|Home | About | Journals | Submit | Contact Us | Français|
To assess the relationship of psychological, environmental, and behavioral factors with pressure ulcers (PUs) in persons with spinal cord injury (SCI).
A total of 1,549 participants from a large rehabilitation hospital in the southeast United States answered questions regarding outcomes after SCI. Variables from each set of factors were entered sequentially into the model: (1) psychological and environmental, and (2) behavioral.
Forty-eight percent of participants reported having a PU in the past year. After entering behavioral variables into the model, all environmental and psychological variables became nonsignificant. Odds of having a PU increased 28% with each psychotropic medication taken weekly. Persons who smoked one or more packs of cigarettes daily had 2.82 times the odds of having a PU than persons who did not smoke. Increased hours out of bed were protective against PUs.
This study demonstrated the importance of health behaviors in the occurrence of PUs after SCI. These health behaviors provide important targets for intervention for health care providers.
Pressure ulcers (PUs) are a serious secondary condition that can affect the lives of persons with spinal cord injury (SCI) in many ways and may ultimately result in death.1 Krause2 found persons with no sores reported significantly higher subjective well-being and a higher frequency of social outings than those with PUs. Persons with PUs have been found to be less likely to be working.2,3 Additionally, among persons who reported PUs, 27% reported having to reduce their sitting time because of a PU.2 Several studies of outcomes after SCI have reported the annual prevalence of PUs to be 32% to 50%.2,4,5
In 1996, Krause6 developed a theoretical risk model to study the relationships between different sets of factors with morbidity and mortality after SCI. This model predicts demographic and injury characteristics are the most basic level of predictors of health, followed by environmental and psychological factors, and finally behavioral characteristics. According to the model, inclusion of psychological factors and environmental factors will significantly enhance the prediction of health beyond that of demographic/injury factors, and behavioral factors will enhance prediction beyond psychological/environmental factors. Previous research findings have supported this model in the prediction of mortality.7-11
Several studies have investigated risk and protective factors of PUs after SCI. Demographic variables, including age, gender, and race, have been found to be related to PUs after SCI, although not consistently across studies.3-5,12-14 Being married or cohabitating is protective against PUs.15 Saladin, Krause, and Adkins5 assessed barriers to treatment for PUs among a group of racially/ethnically diverse persons with SCI. Blacks were more likely than whites to report not having financial resources or transportation as barriers to treatment for PUs. All races compared to Hispanics were more likely to report not having enough money for attendant help as a barrier. A higher educational level has been linked with a lower risk of having a PU within the past year,13,14,16 as has employment.14,17 In general, access to primary care has been found to be protective in the development of PUs.15 Very few studies have assessed psychological factors in relation to PUs. Elliot and colleagues18 found that PUs in the first 3 years after injury were associated with problem-solving ability. Correa and colleagues19 found that having a personality disorder was related with PUs.
Health behaviors are important as they provide a target for intervention. Krause and Broderick20 used a bi-dimensional model of risk and protective factors in their assessment and investigated the relationship of individual factors with PU history, controlling for demographic and injury characteristics. They found healthy lifestyle, healthy eating, and exercise to be protective factors and smoking and taking medication for sleep to be risk factors for recurrent PUs. Other studies have also found smoking to be a risk factor for PUs.21,22 Similarly, substance abuse has been related to PUs after SCI.17,23,24 In a study of 86 persons who had received inpatient rehabilitation, persons with PUs had a lower number of daily transfers than persons without.3 However, no difference was found in the use of prevention techniques, hours out of bed, frequency of weight shifts, alcohol consumption, or smoking. Schryvers and colleagues17 found persons who were independent in self-care were more likely to have PUs than those who were not. Looking at PUs in a broader context, not just in persons with SCI, food intake, mobility, and activity were associated with PUs.25
Although these studies give insight as to factors associated with PUs after SCI, they have not utilized a theoretical model to assess stages of factors related to PU occurrence in persons with SCI. Additionally, very few studies have considered the relationship between psychological factors and PUs after SCI. The objective of this study is to assess three levels of factors and their relationship with PUs using the theoretical risk model developed by Krause in 1996.6 We hypothesize behavioral variables will mediate the relationship of psychological and environmental factors with PUs. Consistent with mediation of models as proposed by Baron and Kenney,26 we would expect that psychological and environmental factors would be significant predictors of PUs but that these relationships would disappear after consideration of behavioral factors.
