Introduction

Pressure injury (PI) is defined as localized damage to the skin and the underlying soft tissue, usually over a bony prominence, or related to medical or other devices. It can present as a red spot on intact skin or as an open ulcer and can be painful. It occurs as the result of intense or prolonged pressure, or pressure in combination with shear [1]. A systematic review published in 2013 concluded that overall there is no single factor, which can explain the risk of PI, but rather a complex interplay of factors that increase the probability of the development [2]. Impaired motor and sensory function, altered perfusion and circulation, moisture, and incontinence were found to be factors that significantly increased the risk of PI [2].

Due to paralysis, sensory loss, and prolonged exposure to moisture, individuals with spinal cord injury (SCI) are at particular risk for PI [2,3,4,5].

The occurrence of PIs during hospitalization of persons with SCI has shown to vary between 6 and 52% [6, 7], and comorbidities can affect the development of severe PI [6].

Persons with SCI are a heterogeneous group, and therefore risk factors may vary in specific subgroups [2, 8, 9]. Verschueren found that both complete injury and tetraplegia were significant risk factors for PI during the acute rehabilitation period [7], while Grigorian et al. found that a higher SCI level was associated with increased risk for PI as compared with the lower level [10]. The actual number of individuals with SCI and PIs in Norway is unknown; moreover, we do not know if the risk factors for PIs in the Norwegian population correspond to risk factors reported in other studies [2,3,4,5]. Therefore, a research program (NORSCIPI) at all three spinal cord units in Norway (NSCUs) was conducted to identify characteristics associated with PI development in the hospitalized population of persons with acute SCI [11]. The first study of this research program aims to investigate the prevalence of PI in the population, and further to investigate potential risk factors and associations for PI in these individuals from admission to and discharge from the NSCU [11].

Methods

Setting and population

All individuals acquiring either a traumatic SCI (TSCI) or nontraumatic SCI (NTSCI) during 2004–2014 and admitted to one of the three specialized NSCUs for acute rehabilitation after the injury, were included in the study. The acute rehabilitation period is defined as the continuous time period from admittance to the NSCU and to final discharge from the hospital. The electronic medical record (EMR), at each of the three NSCUs, was used to identify individuals and retrieve data. The system of care for persons with SCI in Norway has been described in a recent publication [12]. Because of the strict legislation regarding privacy, and the data collecting permission from the Ethical Committee [13,14,15,16], available information from the acute care hospitals is dependent on the information given in the transfer letters from the acute care hospitals to the NSCUs. These transfer letters do not include any information regarding the time from injury to the arrival at the acute care hospitals, neither any information regarding immobilization during transfer, mode of transport, or use of pressure-relieving devices or interventions.

Study design

We conducted a national, retrospective, cross-sectional study, with the aim to estimate the period prevalence of PI, and investigate potential risk factors for PI during the period between admission to and discharge from the acute rehabilitation. Available information from the EMR at the NSCUs was evaluated to retrieve potential risk factors.

Study variables

Study variables were recorded as “yes” if present, “no” if not present, and “unknown” if the information was missing. The term “PI” was used to describe pressure ulcers/wounds, according to the newest recommendations [1].

The International Standards for Neurological Classification of SCI (ISNCSCI) was used, including the clinical findings standardized by the American Spinal Injury Association Impairment Scale (AIS) [17]. Relevant information recorded from the EMR at the NSCUs were gender (male/female), date of birth, date of injury, marital status (single and living alone, single but living with parents/children, cohabitant, partner/married, divorced, widow/widower, and unknown), level of education (primary school, high school, college/university, and unknown) occupational status (full time, part time, social welfare benefits, retired, and unknown), etiology of the injury (traumatic and nontraumatic), neurological level of the injury (cervical, thoracic–sacral, and cauda equina), and any associated injury (brain injury and multitrauma). A complete examination of the skin was recorded within the first few weeks after the admission to the NSCU. Occurrence of PI, as well as use of alcohol and tobacco, and all abuse of drugs and SCI-associated problems, such as incontinence and ventilator dependence, in conjunction with premorbid comorbidities, such as hypertension, cardiac disease, diabetes mellitus (DM), clinically diagnosed depression, allergy, and skin disease, were recorded from information in the EMR. In addition, attention deficit/hyperactivity disorder (ADHD/ADD), which was previously not evaluated as a potential risk factor for PIs, was recorded if diagnosed before admittance to the NSCUs, and recorded in the EMR. The EMR documentation used in this study does not specify any diagnostic tools regarding depression or ADHD/ADD; the variables recorded were “yes,” “no,” or “missing,” depending on the information given in the EMR.

