-re- diminished level of conlated discomfort sciousness or sedation. OR limited ability to feel pain over most of body Surlace.
IAGS-DECEMBER 1992-VOL. 40, NO. 1 2
2. V a q UuIW Responds only to painful stimuli. Cannot communicate discomfort except by moaning or restlessness.
3. tll#MIlUUltJ:
4. hlapahent
Responds to verbal commands. but cannot always communicate discomfort or need to be turned.
Responds to verbal commands. Has no sensory deficit which would limit ability to feel or voice pain or discomfort.
OR
OR has a sensory impairment which limits the ability to feel pain or discomfort over 112 of body.
has some sensory impairment which limits ability to feel pain or discomfort in 1 or 2 extremities.
2. Vtryldrt Skin is often. but not always moist. Linen must be changed at least once a shift.
3. ossulrrlly ud8t Skin is occasionally moist. requiring an extra linen change approximately once a day.
4. lhnll Uabt
1. Wat Confined to W
2. rnldwt Ability to walk severely limited or non-existent. Cannot bear own weight andlor must be assisted into chair or wheelchair.
3. W l l b accaldly: Walks occasionally during day, but for very short distances. with or without assistance. Spends majority of each shift in bed or chair.
4. w m Fmmaar Walks outside the room at least twice a day a d inside room at least O m every 2 hours during waking hours.
mourn
1. Cwphmly larllk:
2. bq UmltJ:
ability to change and control body position
Doas not make even slight changes in body or extremity position without assistance.
Makes occasional slight changes in body or extremity position but unable to make frequent or signficiant changes independently.
3. r l l ~ U vm M 4. lhUalb(lr+ Makes frequent though Makes major and frequent slight changes in body or changes in position withextremity position indeout assistance. pendently.
UOISTURE
degree to which skin is exposed to moisture
AcTlvm
1.
cmaaul Ud8t
Skin is kept moist almost constantly by perspiration. urine. etc. Dampmss is detected every time patient is moved or turned.
degree of physical activity
I. vwhw: Never eats a complete usuulfood intake pattern meal. Rarely eats more than 113 of any food offered. Eats 2 servings or leSs of protein (meat or dairy products) per day. Takes fluids poorly. Does not take a liquid dietary supplement.
MUTRnlOM
FRICTIOW AN0 SHEAR
2. PmbIbly Inrdqultt Rarely eats a complete meal and generally eats only about 112 of any food offered. Protein intake includes only 3 servings of meat or dairy products per day. Occasionally will take a dietary supplement. OR OR is NPO andlor maintained receives less than option clear liquids or IV’s for mum amount of liquid diet more than 5 days. or tube feeding.
3. Adr*lmc Eats over half of most meals. Eats a total of 4 servings of protein (meat. dairy products) each day. Occasionally will refuse a meal, but will usually take a supplement if offered.
2. PSwamI m1.c Moves lecbly or requires minimum assistance. During a move skin probably slides to some extent against sheets, chair, restraints. or other dovices. Maintains relatively good position in chair or bed most of the time but occasionally slides down.
3. M. &mlt P M r n Moves in bed and in chair independently and has sufficient muscle strength to lift up completely during move. Maintains good position in bed or chair at all times.
1. frabhc Requires moderate to maximum assistance in moving. Complete lifting without sliding against sheets is impossible. Frepwntly slides down in brd or chair. requiring frequent repositioning with maximum assistance. Spasticity. conbactures or agitation leads to almost constant friction.
Copyright Barbara Bndrn and Nancy Lrgstrom. 1988
Skin is usually dry. linen only requiries changing at routine ifltervalS.
4. wkrt
Eats most of every meal. Never refuses a meal. Usually eats a total of 4 or more servings of meat and dairy products. Ocasionally eats between meals. Does not require supplementation.
OR is on a tube feeding or
TPN regimen which probably meets most of nutritional needs.
Total Scorc
FIGURE 2. Braden Scale for Predicting Pressure Sore Risk (reprinted by permission).
receiving a nursing preventive intervention if either a repositioning schedule was documented in the treatment cardex or a pressure reducing mattress was found on the bed. Demographic information was obtained from the subject’s medical record at the time of risk assessment. This information included the subject’s age, gender, and race. The presence or absence of a pressure ulcer was recorded.
