Pressure ulcers (PrUs) are a significant health problem among older adults in all care settings. Estimates of PrU incidence range from 0.4% to 40% in hospitalized patients, 2.3% to 23.9% in nursing home patients, and 0% to 17% among those receiving home health care. 1–6 Pressure ulcers are associated with increased morbidity, elevated health-related costs, and reduced quality of life. 7–9 However, most PrUs can be prevented, with the first step being to identify those at risk. 10
Research on PrUs has shown that a number of factors predispose hospitalized and nursing home patients to PrUs, including impaired mobility, limitation in activity, sensoryperception alterations, moisture, friction/shear, older age, impaired nutrition, hemodynamic alterations, select comorbid conditions, and stress. 11,12 Tools that quantify a number of these factors, such as the Norton Scale 13 and the Braden Scale, 14 have also been used to assess or predict PrU risk in hospital and nursing home settings.
Little is known about predicting PrU development in home health care patients. The few studies conducted in this setting found that urine or stool incontinence, 15–17 altered levels of activity and mobility, 15 recent discharge from an institutional setting, or more functional impairment 17,18 were associated with the presence of a PrU. Because factors associated with PrU presence may not predict their formation, a retrospective cohort study of risk factors for Stage II to IV PrU development was conducted among 1711 older adults receiving home health care services. 19 The study showed that limitation in activity to a wheelchair, needing assistance with dressing, bowel/bladder incontinence, a Braden Scale 14 mobility subscore of “very limited mobility,” anemia (indicated by ICD-9 code), having an adult child as primary caregiver, male gender, a recent fracture (indicated by ICD-9 code), oxygen use, and skin drainage predicted Stage II to IV PrU development by Cox multivariate regression analysis (P ≤.05).
Few studies have estimated PrU risk in home health care using the Braden Scale. Ramundo 6 found that a score of 18 identified home health care patients at risk for incident PrUs (100% sensitivity, 34% specificity). However, the generalizability of the results was limited because the sample size was small and the study included only subjects who were wheelchair-bound and bedridden. The validity of the Braden Scale was further examined among 1696 older adults receiving home health care. 20 This study revealed that a Braden Scale score of 19 or less on admission identified individuals at risk for a Stage I to IV PrU with 61% sensitivity and 68% specificity. The scale most accurately predicted PrUs that occurred within the first week following admission to the home health care agency. However, the critical cutoff score of 19 was higher than the 16 to 18 cutoff score recommended for acute and long-term care and the sensitivity and specificity were less than optimal. In addition, no admission score reliably predicted PrUs that developed after the first week. These findings leave unclear whether the Braden Scale score obtained on admission is the best method for identifying older adults at risk for pressure ulceration in the home health care setting.
The recent introduction of the Outcome and Assessment Information Set (OASIS) provides an opportunity to use routinely collected admission data to identify home health care patients who may be at risk for PrU formation. The OASIS is a standardized data set that was developed to measure patient outcomes and improve the effectiveness of home health care. 21,22 Medicare-certified home health care agencies are now required to collect OASIS data under a mandate from the Centers for Medicare and Medicaid Services (CMS) for prospective payment of home health care services. Included in the OASIS data set are 79 items that must be completed at the start of care. Nearly all of these items assess a patient’s condition and care needs; however, collectively, they do not constitute a comprehensive assessment instrument. 22 Many of the assessment items include known risk factors for PrU development in hospitalized and nursing home patients; therefore, the admission OASIS assessment may provide a method for identifying older adults at risk for PrU development in home health care. No study has examined the relationship between data obtained on the admission OASIS assessment and development of a PrU. The purpose of this study was to determine whether admission data routinely collected on the OASIS might be used to identify the older adult at risk for a PrU in the home health care setting.
Design, setting, and subjects
This study is a secondary analysis of data from a retrospective cohort study of risk factors for PrU development in older adults receiving home health care. 19 The study was conducted in a large, midwestern, urban home health care agency. Approval for the study was obtained by the university’s institutional review board and the agency’s professional practice committee.
The medical records of 1820 nonhospice, nonintravenous therapy patients were reviewed for the presence of a PrU. Patients were 60 years and older and were admitted to the intermittent skilled nursing division of the home health care agency between January 1995 and March 1996. Only the first admission was considered for patients admitted more than once during the study period. Patients admitted with a Stage I to IV PrU (n = 109) were excluded from the study. The 1711 patients who were free of PrUs on admission were included in the cohort sample.
