THE COMPLICATIONS OF traumatic spinal cord injury

Transcription

1 87 Nontraumatic Spinal Cord Injury Rehabilitation: Pressure Ulcer Patterns, Prediction, and Impact Peter W. New, MBBS, FAFRM (RACP), H. Barry Rawicki, MBBS, FACRM, FAFRM (RACP), Michael J. Bailey, MSc ABSTRACT. New PW, Rawicki HB, Bailey MJ. Nontraumatic spinal cord injury rehabilitation: pressure ulcer patterns, prediction, and impact. Arch Phys Med Rehabil 2004; 85: Objective: To investigate the characteristics, predictors, and consequences of pressure ulcers in patients with nontraumatic spinal cord injury (SCI). Design: Retrospective, 3-year, case series. Setting: Tertiary medical unit specializing in SCI rehabilitation. Participants: Consecutive sample of 134 adult inpatient referrals with nontraumatic SCI. Patients requiring initial rehabilitation or readmission were included. Intervention: Chart review. Main Outcome Measures: Primary outcome measures were presence of pressure ulcers on admission to rehabilitation, incidence of new pressure ulcers developing during hospitalization, and any complications attributable to pressure ulcers during inpatient rehabilitation. Secondary objectives were to examine the predictability of risk factors for pressure ulcers, to assess the usefulness of a model previously developed for predicting pressure ulcers in patients with chronic SCI, and to estimate the effect of pressure ulcers on rehabilitation of nontraumatic SCI. Results: Prevalence of pressure ulcers among admissions was 31.3% (n 42). Only 2.2% (n 3) of patients developed a new pressure ulcer after admission. The length of stay (LOS) of patients admitted with a pressure ulcer was significantly longer than that of those without a pressure ulcer (geometric mean, 62.3d for pressure ulcer vs 28.2d for no pressure ulcer, P.0001). Many previously identified risk factors for pressure ulcers in SCI patients did not apply to our nontraumatic SCI patients. It is estimated that the inpatient LOS for those patients with a significant pressure ulcer was increased by 42 days. Conclusions: Pressure ulcers are a common complication for people with nontraumatic SCI who are admitted for rehabilitation, and they have a significant impact on LOS. Key Words: Decubitus ulcer; Rehabilitation; Spinal cord injuries; Spinal cord diseases. From the Head of the Spinal Rehabilitation Unit, Caulfield General Medical Centre, Bayside Health, Melbourne, Victoria (New); Departments of Epidemiology & Preventive Medicine (New, Bailey) and Medicine (Rawicki), Monash University, Melbourne, Victoria, Australia. Presented as a poster at the American Academy of Physical Medicine and Rehabilitation s 61st Annual Assembly and the 13th World Congress of the International Federation of Physical Medicine and Rehabilitation, November 13, 1999, Washington, DC. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Correspondence to Peter New, MBBS, Head Spinal Rehabilitation Unit, Caulfield General Medical Center, 260 Kooyong Rd, Caulfield 3162, Melbourne, Victoria, Australia, p.new@cgmc.org.au. Reprints are not available from the author /04/ $30.00/0 doi: /s (03) by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation THE COMPLICATIONS OF traumatic spinal cord injury (SCI) have been well studied. 1-7 There has been comparatively little research into the complications that occur in patients with nontraumatic SCI. 8,9 Pressure ulcers are a relatively common and potentially serious complication for all persons with SCI, with a reported lifetime likelihood of up to 85%. 10 Many persons with SCI develop pressure ulcers during the acute hospitalization phase after their injury, with the prevalence in this setting reported to range from 20% to 60%. 11 A recent study 12 reported a 17.1% prevalence of pressure ulcers among nontraumatic SCI patients on admission to rehabilitation. In a study involving only patients with cancer-related SCI, 8 the prevalence was 9.4%. There is little published research that quantifies the impact of pressure ulcers on the length of stay (LOS) of traumatic SCI patients in rehabilitation. 13 Studies of chronic traumatic SCI suggest that the risk of pressure ulcers increases with time after injury, with patients who are more than 10 years after traumatic SCI being at an increased risk. 11 It is not known, however, whether this also applies to persons with nontraumatic SCI. Many risk factors for pressure ulcers have been identified, 14,15 but most tools for predicting pressure ulcers have been developed in geriatric populations. 