Agreement in assignment of at-risk status between the two assessments and receipt of a preventive intervention was analyzed using Cohen’s Kappa. Cohen’s Kappa reports the percentage agreement between different methods of assessment after correcting for chance agreement. An additional analysis was performed in order to determine if any of the items of the Norton or Braden Scales for Predicting Pressure Sore Risk were used
PRESSURE ULCER RISK VS PREVENTIVE INTERVENTION
IAGS-DECEMBER 1992-VOL. 40, NO.12
TABLE 1. BASELINE CHARACTERISTICS OF STUDY RESIDENTS (n = 478) Age, yrs. Male Caucasian Underlying Medical Conditions Diabetes Malignancy Organic Heart Disease Stroke Hemiparesis/hemiplegia Quadriplegia/paraplegia Dementia Arthritis
COPD Activity level Bed/chair bound Ambulant withlwithout help Incontinence none occasionally of urine usually of urine Both urine and fecal
73 k 12 82% 97% 79 48 187 140 82 12 156 135 121 160 318 234 88 32 124
TABLE 2. CLASSIFICATION OF SUBJECTS’ PRESSURE ULCER RISK STATUS BY THE BRADEN SCALE AND THE NORTON SCALE COMPARED WITH THE PRESENCE OF PREVENTIVE NURSING INTERVENTION IN A STUDY OF 478 SUBJECTS Nurse At Risk At Risk Preventive by Braden by Norton Number of Subiects Intervention Scale Scale Yes Yes Yes Yes No No No No
Yes Yes No No Yes Yes No No
Yes No Yes No Yes No Yes No
106 1 26 80 23 0 28 214
intuitively by the staff nurses to assign risk. The item scores on the Norton and Braden scales were considered as independent variables in a stepwise logistic regression with the presence or absence of a preventive intervention as the dependent variable.
RESULTS A total of 504 subjects were assessed for risk of pressure ulcer formation. Of these, 22 (4%) currently had pressure ulcers and were excluded. Four additional subjects either left the facility or died before all risk assessments were completed. This resulted in a total sample of 478 subjects. The mean age of the subjects was 73 f 12 yrs (range 35-101). The subject group was 82% male and 97% Caucasian. Functional and physical characteristics of the study subjects are presented in Table 1. Staff nurse preventive interventions were found on 45% of all patients, while the Norton scale identified
1253
38% and the Braden scale identified 27% as at-risk for pressure ulcer development. The number of subjects defined as at-risk by each scale and the number of subjects receiving preventive intervention is presented in Table 2. Cohen‘s Kappa showed that overall agreement above chance among all three was 0.53. Cohen’s Kappa for agreement in patient classification between the Braden and Norton scales was 0.73. Agreement between the Braden scale classification and receipt of a nursing preventive intervention was 0.41. Agreement between the Norton scale and the nursing intervention was 0.43. Stepwise logistic regression analysis revealed that two items from the Braden scale significantly contributed toward the predictability of the staff nurse implementing a preventive intervention. These were the mobility item (BMOB) (x2 = 27.3, P < 0.01) and the friction and shear item (BFS) (x2= 17.2, P < 0.01). The nurses were 21 times more likely to have implemented a preventive intervention if a patient’s Braden mobility score was 1 (the minimum score) than if it was 4 (the maximum score) (odds ratio = 2.74 times more likely for each unit decrease in score). The staff nurses were 11 times more likely to implement a preventive intervention if a patient’s Braden friction/shear score was 1 (the minimum score) than if it were 3 (the maximum score) (odds ratio = 3.29 times more likely for each unit decrease in score). The model using these two items to predict the nurse assessment of patient risk is expressed by the following equation: In{prob(x)/l - prob(x)j = 6.03 - 1.01 (BMOB) - 1.19 (BFS)
where: prob(x) = the probability that the staff nurses implemented a preventive intervention BMOB = the patient score on the Braden mobility item BFS = the patient score on the Braden friction and shear item prob(x)/l - prob(x) = the odds of a patient having received a preventive intervention The probability that an individual subject had received a preventive intervention from the staff nurses was calculated for each subject in the study using the regression model above. Probabilities ranged from .OO to 1.00. Using 0.50 to divide “no intervention present” and “intervention present,” the model correctly predicted staff nurse action for 75% of the subjects.
