An Instrument to Measure Healing in Pressure Ulcers: Development and Validation of the Pressure Ulcer Scale for Healing (PUSH)

Journal of Gerontology: MEDICAL SCIENCES 00, Vol. 56A, No., M795 M799 Copyright 00 by The Gerontological Society of America An Instrument to Measure Healing in Pressure Ulcers: Development and Validation of the Pressure Ulcer Scale for Healing (PUSH) Nancy A. Stotts, George T. Rodeheaver, David R. Thomas, Rita A. Frantz, Alfred A. Bartolucci, 5 Carrie Sussman, 6 Bruce A. Ferrell, 7 Janet Cuddigan, 8 and JoAnn Maklebust 9 for the PUSH Task Force School of Nursing, University of California San Francisco. Plastic Surgery Research, University of Virginia Medical Center, Charlottesville. Division of Geriatrics, Saint Louis University, Missouri. College of Nursing, University of Iowa, Iowa City. 5 University of Alabama Birmingham. 6 Sussman Physical Therapy, Inc., Torrance, California. 7 University of California Los Angeles. 8 College of Nursing, University of Nebraska, Omaha. 9 Harper Hospital/Detroit Medical Center, Detroit, Michigan. Background. Currently, there is no instrument that provides an accurate and simple method of monitoring pressure ulcer healing in clinical practice. This article reports the two studies that were conducted to assess the validity of the Pressure Ulcer Scale for Healing (PUSH) as a tool to monitor healing of stage II IV pressure ulcers. Methods. Subjects in both studies (N 0 and N 69) were elderly (mean Study, 75 years, mean Study, 80 years), and the majority were women (Study, 5%, Study, 70%). Study data were extracted from patients permanent records. Results. Principal components analysis confirmed that the PUSH tool accounted for 58% to 7% of the wound healing variance over a 0-week period in Study and 0% to 57% of the wound healing variance over a -week period in Study. In addition, multiple regression analysis, used to measure the sensitivity of the model to total healing, showed PUSH accounted for 9% of the variance in 6 weeks and % of the variance over weeks (p.00; Studies and, respectively). Conclusions. Data from these two studies confirmed that the PUSH tool, with the components of length times width, exudate amount, and tissue type, is a valid and sensitive measure of pressure ulcer healing. It is a practical approach that provides clinically valid data regarding pressure ulcer healing. Further testing is needed to confirm these findings. P RESSURE ulcers are common ( ), costly (,5), and associated with pain, suffering, and decreased quality of life (6). Moreover, pressure ulcers are extremely slow to heal. Measuring progress toward healing has been recommended to occur on a weekly basis (7). Currently, there is no instrument that provides a uniform, simple method of tracking pressure ulcer healing. Data indicate that changes in ulcer size over a -week period are associated with the ulcer s likelihood of healing completely (8,9). However, the change in size of an ulcer alone may not be a sufficiently sensitive indicator of progress toward healing. Other characteristics that are valid indicators of ulcer healing include the quantity or quality of exudate, presence of necrotic tissue, and an increase in the amount of granulation tissue (0,). Measurement of progress in healing requires the wound be evaluated at more than one setting and that improvement or deterioration in wound status be identified. Although tools exist that allow for overall wound assessment (,), no tool exists to track healing that is accurate, easy to use, and sensitive to change (). Thus, the National Pressure Ulcer Advisory Panel developed the Pressure Ulcer Scale for Healing (PUSH) tool to track healing in stage II to IV pressure ulcers (). The PUSH tool consists of three parameters: length times width, exudate amount (none, light, moderate, and heavy), and tissue type (necrotic tissue, slough, granulation tissue, epithelial tissue, and closed). Each parameter is scored, and the sum of the three yields a total wound status score. Observation of the changes in the direction and magnitude of the score over time indicates whether wound healing is occurring. When developed, the PUSH tool explained 55% to 60% of the variance (p.0) at Weeks through 8. Because the PUSH tool involves only three parameters, it is easy to use and takes less than a minute to complete. The PUSH tool was presented at a national conference, Monitoring Pressure Ulcer Healing: An Alternative to Reverse Staging. Participants provided input regarding parame- M795