10 A few studies 10,16,17 have reported how pressure ulcer risk factors apply to patients with traumatic SCI. These factors, however, may not apply to patients with nontraumatic SCI because their demographic characteristics particularly regarding their increased age 9,12,18 and comorbid diagnoses 9 differ significantly. The Spinal Rehabilitation Unit at Caulfield General Medical Center, Caulfield, Victoria, Australia, is a 9-bed inpatient rehabilitation unit that provides specialist interdisciplinary rehabilitation, principally for adults with nontraumatic SCI. The patients are usually residents of Melbourne, but residents who live elsewhere in the State of Victoria are also accepted. Patients are admitted into the program once they are medically stable. The unit is in a unique position to research this understudied group of patients. Our primary objectives in this study were to measure the following: the prevalence of pressure ulcers in patients with nontraumatic SCI who were admitted to a rehabilitation unit; the incidence of new pressure ulcers that develop during a rehabilitation admission; the incidence of complications from pressure ulcers that occur during rehabilitation; and the grade and location of pressure ulcers in patients with nontraumatic SCI. One secondary objective was to examine the utility of a previously developed model with which to predict the development of pressure ulcers in patients with nontraumatic SCI. We also wanted to quantify the effect of pressure ulcers on the rehabilitation of such patients and to assess the influence of the grades of pressure ulcers on LOS and the time spent to heal

2 88 NONTRAUMATIC SCI PRESSURE ULCERS, New them. Among the readmission group of patients, we wanted to examine the role of time after injury in the development of pressure ulcers. METHODS The medical records of 134 patients who were admitted to the Spinal Rehabilitation Unit between January 1, 1995, and December 31, 1997, were reviewed by the first author (PWN). Seventy of the 134 were admitted for the first time (initial admission) for rehabilitation of a nontraumatic SCI; 64 were readmitted (readmissions) to the unit because of postdischarge problems (mean standard deviation [SD], y; range, 3mo to 56y after). Twenty-four (34.3%) of the initial admissions had an existing ulcer, as did 18 (28.1%) of the readmissions. The following information was extracted from the medical records: demographic characteristics, neurologic level of injury, and completeness of nontraumatic SCI; etiology of nontraumatic SCI; and comorbidities and complications, with a focus on pressure ulcers and factors believed to influence pressure ulcer development and their healing in SCI patients. The diagnosis of complications was based on both clinical features and relevant investigations. Patients were excluded if they did not have a nontraumatic SCI or if they had a peripheral nerve injury that was not associated with myelopathy or cauda equina involvement. The completeness of nontraumatic SCI was assessed according to the International Standards for Neurological and Functional Classification of Spinal Cord Injury. 19 Based on a review of risk factors for pressure ulcers in SCI, 10 we categorized patients in a binary fashion for the following risk factors: living in a nursing home or hostel, tetraplegia, American Spinal Injury Association (ASIA) class A, B, or C, diabetes, current smoker, ischemic heart disease or congestive cardiac failure, chronic obstructive pulmonary disease, hypoalbuminemia (serum albumin, 3.0g/dL), anemia (hemoglobin, 10g/dL), and significant spasticity. We also included the following factors that we believed would be relevant to the nontraumatic SCI population: male, living alone, initial rehabilitation admission after acute injury, not working before admission, cerebrovascular or peripheral vascular disease, and urinary tract infection (UTI) during admission. Age was examined as a continuous variable. The above factors were examined as possible predictors of pressure ulcers. Salzberg et al 20 developed a model to predict pressure ulcers in a group of chronic and mostly traumatic SCI patients. They identified the following factors as being important: level of activity (ambulatory, wheelchair, bed), mobility (full, limited, immobile), complete SCI, urinary incontinence or continually moist, autonomic dysreflexia or severe spasticity, age ( 35y, 35 64y, 64y), smoking status (never, former, current), pulmonary disease, cardiac disease, diabetes, renal disease, impaired cognitive function, nursing home or hospital resident, serum albumin less than 3.4g/dL or protein less than 6.4g/dL, and hematocrit less than 36% or hemoglobin less than 12g/dL. Patients are given a weighted score for each factor, and the total score is used to predict development of pressure ulcers. Total scores in the model can range between 0 and 25; a cutoff point of 6 or higher is the most predictive of pressure ulcers. The model predicted who would and would not develop a pressure ulcer with the following accuracy: sensitivity, 75.6%; specificity, 74.4%; positive predictive value, 92.4%; and negative predictive value, 42.7%. We used the Salzberg