After receiving approval from the institutional review board, we identified participants through three different sources of records at a large specialty hospital in the southeastern United States: (a) Model SCI Systems patient database, (b) Model Systems registry, and (c) outpatient directory. There were three inclusion criteria: (1) traumatic SCI, (2) 18 or older at assessment, and (3) minimum of 1 year post injury. Out of 2,480 potential participants meeting these criteria, 1,549 returned their survey (62.5% response rate). Because this article is focused on PUs, we eliminated persons who said they had made a full recovery from their injury or who were ambulatory (n = 496), 1 person who was less than 1 year post injury, and 2 people with missing information on PUs in the previous year, leaving 1,050 persons for analysis. There were 42 people (4.0%) who were missing information for the final model. Persons with missing data were more likely to be from a minority race (P = .0189). There were not differences in missing data by gender, injury severity, or having a PU in the past year.
Five weeks prior to receiving an initial packet of study materials, participants received preliminary letters that described the intended research and the method by which they would receive the materials. The initial packet included a letter that described the study and served as implied consent and the survey to be completed and mailed in. If the initial packet was not returned, a second mailing was set in place for all nonrespondents. If a potential participant did not respond, a phone call was made to engage in active conversation. A third mailing was then used for those participants who had misplaced or accidently discarded the materials but consented by phone to receive another packet.
Measurements were taken using a mail-in survey tool. We used a subset of measures from a larger study of protective and risk factors associated with the onset of multiple types of adverse health outcomes and secondary conditions among a large sample of individuals with SCI.27 Our outcome in this study was having PUs in the previous year. PUs were described as “open sores in pressure areas, such as your tailbone, ischium, heel, or elbows. They are usually caused by pressure, but may also be caused by friction or shearing (rubbing), moisture, burns, or falls.” Participants were asked, “All totaled, how many different open pressure sores have you had in the past year?” This was dichotomized as none or one or more as we were interested in the odds of having PUs at all. We only asked participants to report open PUs as it might be difficult for some participants to recognize a less severe PU, especially persons with pigmented skin.28 Therefore, all results in this study refer to open PUs. This measurement has been used in previous studies of persons with SCI.
We measured 4 specific levels of factors in this study; demographic/injury, psychological, environmental, and behavioral. Demographic variables included race (white, black, other), gender, age, years post injury, and age at injury. Injury level was categorized as C1-C4, C5-C8, and noncervical.
We assessed psychological status using the Zuckerman-Kuhlman Personality Questionnaire (ZKPQ).29 The ZKPQ is a 99-item measure of personality with 5 scales: (a) impulsive sensation seeking, (b) neuroticism-anxiety, (c) aggression-hostility, (d) sociability, and (e) activity. Impulsive sensation seeking measures a lack of planning and the tendency to act impulsively and serves as a proxy for reckless and dangerous behavior. Neuroticism-anxiety measures tension, worry, and fearfulness. Aggression-hostility reflects items that express rude, thoughtless, or antisocial behavior. Activity assesses a need for high-energy activity. Finally, sociability measures social contacts and friends. The ZKPQ is a highly reliable instrument, with test-retest reliability ranging from .70 to .86.29
Several variables were used to assess environment. There were two indicators of socioeconomic status (SES), education and income. The participants were asked the exact number of years of education they had completed; this was broken down into 4 categories: <high school, high school or GED, some college, and a bachelor’s degree or higher. Participants were asked their annual household income, which was re-coded into 3 categories: <$25,000, $25,000-74,999, and ≥$75,000. Three variables were used to assess health care access. Participants were asked if there was a time in the past 12 months they could not see a doctor because of cost.
