Critical Evaluation of the Jackson/Cubbin Pressure Ulcer Risk Scale — A Secondary Analysis of a Retrospective Cohort Study Population of Intensive Care Patients
Although the Jackson/Cubbin pressure ulcer (PU) risk scale performs best among risk scales used in intensive care units (ICUs), its performance was not fully satisfactory. In 2010, a minimally modified Jackson/Cubbin (mJ/C) PU risk scale was introduced to formalize PU risk assessment in a large medical-surgical ICU in Finland. The purpose of this secondary analysis was to examine whether individual categories of the mJ/C scale have similar weight and whether the scores within each category (from 1 to 4; 1 equaling highest risk and 4 equaling lowest risk) are linear, as is assumed for the original and modified scales.
Using data from a cohort of 1,616 consecutively admitted patients retrieved from the ICU database, a detailed secondary analysis of each of the 12 main scoring categories of the Jackson/Cubbin risk scale was performed using logistic regression and analysis of linearity and weight. Of the 1,616 admitted patients, 168 developed a PU during their ICU stay. Among the risk categories, body mass index, nutrition, respiration, age, and transportation during the 48 hours before scoring did not contribute significantly (P >0.05) to the total risk score or the actual development of a PU. The 7 other main categories — incontinence, mobility, medical history, oxygen requirement, need for assistance with hygiene, hemodynamics, and general skin condition — were the main risk contributors. Although only the linearity of the different categories correlated significantly with the predictive value of the categories, the linearity as well as the weights of the categories were at variance from what was assumed originally. The mJ/C scale needs refinement to be a more accurate instrument for PU risk assessment of ICU patients. Not all mJ/C categories were found to contribute to the risk and, when they do, their weight and linearity vary from what has been assumed. The categories respiration and oxygen requirement and the categories mental condition, mobility, and hygiene may overlap. The importance of the incontinence category depends on the frequency of urinary and fecal incontinence management system usage. A simpler, more valid and more sensitive risk assessment scale than the current Jackson/Cubbin scale is needed for ICU patients.
Patients in intensive care units (ICUs) are at high risk of developing pressure ulcers (PUs).1-5 Several risk scales have been developed for general PU risk assessment.6,7 The Braden Scale often is used in ICUs.3,5,8 A survey by VanGilder et al5,8 comprising ~11,000 patients in more than 500 ICUs showed PU incidence varied between 8.8% and 12.3% and close to half of the patients had a Braden score above 14, indicating low to medium risk.
A specific risk scale for intensive care patients, the Jackson/Cubbin (J/C) pressure area risk calculator9 includes 4 out of the 5 Norton and 5 out of the 6 Braden categories (the missing Braden category involves friction and shear). The J/C pressure area risk calculator consists of 12 categories, each of which is graded linearly from 1 point (highest risk) to 4 points (lowest risk) to describe the clinical risk of PU for ICU patients9 (see Table 1). The lower the score, the higher the PU risk — a total score 29 signifies extremely high risk. Among the additional main categories in the J/C scale are age, weight, past medical history, and skin condition plus ICU-specific categories such as respiration, oxygen requirements, and hemodynamics.
The J/C scale has not gained wide acceptance, although its receiver operating characteristic (ROC), curve and sensitivity, and specificity have been the best of several scales proposed for ICU use.5,10-12 Specifically, the validation study by Seongsook et al10 comparing the J/C, Braden, and Douglas Scales included 112 ICU patients and found the specificity (61%) and positive predictive value (51%) of the J/C scale to be better than those of the Braden (26% and 37%, respectively) and Douglas scales (18%, 34% respectively). The review by Shanin et al11 compared the J/C, Braden, Douglas, Waterloo, and Norton scales and included 1,150 ICU patients in 7 studies; it concluded only the J/C scale was developed for assessment of PU risk in ICU patients, and the assessment of validity of different risk scales varied. The systematic review by García-Fernández et al12 presented 15 scales and 5,187 patients in 26 studies; the Braden scale was examined in 13 studies (1,458 patients) and J/C scale in 5 studies (629 patients). The authors recommended the Braden scale for use in PU risk assessment in ICUs. Both the Braden Scale and J/C risk scale are based on the Norton scale.13
In 2010, a retrospective research program utilizing the ICU database was launched by the present study team to study the epidemiology and risk factors related to PUs in a large medical-surgical ICU. The program included introduction of a modified version of the J/C (mJ/C) risk scale in an attempt to formalize PU risk assessment14 based on previous studies.10,11 Briefly, the previous results implied the J/C risk scale and its predefined cut-off score of 29 do not sufficiently identify extremely high-risk patients because 55.8% of patients with PUs fell into the ≤29 score group, while 46.6% patients without PUs had a similar score.13 The sensitivity of the mJ/C scale was 55.8%, the specificity was 53.4%, the positive predictive value was 13.4%, and the negative predictive value was 90.3%. The total incidence of PUs was 11.1% (181/1,629); 13 patients with exclusively nasal PUs were excluded from the analysis.