model s cutoff point and the same definitions of the risk factors, except for serum albumin and hemoglobin, where we used the cutoffs mentioned immediately above. For both the initial admission and readmission patients in our sample, we assessed the sensitivity, specificity, positive predictive value, and negative predictive value of the model. We also compared the median values in our sample among patients in both the initial and readmission groups who did and did not develop a pressure ulcer. We expected that the model would have greater utility in predicting pressure ulcers in the readmission group than in the initial rehabilitation group of patients. We based our grading of pressure ulcers on the description in the patient s medical file, plus the results of any relevant investigations that were made during the patient s stay in the rehabilitation unit. The classification system we used was based on the National Pressure Ulcer Advisory Panel guidelines for staging pressure ulcers. 21 The impact of pressure ulcers on inpatient rehabilitation was primarily assessed by estimating their influence on LOS. We used 2 methods. First, the LOS of patients with and without pressure ulcers was directly compared. Second, the amount of time needed to heal significant pressure ulcers was estimated. This was done only with patients who were confined to bed and were turned regularly to facilitate healing. This applied particularly to patients with pressure ulcers involving the sacrum, greater trochanter, or ischial tuberosity. We recorded number of days before patients were able to sit for 2 hours, 3 times a day, after completing a gradual graded mobilization program designed to increase the time spent sitting on the healed pressure ulcers. This information is routinely documented in the patient s medical file. This cutoff was chosen because we believe that it represents an adequate skin healing and sitting tolerance that is sufficient for patients to progress to addressing their key rehabilitation goals, such as dressing and transfer training. If patients were admitted with a pressure ulcer that healed but subsequently broke down again in the same location during the mobilization program, the breakdown was not considered a new pressure ulcer but a continuation of the healing process of the original ulcer. In the initial rehabilitation group of patients, additional information was extracted from the medical file about outcomes and functional status (unpublished data). The potential impact of pressure ulcers on inpatient rehabilitation was explored further in the initial rehabilitation group in several ways. We examined the influence of ulcers on patients being discharged home, on patients dying, on the total number of complications unrelated to pressure ulcers, and on patients disability. Physical disability on admission and discharge was assessed with the motor component of the FIM instrument. 22 FIM scores were Rasch transformed, 23,24 using established guidelines, 25 to facilitate statistical comparison between patients with and without pressure ulcers. The relevant information was extracted from patients medical files, recorded on a standardized data collection form, and subsequently entered into a database. All information in the database was double-checked against the data collection forms. We rechecked a random sample of 10% of the medical records and compared them with the database to detect any errors. If we found an error rate of greater than 5%, the study protocol required rechecking another random sample of 40% of the patient files. If discrepancies were found, they were corrected. The data were deidentified to maintain patient privacy. Our local ethics committee approved the study. Initial database entry and descriptive analysis was done with Excel 97 software. a Analytical statistical analysis was performed using SAS, version 8.0. b Demographic characteristics, patient etiologies, and comorbid factors were compared be-

3 NONTRAUMATIC SCI PRESSURE ULCERS, New 89 tween the pressure ulcer and non pressure ulcer groups, using a chi-square test for equal proportions. LOS and immobilization time followed a log-normal distribution and were logtransformed to give a continuous outcome that was normally distributed. LOS and age were compared with a 2-sample t test. We used multiple regression and multiple logistic regression, where appropriate, for our multivariate analysis and a stepwise selection procedure for multivariate models. We then validated the results by using a backward selection procedure. P values of.05 or less were deemed statistically significant. RESULTS Characteristics of the Patient Sample Patients demographic characteristics, etiologies, comorbidities, and non pressure ulcer complications have been reported elsewhere. 