Last, social support was measured using the Berlin Social Support Scale (BSSS),30 which was developed in 2000 for use with an adult cancer patient population. There are 6 subscales with 5 to 15 items each. The subscales are: perceived available support (emotional and instrumental), need for support, support seeking, actual received support, provided support (for the provider, not the patient), and the protective buffering scale. We used all but the last 3 subscales. Respondents indicated their agreement with statements on a 4-point Likert-type scale. Scores are computed by totaling item responses. The BSSS has been validated with several subpopulations. Internal consistency alphas ranged from 0.63 for need for support to 0.83 for perceived available support and received social support.
We assessed behavioral factors. Participants responded about the number of cigarettes they smoked each day (none, < 1 pack, 1 + packs). Overall fitness was reported as poor, good/fair, and very good/excellent. Overall diet health was reported as poor/fair, good, and very good/excellent. These items were proxy variables for health-related behaviors. Participants were asked how often they took prescription medications for pain, sleep, spasticity, and stress. We used the total number of prescriptions taken for any of these conditions (maximum of 4). Binge drinking was measured as the number of days over the past month 5 or more drinks were consumed in a sitting. Participants reported on the typical number of hours spent out of bed each day and the typical number of days they got out of the house each week.
A 2-stage hierarchical model building strategy using logistic regression was used to identify the association of psychological, environmental, and behavioral factors with PUs in the past year. During the first stage of analysis, a base model consisting of demographic (race, age at survey, and years lived since injury) and injury level was specified. Gender was not significant in the preliminary analysis; therefore, it was excluded from further model building. Single variables were added to the base model as a means of “screening” potential predictors for the final stage model. All variables significant at the P < .15 level were considered for subsequent modeling.
In the second stage, variables retained from the preliminary screening were added to the base model sequentially as a group by factor (ie, psychological, environmental, behavioral). Three intervening models were generated based on the theoretical risk model.6 The first intervening model added the psychological and environmental factors simultaneously to the base model. The second intervening model added the behavioral factors to the first intervening model. At each intervening model, backwards elimination, with P < .15 as the significance level, was employed. Once the final model was established, all pair-wise interaction terms were included to further assess goodness of fit; however, no interactions were significant and therefore were not retained in the model. Hosmer-Lemeshow and global chi-square tests were used to assess goodness of fit of the model.31 The C statistic, measuring area under the receiver operating characteristic (ROC) curve, was used to assess discriminatory ability.31 Odds ratios (ORs) with 95% confidence intervals (CIs) are reported.
Seventy-five percent were male, and 72.5% were white. Cervical injuries were reported by 49.8%. The average age at survey was 44.5 (13.0) years, and participants were on average 13.8 (2l.6) years post injury. High income (≥$75,000) was reported by 2l.0% and high education (bachelor’s degree or higher) by 25.7%. Just fewer than half of participants reported having a PU in the past year (48.5%).
In the first stage of modeling, we created a base model using the demographic and injury characteristics, which included race, injury severity, years post injury, and age at survey (Table 1). After the base model was formed, we added variables to the model as a means of “screening” potential predictors (Table 2).
The second stage of modeling assessed groups of factors and their relationship with PUs in the past year (Table 3). First, psychological and environmental factors were entered together (stage IIa). Three psychological (neuroticism/anxiety, impulsive sensation seeking, activity) and 3 environmental factors (income, cost of treatment, need for support) were retained. Next, behavioral factors were entered into the model with the retained psychological and environmental variables (stage IIb). At this stage, all psychological variables were dropped from the model as they were no longer significant. Additionally, cost as a barrier to treatment was also dropped. Income and the need for support were retained. Of the behavioral variables entered, all were significant except for number of days out of the house and diet.
Other than the demographic and injury characteristics, only 3 behavioral factors were statistically significant in the model. Each increase in the number of prescription medications taken resulted in a 24% increase in the odds of having had a PU in the past year (OR, l.24; 95% CI, 1.11-l.38). Persons who smoked a pack of cigarettes or more per day were 2.82 times as likely to have had a PU in the past year than nonsmokers. Increased number of hours out of bed was protective against PUs (OR, 0.93; 95% CI, 0.89-0.97). Additionally, an increased number of years post injury resulted in higher odds of PUs in the past year (OR, 1.02; 95% CI, 1.01-1.04). Last, there was a trend toward increased risk of PUs with lower income (P = .0547).