Data collection

A selection of 84 EMR diagnoses was scrutinized for SCI, and only individuals with acquired traumatic or non- traumatic SCI between January 1st 2004 and January 1st 2014 were included. Based on data obtained from the EMR, neurological level of injury and the AIS were examined and recorded during the first 3 weeks after admission to the NSCUs [12]. In some cases, the degree of impairment was not registered, but the EMR described the sensory and motor grade and level, as well as the sphincter tonus. In these cases, the impairment was graded by the first author (II) in accordance with the ISNCSCI [17,18,19].

Ethics

The Norwegian Data and Telecommunications Authority’s requirements for safe information flow were followed [14]. The study was approved by the National Regional Ethical Committee (2014/684/REK-Nord) [15, 16].

Statistical analyses

Potential risk factors diagnosed before the occurrence of the PI were included in the analyses. Continuous variables are presented as mean with standard deviation (SD). Categorical variables are presented as counts and percentages.

The categorization of age into age groups is performed, according to the newest recommendations [17, 18].

Participants’ demographics and injury characteristics are analyzed descriptively. The term “period prevalence” refers to the 10-year period between 2004 and 2014. To identify factors associated with PI occurrence, potential risk factors were entered into a binary, logistic regression model. Crude and adjusted (for gender and age) odds ratios (ORs) were calculated along with 95% confidence intervals (CIs). p values less than 0.05 were considered significant. The common confounding variables age and gender were adjusted for in the analyses. The adjusted results will be reported and discussed in the paper. As a sensitivity analysis, we also performed logistic regressions where missing values on the PI variable were taken as “no PI.” Our reasoning was that if there was no PI during acute rehabilitation, PI would not be mentioned in the EMR.

IBM SPSS Statistics, version 23, was used for all statistical analyses.

Results

Description of the population

After reviewing data from 1488 EMRs at the three NSCUs, 1012 individuals, 742 men (73%), and 270 women (27%) were included in the study. Demographics are presented in Table 1.

Table 1 Demographics.
Full size table

Period prevalence and location of the pressure injuries

We had information about PI in 891 of the individuals, and the period prevalence of PI in the studied population was 144/891 (16%, 95% CI = 0.14–0.19). We found that 61% of the population with a known number of PIs had a single PI (86/142 individuals), while 39% (56/142 individuals) had two or more.

The total number of PIs recorded from the EMR were 373. Most of the PIs were located at the coccyx (33%) (Fig. 1).

Fig. 1: Location and frequency of the PIs in the studied population.
figure 1

Most of the PIS were located at the seat (n = 168, hereof the coccyx n = 122 and the sit bones n = 46), and heels (n = 67). The feet (n = 26), ankle (n = 14), calf/ knee (n = 18), thigh (n = 11), hip (n = 11), column (n = 14), elbow (n = 7), shoulder (n = 9), ear (n = 7). The group other (n = 21) consists of PI at the chest/ abdomen (n = 3), face/nose (n = 2), occipital (n = 2), neck (n = 2), penis (n = 4) and unknown location (n = 8).

Full size image

Factors associated with pressure injury development, classified by categories

A detailed overview of factors associated with PI in our population is provided in Table 2.

Table 2 Pressure injury associations.
Full size table

Gender

Men had an overall period prevalence of PIs of 19% (95% CI = 0.16–0.22), compared to 9.0% (95% CI = 0.05–0.13) among women.