9 Table 1 summarizes these details for both initial and readmission rehabilitation patients with and without pressure ulcers. There was no statistically significant difference in the prevalence of pressure ulcers on admission between the initial rehabilitation and readmission patients (P.56). Among readmission patients with a pressure ulcer, 23.8% had a previously documented pressure ulcer. Three patients (2.2%) developed a pressure ulcer after being admitted to the unit. One initial rehabilitation patient was admitted with a grade I sacral pressure ulcer, which fully healed, and the patient successfully completed a remobilization program. Later, in their admission, the patient developed a grade III pressure ulcer at the same site, which took 30 days to heal and complete the full mobilization according to the above criteria. Two patients (1 initial, 1 readmission) developed new pressure ulcers during their rehabilitation stay. These pressure ulcers (sacral, grade II; trochanter, grade II) had minimal impact on the patients LOS because they spent no days confined to bed while their pressure ulcers healed. There was a statistically significant difference in the LOS of patients admitted with a pressure ulcer compared with that of patients admitted without (geometric mean for pressure ulcers, 62.3d; 95% confidence interval [CI], ; interquartile range [IQR], vs geometric mean for no pressure ulcers, 28.2d; 95% CI, ; IQR, ; P.0001). Of patients with a pressure ulcer present on admission, 24 (57%) developed them in an acute care hospital before admission to the rehabilitation unit, 8 (19%) developed ulcers while living in the community, and 2 (5%) developed ulcers while living in a hostel. The setting of 8 (19%) could not be determined because of incomplete documentation in the medical file. Fortunately, only 3 (6.8%) of the 44 patients who had a documented pressure ulcer at any time during rehabilitation stay had significant complications. All 3 were readmission patients. One patient developed osteomyelitis, 1 had septicemia, and the third had severe wound cellulitis. Pressure Ulcer Grade and Location Table 2 lists the grades of pressure ulcers that were present on admission, and table 3 reports the location of ulcers that were present on admission or that were documented during rehabilitation. There was no statistically significant difference between the initial and readmission patients regarding the grades or the distribution of the ulcers. The pelvic region was the site of 59.7% of pressure ulcers and 30.6% were in the foot region. Other sites were the occiput (n 3), shoulder (n 2), and knee (n 2). Of the patients with pressure ulcers present on admission, 17 patients (40.5%) had more than 1 pressure ulcer: 9 patients (21.4%) had 2 pressure ulcers, 4 patients (9.5%) had 3, and 4 patients (9.5%) had 4 or more pressure ulcers. Prediction of Pressure Ulcers in the Study Sample The increased risk of pressure ulcers associated with various demographic characteristics, causes of nontraumatic SCI, comorbid conditions, and complications was initially examined by univariate analysis (table 1). The initial and readmission groups had different risk factors regarding association with pressure ulcers that were statistically significant at a univariate level. No risk factor was common to both groups. Age, level of nontraumatic SCI, hypoalbuminemia, other infections (nonurinary tract), and psychologic factors were statistically significant for pressure ulcers among the initial rehabilitation patients. Only ASIA classification and anemia were statistically significant among the readmission patients. Multiple logistic regression showed age to be the only factor that significantly increased the risk of pressure ulcers in the initial rehabilitation patients (odds ratio [OR] 1.042; 95% CI, ; P.04). For every 1 year increase in the age of these patients, the risk of pressure ulcers increased by 4.2%. Depression was strongly associated with pressure ulcers, but the result was not statistically significant (OR 8.94; 95% CI, ; P.059). In readmission patients, multiple logistic regression showed both ASIA grade (P.012) and anemia to be the only significant risk factors (table 4). After applying to our study the model previously developed to predict pressure ulcers in chronic SCI patients, 20 we found that it had much better precision in our readmission patients than in the initial rehabilitation patients (table 5). The latter group tended to have higher scores. There was a statistically significant difference in prediction scores between patients with and without pressure ulcers in both the initial and readmission groups. Pressure Ulcers Impact on Rehabilitation Immobilization data was available for 42 of 44 patients (95.5%) who required bedrest to heal their pressure ulcers. Thirteen patients did not require immobilization that would prevent them from sitting, lying, or participating