Due to the limited mJ/C scale score separation of PU from nonPU patients and to the modest sensitivity and specificity, this retrospective secondary analysis was conducted to examine whether the individual categories of the mJ/C scale have similar weight (ie, each subcategory is scored from 1 to 4) and whether the scores within each category from 1 to 4 (1 meaning highest risk and 4 meaning lowest risk) are linear, as is assumed for the J/C scale. The hypothesis was not all factors contribute equally to PU development, as originally described by Jackson in 1999.9 A specific examination was conducted to determine more specifically what factors are related to the development of PUs in the ICU.
Patients. The Turku University Hospital serves a population of 700,000. The adult ICU has 24 beds. The hospital serves as a national center for hyperbaric oxygen therapy. All surgical and medical intensive care patients in the region are treated in this hospital, except patients with major burns and patients undergoing solid organ transplantation, who are treated elsewhere. Based on the treatment needs on admission, patients are classified as ICU-patients or as patients needing high-dependency (ie, step down) care.
Patient data from all 2010 ICU admissions (1,629 patients >18 years old, mean age 60.5 years [range 18–95 years], average length of stay 3.7 days) were retrospectively analyzed to study the occurrence of PUs and their risk factors in medical and surgical patients.14 Patients with nasal PUs only (n = 13) were excluded from the analysis, leaving 1,616 patients for the current secondary analysis.
The main categories of the J/C scale are presented in Table 1. In addition to these 12 main categories, 3 additional assessments are made that may lower the PU risk score (see Table 1, “deduct points”). The J/C risk scale9 from 1999 was back-translated and forward-translated into Finnish for validation purposes, and some of the categories were more precisely defined (mJ/C risk scale) to improve the clarity and reproducibility of the assessments (revised sections are marked as bolded in Table 1). The age and body mass index (BMI) values are automatically calculated by the clinical documentation and information system used by the ICU (Clinisoft, GE Healthcare, Buckinghamshire, UK). It takes 2 to 3 minutes for an ICU nurse to calculate risk and enter the information into the ICU database.
All data used in this study were retrieved from records, abstracted (ie, checked by computer for their correct format), and entered into a database for analysis.
The study plan was approved by the Ethics Committee of the Hospital District of Southwest Finland.
Statistical methods. Statistical evaluation of the main categories of the mJ/C risk scale was based on exploratory analysis and regression modeling. Linearity was scored from +4 (arbitrary unit) when linearity was as described in the J/C scale9 (see Table 1) to -4 if linearity was reversed. The relative weight of each individual J/C category was determined by multiplying each subcategory score point by the number of patients in that subcategory and then summing the 4 scores together and dividing the sum by the total number of patients in that category.
Logistic regression modeling was applied to evaluate the association between PUs and the categories of the J/C PU risk calculator. A proportional-odds cumulative logit model15 was used because all of the J/C PU risk categories are ordinal (1, 2, 3, 4) and the risk for PUs is expected to increase with lower scores (ie, 1 point represents the highest risk value of an individual risk category9). Separate analyses were performed for all risk categories to evaluate the marginal independence/association of PUs and individual risk categories. Odds ratio (OR) and the Wald chi-square (2) test, based on logistic regression modeling, were used to evaluate the association between PUs and the risk categories. Wald 2 is calculated as:
MLE (PU) is the maximum likelihood estimated logistic odds of PUs effect. OR, the Wald 2-test, and the related P values (same for both) were used to evaluate the statistical null hypothesis, H0. The distribution of mJ/C PU risk category values in a population with PUs was equal to the distribution in a population with no PUs. A P value <0.05 indicated sufficient evidence in favor of the alternative hypothesis, H: values of a mJ/C PU risk category were statistically significantly higher in a population with PUs than in a population with no PUs. Fisher’s exact test16 and the 2 test were used to test statistical significance of the contingency tables. Both tests were used to determine whether the incidence rates of PUs were the same in 2 groups that were compared (ie, if the incidence of PUs was independent of the risk factor or of other predictive variables15,16).
