OUP user menu

Tracking quality over time: what do pressure ulcer data show?

Lena Gunningberg, Nancy A. Stotts
DOI: http://dx.doi.org/10.1093/intqhc/mzn009 246-253 First published online: 7 April 2008


Objective To compare the prevalence of pressure ulcers and prevention before and after a quality improvement program; determine whether patient characteristics differed for those who did and did not develop pressure ulcers; identify pressure ulcer prevention implemented at admission and whether prevention and risk factors varied by pressure ulcer severity.

Design Descriptive comparative study based on two cross-sectional pressure ulcer surveys conducted in 2002 and 2006, complemented with a retrospective audit of the electronic health record and administrative system for patients identified with pressure ulcers.

Setting 1100-bed Swedish university hospital.

Participants 612 hospitalized patients in 2002 and 632 in 2006.

Main outcome measures Prevalence of pressure ulcers and prevention (pressure-reducing mattresses; planned repositioning; chair, heel and 30° lateral positioning cushions).

Results Pressure ulcer prevalence was 23.9% in 2002 and 22.9% in 2006. When non-blanchable erythema was excluded, the prevalence was 8.0 and 12.0%, respectively. The use of pressure-reducing mattresses increased while planned repositioning decreased. Those who developed ulcers were older, at-risk for ulcers, incontinent and had longer length of stay. Little prevention was documented at admission. Some prevention strategies and risk factors were related to severity of ulcers.

Conclusions Pressure ulcer prevalence did not decrease, despite a comprehensive quality improvement program. Special attention is needed to provide prevention to older patients with acute admission. Skin and risk assessment, as well as prevention, should start early in the hospitalization. Identifying those persons with community-acquired versus hospital-acquired ulcers will strengthen pressure ulcers as an accurate marker of quality of care for hospitalized patients. If possible, data should be reported by ward level for comparison over time.

  • hospitals
  • pressure ulcer
  • prevention
  • quality indicators
  • risk assessment


Patient safety and quality of care are high on the healthcare agenda [1]. Pressure ulcers have long been used as a quality indicator of nursing care. Pressure ulcer prevalence studies are being currently used in many institutions around the world to monitor quality of care [2, 3]. The prevalence of pressure ulcers is high in hospitalized patients. In the United States, large datasets (n = 17 510 to n = 31 969) show a pressure ulcer prevalence between 14 and 17% [4]. A Canadian study reports a prevalence of 25.1% in acute care (n = 4831) [5] and across European settings (Belgium, Italy, Portugal, UK and Sweden), the prevalence is 18.1% in hospitals (n = 5947) [6]. Pressure ulcers are a problem because they cause suffering [7, 8] and increase healthcare costs [9, 10]. Efficient comprehensive improvement work is needed to reduce the prevalence of pressure ulcers in hospitalized patients [10].

In Sweden, pressure ulcers have not routinely been a hospital-level quality indicator. However, pressure ulcer prevalence was evaluated in 2002 in a Swedish university hospital using the European Pressure Ulcer Advisory Panel methodology [6]. Of the patients surveyed (n = 612), 23.9% had pressure ulcers [11]. A comprehensive hospital-wide quality improvement plan was developed and undertaken that addressed pivotal aspects of pressure ulcer prevention (Table 1). The prevalence was subsequently reevaluated in 2006. This paper provides a snapshot of the pressure ulcer status prior to and after the quality improvement program and provides insights into factors that contribute to the post-implementation pressure ulcer status.

View this table:
Table 1

Pressure ulcer prevention after prevalence survey 2002

Information and education
 Information to all head nurses.2002–06
 Educational program and seminars for registered nurses and nurse assistants.2002–06
 Networking for pressure ulcer nurses.2002–03
 Web-based program (PUCLAS) for pressure ulcer classification for registered nurses and nurse assistants.2003–06www.epuap.org
 Risk and skin assessment mandatory for nursing students.2002–06[25]
Development of clinical guidelines
 Guidelines for purchase and allocation of pressure-reducing mattresses.2002
 Multidisciplinary clinical guidelines developed by a work group for the county (university hospital, county hospital, primary care and community settings).2005–06www.akademiska.se
 Comparison of the accuracy and quality of the documentation of pressure ulcers between physical examination of patients and audit of patient record content.2003[26]
 Templates for risk assessment, pressure ulcer grading and standard care plans were developed to facilitate adequate documentation in the electronic health record.2003
 Mandatory use of the templates for pressure ulcer grading in the electronic health record both in admission and discharge notes in nine surgical wards.2005[27]
Quality indicator–improvement
 The EPUAP prevalence survey was repeated in the three departments with highest prevalence (orthopedic/surgery, medical, geriatric). Fast feedback of results.2004[28]
 Pressure ulcer identified as a quality indicator on hospital level. Mandatory annually reporting of pressure ulcer prevalence.2004
 Workgroup commissioned by the County council to develop a model for feedback of pressure ulcer incidence from the electronic health record.2005–06
 Actions (Plan-Do-Study-Act) performed on department level.2002–06