in a rehabilitation program. The 29 patients who needed to be confined to bed required between 5 and 282 days before they were able to participate fully in a rehabilitation program under the criteria detailed above. The length of time required to heal a pressure ulcer and to remobilize the patient to an adequate level of sitting on the healed skin followed log-normal distribution, with a geometric mean of 41.7 days (95% CI, ; IQR, ) and a median of 42 days. The amount of time spent healing and mobilizing pressure ulcers increased with the grades of the ulcers, and there was a moderately strong correlation between the grades and the time spent immobilizing (table 6). Among all patients with nontraumatic SCI admitted to the unit during the study period, 19.8% of inpatient bed days were spent healing and mobilizing patients who had pressure ulcers. Among the initial rehabilitation group of patients, there was no statistically significant difference between those with and without a pressure ulcer documented during their stay, and in the following outcomes: discharge home ( , P.10), patients who died ( , P.42), and the total number of other non pressure ulcer complications ( , P.26). There was no statistically significant difference (Student t test, t , P.09) in the Rasch-converted FIM motor scores on admission for patients with (mean, ) and without (mean, ) a pressure ulcer. Nor was there a statisti-

4 90 NONTRAUMATIC SCI PRESSURE ULCERS, New Table 1: Univariate Analysis of the Characteristics of Initial* and Readmission Patients With and Without Pressure Ulcers Initial Rehabilitation: Pressure Ulcer Initial Rehabilitation: No Initial Pressure Ulcer P Readmission: Pressure Ulcer Readmission: No Pressure Ulcer Number Demographics Men 41.7% 47.8% % 32.6%.20 Age (mean SD) % CI Range Lives alone 29.2% 50% % 32.6%.96 Partner/family 66.7% 47.8% % 50.0%.69 NH/hostel 4.2% 0% % 11.1%.61 Pension 62.5% 60.9% % 69.6%.52 Working 20.8% 19.6% % 4.4%.10 Not working 16.7% 19.6% % 23.9%.89 LOS geometric mean (d) % CI Level Tetraplegia 16.7% 41.3% 27.8% 34.8% Paraplegia 83.3% 58.7% 72.2% 65.2% ASIA class A 12.5% 6.5% 38.9% 4.4% B 8.3% 2.2% 0.0% 0.0% C 25.0% 19.6% 22.2% 28.3% D 54.2% 69.6% 22.2% 56.5% Unknown class 0.0% 2.2% 16.7% 10.9% Etiology Tumor 41.7% 28.3% % 2.2%.064 TVM 4.2% 13.0% % 10.9%.31 Degeneration 25.0% 26.1% % 13.0%.67 MS 0.0% 4.4% % 37.0%.89 Vascular 16.7% 13.0% % 13.0%.17 Other 12.5% 15.2% % 23.9%.23 Comorbidity Diabetes 29.2% 13.0% % 19.7%.79 IHD or CCF 12.5% 19.6% % 13.0%.45 Vascular 4.2% 4.4% % 0.0% 1.0 COPD 16.7% 8.7% % 0.0%.076 Smoker 4.2% 8.7% % 26.1%.15 Anemia 62.5% 45.7% % 6.5%.034 Hypoalbuminemia 70.8% 43.5% % 6.5%.34 Complications UTI 54.2% 41.3% % 13.0%.062 Other infection 33.3% 13.0% % 6.5%.34 Spasticity 16.7% 10.9% Pain 25.0% 21.7% % 8.7%.21 Psychologic 16.7% 2.2% % 8.7% 1.0 Abbreviations: CCF, congestive cardiac failure; COPD, chronic obstructive pulmonary disease; IHD, ischemic heart disease; MS, multiple sclerosis; NH, nursing home; Psychologic, depression or adjustment difficulties; TVM, transverse myelitis; Vascular, cerebrovascular or peripheral vascular disease. *t 68 test; t 62 test; Fisher exact test. P cally significant difference in the Rasch-converted FIM motor scores at discharge (Student t test, t , P.15) for those with (mean, ) and without (mean, ) a pressure ulcer. Pressure Ulcers in Readmission Patients Among the readmission patients, there was no statistically significant relation between the time after nontraumatic SCI and the occurrence of a pressure ulcer (Mann-Whitney test, z 1.6, P.11). Error Rate When we rechecked the random sample of 13 (9.7%) medical records against the data collection forms, we compared 163 separate data entries. Only 2 errors (1.2%) were found. As detailed previously, 9 these were both minor and did not relate to the results reported here. DISCUSSION Pressure ulcers were common in both the initial rehabilitation and readmission groups of patients. Ulcer location and our

5 NONTRAUMATIC SCI PRESSURE ULCERS, New 91 Table 2: Grade of Pressure Ulcers Present on Admission to Rehabilitation Pressure Ulcer Grade Initial Readmission n (%) n (%) I 6 (12.5) 4 (15.4) II 15 (31.3) 9 (34.6) III 20 (41.7) 8 (30.8) IV 7 (14.6) 3 (11.5) Unknown 0 (0) 2 (7.7) Total 48 (100) 26 (100) NOTE , P.87. finding of multiple pressure ulcers in many of our patients were similar to results reported by others. 11,26,27 In contrast to studies of traumatic SCI, where initial rehabilitation patients tend to have more sacral and fewer trochanter or ischial pressure ulcers than do readmission patients, 27 there were no statistically significant differences in the locations of pressure ulcers in our groups. Our results are also similar to previous reports that many pressure ulcers after SCI occur while patients are still in acute care hospitals and before they are admitted to rehabilitation units. 