M796 STOTTS ET AL. ters, which included definitions, scoring, and directions for use during a structured forum. The recommendations included the following: the addition of tunneling, undermining, and volume; labeling size as length times width; describing exudate characteristics, as well as quantifying the amount; and scoring a healed ulcer as a 0 (5). The concept of PUSH was accepted with the understanding that an additional validation study with a larger number of patients would be conducted to address the recommended modifications. Data files of patients with pressure ulcers containing routine quantitative assessment of wound parameters were solicited for a validation study of the tool. This study reports the additional testing that has been conducted to validate the PUSH for tracking healing of stage II IV pressure ulcers. METHODS AND RESULTS Study After the study was approved by the Committee on Human Research at the University of California-San Francisco, a convenience sample was recruited for this study that included subjects who were adults with a stage II IV pressure ulcer, about which at least weeks of data were available. Subjects were recruited from 0 sites that included clinics, nursing homes, research study sites, and home care settings. One ulcer per patient was selected. A standardized data collection tool was created that included demographic variables and a weekly wound assessment. Study terms were operationally defined and provided in the instructions for the data collection tool. The staging system recommended by the Agency for Health Care Policy and Research s Pressure Ulcer Guideline Panel was utilized (6). Study forms were completed and submitted by each site. Data were scanned or manually entered into a computer program, and statistical analyses were performed. Data were analyzed using SAS for Windows, version 6. (Statistical Analysis Institute, Inc, Cary, NC). Principal components analysis was used to answer the question of whether the PUSH items of surface area, exudate amount, and predominant tissue provided the most efficient measure of healing of pressure ulcers over time (7). Multiple regression was utilized to determine the sensitivity of the tool to measure healing over a -week period. Study : Findings The sample was composed of 7 subjects from 0 sites in eight states (Alabama, California, Connecticut, Iowa, Oregon, Missouri, Nebraska, and Pennsylvania) whose mean age was 7.6 years (SD, 5.0), about half of whom were women (50.7%). Eighty-one percent of the subjects were white, 7% were black, and % other. Complete data for 0 weeks or more were available for 0 subjects and were used for testing the PUSH tool. At baseline, stage III ulcers were recorded most frequently (7.%), followed by stage II (.9%) and stage IV (.0%). A small portion (6.6%) of ulcers could not be staged. The mean ulcer area was 7. cm (SD,.8), and the mean ulcer volume was 5.5 cm (SD,.8). The variables available for analysis were as follows: surface area, surface appearance, wound volume, exudate amount, exudate quality, tunneling, undermining, and wound edge characteristics. Principal components analysis confirmed the PUSH variables of surface area, exudate amount, and surface appearance provided the best model of healing and accounted for 58% to 7% of the variation across the 0 weeks. Insertion of the other variables into the analysis did not significantly enhance the model. To examine how well the model identified changes suggestive of healing over time, pairwise comparisons were done. Healing at baseline, measured by total PUSH score, was significantly different (p.05) from the total scores at each week (Weeks 0), as well as Week versus Weeks to 0, Week versus Weeks to 0, Week versus Weeks 5 to 0, Week versus Weeks 5 to 0, and Week 5 versus Weeks 7 to 0. Scores at Weeks 6 through 0 were not statistically different from each other, indicating the ulcers were healed or nearly healed. Multiple regression techniques indicated that the three variables explained 9% of the variance in healing in the first 6 weeks (p.00). Pilot Testing After validation in Study, the PUSH tool was pilot tested in a long-term care facility. The results indicated that the existing five categories of wound size did not accurately reflect the larger pressure ulcers that are routinely encountered. In addition, the existing definition for scoring predominant tissue was not sensitive to the clinical concerns about the presence of any necrotic tissue. To address the size issue, the decision was made to expand the size variable into 0 categories and test it to see whether these divisions increased the tool s sensitivity. To address the issue of the sensitivity of tissue type, this parameter was modified so a higher score was given for the presence of any necrotic tissue. These new definitions are found in the Appendix. Thus, the original tool has been changed in three ways: (i) two instrument item titles have been refined (size was renamed length width, predominant tissue type was renamed tissue type ); (ii) the length-times-width category has been expanded from five to 0 categories; (iii) the scoring has been simplified (the categories are