Fisher’s exact test is designed to test the significance of contingency tables when the sample size is small or some cells of a contingency table number <5. This test is more conservative than the 2 test and may yield higher P values.16
The results of the secondary analysis of each individual risk category of the mJ/C scale showed mobility, medical history, oxygen requirements, hygiene, hemodynamics (linearity score +2, relative weight 2.0; OR 2.232, P <0.001) and general skin condition (linearity score +4, relative weight 3.6; OR 3.272, P <0.001) significantly predicted PU development (see Tables 2 and 3). Incontinence (linearity score 0, relative weight 4.0; OR 3.910, P = 0.0126) and mental condition (linearity score +1, relative weight 2.2; OR 1.481, P = 0.0116) were significant predictors of PU, while BMI, nutrition, respiration, and age were not associated with the development of PUs (see Tables 2 and 3).
A trend was noted toward a linearity of the risk score point distribution of the categories that were significantly associated with the PU risk (see Table 2, Table 3, Figure 1a). The linearity of the incontinence category could not be determined due to too few patients in the score point groups 1 to 3 (see Table 3). Only general skin condition contributed linearly to PU risk, as defined in the J/C risk scale (see Table 1, Table 3, Figure 1a). Oxygen requirements and hygiene contributed almost linearly, and BMI, nutrition, and respiration were almost the opposite (ie, contralinear) (see Table 3, Figure 1a). A significant correlation was noted between linearity and the predictive value of the categories (OR): rs = 0.694 (P = 0.0123, Spearman’s rank correlation17).
The J/C risk scale assumes the relative weight of each J/C category is 2.5. The weights of the different risk categories that contributed to the PU risk differed from 2.5 and from each other (see Table 3, Figure1b). The relative weight of the hygiene category (1.6) was 3 times that of the nutrition (3.9) and incontinence (4.0) categories (see Figure 1b). The weight contribution of the other categories fell somewhere between these values. No correlation was noted between the relative weights of the categories and their predictive value (OR): rs = 0 (P = 1, Spearman’s rank correlation17).
Two thirds of patients were transported within the hospital either to surgery or for examinations, but only a small number of patients (n = 20) qualified for the transportation group, because transportation had to have occurred within 48 hours of mJ/C scoring (see Table 4). All of the 20 patients who had been transported developed a PU (P <0.001, Fisher’s exact test) compared to the nontransport (PU n = 43, 8.5% of that group) and transport group (PU n = 105, 10% of that group). Of the patients who had been transported, 14 were in the high-risk group, as defined by the current J/C risk score (29), 3 patients were very close to the cut-off limit with a J/C score of 30–31, and 3 had scores of 34, 37, and 39 when transportation and other conditions requiring application of the deduction of points were considered (see Table 1).
Because development of PUs is a multicausal problem, an adequate PU risk scale combined with clinical evaluation is important for overall PU risk assessment. The J/C risk scale is specifically designed for intensive care and has performed optimally for ICU patients.10,11 However, its performance in an unselected ICU population is satisfactory only to some extent, as shown by the previous study14: the separation of low-risk and high-risk patient groups and its sensitivity and specificity were suboptimal.14 Therefore, in this study, each of the scoring categories9 of the J/C risk scale were analyzed critically.