11 The incidence of pressure ulcers that occurred in our patients during rehabilitation was low when compared with other spinal rehabilitation units in which the incidence has been reported to be between 7.5% 11 and 31.7%. 1 This may be because the retrospective nature of this study did not detect all pressure ulcers, because of differences in patient risk factors or because of increased awareness and implementation of prevention strategies by our unit. It is likely that some grade I or II pressure ulcers occurred in our study sample but were not documented in the medical files. We believe, however, that no significant pressure ulcers of higher grades would have missed documentation. Not unexpectedly, pressure ulcers increased patients LOS in rehabilitation, with the deeper wounds having a greater effect. The estimated impact on LOS was similar for each of the 2 methods used. First, by comparing the LOS for patients with and without pressure ulcers, there was a statistically significant difference of 36.1 days between the means for LOS in rehabilitation. Second, by calculating the time spent healing and then mobilizing patients with pressure ulcers, we estimated that these patients required an additional 42 days of inpatient rehabilitation. This latter group represents a smaller and more severe subgroup, and this is reflected by their results. We believe that the overall similarity of the 2 methods indicates that these results are a reasonably accurate estimate of the effect of pressure ulcers on patient LOS. Table 3: Location of Pressure Ulcers Pressure Ulcer Location Initial Readmission n (%) n (%) Sacrum 18 (38) 8 (31) Ischium 8 (17) 4 (15) Trochanter 4 (8) 3 (12) Heel 9 (19) 9 (35) Malleolus 3 (6) 1 (4) Others 6 (13) 1 (4) Total 48 (100) 26 (100) NOTE , P 0.6. Table 4: Multiple Regression of Risk Factors for Pressure Ulcers in Readmission Rehabilitation Patients With Nontraumatic SCI Risk Factor OR 95% CI P ASIA class A vs D ASIA class C vs D Anemia A previous study 13 of patients with traumatic SCI examined the effect of pressure ulcers on time from the onset of injury to discharge from rehabilitation. In that study, the total LOS in patients with a severe pelvic pressure ulcer (median 7.0mo; mean, mo) was much greater than it was for patients with no pelvic pressure ulcers (median 5.0mo; mean, mo) or no pressure ulcers (median 3.0mo; mean, mo). Because the study included the acute medical management period and the rehabilitation period, it is not possible to compare directly these results with our results. Among initial rehabilitation patients, a pressure ulcer did not influence the rate of patients who were discharged home or died, the occurrence of other complications, or physical disability on admission and discharge, as measured by the FIM motor subscale. We expected that the admission motor FIM scores for patients with a pressure ulcer would be lower than the scores for patients without a pressure ulcer. Among readmission patients, there was no relation between time after nontraumatic SCI and the occurrence of a pressure ulcer. Different factors for pressure ulcers were identified by multiple logistic regression in both groups of patients. Age was the only significant factor in the initial rehabilitation group. Although depression was not statistically significant in the initial rehabilitation patients, the OR was high, which indicates a potentially large influence. This probably reflects an effect of a pressure ulcer, rather than a cause, and reinforces the importance of addressing the psychologic needs of this particular group of patients. ASIA grade and anemia were the only statistically significant factors in the readmission group of patients. The difference between ASIA class A and ASIA class D patients was much greater than that between ASIA class C and D patients, which was not statistically significant but was still included in the regression model. A recent study of patients with nontraumatic SCI reported that pressure ulcers were more likely to develop in men and in patients with a more complete injury on admission. 12 Multi- Table 5: Traumatic SCI Pressure Ulcer Prediction Model Applied to Patients With Nontraumatic SCI Initial Readmission Sensitivity 100% 100% Specificity 34.1% 69.6% PPV 47.3% 56.3% NPV 100% 100% Pressure ulcer patients median PS (IQR) 13 (12 14) 12 (10 13) Non pressure ulcer patients median PS (IQR) 8* (5 9) 3 (2 7) Abbreviations: NPV, negative predictive value; PPV, positive predictive value; PS, prediction score. *Two-sample Wilcoxon rank-sum (Mann-Whitney) test difference in pressure ulcer prediction scores for patients with and without pressure ulcers (z 6.7, P.00001). Two-sample Wilcoxon rank-sum (Mann-Whitney) test difference in pressure ulcer prediction scores for patients with and without pressure ulcers (z 5.7, P.00001).