no longer weighted, so the potential error in calculating a score has been reduced; the tissue-type item provides a mechanism to acknowledge the importance of any slough or necrotic tissue, not just when it covers more than 50% of the wound). To validate that the modified PUSH tool with increased size categories and refined scoring for tissue type was sensitive to changes in wound status, another group of patients with pressure ulcers and quantitative wound data was obtained, analyzed, and reported here as Study. Study The sample was composed of nursing home residents who were participating in the National Pressure Ulcer Long-term Care Study (NPULS) (8). This study was a secondary analysis of an existing data set and was utilized with no subject identifiers and so was exempt from review by the institutional review board. This study included residents who met the following criteria: aged 8 years or older, a resident of a long-term care facility for at least days, and either at risk for or currently experiencing a pressure ulcer. One ulcer per subject was utilized. Only residents who had

PUSH: AN INSTRUMENT TO MEASURE HEALING M797 at least three complete ulcer assessments within weeks were included in our study. Dedicated personnel were trained in data extraction from the NPULS. Agreement among the personnel was greater than 95%. Data from NPULS were electronically downloaded to a central computer, and data for the PUSH analyses were selected. These data were analyzed using SAS for Windows, version 6. (Statistical Analysis Institute, Inc, Cary, NC). Principal components analysis was used to answer the question of whether the PUSH components as modified were validated in this sample. Multiple regression was utilized to determine the sensitivity of the model to measure healing over a -week period. Study : Findings The NPULS included 90 residents, of whom 7 had a pressure ulcer at initial assessment or developed a pressure ulcer during the subsequent -week study period. The mean age of the sample was 80 years (range, 8 0 years). There were 76 (70%) women and 7 (0%) men. Residents came from long-term care facilities. At baseline, stage II ulcers were reported most frequently (n 96), followed by stage I (n 6), stage III (n 9), and stage IV (n 0). The remainder were not staged. Distribution of ulcer size is shown in Figure. These ulcer sizes were used to increase the size categories in the PUSH tool from 5 to 0. The ulcers were located on the sacrum/coccyx (7%), heels (%), buttocks (9%), and other sites (%), including the iliac crests and trochanters. Two hundred sixty-nine subjects met the PUSH study inclusion criteria and were the sample for this testing. Utilizing one ulcer per subject (n 69), 7 healed, and did not heal. Using the increased categories of ulcer size and the revised scoring system for tissue type, principal components analysis showed the PUSH variables of surface area, exudate amount, and tissue type provided the best model of healing and accounted for 9% to 57% of the variation over time for the entire group. This analysis was performed using the revised definition for tissue type and tested in dectiles for surface area. Figure. Percentage of National Pressure Ulcer Long-term Care Study ulcers in each dectile by size (cm ). Pairwise comparisons were performed to explore how well the model discriminated healing over time. Findings showed the total PUSH score at baseline was significantly different (p.05) from the total score at the following: all weeks, Week versus all, Week versus Weeks to, Week versus Weeks 5 to, Week versus Weeks 5 to, and Week 6 versus Weeks 0 to. The other scores were not statistically different from each other, suggesting that the PUSH score in smaller ulcers or at terminal healing velocity is not as sensitive to change as it is in larger ulcers. Multiple regression analysis was used to determine the sensitivity of the tool to measure healing over the -week period. The total PUSH score was the dependent variable, and healed and not healed were the independent variables. Data showed that PUSH accounted for % of the variation (p.00). DISCUSSION A tool to measure healing must accurately reflect clinical reality. The tool must be consistent with outcome, that is, its score increases in wounds that deteriorate and decreases in wounds that heal. The tool must be sensitive to change, that is, the observer can see the progression of healing over time. The tool must demonstrate validity when tested on different populations. To be clinically useful, the tool must be simple to use. The PUSH tool has demonstrated these characteristics in the initial testing and has been revalidated in this study. The PUSH model provides an accurate, simple, and clinically useful way to measure progress toward wound healing. Assessment of the ulcer characteristics of surface