The aims of the study were to:

  1. compare pressure ulcer prevalence and prevention in a university hospital before and after implementation of a pressure ulcer quality improvement program,

  2. determine whether the patient characteristics differed for those who did and did not develop pressure ulcers,

  3. identify pressure ulcer prevention implemented at admission, and

  4. determine whether prevention and risk factors varied by pressure ulcer severity.



The data for this descriptive comparative study come from two cross-sectional pressure ulcer surveys conducted in 2002 and 2006. Both datasets were utilized to address the first study aim. The 2006 survey data were used to examine the remaining study aims, complemented with a retrospective audit of the electronic health record and the administrative system for patients identified with pressure ulcers.


The sample included all patients, 18 years and older, admitted to a 1100-bed university hospital before midnight the day of the survey. All inpatient areas were surveyed except psychiatry, day care, maternity and hospice. In addition, 120 eligible patients (16.4%) in 2002 and 92 patients (12.7%) in 2006 were not included because they were not available for inspection in the ward or they refused to participate.


Prevalence was defined as the number of persons with pressure ulcers detected by physical examination on the survey day. The EPUAP prevalence methodology was used for the physical examination [6]. Pressure ulcers were defined using the EPUAP criteria [6].

  1. Grade 1: non-blanchable erythema of intact skin;

  2. Grade 2: partial-thickness skin loss involving epidermis, or dermis, or both;

  3. Grade 3: full-thickness skin loss involving damage to or necrosis of subcutaneous tissue that may extend down to, but not through, underlying fascia; and

  4. Grade 4: full-thickness skin loss with extensive destruction, tissue necrosis, or damage to muscle, bone or supporting structures. Necrotic ulcers are classified as Grade 4 [6].

Pressure ulcer severity increases from Grade 1 to 4.

Prevention was defined as the use of pressure-reducing mattresses, chair cushions and planned repositioning in bed and chair observed at the time of the physical examination. In the 2006 survey, cushions for 30° lateral positioning and heel cushions were added. In the retrospective audit of records, documented information regarding prevention was used.

Pressure ulcer risk was assessed with the Braden Scale and the incontinence subscale of the Norton scale [6]. A total Braden score <17 was defined as at risk for pressure ulcer development. In the audit of records, risk assessment was defined as either documentation of clinical judgment ‘at-risk’ or findings of risk using a validated risk assessment tool.

Patient characteristics in the survey included age, gender, expected length of stay and department. In the retrospective audit, information was abstracted from the administrative system on primary and secondary diagnoses, admitted from home or not, acute or elective admission (acute admission was through the emergency department), not cared for in usual ward, surgery, length of stay in the hospital, and whether the patient died within 7 months. Data were obtained from the electronic health record on the hemoglobin, blood pressure (BP) and time in the emergency department. For documentation of pressure ulcer, notes from all professionals were searched with the key word ‘Skin’, ‘Pressure Ulcer’ or ‘Ulcer’.

A standardized data-collection form was utilized in the survey [6]. An additional one-page data-collection form was developed for the retrospective audit of the electronic health record and the administrative system.


Permission for the study was obtained from the medical director at the hospital. The patients received verbal and written information about the study and gave verbal consent. All data were treated confidentially. Participants were free to withdraw at any time. Approval was obtained from the Research Ethics Committee of the Faculty of Medicine at Uppsala University (No. 01-502).

For both prevalence studies, each patient was visited by a team of two registered nurses, i.e. a specially trained data collector (non-ward nurse) and a staff nurse (ward nurse). The patient's skin was assessed, the Braden scale was completed, and preventive strategies were recorded. If there was a disagreement about the pressure ulcer grade, the decision was made by the non-ward nurse.

Prior to each survey day, all nurses participating in the prevalence survey attended a half-day training on the survey procedure. Each nurse graded 10 color photos of pressure ulcers. Inter-rater reliability examined with Cohen's kappa was 0.82 (n = 22) in 2002 and 0.78 (n = 52) in 2006, which was judged to be excellent agreement [12]. After the inter-rater reliability was tested, additional education was provided by reviewing each photograph and discussing the criteria for its pressure ulcer grading.