DISCUSSION The overall level of agreement among the three measures of patient pressure ulcer risk was not high (approximately 50%). However, agreement between the Norton and Braden risk assessment scales was close, and the discrepancy between the two risk scales was due almost entirely to the Norton scale classifying patients as at-risk while the Braden scale classified them as not-at-risk. The poor agreement between the
1254
XAKELLIS ET AL
patients receiving a nursing preventive intervention and their being classified as at-risk by either risk assessment scale was due to the fact that nearly 50% of patients were receiving a preventive intervention, although many of these patients were classified as notat-risk by the risk assessment tools. At the same time, however, the study facility maintained a relatively low prevalence (496-576) of patients with pressure ulcers. Whether this low prevalence was due to the high frequency of preventive interventions or some other inherent patient characteristic is not known. Regular use of the Braden scale in place of the nurses' intuitive clinical assessment as a guide to the use of preventive interventions would most likely reduce the number of patients receiving these interventions and their associated costs. These study findings suggest that for facilities just initiating routine risk assessment, the Braden scale offers a method for systematic assessment of risk factors, while minimizing the number of patients at risk and requiring preventive measures. The success of such an approach could be evaluated by following serial point-prevalence rates. This study attempted to make explicit the intuitive expertise of the nurses in assessing risk for pressure ulcer formation. Since this facility has a low pressure ulcer prevalence (4%-5%), these nurses could be considered adept in pressure ulcer prevention. If their assessment process could be explicitly described, it might serve as a model for other facilities. Stepwise logistic regression created a model using two of the Braden items (Mobility and Friction/shear) which correctly predicted the presence or absence of an intervention in 75% of patients. While this was a reasonably good result for such a model, 25% of patients were misclassified. Other unidentified factors must have contributed to the staff nurses' intuitive judgement as to when and when not to implement an intervention. These additional factors will need to be identified before the model of staff nurse clinical judgement can be further refined. This study had several limitations. Since it was conducted in a single long-term care facility with a predominately male population, findings may not be generalizable to other long-term care facilities. In addition, the study included only two types of interventions in the classification of "preventive intervention present." There are other types of preventive interventions, such as pressure-reducing wheelchair cushions, which were not included. Expanding the definition of a preventive intervention would have increased the discrepancy between the number of patients receiving a preventive intervention and the number classified as being at-risk by the scales. Also, some subjects with a pressurereducing mattress might have received this intervention for a reason other than the staff nurses' concern about pressure ulcer formation, although this was probably unlikely. Additionally, the study did not allow for the possibility that the patient's condition had improved between the time the mattress was placed
IAGS-DECEMBER 1992-VOL. 40, NO. 12
on their bed and the time of the study assessment. It is possible that some of the patients with pressurereducing mattresses experienced an improvement in their clinical status, eliminating the need for preventive interventions, although this would not be a common occurrence in long-term care. Finally, this study did not attempt to determine which method of risk assessment would prevent the largest number of pressure ulcers. Rather, it compared the number of patients who would be defined as at-risk by tools which are being recommended for widespread implementation. Consequently, this study needs to be seen as providing a look at the impact of implementing a governmentsponsored clinical guideline. The time has come when pressure ulcer risk assessment is viewed as an essential part of patient assessment in the long-term care facility. The implication of routine risk assessment is that a preventive intervention will be used for at-risk patients. These interventions will not be without cost, and the magnitude of the cost incurred will be directly related to the number of patients defined as at-risk. Further study is needed to define the cost, efficacy and related cost effectiveness of routine pressure ulcer risk assessment.
REFERENCES 1. National Pressure Ulcer Advisory Panel. Pressure ulcer prevalence, cost,
and risk assessment: Consensus development conference statement. Decubitus 1989;2:24-28. 2 Allman RM. Pressure ulcers among the elderly. N Engl J Med 1989;320: 850-853. 3. Andersen KE, Jensen 0, Kvorning SA, Bach E. Prevention of pressure sores by identifying patients at risk. Br Med J 1982;284:1370-1371. 4. Maklebust J, Sieggreen M. Pressure Ulcers: Guidelines for Prevention and Nursing Management. West Dundee, ILS-N Publications, 1991. 5. Braun JL, Silvetti AN, Xakellis GC. Decubitus ulcers: What works best. Patient Care 1988;22(16):22-23. 6. Gosnell DJ. An assessment tool to identify pressure sores. Nurs Res 1973; 22:35-59. 7. Gosnell D. Pressure sore risk assessment: A critique. Part 1. The Gosnell Scale. Decubitus 1989;2:32-38. 8. Waterloo J. A risk assessment card. Nurs Times 1985;81:49-55. 9. Towney AP, Erland SM. Validity and reliability of an assessment tool for pressure ulcer risk. Decubitus 1988;1:40-48. 10. Norton D, McLaren R, Exton-Smith NA. An Investigation of Geriatric Nursing Problems in Hospitals, London: National Corporation for the Care of Old People, 1962. 11. Norton D. Calculating the risk Reflections on the Norton Scale. Decubitus 1989;2:24-31. 12. Bergstrom N, Braden BJ, Laguzza A, Holman V. The Braden Scale for Predicting Pressure Sore Risk. N u n Res 1987;36:205-210. 13. Bergstrom N, Demuth PJ, Braden B). A clinical trial of the Braden Scale for predicting pressure sore risk. Nurs Clin North Am 1987;22(2): 41 7-428. 14. Braden BJ. Clinical utility of the Braden Scale for Predicting Pressure Sore Risk. Decubitus 1989;2:44-51. 15. Goldstone LA, Goldstone J. The Norton score: An early warning of pressure sores? J Adv Nurs 1982;7:419-426. 16. Ek AC. Prediction of pressure sore development. Scand J Caring Sci 1987; 1 :77-84. 17. Panel for the Prediction and Prevention of Pressure Ulcers in Adults. 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, US Department of Health and Human Services, May 1992. 18. Lincoln R, Roberts R, Maddox A, et al. Use of the Norton pressure sore risk assessment scoring system with elderly patients in acute care. J Enterostom Ther 1986;13:132-138. 19. Roberts BV, Goldstone LA. A survey of pressure sores in the over-sixties on two orthopaedic wards. Int J Nurs Studies 1979;16(4):355-364.