Data on factors associated with PrU development in hospital and nursing home settings and those considered relevant to the home setting were extracted from each patient’s home health care record for the original study. 19 These data were located on 3 forms that had been completed by a registered nurse when the patient was admitted to the home health care agency and included the Patient Information Summary, the Health Care Financing Administration (HCFA)-485 Admission Form, and the Nursing Admission Assessment Form. 23
Data identical to items on the admission OASIS assessment (OASIS-like) were selected for the current study. Categoric variables included gender, type of medical condition, type of prescribed medication (steroids, analgesics, sedatives/hypnotics, antidepressants), type of residence (single family, multiple family), primary caregiver, level of activity (bedrest, wheelchair, up as tolerated), use of oxygen, presence/absence of pain, dyspnea, urinary incontinence, bowel incontinence, bowel/bladder incontinence, a Foley catheter, depression, anxiety, confusion, unresponsiveness, short-/long-term memory deficit, lack of concentration, impaired reasoning ability, and activities of daily living (ADLs). Each ADL (bathing, dressing, feeding, transferring, and toileting) was categorized as either independent, needs assistance, or dependent to determine which level of dependence was more strongly associated with PrU development. Continuous variables included age, total number of medical conditions, total number of prescribed medications, and the number of hospitalizations or nursing home residencies within 1 year of admission to the home health care agency. Data on length of stay, number of home health care visits, and reason for discharge were also included.
Patient records were followed forward chronologically to 1 of 2 outcomes: PrU development or no PrU development. Pressure ulcer development was defined as the first PrU that developed following admission to the home health care agency and its stage (I to IV) on discovery. The outcome of no PrU development was defined as the absence of a PrU on discharge from the agency, on admission to an institutional health care facility, at death, or at the end of the study period.
Development of a PrU was determined from the nursing visit report or documentation on the agency’s wound assessment form. A nursing visit report was completed by the registered nurse with each home visit. Space for documenting nurse-observed alterations in integumentary status was available on the report.
The home health care agency used the following criteria for staging PrUs: Stage I, nonblanchable erythema of intact skin; Stage II, partial-thickness skin loss involving epidermis and/or dermis; Stage III, full-thickness skin loss involving damage or necrosis of subcutaneous tissue; Stage IV, full-thickness skin loss with extensive destruction, tissue necrosis, or damage to muscle, bone, or supporting structures. 10
Differentiating reactive hyperemia from nonblanchable erythema during a home health care visit lasting an hour or less may be difficult. Therefore, a Stage I PrU was defined as the presence of a Stage I PrU at the same site for 2 consecutive home visits longer than 24 hours apart. A Stage I PrU that progressed to a Stage II or greater PrU at a later home visit was recorded as a Stage I PrU. Location of the PrU was also noted.
Criterion-related validity of PrU presence, stage, and location was established during the original study 19 by the principal investigator who accompanied approximately 5% of staff nurses on home visits to 11 patients with PrUs. The percentage of agreement for PrU presence and location was 100% and the percentage of agreement for PrU stage was 91%. In the 1 case of disagreement, the PrU was recorded as a Stage IV PrU (healing) in the patient record; the principal investigator, who had no information about its history, classified it as a Stage III PrU. Interrater reliability of 96% was established by randomly selecting 5% of the patient records for data extraction by a research assistant. Another 4.5% of randomly chosen charts were recoded by the principal investigator to establish an intrarater reliability of 97%.
Data analyses were conducted using SPSS for Windows (version 11.0, 2001; SPSS Inc, Chicago, IL). Descriptive data were analyzed and differences between subjects with and without PrUs were tested using independent sample t tests. Because Stage I PrUs are often underreported 24 and research lacks consensus about their inclusion in data analyses, PrU-positive subjects were divided into 2 groups: those who developed a Stage I PrU and those who developed a Stage II or greater PrU. Chi-square tests (Pearson) and independent sample t tests were used to compare subjects who developed a Stage I PrU with subjects who developed a Stage II or greater PrU, with a P value of .01 or less considered statistically significant.