Age

The mean age at injury was 48 years, SD 19 (minimum 0.47 years–maximum 88 years). The age group of 45–59 years had significantly decreased odds of PI (OR = 0.5, 95% CI = 0.3–0.9) compared with the reference group (15–29 years).

Marital status, education, and occupational activity

We did not find any significant variation in the occurrence of PI concerning marital status, level of education, or occupational activity at the time of injury.

Cause and severity of the spinal cord injury at admission

The occurrence of PI was higher among those who were injured outside Norway.

There was no significant difference in the occurrence of PI, based on having a traumatic or nontraumatic injury, or based upon the neurological level.

Individuals with AIS D had a 90% decreased odds of PI, compared to individuals with AIS A (OR = 0.1, 95% CI = 0.1–0.2, p < 0.001). The decrease in odds of PI for individuals with AIS C was 70%, compared to individuals with AIS A (OR = 0.3, 95% CI = 0.2–0.5, p < 0.001). A test for trend in the AIS categories showed a significantly decreasing trend (p < 0.001) (Fig. 2).

Fig. 2: The trend between the AIS grade and the PI risk in the studied population.
figure 2

The figure shows estimated odds ratios with corresponding 95% CI for AIS grades B, C and D compared to the reference grade AIS A.

Full size image

Spinal cord injury sequelae

A significantly higher occurrence of PIs was observed among individuals with bladder and bowel dysfunction related to the SCI, compared to no dysfunction. A corresponding pattern was found regarding the need for ventilator support before or at admission to the NSCUs. The occurrence of multitrauma together with the SCI did not associate with the occurrence of PI; however, having a concomitant traumatic brain injury did.

Comorbidity, acquired prior to the spinal cord injury

For patients diagnosed with diabetes mellitus type 1 (DM1) ahead of the SCI, there was approximately an eight-time increased odds of PI, compared with individuals with no DM1 diagnosis (OR = 7.9, 95% CI = 2.4–26, p = 0.001); however, we did not find any increased PI occurrence for DM2. Other comorbidities, such as hypertension (OR = 3.7, 95% CI = 2.3–5.9, p < 0.001) and cardiovascular disease (OR = 3.6, 95% CI = 2.3–5.9, p < 0.001) also significantly increased the odds of PI.

Clinically diagnosed depression was present in 285 (28%) of the total population during the acute rehabilitation, and there was a higher PI occurrence in those with depression, than in those without. ADHD/ADD diagnosed before the SCI did not show any association with PI occurrence.

Stimulants

Abuse of illegal or prescribed drugs before the SCI and registered in the EMR seemed to be associated with an increased occurrence of PI (OR = 3.0, 95% CI = 1.5–6.9, p = 0.002), while being under the influence of alcohol or drugs at the time of the injury (20% of the population) did not. Regular use of tobacco and alcohol did not show any significant increase in the risk of PI.

Discussion

NORSCIPI is the first national study of PI in the SCI population in Norway. Our study population was representative and comparable with previous studies in Norway [20, 21]. An important finding in our study was the association between psychological impairments and the risk of PI, which is in accordance with previous studies [2, 22].

It should be noted that no standardized instrument to assess depression was applied in this study, since we build on information retrieved from the EMR. There is some ambiguity in the recorded depression diagnoses; however, we refer to it as depression, based on clinical evaluation.

