Pediatric Pressure Ulcer Prevalence: A Multicenter, Cross-Sectional, Point Prevalence Study in Switzerland
Pressure ulcers (PUs) are a common concern for hospitalized children and adults, but knowledge about PU risk factors, prevalence, and incidence rates among children remains limited. To assess the prevalence of and risk factors for PUs in pediatric care settings, a 1-day cross-sectional study was conducted among all hospitalized children ages 0 to 18 years in all 14 pediatric hospitals in the German-speaking part of Switzerland. Data collection involved a direct systematic inspection and assessment of the skin. A standardized data collection instrument was used, and each patient was assessed by a previously instructed rater pair. The total number of participating children was 412 (75% of all hospitalized children). An overall PU prevalence of 35% (including European Pressure Ulcer Advisory Panel category 1 ulcers) was observed. Most patients with PUs (80%) had category 1 ulcers. The prevalence rate was highest among patients in the pediatric intensive care unit (PICU) (16/36, 44%), followed by the department of neonatology (47/109, 43%). The presence of a PU was significantly higher among patients with a medical device, who were young (<1 year old), had a longer length of stay, and low Braden scale score (P <0.05). Rates also varied by institution (P <0.05). Department, patient age, Braden scale score, and institution explained 25% of the variance in PU prevalence. The prevalence of PUs in pediatric patients is higher than expected, and the rate of category 1 PUs suggests that interventions to prevent PUs are needed, especially in the high-risk patients identified. Future studies are needed to further assess these risk factors, especially for patients in PICUs.
Potential Conflicts of Interest: none disclosed
A pressure ulcer (PU) is a localized injury to the skin and/or underlying tissue as a result of pressure or pressure in combination with shear forces.1 Although this care problem has gained a great deal of attention in adults, far less is known about PUs in children and neonates.2 Multimorbidity is limited to a small percentage of children only. Because survival rates of both critically and chronically ill infants and children have improved dramatically in recent years, new challenges for medical and nursing care have been introduced; one of these is an increase in PU risk.3 In a retrospective, exploratory study of 50 children by Samaniego,4 as well as in a systematic literature review aimed at identifying factors contributing to the development of PUs in pediatric patients by Cockett,5 several additional PU risk factors in children are described, including the use of additional medical and therapeutic aids, such as wheelchairs, unadjusted ortheses, and prostheses. The consequences of immobility and decreased skin sensitivity are described in a prospective study of 347 pediatric patients by Suddaby et al,6 as well as in a multicenter survey by Willock et al7 that included 54 children ages 0 to 18 years, conducted over 18 months. The latter study’s goal was to identify characteristics of children with PUs. In an earlier prevalence study by Schlüer et al8 in 155 pediatric patients, as well as in a prospective matched case-control study of 271 consecutive admission patients in a pediatric intensive care unit (PICU) setting by Zollo et al,9 risk factors related to equipment such as tubes, IV catheterization, and airway devices were described. These risk factors are also observed in the study of Samaniego,4 in a point prevalence study conducted by Noonan et al10 involving 283 inpatients between the ages of 6 months and 11 years, and in a multicenter prevalence study by McLane et al.11 The latter involved 1,064 children in nine children’s hospitals. In the study by Willock et al,12 in a multisite prospective cohort study by Curle et al13 of 322 pediatric patients, and a literature review by Willock and Maylor,14 the limited age-related communication skills of infants and toddlers or of mentally handicapped children also were found to contribute to an increased PU risk (see Table 1).6,8-11,13,15-17
The skin may be affected by external influences such as increased pressure, friction, and shear forces due to intervention with external devices. Patients in PICUs, in particular those on oscillation and extracorporeal membrane oxygenation, are affected by pressure exerted by device tubes. In addition, a decreased tissue tolerance in these patients due to their critical life conditions increases the risk of skin failure and makes them most vulnerable.2,13
Recent investigations indicate that PUs are common in infants and children: a review of the literature18 noted that PU prevalence rates including all PU categories range from approximately 3% to 28% in hospitals. The previously mentioned study8 of 155 pediatric patients found the PU prevalence within the group of hospitalized children assessed as being at risk was 35%. Published incidence and prevalence study data are limited and rates vary widely,14,18 making any evidence-based statements about PU frequency in pediatric populations difficult. Often, basic reporting issues such as descriptions of samples, inclusion and exclusion criteria, or provision of clear PU definitions are missing in studies or vary from studies conducted earlier. Due to this lack of methodological descriptions of PU rating scales and to missing variables such as age or care setting, the results of these earlier studies are hardly comparable.18 Moreover, a deficit of research exists on pediatric PU risk assessment and age-specific PU interventions14; the magnitude of this health problem in the pediatric population is, in fact, largely unknown. However, a clear understanding of PU incidence and prevalence is important for the evaluation of the effectiveness of PU prevention, resource requirements, and resource allocation for this group.