Bivariate associations between each OASIS-like variable and PrU development were examined using simple Cox proportional hazards regression procedures. Each OASIS-like variable was entered as the only independent variable in a model that examined the strength of its relationship to PrU development. The time variable was the number of days to outcome, with a P value less than or equal to .01 considered statistically significant. The relationship of each OASIS-like variable to PrU development was also analyzed by PrU stage. For analyses of variables to Stage I PrU development, Stage II and greater PrUs were removed from the calculations. Likewise, Stage I PrUs were removed from the analyses of variables to Stage II and greater PrU development.
The Cox proportional hazards model is a type of regression procedure that takes into account the length of time to an outcome, such as PrU development or no PrU development. 25 In the current study, the length of time to PrU development or no PrU development varied for each subject. Computation of a simple Cox proportional hazards model produces an unadjusted hazard ratio, or estimated relative risk; 95% confidence interval; and significance level for each variable. Multivariate models produce adjusted hazard ratios relative to other variables in the model.
Multivariate analyses of factors significantly associated with PrU development in bivariate analyses were conducted using forward stepwise Cox regression procedures. The likelihood-ratio statistic, based on the maximum partial likelihood estimates, was used for variable entry and removal. Models were constructed predicting Stage I, Stage II and greater PrUs, and Stage I plus Stage II and greater PrUs. For the model predicting Stage I PrUs, Stage II and greater PrUs were removed from the calculations. Similarly, Stage I PrUs were removed from the model predicting Stage II and greater PrUs.
A significance level of .001 or less was used to account for the number of variables involved in the multivariate comparisons. Significantly correlated variables (r ≥0.3) were tested in separate Cox regression models. Models reported for Stage I, Stage II and greater, and Stage I plus Stage II and greater PrUs were those that best fit the data (lowest minus 2 times the log likelihood [LL]). Interaction was tested by adding cross product terms to the model.
The 1070 (62.5%) female and 641 (37.5%) male subjects ranged in age from 60 to 101 (mean [m] = 76.4, standard deviation [SD] = 8.6). Common medical conditions included heart disease (51.5%), hypertension (34.0%), musculoskeletal and connective tissue diseases (26.8%), diabetes (23.3%), chronic obstructive pulmonary disease (19.3%), and cognitive disorders (17.5%).
Following admission to the home health care agency (n = 108; incidence = 6.3%), 53 subjects (49.1%) developed a Stage I PrU, 54 (50.0%) developed a Stage II PrU, and 1 subject (0.9%) developed a Stage III PrU. No subject developed a Stage IV PrU. Data on PrU worsening was inconsistently available. Pressure ulcers were primarily located on the pelvic area (n = 56, 52%) and heel (n = 15, 14%). Most ulcers (n = 58, 54%) developed within the first 4 weeks of admission (m = 54.7; SD = 73.6). The average number of days to discharge, institutional admission, death, or the end of the study period was 60.9 (median = 26; SD = 95.2). There was no significant difference between PrU-positive and PrU-free patients in the average length of time to a study outcome (t = -0.68, P = .50). There were no statistically significant differences between subjects who developed a Stage I PrU and subjects who developed a Stage II or greater PrU with respect to age, gender, number of medical diagnoses, time to PrU discovery, number of nursing home visits, and other potential risk factors.
Pressure ulcer-positive subjects received an average of 11.6 home health care visits (SD = 10.0) from a registered nurse prior to PrU development. Subjects who were PrU-free received an average of 12.5 home visits (SD = 14.8). There was no significant difference in the number of home visits (t = -0.69;P = .49) between subjects who developed a PrU and subjects who remained PrU-free. There was no significant difference in the intensity of home visits (number of patient care days/number of home visits) between subjects who developed a PrU (m = 4.4, SD = 5.2) and subjects who did not (m = 4.9; SD = 8.5) (t = -0.65;P = .52). Although subjects with a Stage I PrU averaged more home health care visits (m = 12.0; SD = 10.4) than subjects with a Stage II or greater PrU (m = 10.5; SD = 8.4), there was no significant difference between the 2 groups in the number of home visits (t = 0.76;P = .45) or their intensity (t(98) = 1.4;P = .17).
Predictors of Stage I PrUs
Bivariate analyses showed that 16 OASIS-like variables were significantly associated with Stage I PrU development (P ≤.01;Table 1). These variables were entered into a forward stepwise Cox regression procedure. Dependence in bathing was significantly correlated with dependence in dressing, feeding, toileting, and transferring (r ≥0.3) and needing feeding assistance was significantly correlated with needing toileting or transferring assistance (r ≥0.3); therefore, each of these variables was tested in separate multivariate regression procedures. Limitation in activity to bed, dependence in dressing, urinary incontinence, and needing assistance with transferring predicted Stage I PrU development (P ≤.001;Table 2). No cross product terms were statistically significant.