We found that the level of injury was insignificant regarding the risk of having a PI, while individuals with AIS A and B were more at risk of having a PI, compared with AIS C and D. Previous findings are inconsistent when it comes to the association between the degree of impairment, evaluated by the AIS grade and PI [23,24,25,26]. However, our results are in accordance with previous studies, where the completeness of the injury determines the risk of having a PI [2, 7, 23, 24]. In our study, the occurrence of PI was 16%, which is lower than previous studies [6, 27, 28], and lower than the occurrence of PI in the general inpatient population in both Norway and other comparable countries [6,7,8,9, 27,28,29,30]. Mawson et al. postulated that the most likely time for the development of PI is the immediate postinjury period of spinal shock, and that some of the PIs appearing during acute care may be the visible result of ischemic injuries occurring prior to acute admission [26]. Unfortunately, we could not obtain documentation on how patients were immobilized during transfer to the acute care hospitals or the NSCUs, the mode of transport, transfer surfaces used, or whether pressure-relieving devices or interventions were utilized during transportation and hospitalization [28]. In our study, 5% of the population were injured abroad, and the occurrence of PI among them was significantly higher, compared with those injured in Norway. We believe that delay in admittance to the NSCUs might explain the findings of the increased occurrence of PI in those injured abroad. Because of strict Norwegian guidelines, regarding the prevention of multiresistant bacteria, patients injured abroad are isolated at home or at the local hospital, and not admitted to the NSCUs until their infection status is clarified [31]. PI-preventing routines for the transportation, in-hospital preventing care at both the acute care wards, and the NSCUs are important issues in the future PI-preventing recommendations. Knowledge about PI prevention should be a part of the education and training for all staff members, as well as newly injured individuals, and their relatives [32]. The Norwegian “In safe hands” program (https://www.pasientsikkerhetsprogrammet.no/om-oss/innsatsomr%C3%A5der/_attachment/3304?_download=false&_ts=14e26104012) could be implemented as a simple way to identify patients at risk of developing PI, by asking three questions for risk assessment immediately after admittance to hospital:

  1. (1)

    Does the patient have PI now?

  2. (2)

    Does the patient need assistance in position changing?

  3. (3)

    Is the patient at risk of developing PI during the hospitalization?

If the answer to any of these questions is “yes,” an action plan should be initiated, with the aim to prevent the occurrence of any PI, or to treat an already-existing PI. The flowchart in Fig. 3. provides a visual overview of the action plan and recommended measures to achieve the “No PI” goal (Fig. 3). Checklists should be used to record this information in the EMR.

Fig. 3: Suggested action plan for prevention of PI.
figure 3

This should start with a risk assessment of each patient at admission to the spinal cord unit, followed by daily observation and re- evaluation of the risk. The patients, together with their families should be included in all parts of the prevention and treatment at all stages of the rehabilitation stay. Nutrition review and nutrition measures should be provided to all hospitalized at- risk patients, together with pressure distributing equipment and optimal skin care.

Full size image

In NORSCIPI, the occurrence of PI was more than double among men, compared with women. The association between gender and PI has been studied with mixed results in previous studies [24, 26, 33]. We speculate that individuals with risk-taking behavior may continue this behavior into rehabilitation, and if there is more risk-taking behavior in the male population in our study, they may be more vulnerable to PI? This question requires further investigation. Another possibility is the difference in fat distribution in women versus men, as women often have increased adipose tissue at the buttocks and thighs, two areas that are especially vulnerable to PI [34]. Even if the cause for the gender difference is not sufficiently explored, it highlights the need for repeated information about prevention actions in vulnerable individuals, and that staff planning the rehabilitation are assessing each patientʼs risk for PI individually.

It is known that aging causes reduction of the microvascularization and of the proliferative activity of the dermis, as well as changes in the elasticity of the skin, enhancing the effect of local pressure and stretch on the skin, and thus increasing the risk of PI [28]. Nevertheless, previous studies show contradictory results concerning the association between age and PI [23, 24, 26]. In the present study, the age group of 45–59 years actually had a 50% reduced odds of PI compared with the reference group (15–29 years). An analysis regarding differences in the age groups identified a higher occurrence of depression and AIS grades A–B in the age group of 15–29 years, and we believe that these were the reasons for the increased PI odds. These findings reinforce the need to focus on particularly at-risk individuals, or subgroups, during rehabilitation.

We did not find any association between the occurrence of PI and level of education or occupational activity. The social welfare system in Norway gives everyone the same opportunity for health care, regardless of education, occupation, or income [35], and this may influence the results in our study, compared with other reports [2, 23, 24].