area, exudate amount, and surface appearance will provide the clinician with an indication of the healing of pressure ulcers over time. As scores decrease, healing is taking place. The validated PUSH tool is shown in the Appendix. Surface area is measured as length times width. Surface area is expressed in dectiles, rather than as a continuous measure, as this reduces error due to slight differences in measurement. Exudate is estimated as the amount present, scored as none, light, moderate, or heavy. Tissue type refers to the types of tissue present in the ulcer bed. If there is any necrotic tissue present, it is scored as a, and if necrotic tissue is absent but any slough is present, it is scored as a. Granulation tissue receives a score of, and a superficial wound that is reepithelializing receives a. The closed wound receives a score of 0. Other wound characteristics, such as exudate quality, initial ulcer stage, and wound edge status, have been suggested as markers for wound healing. Wound undermining or tunneling has been postulated to give additional information about progress toward healing. However, these variables were not present frequently enough that they added to the sensitivity of the tool. Further work may be needed to validate how often they occur and whether they contribute to the explained variance in healing. The validation sample may have certain limits. Retrospective data were used in these analyses, and the findings are limited by what was documented. In many ways, these analyses represent real-world data and may more closely approximate clinical practice than that of highly controlled situations. Also, the tool is more sensitive to change early in

M798 STOTTS ET AL. healing but is accurate for monitoring healing throughout the repair trajectory. As wounds heal, there is a greater change in surface area initially than when the wound is almost completely closed (8,9). The total score loses some of its discriminatory function as the ulcer closes. This expected loss of sensitivity is due to the difficulty of distinguishing small changes with integer numbers. Further, preliminary analyses using dectiles and revised definitions show that these modifications did not detract from the ability of the original PUSH model () to explain healing. It is conceivable that the failure of PUSH version.0 to provide more explanatory power is due to the short period of time in which healing was monitored. The wounds in Study were large (by design), and large wounds require a longer time to reach complete closure than do small wounds. More power may be obtained when large ulcers are studied if they are followed for a longer period. The PUSH tool demonstrates the best combination of pressure ulcer characteristics that measure progress toward healing. This tool, as any other, is not intended to replace clinical judgment. The PUSH tool does allow a consistent, evidence-based methodology for communicating wound status toward healing among health care providers. It is important to note that the PUSH Score is not designed as a comprehensive assessment paradigm for pressure ulcers and should not be used for that purpose. It is intended to provide a method for ongoing monitoring of healing. Other dimensions of assessment may be the basis for treatment decisions (9). Furthermore, the PUSH tool is not a research tool for measuring healing. It is a clinical tool to assist clinicians with routine evaluation of healing in dayto-day practice. We recommend that PUSH be used to document healing of pressure ulcers as an alternative to reverse staging. Further refinement of the PUSH tool will likely result from prospective clinical studies. However, in its present form, it represents a practical alternative to reverse staging that will provide clinically valid data regarding pressure ulcer healing. Additional testing is needed to confirm these preliminary findings. Acknowledgments This study was funded by the National Pressure Ulcer Advisory Panel. We thank the following individuals for providing subjects for this study: Janet Cuddigan and Joyce Black, Omaha, Nebraska; Catherine Eager and Deann Edgers, McMinnville, Oregon; Rita Frantz, Iowa City, Iowa; Gloria Lott, Hazelwood, Missouri; Cathy Milne and Lisa Corbett, Bristol, Connecticut; Patricia Pasceri, Norristown, Pennsylvania; Ginny Paulick, Bethel Park, Pennsylvania; Fran Schuda, Philadelphia, Pennsylvania; Jan Stanfield, Santa Clarita, California; David Thomas, Saint Louis, Missouri; Anne Voss and the Ross Products Division, Abbott Laboratories, Columbus, Ohio; and the National Pressure Ulcer Long Term Care Study Group. Address correspondence to Nancy A. Stotts, RN, EdD, FAAN, University of California San Francisco, School of Nursing, Koret Way, Room 6, Box 060, San Francisco, CA 9-060. E-mail: nancy.stotts@nursing.ucsf.edu References. Berlowitz DR, Brandeis GH, Anderson J, Brand HK. Predictors of pressure ulcer healing among long-term care residents. J Am Geriatr Soc. 997;5:0.. Bergstrom N, Braden B, Kemp M, Champaign M, Ruby E. Predicting pressure ulcer risk: a multisite study of the predictive validity of the Braden scale. Nurs Res. 998;7:6 69.. Brandeis GH, Berlowitz DR, Hossain M, Morris JN. Pressure ulcers: the minimum data set and the resident assessment protocol. Adv Wound Care. 996;9:8 5.. Miller H, Delozier J. Cost Implications of the Pressure Ulcer Treatment Guideline. Columbia, MD: Center for Health Policy Studies; 99:7. Contract 8-9-0070. 