Predictors of Stage II and greater PrUs
In this study, only 1 subject developed a Stage III PrU and no subject developed a Stage IV PrU. Therefore, analysis of variables was essentially limited to Stage II PrUs. Bivariate analyses showed that 12 OASIS-like variables were significantly associated with Stage II and greater PrU development (P ≤.01;Table 1). These variables were entered into a forward stepwise Cox regression procedure. Because dependence in bathing was significantly correlated with dependence in dressing, toileting, and transferring (r ≥0.3), each variable was tested in separate multivariate regression procedures. Bowel/bladder incontinence, oxygen use, a current fracture, and dependence in dressing predicted Stage II and greater PrU development (P ≤.001;Table 3). The cross product terms representing interaction between bowel/bladder incontinence, oxygen use, fracture, and dependence in dressing were not statistically significant.
Predictors of Stage I plus Stage II and greater PrUs
Bivariate analyses showed that 20 OASIS-like variables were significantly associated with Stage I plus Stage II and greater PrU development (P ≤.01;Table 1). These variables were entered into a forward stepwise Cox regression procedure. Significantly correlated variables (r ≥0.3) were tested in separate multivariate regression procedures. Limitation in activity to bed, dependence in dressing, a current fracture, oxygen use, needing assistance with transferring, and urinary incontinence predicted Stage I plus Stage II and greater PrU development (P ≤.001;Table 4). The cross product terms representing interaction between limitation in activity to bed, dependence in bathing, a current fracture, oxygen use, needing assistance with transferring, and urinary incontinence were not statistically significant. A correlation matrix of regression coefficients for the predictor variables is presented in Table 5.
Most of the ulcers that developed in this study were either Stage I or Stage II PrUs. The 6.3% incidence of PrUs contradicts clinical belief that few PrUs develop after admission to home health care. Most ulcers developed within the first 4 weeks of admission to the agency, highlighting the need for early risk identification and prevention among older adults receiving home health care.
Predictors of Stage I PrUs
This study found that limitation in activity to bed, dependence in dressing, urinary incontinence, and needing assistance with transferring independently predicted Stage I PrUs. Subjects who were limited in activity to bed were at high risk for Stage I PrU development. This finding is consistent with results from studies in acute and long-term care 3,26,27 and verifies the importance of this risk factor to early PrU formation in the home health care setting.
Urinary incontinence, rather than bowel incontinence, predicted Stage I pressure ulceration in this study. Contrary to several previous studies with Stage II and greater PrUs, 28,29 bowel incontinence was not included in the final regression model, nor was it highly correlated with urinary incontinence. More alterations in ADLs were associated with Stage I PrU development than were associated with Stage II and greater PrU development. Inclusion of dependence in dressing and needing assistance with transferring in the model of Stage I PrU predictors is consistent with findings from a number of studies reporting a significant association between PrU development and increasing dependence in ADLs, such as bathing, transferring, and toileting. 3,30,31
Predictors of Stage II and greater PrUs
Independent predictors of Stage II and greater PrU development included bowel/bladder incontinence, oxygen use, a current fracture, and dependence in dressing. The significant association between bowel/bladder incontinence and Stage II and greater PrUs was most likely the result of urinary incontinence because only 2 subjects with Stage II and greater PrUs who were incontinent of urine were also incontinent of stool. Urinary incontinence and dependence in dressing predicted both Stage I and Stage II and greater PrUs. Oxygen use and a current fracture were distinctive to the model predicting Stage II and greater PrUs. The difference in predictors of Stage I PrUs and Stage II and greater PrUs may be related to unmeasured subject characteristics. For example, elderly patients requiring supplemental oxygen may be at an increased risk for a PrU from inadequate tissue oxygenation, or the variable itself may be a surrogate measure of severity of illness. Although fractures were prevalent among subjects who developed a Stage I PrU, inclusion of this diagnosis in the model predicting Stage II and greater PrUs may suggest differential effects of fracture location on mobility. Because little is known regarding differences in predictors for Stage I PrUs versus Stage II and greater PrUs, further research in this area is needed.