Because information about the relevance and the extent of the PU problem in pediatric patients is limited, and to improve the quality of PU care, a multicenter, point prevalence study was conducted to: 1) assess PU prevalence in different pediatric hospitals according to care settings and age groups, and 2) assess risk factors and characteristics among this population of PU patients.
Methods and Procedures
Study design. A multicenter, cross-sectional, point prevalence study was conducted in all 14 pediatric hospitals in the German-speaking part of Switzerland on 1 day in June 2009. These hospitals have a total of approximately 70 wards. All pediatric disciplines at the participating sites were included. These involved children from all departments including PICU, neonatal intensive care units (NICU), all surgical disciplines, all pediatric medical disciplines, and departments for pediatric rehabilitation care. All hospitalized children ages 24 hours (including premature babies) up to 18 years of age were included in the study. Children had to have been hospitalized for at least 1 day in the respective institutions. Exclusion criteria were 1) hospitalization in psychiatric wards, 2) children whose legal representatives did not allow participation, and 3) children who refused to participate. The ethics board of each hospital approved the study where necessary, as well as all cantonal ethics committees. After the patients and their families had been verbally informed about the study’s content and aim, they were given the same information in written form along with an informed consent form to sign. The information letter, as well as the informed consent form for the parents, was available in eight different languages (German, French, Italian, English, Portuguese, Albanian, Serbian, and Turkish). Children 10 years of age and older were asked to give their own written consent.
Measurements. To assess the prevalence, severity, and predisposing factors for PUs, the Dutch National Prevalence Measurement Care Problems (LPZ)19 tool was utilized. This instrument is widely used in Europe and has been shown to be reliable and valid.20 The following data categories are collected: 1) patient characteristics (demographic and clinical data), 2) assessment of PU prevalence and severity, 3) PU risk assessment using the Braden scale, 4) predisposing factors for PU development, 5) preventive interventions, and 6) therapeutic interventions. In addition to factors included in the original instrument, the authors also assessed the presence of medical devices on the patient’s body at the time of the assessment (eg, tubes, IV catheterizations, continuous positive airways pressure [CPAP], splints, and other installations). These adaptations were incorporated based on the findings of a previous study8 and were verified by the clinical nurse specialists in the participating hospitals.
Data collection related to PU assessments involved a direct and systematic inspection and judgment of the skin per the European Pressure Ulcer Advisory Panel (EPUAP)1 four-category system. Demographic and clinical data, such as date of birth, weight and size, body mass index, and diagnosis, were collected from the patient charts.
PU risk assessment was completed using the original Braden Scale.21 The Braden Scale is one of the most well known and widely used tools for evaluating PU risk in adults; its psychometric properties in assessing risk have been verified. The Braden Scale also has been used previously to assess PU risk in pediatric patients.4,6,8 It includes six subscales, five of which have four and one of which has three answer categories. All data were recorded on a paper case report form and entered by the researcher for data analysis. A second research assistant entered the data again. A validation of these two files (original and duplicate) was conducted to achieve concordance.
Previously trained rater pairs gathered data from each patient, one of whom did and one who did not work on the ward surveyed. A total of 35 rater pairs were involved in the inquiry. If the pair disagreed on their assessment, they were asked to find a consensus. If no consensus could be reached, the rater not working on that ward was given the right of decision. To be selected, raters were required to have more than 2 years of nursing practice with children. In all participating hospitals, nurses who worked either as wound consultants in their units or as clinical nurse specialists acted as raters. Preparatory training included methodological aspects, detailed information about data collection, the role and responsibilities of the raters, a detailed introduction to the measurement instruments, and special training in PU grading and use of the Braden Scale in infants and children. Pilot studies at all sites with more than two rater pairs were conducted to assess the inter-rater reliability for grading and risk assessment23 using a balanced incomplete block design. A data consistency of 95% as to whether the patient had developed a PU was observed. Overall, interrater reliability (Kappa 0.79) for the grading of PUs was sufficient.