For the 2006 survey, two experienced Quality Coordinators, prior head nurses with special training in the electronic record use and the administrative system, conducted the retrospective audit for patients identified with pressure ulcers.

Data analyses

Study data were analyzed using SPSS (version 14.0) and were explored descriptively. To compare groups, Student's t-test was used for continuous variables, Mann–Whitney U-test for ordinal scale variables, and Chi-square for dichotomous variables. A P-value of <0.05 was considered statistically significant.


Comparison of pressure ulcer prevalence and prevention: 2002 and 2006

Over 600 patients were included in each prevalence study (n = 612 in 2002; n = 632 in 2006). There were no significant differences between the groups in gender, age, type of unit or risk status (Braden subscales or total score). Expected length of stay was significantly shorter in 2006 (P = 0.002) (Table 2).

View this table:
Table 2

Patient characteristics in 2002 and 2006

2002 (n = 612)2006 (n = 632)P-value
 Missing data152.5
 > 89335.4467.3
 Missing data50.8
Expected hospital stay
 < 6 days14323.420232.00.002
 > 6 days–1 month34356.030748.6
 > 1 month11018.09715.3
Type of unit
 Acute department49681.051281.00.56
 Intensive department304.9243.8
 Geriatric department8614.19615.2
Risk assessment
 Braden score <1713722.415524.50.28

The prevalence of all pressure ulcers (Grade 1–4) was 23.9% in 2002 and 22.9% in 2006. When Grade 1 pressure ulcers were excluded, the prevalence rates were 8.0 and 12.0%, respectively (Table 3), which reveals a significant increase (P = 0.018) from baseline to the second survey. On the other hand, the mean number of ulcers per patient decreased significantly from 1.9 in 2002 to 1.6 in 2006 (P = 0.02). In both years, the sacrum and heels were the most common locations. However, in 2006, ‘other locations’, e.g. elbows, ears and feet, have increased (P = 0.02). The use of pressure-reducing mattresses increased significantly (P <  0.001) from 25.3% in 2002 to 41.1% in 2006. Planned repositioning in the bed or chair as well as the use of a pressure-reducing cushion in the chair were used sparsely in both years and decreased significantly over time, i.e. repositioning in bed (P = 0.02), chair (P = 0.01) and pressure-reducing cushions (P = 0.02).

View this table:
Table 3

Pressure ulcer prevalence and prevention in 2002 and 2006

2002 (n = 612)2006 (n = 632)P-value
Pressure ulcer
 Grade 19715.86910.90.97
 Grade 2203.3487.6
 Grade 3172.8142.2
 Grade 4122.0142.2
Location of most severe pressure ulcer
 Missing data42.785.5
Prevention in bed
 Pressure-reducing mattress15525.326041.1<0.001
 Planned repositioning8513.9619.70.02
Prevention in chair
 Pressure-reducing cushion6610.8447.00.02
 Planned repositioning376.0172.70.01

Characteristics of those who did and did not develop pressure ulcers in 2006

Patients with pressure ulcers (n = 145) were compared with those without ulcers (n = 487). Patients with pressure ulcers were significantly older (mean age 77 versus 66 years; P < 0.001), had lower scores on all Braden subscales (P < 0.001), total Braden score (P < 0.001), and more incontinence (P < 0.001) than patients without pressure ulcers. They also had significantly longer hospital stay prior to the survey (mean number of days 16 versus 10 days on that specific ward; P < 0.001). Patients with a pressure ulcer received pressure-reducing mattresses (P < 0.001) and planned repositioning (P < 0.001) more frequently than patients without pressure ulcers.

Most of the patients (77.2%) with pressure ulcers had an acute admission and were admitted from home. One-third of the patients in the emergency department and in the operating room stayed there more than 4 h. Although the majority of patients were older (75.2%, >70 years), 25% of the patients with pressure ulcers were younger. Patients with pressure ulcers were present on all of the wards. Twenty-eight patients (20.5%) died within 7 months after the survey. At hospital admission, mean systolic BP was 132.7 (SD 28.6), mean diastolic BP was 73.8 mmHg (SD 13.7) and mean hemoglobin was 124.6 mg/l (SD 19.4).

Patients with pressure ulcers were admitted for varied reasons. Most common were admission for rehabilitation (n = 27), neurological conditions (n = 21), fractures (n = 17), circulatory conditions (n = 14), infection (n = 12) and malignancy (n = 11). Only two patients were admitted primarily for pressure ulcer treatment. Many patients had multiple secondary diagnoses (range 1–10), with the mean and median being 4.0.