In this study, 48% of respondents reported having a PU in the previous year, which is similar to what previous studies have found.2,4,5 These results expanded on previous findings by using a theoretical risk model6 to assess groups of risk and protective factors of PUs. Our hypothesis was supported with these data such that, in our final model, environmental and psychological factors were no longer significantly related to having PUs in the past year.
These data met the 3 criteria necessary for mediation: (a) a significant relationship between environment and psychological factors with PUs, (b) a significant relationship between behavioral factors and PUs, and (c) the relationship between environmental and psychological factors with PUs disappeared after the behavioral factors were added. During the preliminary modeling stages, several environmental and psychological factors, including personality and income, were related to PUs in the past year. However, these factors became nonsignificant after the behavioral factors were entered into the model. These results support the framework of the theoretical model and, ultimately, 2 risk behaviors and 1 protective behavior were significantly associated with PUs. We found use of substances, specifically cigarettes and prescription medications, resulted in increased odds of PUs in the past year. These findings support previous results linking smoking,20–22 and substance use17,23,24 with PUs after SCI.
We also found persons who spent more time out of bed each day were less likely to have had PUs in the past year, Spending more hours out of bed suggests higher participation in life.32 While hours out of bed is a measure of participation, it could also be representative of concomitant morbidity where someone who is ill might be staying in bed more hours of the day. Previous studies have found comorbid disease to be associated with PUs.14,22 Additionally, spending more hours in bed has been linked with depression in persons with SCI.33
There are several implications of these findings for clinical practice and rehabilitation. Intervening at the behavioral level is necessary to modify existing behaviors associated with an elevated risk of PUs. However, consistent with the theoretical risk model,6 it is preferable for the intervention to occur as early as possible within the chain of risk factors. Therefore, even though psychological and environmental factors were no longer significant after consideration of behavioral factors, they represent important targets for intervention prior to the development of the risk behaviors. Specifically, certain personality traits are associated with an elevated risk of behaviors that, in tum, relate to a higher risk of PUs. Working with individuals who have these traits may help to circumvent behavioral patterns.
There were several limitations of this study. First, these data are self-report and could be subject to recall bias. However, the questions involve a limited time frame (within the past year) to minimize this bias. Additionally, we only asked about open PUs so all persons, regardless of skin pigmentation, would be able to answer the question. Although this minimized recall bias, it also is a limitation to this study. We do not have information on PU grade, severity, or location and do not have information on less severe (not open) PUs. Third, these data were cross-sectional; therefore, causal sequence could not be established, only associations could be made. For example, we could not establish whether the amount of hours out of bed was limited because of a PU or whether limited hours out of bed caused a PU. Also, we asked about PUs in the past year, so it is not possible to delineate between PUs beginning in the past year and those beginning in the previous year but continuing to the past year. We were limited in the number of environmental, psychological, and health behavior questions we asked. Future research should expand upon the types of variables assessed.
This study has demonstrated the importance of health behaviors in the occurrence of PUs after SCI. These health behaviors provide important targets for intervention for health care providers. Future research should expand on the types of factors collected with respect to PUs, specifically health behaviors, Additionally, future work should attempt to further classify PU grades and identify persons with all levels of severity.
The contents of this article were developed under grant H133G050165 from the National Institute on Disability and Rehabilitation Research (NIDRR), Office of Special Education and Rehabilitation Services, US Department of Education. grant H133G050165. These contents do not necessarily represent the policy of the US Department of Education, and endorsement by the US Federal Government should not be assumed.
This article was made possible by National Institutes of Health grant number IROI NS 48117. The contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Institutes of Health.
Lee L. Saunders, Department of Health Sciences and Research, College of Health Professions, Medical University South Carolina, Charleston, South Carolina.
James S. Krause, Department of Health Sciences and Research, College of Health Professions, Medical University of South Carolina, Charleston, South Carolina.