Data analysis. Predictive Analytics Software (PASW) (version 18, formerly SPSS, Chicago, IL) was used to analyze the study data incorporating descriptive and univariate statistical methods. Distributions and frequencies were calculated. The relations between dependent variables and influencing factors were explored with cross tabs. Group differences were due to the type of data tested with nonparametric tests for categorical and not normally distributed data. The relationship between PUs and possible influencing and confounding factors was studied using logistic regression analysis. To find the most suitable model for the analysis of variance as well as logistic regression, the following variables were applied to the model: age, gender, operation (ie, if the patient had an operation in the 2 weeks before assessment), body mass index (BMI), preventive interventions, Braden score, institutions, and departments. The data were entered into the model followed by a forward stepwise selection of the important variables according to the use of likelihood quotients. The entry and removal probabilities were 0.05 and 0.10. A P value of <0.05 was considered statistically significant.
Demographics. The overall sample size of potential study participants who met the inclusion criteria was 560 (100%). The total number of participants was 412 (75%); 141 (25%) dropped out before the assessment. In 67 cases (12%), either the parents or the child refused to participate. Sixty patients (60, 11%) were excluded due to unexpected discharge or because the patients were undergoing tests or operations that lasted longer than the survey. Four patients were not assessed due to an unexpected change in their condition to a critical level. Only five patients withdrew related to language problems of the parents.
In the 14 clinics, the participation rate ranged from six to 97 patients (43% to 100%). The average length of stay was 25 days (range 1–497 days), with a median of 7 days. Of the 412 assessed patients, 275 (67%) had been hospitalized for <14 days. In this subcategory, the average length of stay was 5 days, with a median of 3 days. Patients with a diagnosis related to oncology, acute gastroenteritis, pneumonia, or a cardiology diagnosis comprised the largest group (135, 33%) and represented pediatric medical patients, followed by neonates (109, 27%) and young patients in the surgery disciplines (99, 24%), including plastic and burn surgery, visceral surgery, neurosurgery, and orthopedic and trauma patients. The number of PICU patients was 36 (9%). Additionally, 33 patients (8%) were hospitalized for rehabilitation problems.
The average age of all patients was 4 years with a median of 1 year. Approximately 50% (203) patients were <1 year old on the day of inquiry. In this subgroup, the average age was 6 weeks, with a median of 3 weeks (see Table 2).
PU prevalence and patient characteristics. Of the 412 patients taking part in this study, a total of 142 had one or more PUs on the day of the study, a PU prevalence rate of 35%. Of these 142 patients, 77 (54%) had one, 35 (25%) had two, 17 (12%) had three, six had four (4%), two had five (2%), three had six (1%), one had nine, and one patient had 10 PUs. This yields a total number of 269 PUs (see Table 3).
The most frequent type of PU was a category 1 PU (121, 94%). Sixteen patients (16, 3%) had category 2 and higher, three had category 3, and two had category 4 where assessed. The three children with a category 3 PU were all 16 years of age with multiple diagnoses and treatment for orthopedic surgery. One category 3 PU had occurred within the last 2 weeks on the ward where the patient was hospitalized, while the two other category 3 PUs had existed for a longer time. These two category 3 PUs had existed for 3 to 6 months and had developed in another hospital before these patients were transferred to the clinic where they were assessed. The two children with a category 4 PU were 9 and 16 years of age. The younger child was being treated in the surgery department after a multiple trauma incident and had developed the category 4 PU within the previous 2 weeks, while the other patient was being treated for an orthopedic surgery indication. The PU in this patient had already existed for more than 1 year and had developed in another care setting.
The PU prevalence differed between the departments (X2: 13.8, df: 5, P = 0.002), with the highest PU prevalence (44%) seen among patients in PICU (see Table 3).
The prevalence of PUs for patients with an external device was 40% (see Table 3). By age category, the PU prevalence for patients <1 year of age was 43%, followed by 31% for patients >12 years of age. In terms of length of stay before measurement, patients with a PU had been in the hospital slightly longer than patients without a PU (P = 0.036) (see Table 3).
Within the group found to be at risk for PUs (ie, patients with a Braden Scale score ≤20), 93 patients presented with a PU (49%), whereas among patients not at risk 46 (24%) had a PU (X2: 26.4; df: 1; P <0.001). Classified by departments, the PU prevalence for patients at risk was 50% for a patient after a surgical intervention as well as for neonates, 46% for pediatric medical patients, 25% for patients in a rehabilitation setting, and 60% for patients in the PICU. Of all patients with at least one PU, 76% were assessed as being at risk (sensitivity 67%). The number of patients assessed as being at risk who did not have a PU (specificity) was 95 (40%).