Among the general population and ICU patients, the prevalence of PUs increases with age.2,5,18 This observation is corroborated by the results of the present study among patients older than 70 years, but, as a whole, the importance of age as a predictor is negligible. In this study, low BMI values (<18 kg/m2) contributed to the PU risk and high BMI values did not, even though both are linked to PU development in the general population.19-21 A low BMI may be a PU risk factor (see Table 3) because it is associated with a minimal amount of subcutaneous fat, which may increase the deleterious effects of pressure from underlying osseous protuberances on the tissues.21,22 The mental condition, mobility, and hygiene categories contained definitions that are not sufficiently stringent; they are partly interchangeable or superfluous and this impairs their value.9,21,23 Still, it is understandable that mobility does contribute to PU risk because mobility is a crucial consideration in general risk assessment scales.20,23 ICU patients are rarely able to take care of themselves and need much assistance from staff due to their limited mobility and limited ability to independently maintain hygiene. Not surprising, most PUs develop in patients that need assistance when friction and shear on the skin can occur if the procedures to assist the recumbent patient are not performed correctly (see Table 3).
Moist skin, especially due to urine and feces, is a major risk factor for PUs.6,23,24 In the ICU of the Turku University Hospital, most patients have urinary catheters, and fecal catheters are used when needed to keep the skin as dry as possible. This is represented in the point distribution in the incontinence category (see Table 3). Despite a wide confidence interval, the incontinence category reached significance (see Table 2).
The point distribution of the general skin condition category correctly reflects the PU risk: the risk is lowest when the skin is intact and highest when deep or secreting wounds are present9 (see Table 3). This is in accordance with what is known about the pathophysiology of PUs, because wounds are associated with an inflammatory reaction and often with infection and fever, release of free radicals, and pro-inflammatory cytokines.25-27 Wounds often produce high volumes of fluid, which keeps the skin moist and severely impairs positioning therapy because patients cannot be positioned on their large wounds.
Although definitions regarding respiration and oxygen are partly interchangeable or superfluous (see Table 1), both are related to tissue oxygenation, which has been examined only minimally as a risk factor for PU development.28 The contribution of respiration and oxygen requirements to PU risk differed significantly from each other (see Tables 2 and 3), probably because the respiration mode does not correlate with tissue oxygenation in practice. Oxygen requirements contributed to the risk of PU development in this study, especially the highest risk (1 point = >60% O2) (see Table 3). This probably reflects the situation where lower oxygen concentrations keep the tissue oxygen levels sufficiently stable, while patients needing a great deal of oxygen are so ill their tissue oxygen tension is insufficient to support tissue viability and prevent PUs. This assumption is supported by the findings in the hemodynamics category. Although the current scoring order was not optimal (see Table 3), the incidence of PUs is higher in patients who are unstable regardless of whether they are on or off inotropes (see Table 3).
With respect to PU risk score point deduction, anemia and hemoglobin concentration scoring may replace the need for blood products as an assessment category.6,29 Although 20 patients underwent transportation within 48 hours (see Table 4) before scoring, most of them also were otherwise classified as high risk based on their mJ/C score (29). Because, overall, two thirds of the patients were transported (most earlier than 48 hours before scoring), the contribution of this part of the mJ/C score seems to be insignificant relative to the overall score.
Although only the linearity of the mJ/C categories correlates significantly with the predictive values of PU development, the results show point order (1 as highest and 4 as lowest risk) is not correct in most categories. The J/C risk scale assumes the relative weight of each of the categories is 2.5, which was not the case, because the patients are distributed differently within each of the categories. A 3-fold difference was noted between the highest and lowest relative weight categories, and no correlation was found between the relative weights and predictive values of the mJ/C categories, meaning differences exist among the various categories of the mJ/C scale but this is not a reflection of the true weights of the different categories. This matter requires additional research.
The retrospective retrieval of the data infers individual contributions of the nurses to the actual risk assessment could not be controlled. However, this is the largest single study of a consecutive ICU patient population. Also, the number of PUs was only 168, which leaves room for some chance effects in the results, because the number of both PU and nonPU patients is quite small in some of the mJ/C subcategories.