Predictors of Stage I plus Stage II and greater PrUs
Predictors of Stage I plus Stage II and greater PrUs included those from each of the individual models; specifically, limitation in activity to bed, dependence in dressing, a current fracture, oxygen use, needing assistance with transferring, and urinary incontinence. Because level of ADLs predicting PrU formation was often significantly correlated, increasing dependence in ADLs—rather than increasing dependence in any one ADL—may indicate PrU risk among older adults receiving home health care.
Implications and limitations
Overall, these findings suggest that the admission OASIS assessment may provide a method for identifying older adults receiving home health care who are at risk for Stage I and Stage II PrU development. Older adult home health care patients in the current study who were limited in activity to bed, were incontinent of urine, had a recent fracture, used supplemental oxygen, needed assistance with transferring, or were dependent in dressing on admission were at risk for PrU development. The risk factors identified are among the 79 OASIS assessment items that must be completed for each patient at the start of home health care. Study results suggest that the presence of 1 or more of these risk factors on completion of the admission OASIS assessment indicates PrU risk and the need for a PrU prevention plan of care. Progressive home health care agencies may even develop mechanisms to electronically alert nursing staff to the presence of these risk factors and link to appropriate PrU prevention strategies. Because variables included in the current study were OASIS-like, research examining the relationship between OASIS data obtained on admission and PrU development is needed to support these suggestions.
Using routinely collected admission data may be a more efficient method of identifying those at risk than completing a supplemental form or questions. Whether or not risk established on admission predicts PrU development during the length of home health care stay should be studied but may be possible because prospective payment has generally promoted earlier patient discharge. Future research might also explore which additional factors are important to precise prediction of risk in the home health care setting.
There were several limitations to this study. Data collected from existing records limited the measurement of select variables, including hospital length of stay prior to home health care admission, severity of illness, and employed preventive strategies. Data analysis was essentially limited to Stage I and Stage II PrUs. The more rigorous definition of Stage I PrUs used in this study to minimize overreporting of Stage I PrUs may have contributed to underreporting their development. In addition, patients who did not report a suspect or former lesion during the home health care visit and nurses who were unfamiliar with assessing darkly pigmented skin may have contributed to underreporting Stage I PrU development.
The problem of PrUs among older adults receiving home health care rivals the problem of PrUs among older adults in hospitals and long-term-care facilities. Results from this study suggest that the admission OASIS assessment may provide a method for identifying those patients at risk for Stage I and Stage II PrUs, which must be validated by subsequent research findings. Prevention interventions can then be directed to at-risk individuals to alleviate or eliminate the particular factors that place them at risk and reduce the incidence of PrUs in the home health care setting.
1. O’Sullivan KL, Engrav LH, Maier RV, Pilcher SL, Isik FF, Copass MK. Pressure sores in the acute trauma patient: incidence and causes. J Trauma 1997; 42:276–8.
2. Jesurum J, Joseph K, Davis JM, Suki R. Balloons, beds and breakdown. Effects of low-air loss therapy on the development of pressure ulcers in cardiovascular surgical patients with intra-aortic balloon pump support. Crit Care Nurs Clin North Am 1996; 8:423–40.
3. Berlowitz DR, Brandeis GH, Morris JN, et al. Deriving a risk-adjustment model for pressure ulcer development using the Minimum Data Set. J Am Geriatr Soc 2001; 49:866–71.
4. Bergstrom N, Braden B, Kemp M, Champagne M, Ruby E. Predicting pressure ulcer risk: a multisite study of the predictive validity of the Braden Scale. Nurs Res 1998; 47:261–9.
5. Langemo DK, Olson B, Hunter S, Hanson D, Burd C, Cathcart-Silberberg T. Incidence and prediction of pressure ulcers in five patient care settings. Decubitus 1991; 4( 3):25–26, 28, 30.
6. Ramundo JM. Reliability and validity of the Braden Scale in the home care setting. J Wound Ostomy Continence Nurs 1995: 22:128–134.
7. Rosen A, Wu J, Chang BH, et al. Risk adjustment for measuring health outcomes: an application in VA long-term care. Am J Med Qual 2001; 16:118–27.
8. Allman RM, Goode PS, Burst N, Bartolucci AA, Thomas DR. Pressure ulcers, hospital complications, and disease severity: impact on hospital costs and length of stay. Adv Wound Care 1999; 12( 1):22–30.