PU influencing factors. The risk factors age, gender, surgery (ie, the patient had at least one surgical intervention during the present hospitalization), BMI, preventive intervention, Braden scale score, institution, and department were examined with a multiple logistic regression analysis. The factors department, age, Braden score, and institution contributed significantly to fit in the model. The Hosmer Lemeshow test showed an adequate goodness of fit for this model (Nagelkerke R2 = 0.248) (see Table 4). This means that those remaining four variables explain 25% of the variance when PUs are the dependent variable.
The current study reveals a very high PU prevalence rate (35%) in the 14 pediatric hospitals. However, the prevalence rate of category 2 and higher PUs is only 3%. In particular, the age of the patients, the Braden risk score for PU development, and the institution in which the patients were hospitalized appear to be related to the development of PU.
In this study, the patients were assessed by a trained rater pair. From the good interrater reliability and data consistency in the pretest, one can assume the results presented are of sufficiently good quality. The prevalence rate of 35% is higher than previously found prevalence rates in worldwide pediatric care settings, which varied from 3% to 28%.8
The high prevalence rate demands critical appraisal. First, an already well-known problem is the diagnosis of category 1 pressure ulcers. This problem was first described in a study by Halfens et al.24 In a cross-sectional mail survey of coordinators of the Dutch National Pressure Ulcer Prevalence Survey in 1998, the authors distinguished several factors that have an impact on the assessment of PUs in institutions. One of these factors is the difficulty in diagnosing category 1 PUs. It should be noted that a category 1 PU can be misidentified because nurses might diagnose a blanchable erythema as a category 1 PU. This may lead to an overdiagnosis of PUs.26,27 Furthermore, according to the study by Halfens et al,24 most category 1 PUs are reversible. Therefore, several authors have recommended defining PU prevalence by starting the category system at category 2.26,27 The raters for this study were prepared and especially trained in diagnosing category 1 PUs. In the current study interrater pretest, 95% agreement was achieved, suggesting that findings are reliable. Also, even if category 1 PU is not defined as a PU, the presence of a category 1 PU can at least be considered the most important risk factor for PU,27 which subsequently can be interpreted for this study as meaning that a high percentage of pediatric patients are at high risk of developing PUs.
The prevalence of category 2 PUs and above accounts for 3% of the total. This is lower than the prevalence of PUs in the studies by Suddaby et al6 and Groene-veld et al16 (both 5.1%) and also lower than the PU prevalence in an earlier study by Schlüer et al8 (4.5%). In Groene-veld et al’s 2004 study,16 prevalence was assessed in 97 inpatients; prevalence among children was found to be 13.1% (including category 1). On the basis of Groene-veld et al’s16 and Schlüer et al’s8 results, one can conclude that while many patients are vulnerable to PU, the progression to a higher category occurs rather infrequently. Nevertheless, this implies that the diagnosis of a category 1 PU requires a related preventive intervention.
The data collection in all participating clinics took place on 1 day within the same time span from the morning (7:30 am) up to 3:00 pm in the afternoon. This is important insofar as it is known from adult patient populations that approximately 50% of all category 1 PUs disappear during the day, whereas approximately 22% of all category 1 PUs worsen during the day.24 No study related to this particular phenomenon has been undertaken so far for the pediatric setting.
The most widely affected patient group in this study involved very young patients in the PICU setting, who presented with mostly category 1 PUs. This is in line with results of Curley et al.13 The PU prevalence of nearly 45% in this patient group with 60% at risk is disconcerting and needs urgent attention. Patients in the NICU care setting had an overall prevalence of 43%. Children treated in a NICU and/or PICU care setting are most vulnerable due to several factors, such as the immaturity of the skin in very young neonates and newborns and anatomical, physiological, and immunological differences.8,13 In general, these children are in a critical clinical condition and often undergo life-saving treatments. These patients often have a great number of external medical devices present that might affect the skin, and high vulnerability to developing a PU seems probable.8,13,18,28 Preterm babies, newborn babies, infants, and toddlers are unable to judge or acclimate to pressure or sensory perceptions from medical devices, which renders them vulnerable to skin compromise.28 Furthermore, their communication abilities and skills are limited due to their age and development status.