Our findings reiterate that risk factors, such as incontinence, lack of sensation, ventilation support, hypertension and cardiovascular disease increase the odds of PI [2, 3, 5, 28, 36, 37].

Patients with DM1 showed a higher occurrence of PI, with an OR close to 8. Although we cannot claim a causal association, PI-preventing actions regarding persons with DM1 who are acquiring a SCI, should be in focus at all terms of postinjury care and follow-up. In contrast to previous research [2, 27, 28], we did not find an association for PI and DM2. There is limited information about the differences in the risk of PI in DM1 compared with DM2, and neuropathic abnormalities, together with poor circulation and immune function changes, each contributing to vulnerable alteration in the tissue among individuals with DM1 and DM2 [38]. One study found that independent risk factors include renal insufficiency [39]. About 30% of individuals with DM1 (juvenile onset), and 10–40% of those with DM2 (adult onset), eventually will suffer from kidney failure [40]. We speculate that renal insufficiency contributed to the differences in the association between DM and PI in our population; however, the population with DM1 in NORSCIPI only consisted of 13 individuals, with a mean age of 42 years, while there were 68 individuals with a mean age of 62 years with DM2. Thus, with this small population, further research is warranted.

Surprisingly, we did not find any association between the use of tobacco or alcohol and PI, while abuse of drugs seems to be associated with PI development. Thus, our findings do not support findings in previous studies related to the use of tobacco or alcohol [41]; however, uncertainty in the number of reported users in the investigated population may partially explain our results.

Study limitations

There are a number of limitations in our study related to the clinical care of patients with SCI in Norway. Individuals with SCI not admitted to one of the NSCUs post injury due to the limited need for third-line rehabilitation, or comorbidity are not included in the study. Clinical transfer protocols for individuals with newly acquired TSCI as compared with NTSCI are well known in Norway [12], but acute rehabilitation after NTSCI is less well defined. Thus, our NTSCI sample does not include all affected individuals, in contrast to our TSCI sample. Finally, we would optimally have divided the time between injury through acute rehabilitation into two separate periods: accident to acute rehabilitation transfer, and the acute rehabilitation period. Unfortunately, this was not possible to do, given the available information in the EMRs, and this is a limitation in our study.

Because of variable reporting in the EMR, there was missing information about PI in 121 of the individuals. This may reduce the statistical power of the results [42]; however, clinical experience indicates that if there is no information about PIs in the EMR, there is generally not a PI problem. Moreover, performing a logistic regression, by setting the missing PI to “No PI,” did not change the (significance of the) results.

Information about drug abuse at the time of the injury is retrieved from available information in the transfer letter from the acute care hospital. Any missing information in this document will also be missing in the study. The lack of recorded information in the EMR regarding those who use tobacco, alcohol, and/or illegal drugs, and those who do not, results in missing data, and is another limitation of our study.

We have investigated a high number of potential risk factors. Thus, it was infeasible to develop causal models for all of them, and to adjust for all confounding factors. Hence, the identified associations should be taken as indications, worthy of further investigations to clarify casual relationships. We have also performed a high number of statistical significance tests, increasing the risk of type I errors [43]; however, most of our significant findings seem clear and robust; thus, we feel quite confident about our conclusions.

Conclusion

NORSCIPI has a unique design, because variables are recorded over a 10-year period, and data are retrieved from the EMR. The results are unique because they represent the national status of PI in the entire SCI population of Norway. We identified several factors, including DM1 and depression that may be worthy of further research to clarify their role in the causal path to PI. We give recommendations for a simple program on observation and prevention of PIs for health care providers, patients, and next of kin. An increased understanding of factors that associate with PI will allow providers to focus on patients at particular risk. Checklists on factors associated with the occurrence of PI, as well as checklists and better focus on PI prevention should be a part of the acute care SCI rehabilitation. For better outcomes, further research should focus on PI prevention routines and actions during the acute post-injury rehabilitation.