9. Szor J, Bourguignon D. Description of pressure ulcer pain at rest and at dressing change. J Wound Ostomy Continence Nurs 1999; 26:115–20.
10. Panel on the Prediction and Prevention of Pressure Ulcers in Adults. Pressure Ulcers in Adults: Prediction and Prevention. Clinical Practice Guideline, No 3. AHCPR Publication No. 92-0047. Rockville, MD: Agency for Health Care Policy and Research; May 1992.
11. Braden B, Bergstrom N. A conceptual schema for the study of the etiology of pressure sores. Rehabil Nurs 1987; 12:8–12.
12. Allman RM. Pressure ulcer prevalence, incidence, risk factors, and impact. Clin Geriatr Med 1997; 13:421–36.
13. Norton D, McLaren R, Exton-Smith AN. An Investigation of Geriatric Nursing Problems in Hospital. Edinburgh, Scotland: Churchill Livingstone; 1975.
14. Bergstrom N, Braden B, Laguzza A, Holman V. The Braden Scale for Predicting Pressure Sore Risk. Nurs Res 1987; 36:205–10.
15. Ott-Gironomi BA. Pressure ulcer prevalence, incidence and associated risk factors in the community. Decubitus 1993; 6( 5):24–32.
16. Meehan M, O’Hara L, Morrison YM. Report on the prevalence of skin ulcers in a home health agency population. Adv Wound Care 1999; 12:459–67.
17. Ferrell BA, Josephson K, Norvid P, Alcorn H. Pressure ulcers among patients admitted to home care. J Am Geriatr Soc 2000; 48:1042–7.
18. Franks PJ, Winterberg H, Moffatt CJ. Health-related quality of life and pressure ulceration assessment in patients treated in the community. Wound Repair Regen 2002; 10:133–40.
19. Bergquist S, Frantz R. Pressure ulcers in community-based older adults receiving home health care. Prevalence, incidence, and associated risk factors. Adv Wound Care 1999; 12:339–51.
20. Bergquist S, Frantz R. Braden scale: validity in community-based older adults receiving home health care. Appl Nurs Res 2001; 14:36–43.
21. Adams CE, DeFrates DS, Wilson M. Data-driven quality improvement for HMO patients: one agency’s experience with OASIS and OBQI. J Nurs Adm 1998; 28( 10):20–5.
22. Shaughnessy PW, Crisler KS, Schlenker RE. Outcome-based quality improvement in home health care: the OASIS indicators. Top Health Inf Manage 1998; 18( 4):59–69.
23. Martin KS, Scheet NJ. The Omaha System: Applications for Community Health Nursing. Philadelphia, PA: WB Saunders; 1992.
24. National Pressure Ulcer Advisory Panel. Pressure ulcers prevalence, cost and risk assessment: consensus development conference statement. Decubitus 1989; 2( 2):24–8.
25. Friedman G. Primer of Epidemiology. 4th edition. New York, NY: McGraw-Hill; 1994.
26. Olson B, Langemo D, Burd C, Hanson D, Hunter S, Cathcart-Silberberg T. Pressure ulcer incidence in an acute care setting. J Wound Ostomy Continence Nurs 1996; 23( 1):15–25.
27. Eachempati SR, Hydo LJ, Barie PS. Factors influencing the development of decubitus ulcers in critically ill surgical patients. Crit Care Med 2001; 29:1678–82.
28. Allman RM, LaPrade CA, Noel LB, et al. Pressure sores among hospitalized patients. Ann Intern Med 1986; 105:337–42.
29. Brandeis GH, Ooi WL, Hossain M, Morris JN, Lipsitz LA. A longitudinal study of risk factors associated with the formation of pressure ulcers in nursing homes. J Am Geriatr Soc 1994; 42:388–93.
30. Berlowitz DR, Ash AS, Brandeis GH, Brand HK, Halpern JL, Moskowitz MA. Rating long-term care facilities on pressure ulcer development: importance of case-mix adjustment. Ann Internal Med 1996; 124:557–63.
31. Brandeis GH, Berlowitz DR, Hossain M, Morris JN. Pressure ulcers: The Minimum Data Set and the Resident Assessment Protocol. Adv Wound Care 1995; 8( 6):18–25.