This is the first study to document a PU prevalence rate for children with external medical devices. At 40%, this PU prevalence rate is high and demands special attention. Due to this fact and the findings presented here, the question arises whether in the pediatric population a prevalence rate for children with and without external medical devices such as tubes and fixation devices might be an important marker. The impact of pressure and shear forces due to these external devices seems to increase for children more than for adults, because these devices seem responsible for a relatively high number of PUs in pediatric patients.8,28-33 On the other hand, it should be noted that only a few studies have investigated factors favoring PU development in the pediatric population.5,8
At the time of data collection for the present study, the Braden-Q and Glamorgan scales were the only tools available for risk assessment in the pediatric population.13,30-32 The Braden-Q, developed and introduced for immobile patients from 21 days of age, is still not validated for children of >8 years of age, and no validated German version is available. A German version of the Glamorgan scale has not been validated. Because of this lack of alternatives, the original Braden scale was used to assess PU risk in this study. However, the use of the Braden scale for this group is problematic because no defined cut-off point for specifying children at risk is known. Additionally, it had been claimed that the Braden scale does not offer adequate reliability and sufficient predictive performance.34 An adapted, reliable, and valid risk assessment tool, applicable in the pediatric population, is still not available.29 According to a prior study,8 and due to the lack of alternatives, the Braden scale score was used to predict risk assessment.
Although not all PUs are preventable, the high prevalence rate detected in this study indicates a need to improve PU care, especially PU prevention, and to start raising staff awareness of the PU issue in the pediatric healthcare setting; especially in the PICU and/or NICU care setting.8,11,13,18 This need has been described in a systematic literature review on pediatric PU prevalence rates, as well as in a systematic literature review on pediatric PU risk assessment scales, both by Kottner et al.18,29
As other prevalence surveys have shown,24 simply providing deeper insight into a specific care problem may change staff awareness. This may have affected the results of this study, because in four of the 14 hospitals taking part in the survey, a PU prevalence study had been conducted 3 years prior. However, no evidence exists that shows the nurses in these institutions changed their daily practice because of the results from the first PU prevalence survey. In two institutions, PU prevention and treatment guidelines were introduced in 2007 and implemented between 2007 and the beginning of 2009. Repeated training and workshops for the nursing staff were offered along with guideline implementation. It remains unclear whether this influenced the results.
Prevalence studies are always subject to daily variations. However, the prevalence figures over the 2 years are relatively constant.24 Most of the PUs in this study were category 1. To what extent the category 1 PUs are over- or underdiagnosed in this study remains unclear, although the interrater reliability suggest the scores are reliable.
The Braden scale is not a validated risk assessment tool for this population. No details about appropriate cut-off points for patients at risk according to the Braden scale are available. No findings on detailed differences in specific age categories according to risk for the development of a PU or the prevalence of external devices on the skin can be presented. The number of eligible patients was affected by the requirement of a signed informed consent of both parents and all patients older than 10 years of age.
Prevalence, incidence, and prospective studies in different care settings such as PICUs and NICUs are necessary to validate findings and to generalize the results to other pediatric care settings worldwide. Further prospective and case-control studies are needed to establish the special risk factors of the highly vulnerable population in PICU and NICU settings and possibly incorporate these into a target group specific risk assessment instrument. Furthermore, the fact that in this study only older children showed category 3 and 4 PUs should be noted. Detailed research exploring this phenomenon is needed. Until now, there has been a lack of repeated measures over time on the development and deterioration of PUs in pediatric patients in different care settings.
This study confirms the relevance and importance of a high quality of care to target the problem of PUs in pediatric care, especially in PICU and NICU care settings. The PU prevalence rate or, at the very least, the risk for developing a PU, is high. Particularly, the use of medical devices seems to be related to the risk for developing PUs in hospitalized children. Department, patient age, Braden Scale score, and institution have to be considered additional risk factors for PU occurrence. Again, pediatric patients in the PICU and NICU setting appear to be most vulnerable. Therefore, it is important that nursing staff be trained in the early recognition of, risk factors for and evidence-based preventive methods related to PUs in children.
The authors thank all the nursing staff members and the chief nursing officers of all the hospitals involved for their active and interested participation in the study assessment. Further, they thank all of the children and their families for their contribution. In addition, the authors are grateful to Dr. Heather Murray for her support in the preparation of this paper in English.
Ms. Schlüer is a clinical nurse specialist for pediatric wound care and a nurse scientist, Section of Nursing Science, Children’s Research Center, Children’s University Hospital Zurich, Zurich, Switzerland. Dr. Halfens is Associate Professor of Health Services Research, CAPHRI, Maastricht University, The Netherlands. Prof. Dr. Schols is a professor of Old Age Medicine, Department of General Practice and Department of Health Services Research, CAPHRI, Maastricht University. Please address correspondence to: Anna-Barbara Schlüer, MScN, PhD(c), Children’s University Hospital Zurich, Steinwiesstrasse 75, CH-8032 Zurich, Switzerland; email: Barbara.Schlueer@kispi.uzh.ch.