It is not an easy task to identify specific risk indicators for severely ill ICU patients with multiple disorders. The J/C risk scale is based on certain categories that have been assumed to be linear and of equal weight with regard to the risk of acquiring a PU. In this study, these assumptions were found to be invalid. Seven of the categories contribute substantially to the overall score and actual risk: incontinence, mobility, medical history, oxygen requirement, hygiene, hemodynamics, and general skin condition. This also indicates several risk categories of the J/C risk scale correlate and associate among each other, although this specific matter was not examined in this study. Because of this association, some categories measure the same pathophysiological phenomena. In addition, it was found the order of the risk points is suboptimal within the categories of the J/C risk scale. These results should prompt future research aimed at introducing an individualized order and weight of individual risk categories into the assessment of the risk of PUs with the mJ/C scale. Furthermore, numerous other PU risk factors30 such as admission diagnosis, body temperature, Sequential Organ Failure Assessment score, and hemoglobin concentration should be tested either by prospective cohort studies or retrospective analyses of existing large patient cohorts to identify the most important PU risk factors in the ICU setting. These variables were not examined in this study because they are not part of the risk evaluation scale.
There seems to be room for a simpler scale for PU risk assessment and risk prediction for ICU patients. Such a scale should include only those risk indicators that constitute separate, predictive risk categories. This study shows that, currently, the mJ/C risk scale in combination with clinical assessment is a feasible way to assess the risk of PU among ICU patients, but more research is needed to identify the most effective predictors.
1. Takala J, Varmavuo S, Soppi E. Prevention of pressure sores in acute respiratory failure: a randomized, controlled trial. Clin Intensive Care. 1996;7(5):228–235.
2. Bours GJ, De Laat E, Halfens RJ, Lubbers M. Prevalence, risk factors and prevention of pressure ulcers in Dutch intensive care units. Results of a cross-sectional survey. Intensive Care Med. 2001;27(10):1599–1605.
3. Kottner J, Wilborn D, Dassen T, Lahmann N. The trend of pressure ulcer prevalence rates in German hospitals: results of seven cross sectional studies. J Tissue Viability. 2009;18(2):36–46.
4. Terekeci H, Kucukardali J, Top C, Onem Y, Celik S, Öktenli C. Risk assessment study of the pressure ulcers in intensive care unit patients. Eur J Int Med. 2009;20(4):394–397.
5. VanGilder C, Amlung S, Harrison P, Meyer S. Results of the 2008 – 2009 International Pressure Ulcer Prevalence Survey and a 3-year, acute care, unit-specific analysis. Ostomy Wound Manage. 2009;55(11):39–45.
6. National Pressure Ulcer Advisory Panel and European Pressure Ulcer Advisory Panel. Pressure Ulcer Prevention and Treatment: Clinical Practice Guideline. Washington DC: National Pressure Ulcer Advisory Panel 2009. Available at: www.epuap.org. Accessed January 15, 2016.
7. García-Fernández FP, Soldevilla Agreda JJ, Verdú J, Pancorbo-Hidalgo PL. A new theoretical model for the development of pressure ulcers and other dependence-related lesions. J Nurs Scholar. 2014;46(1):28–38.
8. VanGilder C, MacFarlane GD, Meyer S. Results of nine international pressure ulcer prevalence surveys: 1989 to 2005. Ostomy Wound Manage. 2008;54(2):40–54.
9. Jackson C. The revised Jackson/Cubbin Pressure Area Risk Calculator. Intensive Crit Care Nurs. 1999;15(3):169–175.
10. Seongsook J, Ihnsook J, Younghee L. Validity of pressure ulcer risk assessment scales; Cubbin and Jackson, Braden, and Douglas scale. Int J Nurs Stud. 2004;41(2):199–204.
11. Dassen T, Halfens RJ, Shanin ES. Predictive validity of pressure ulcer risk assessment tools in intensive care patients. World Crit Care Nurs. 2007;5(3):75–79.
12. García-Fernández FP, Pancorbo-Hidalgo PL, Soldevilla Argeda JJ, Rodrigez Torres MC. Risk assessment scales for pressure ulcers in intensive care units: a systematic review with meta-analysis. EWMA J. 2013;13(2):7–13.
13. Norton D, McLaren R, Exton-Smith AN. An Investigation of Geriatric Nursing Problems in Hospital. London, UK: National Corporation for the Care of Old People (now Centre for Policy on Ageing);1962. [Available as loan through special libraries.]