5. Xakellis GC, Chrischilles EA. Hydrocolloid versus saline-gauze dressings in treating pressure ulcers. A cost-effectiveness analysis. Arch Phys Med Rehabil. 99;7:6 69. 6. Szor JK, Bourguignon C. Description of pressure ulcer pain at rest and at dressing change. J Wound, Ostomy, Continence Nurs. 999;6:5 0. 7. van Rijswjk L. Frequency of reassessment of pressure ulcers. Adv Wound Care. 995;8:9S S. 8. van Rijswijk L. Full-thickness pressure ulcers: patient and wound healing characteristics. Decubitus. 99;6:6. 9. van Rijswijk L, Polansky M. Predictors of time to healing deep pressure ulcers. Wounds. 99;6:59 65. 0. Maklebust J. Pressure ulcer staging systems. Adv Wound Care. 995; 8():.. Bates-Jensen BM, Vredevoe DL, Brecht ML. Validity and reliability of the Pressure Sore Status tool. Decubitus. 99;5:0 8.. Ferrell BA, Artinian BM, Sessing D. The Sessing scale for assessment of pressure ulcer healing. J Am Geriatr Soc. 995;:7 0.. Thomas D. Existing tools: are they meeting the challenges of pressure ulcer healing? Adv Wound Care. 997;0:86 90.. Thomas DR, Rodeheaver GT, Bartolucci AA, et al. Pressure ulcer scale for healing: derivation and validation of the PUSH tool. Adv Wound Care. 997;0:96 0. 5. Maklebust J. PUSH tool reality check: audience response. Adv Wound Care. 997;0:0 06. 6. Bergstrom N, Allman RM, Alvarez OM, et al. Treatment of Pressure Ulcers. Rockville, MD: U.S. Department of Health and Human Services; 99. AHCPR publication 95-065. 7. Bartolucci AA, Thomas DR. Using principal components analysis to describe wound status. Adv Wound Care. 997;0:9 95. 8. Preventing pressure ulcers in long-term care residents [abstract]. Abstract Book/Assoc Health Serv Res. 999;6:0. 9. Schubert V, Zander M. Analysis of the measurement of four wound variables in elderly patients with pressure ulcers. Adv Wound Care. 996;9:9 6. Received January 5, 00 Accepted February, 00 Decision Editor: John E. Morley, MB, BCh

PUSH: AN INSTRUMENT TO MEASURE HEALING M799 Appendix PUSH (Version.0: 9/5/98, National Pressure Ulcer Advisory Panel) Patient Name: Patient ID#: Ulcer Location: Date: 5 Subscore Directions: Observe and measure the pressure ulcer. Categorize the ulcer with respect to surface area, exudate, and type of wound tissue. Record a subscore for each of these ulcer characteristics. Add the subscores to obtain the total score. A comparison of total scores measured over time provides an indication of the improvement or deterioration in pressure ulcer healing. Length Width 0 0 cm 0. cm 0. 0.6 cm 0.7.0 cm..0 cm..0 cm 6..0 cm 7. 8.0 cm 8 8..0 cm 9..0 cm 0 cm Exudate Amount 0 None Tissue Type 0 Closed Light Epithelial Tissue Moderate Granulation Tissue Heavy Slough Necrotic Tissue Subscore Subscore Total Score Length width: Measure the greatest length (head to toe) and the greatest width (side to side) using a centimeter ruler. Multiply these two measurements (length width) to obtain an estimate of surface area in cm. Caveat: Do not guess! Always use a centimeter ruler and always use the same method each time the ulcer is measured. Exudate amount: Estimate the amount of exudate (drainage) present after removal of the dressing and before applying any topical agent to the ulcer. Estimate the exudate (drainage) as none, light, moderate, or heavy. Tissue type: This refers to the types of tissue that are present in the wound (ulcer) bed. Score as a if there is any necrotic tissue present. Score as a if there is any amount of slough present and necrotic tissue is absent. Score as a if the wound is clean and contains granulation tissue. A superficial wound that is reepithelializing is scored as a. When the wound is closed, score as a 0. - Necrotic tissue (eschar): black, brown, or tan tissue that adheres firmly to the wound bed or ulcer edges and may be either firmer or softer than surrounding skin. - Slough: yellow or white tissue that adheres to the ulcer bed in strings or thick clumps, or is mucinous. - Granulation tissue: pink or beefy red tissue with a shiny, moist, granular appearance. - Epithelial tissue: for superficial ulcers, new pink or shiny tissue (skin) that grows in from the edges or as islands on the ulcer surface. 0 - Closed/resurfaced: the wound is completely covered with epithelium (new skin).