6 92 NONTRAUMATIC SCI PRESSURE ULCERS, New Table 6: Pressure Ulcer Grade and Immobilization Duration for Patients With Pressure Ulcers That Prevented Participation in Rehabilitation Pressure Ulcer Grade n Geometric mean Time Spent Immobilizing (d) Range (d) I 0 NA NA II III IV Total 29 NOTE. Pearson correlation r.62, P Abbreviation: NA, not applicable. variate analysis showed that complete SCI was the only statistically significant variable predictive of pressure ulcers. The model, however, did not use the range of risk factors that we included; it only considered age, gender, etiology, level of injury, and completeness. Patients were typically admitted almost 2 months after nontraumatic SCI, and 17.1% of patients in this group had been discharged home from an acute care hospital before their rehabilitation admission. Both of these factors could have affected pressure ulcer occurrence. The patients in our study were older, more likely to be women, and had a less severe pattern of SCI. There was a similar frequency of tumor in both groups, but our study sample had fewer vascular and more degenerative causes of nontraumatic SCI. When a model previously developed for predicting pressure ulcers in chronic, mainly traumatic SCI was applied to our study sample, we found that there was a significant difference in the scores between patients who did and did not develop a pressure ulcer. This applied to both the initial and readmission patients. As expected, the precision of the model in predicting who would and would not develop a pressure ulcer was much better in the readmission group than in the initial group. When the prediction model was used in the readmission group, its precision was comparable to the original model. These results suggest that a previous model for identifying SCI patients at risk of pressure ulcers does apply to our nontraumatic SCI readmission patients but not to initial rehabilitation patients. Our study results imply that prospective studies of patients with nontraumatic SCI are needed. Future research should ideally involve a larger number of subjects to improve the precision of results. Multicenter research and the inclusion of potentially eligible subjects not captured by our study would improve the generalizability of results. Further work must be done to refine a model with which to predict pressure ulcers in patients with nontraumatic SCI, particularly initial rehabilitation patients with acute SCI. Greater efforts are needed, particularly in acute care hospitals, to prevent pressure ulcers in patients with nontraumatic SCI. Greater vigilance in the acute care setting would prevent some pressure ulcers and reduce the severity of others. This would result in considerable savings in both the human and financial costs of nontraumatic SCI rehabilitation. Pressure ulcer prediction and prevention guidelines have been available for many years, 14,28,29 and recently an Australian version has been developed. 30 Given the prevalence of pressure ulcers among patients admitted from acute hospitals in this study and similar results in previous studies, 11 it would appear that the implementation of pressure ulcers prevention strategies has been suboptimal. The reason for poor implementation of guidelines and strategies to help improve practices have been well described in the literature This is a challenge that must be taken up by administrators, program and clinical directors, and the medical and nursing staffs at all levels of the health care continuum. Study Limitations We acknowledge that the present study has several limitations. Because of its retrospective nature, there may be inaccuracies resulting from incorrect or incomplete documentation in patients medical files. The number of subjects in this study was relatively small, which increases the chance of error. Furthermore, because the results reflect the practice and referral pattern of only 1 center, the generalizability of the findings is limited. Patients with nontraumatic SCI not admitted to our unit would include patients with a very mild SCI who recover quickly at the acute care hospital and do not need inpatient rehabilitation; patients who die at the acute care hospital; patients who have a poor prognosis for rehabilitation and are transferred to a nursing home, hospice, or palliative care unit; and patients who are managed at other institutions. CONCLUSIONS Pressure ulcers are a common complication among nontraumatic SCI rehabilitation admissions and they have a significant effect on LOS. Risk factors for pressure ulcers differ in initial and readmission patients. A previous model for identifying SCI patients at risk of pressure ulcers can be applied to our nontraumatic SCI readmission patients. The model was not as useful with initial rehabilitation patients, and we hypothesize that these patients have additional risk factors that make accurate prediction of pressure ulcers more difficult. Acknowledgments: We thank Michele Hodes for technical help with database management. We acknowledge the assistance of the staff of the medial records department at Caulfield General Medical Center for their help in locating the patient medical files. References 1. Stover SL, DeLisa JA, Whiteneck GG. Spinal cord injury: clinical outcomes from the model systems. Gaithersburg (MD): Aspen; Levi R, Hultling C, Nash