14. Ahtiala M, Soppi E, Wiksten A, Koskela H, Grönlund J. Occurrence of pressure ulcers and their risk factors in mixed medical-surgical ICU — a cohort study. J Intensive Care Soc. 2014;15(4):2–4.
15. McCullagh P. Regression models for ordinal data. J R Stat Soc. Series B (Methodological), 1980;42(2):109–142.
16. Fisher R. A. On the interpretation of 2 from contingency tables, and the calculation of P. J R Stat Soc 1922;85(1): 87–94.
17. Myers JL, Well AD. Research Design and Statistical Analysis, 2nd ed. London, UK: Lawrence Erlbaum Publisher;2003:508.
18. Theaker C, Mannan M, Ives N, Soni N. Risk factors for pressure sores in the critically ill. Anaesthesia. 2000;55(3):221-224.
19. VanGilder C, MacFarlane G, Meyer S, Lachenbruch C. Body mass index, weight, and pressure ulcer prevalence: an analysis of the 2006-2007 International Pressure Ulcer Prevalence Surveys. J Nurs Care Qual. 2009;24(2):127–135.
20. Soppi E, Iivanainen A, Korhonen P. Concordance of Shape Risk Scale, a new pressure ulcer risk tool, with the Braden scale. Int Wound J. 2014;11(6):611–615.
21. Kottner J, Gefen A, Lahmann N. Weight and pressure ulcer occurrence: a secondary data analysis. Int J Nurs Stud. 2011;48(11):1339–1348.
22. Cubbin B, Jackson C. Trial of a pressure area risk calculator for intensive therapy patients. Intensive Care Nurs. 1991;7(1):40–44.
23. Braden B, Bergstrom N. A conceptual schema for the study of the etiology of pressure sores. Rehabil Nurs. 1987;12(1):8–12.
24. Bayón García C, Binks R, De Luca E, Dierke C, Franci A, Gallert E, et al. Expert recommendations for managing acute faecal incontinence with diarrhea in the intensive care unit. JICS. 2013;14(suppl 2):1–9.
25. Friedland JS, Porter JC, Daryanani S, Bland JM, Screaton NJ, Vesely MJJ, et al. Plasma proinflammatory cytokine concentrations, Acute Physiology and Chronic Health Evaluation (APACHE) III scores and survival in patients in an intensive care unit. Crit Care Med. 1996;24(11):1775–1781.
26. Vande Berg JS, Rose MA, Haywood-Reid PL, Rudolph R, Payne WG, Robson MC. Cultured pressure ulcer fibroblasts show replicative senescence with elevated production of plasmin, plasminogen activator inhibitor-1, and transforming growth factor-1. Wound Repair Regen. 2005;13(1):76–83.
27. Bronneberg D. Biochemical markers for early detection of superficial pressure ulcers. Academic dissertation 2007, Technische Universiteit Eindhoven, Eindhoven, Holland.
28. Manzano F, Navarro MJ, Roldán D, Moral MA, Leyva I, Guerrero C, et al. Pressure ulcer incidence and risk factors in ventilated intensive care patients. J Crit Care. 2010;25(3):469–476.
29. Nixon J, Nelson EA, Cranny G, et al on behalf of the PRESSURE Trial Group. Pressure relieving support surfaces: a randomised evaluation. Health Technol Assess. 2006;10(22):1–179.
30. García-Fernández FP, Soldevilla Agreda JJ, Verdú J, Pancorbo-Hidalgo PL. A new theoretical model for the development of pressure ulcers and other dependence-related lesions. J Nurs Scholar. 2014;46(1):28–38.
Potential Conflicts of Interest: Dr. Soppi is the chairman of the board of Carital Group, Helsinki, Finland, a group of companies manufacturing and marketing support surfaces.
Ms. Ahtiala is an intensive care unit RN and an authorized wound nurse, Turku University Hospital, Turku, Finland. Dr. Soppi is a consultant in internal medicine, Eira Hospital, Helsinki, Finland. Mr. Kivimäki is a statistician, StatFinn Ltd, Turku, Finalnd. Please address correspondence to: Maarit H. Ahtiala, RN, Intensive Care Unit, Turku University Hospital, Hämeentie 11, Turku 20520 Finland; email: firstname.lastname@example.org.