MS, Seiger Å. The Stockholm spinal cord injury study: 1. Medical problems in a regional SCI population. Paraplegia 1995;33: Levi R, Hultling C, Seiger Å. The Stockholm Spinal Cord Injury Study: 2. Associations between clinical patient characteristics and post-acute medical problems. Paraplegia 1995;33: Anson CA, Shepard C. Incidence of secondary complications in spinal cord injury. Int J Rehabil Res 1996;19: Johnson RL, Gerhart KA, McCray J, Menconi JC, Whiteneck GG. Secondary conditions following spinal cord injury in a population sample. Spinal Cord 1998;36: McKinley WO, Jackson AB, Cardenas DD, DeVivo MJ. Longterm medical complications after traumatic spinal cord injury: a regional model systems analysis. Arch Phys Med Rehabil 1999; 80: Chen D, Apple DF Jr, Hudson LM, Bode R. Medical complications during acute rehabilitation following spinal cord injury current experience of the Model Systems. Arch Phys Med Rehabil 1999;80: McKinley WO, Conti-Wyneken AR, Vokac CW, Cifu DX. Rehabilitative functional outcome of patients with neoplastic spinal cord compression. Arch Phys Med Rehabil 1996;77: New PW, Rawicki BH, Bailey M. Nontraumatic spinal cord injury: demographic characteristics and complications. Arch Phys Med Rehabil 2002;83: Bryne DW, Salzberg CA. Major risk factors for pressure ulcers in the spinal cord disabled: a literature review. Spinal Cord 1996; 34:

7 NONTRAUMATIC SCI PRESSURE ULCERS, New Hammond MC, Bozzacco VA, Steins SA, Buhrer R, Lyman P. Pressure ulcer incidence on a spinal cord injury unit. Adv Wound Care 1994;7: Celani MG, Spizzichino L, Ricci S, Zampolini M, Franceschini M. Spinal cord injury in Italy: a multicenter retrospective study. Arch Phys Med Rehabil 2001;82: Young J, Burns P. Pressure sores and the spinal cord injured. SCI Digest 1981;3(Fall):9-18, Panel for the Prediction and Prevention of Pressure Ulcers in Adults. Pressure ulcers in adults: prediction and prevention. Clinical Practice Guideline No. 3. Rockville: US Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research; AHCPR Publication No Yarkony GM. Pressure ulcers: a review. Arch Phys Med Rehabil 1994;75: Rodriguez GP, Garber SL. Prospective study of pressure ulcer risk in spinal cord injury patients. Paraplegia 1994;32: Graham A. The development of pressure sores in patients with spinal cord injury. J Wound Care 1997;6: McKinley WO, Seel RT, Hardman JT. Nontraumatic spinal cord injury: incidence, epidemiology, and functional outcome. Arch Phys Med Rehabil 1999;80: Maynard FM, Bracken MB, Creasey G, et al. International Standards for Neurological and Functional Classification of Spinal Cord Injury. American Spinal Injury Association. Spinal Cord 1997;35: Salzberg CA, Byrne DW, Cayten CG, van Niewerburgh P, Murphy JG, Viehbeck M. A new pressure ulcer risk assessment scale for individuals with spinal cord injury. Am J Phys Med Rehabil 1996;75: National Pressure Ulcer Advisory Panel. Pressure ulcers: incidence, economics, risk assessment. Consensus development conference statement. Decubitus 1989;2: Guide for the Uniform Data Set for Medical Rehabilitation (including the FIM instrument), version 5.1. Buffalo: State Univ New York; Linacre J, Heinemann A, Wright B, Granger C, Hamilton B. The Functional Independence Measure as a measure of disability. Chicago: Rehabilitation Services Evaluation Unit, Rehabilitation Institute of Chicago; Research Rep No Heinemann A, Linacre J, Wright B, Hamilton B, Granger C. Relationships between impairment and physical disability as measured by the Functional Independence Measure. Arch Phys Med Rehabil 1993;74: Heinemann AW, Hamilton BB, Granger CV, et al. Rating scale analysis of functional assessment measures. Final report. Department of Education, National Institute on Disability and Rehabilitation Research. Chicago: Rehabilitation Institute of Chicago; Project No. H133H Gunnewicht BR. Pressure sores in patients with acute spinal cord injury. J Wound Care 1995;4: Yarkony GM, Heinemann AW. Pressure ulcers. In: Stover SL, DeLisa JA, Whiteneck GG, editors. Spinal cord injury: clinical outcomes from the Model Systems. Gaithersburg (MD): Aspen; p NHS Centre for Reviews and Dissemination. The prevention and treatment of pressure sores. Effective Health Care 1995;2: Consortium for Spinal Cord Medicine Clinical Practice Guidelines. Pressure ulcer prevention and treatment following spinal cord injury: a clinical practice guideline for health-care professionals. J Spinal Cord Med 2001;24(Suppl 1):S Australian Wound Management Association. Clinical practice guidelines for the prediction and prevention of pressure ulcers. West Leederville (Aust): Cambridge Publishing; Grimshaw J, Russell I. Achieving health gain through clinical guidelines. II. Ensuring that guidelines change medical practice. Quality Health Care 1994;3: NHS Centre for Reviews and Dissemination. Getting evidence into practice. Effective Health Care 1995;5: Thomson R, Lavender M, Madhok R. How to ensure that guidelines are effective. BMJ 1995;311: National Health and Medical Research Council. How to put the evidence into practice: implementation and dissemination strategies. Handbook series on preparing clinical practice guidelines. Canberra (Aust): National Health and Medical Research Council; Suppliers a. Microsoft Corp, One Microsoft Way, Redmond, WA b. SAS Institute, 100 SAS Campus Dr, Cary, NC