Pressure ulcer prevention implemented at admission: 2006

Of the 145 patients with pressure ulcers, 136 records were audited. In nine cases, patients' identity numbers were not correct, and it was not possible to find the records. Skin inspection was recorded in the admission note on 56 patients (41.2%). Twenty patients (3.2%) were identified with pressure ulcers, but only three were described in sufficient detail that pressure ulcer grade could be determined (two Grade 1 and one Grade 3). Risk assessment was documented for a fourth of the patients and only a few used a validated instrument. Only one patient received a pressure-reducing mattress, another 30° lateral positioning cushion, three had a heel cushion, and five had planned repositioning.

During hospitalization and prior to the survey, 18 patients were identified with a new pressure ulcer, thus 38 patients (6.0%) had pressure ulcers when they arrived on the ward where the survey was conducted. On the survey day, 145 (22.9%) had ulcers.

Prevention and risk factors by pressure ulcer severity: 2006

When prevention of those with Grade 1 ulcers was compared with that for more severe ulcers (Grades 2, 3 and 4), those with more severe ulcers had significantly more pressure-reducing mattresses (P < 0.001), cushions for 30° lateral positioning (P = 0.001), and heel cushions (P = 0.01). Interventions rarely used and that did not differ significantly by severity were planned repositioning in bed, chair cushions and repositioning in chair. Seat cushions were used mainly in the geriatric department.

The total Braden scores showed that 80 of 145 patients (55.2%) with pressure ulcers were at risk (Table 4). However, 91 patients (62.8%) were bed or chair fast, 65 (44.8%) had very limited mobility or were completely immobile, and 42 (29.0%) required moderate to maximum help with bed or chair repositioning. Patients with more severe ulcers experienced greater moisture (P = 0.003), were more often bed or chair fast (P < 0.001), immobile (P < 0.001), and had issues of friction and shear (P = 0.001).

View this table:
Table 4

Braden scores and incontinence scores by pressure ulcer grades in 2006

Grade 1 (n = 69)Grade 2–4 (n = 76)Total (n = 145)P-valueNo pressure ulcer (n = 487)
Sensory perception
 1. Completely limiteda2464.10.1171.4
 2. Very limiteda8101812.4306.2
 3. Slightly limited22325437.27214.8
 4. No impairment36296544.837877.6
 1. Constantly moista0442.80.00310.2
 2. Very moista4111510.3173.6
 3. Occasionally moist24315537.97115.0
 4. Rarely moist38266444.138381.0
 1. Bedfasta14324631.70.0016513.4
 2. Chair fasta20254531.05711.8
 3. Walks occasionally25143926.912325.4
 4. Walks frequently95149.723949.4
 1. Completely immobilea3121510.3<0.001102.1
 2. Very limiteda16345034.58216.9
 3. Slightly limited33225537.913728.2
 4. No limitation1582315.925752.9
 1. Very poora7101711.70.59214.3
 2. Probably inadequatea24265034.58417.4
 3. Adequate27325940.717035.1
 4. Excellent1081812.420943.2
Friction and Shear
 1. Problema12304230.00.001367.4
 2. Potential problem31326343.46212.8
 3. No apparent problem25143926.938779.8
Total score <1727438055.2<0.0017516.1
 1. Urine and fecesa9162517.20.11193.9
 2. Usually/urinea55106.9224.6
 3. Occasional48128.3224.6
 4. Not50459565.542087.0
  • aHigh risk for each risk factor.


Pressure ulcers across time were present in about 1 in 4.5 patients, and the prevalence did not decrease despite a comprehensive quality improvement program. When grade 1 ulcers were excluded, the prevalence of ulcers remained high (8% in 2002 and 12% in 2006), although less than that seen in other studies [4]. It is disappointing that the quality improvement program did not result in a decreased prevalence of pressure ulcers. The methodology recommended by the European Pressure Ulcer Advisory Panel is a point prevalence survey and does not gather data for the origins of any pressure ulcers. Thus, these findings may reflect only the nature of patients admitted to the hospital, and not necessarily the care provided during hospitalization. This raises the question of whether prevalence is a good measure of quality of care [10, 13]. Furthermore, the retrospective audit of the electronic health record revealed a lack of documentation of skin assessment on admission to the hospital, thus it was impossible to decide whether the pressure ulcers were hospital-acquired or not. It is crucial to have a methodology that is both reliable and relatively easy to conduct for regular feedback to the clinicians.

A similar study from a 900-bed hospital in the Netherlands found a significant decrease in Grade 1–4 hospital-acquired pressure ulcers from 18 to 11% (P < 0.001) after implementing a hospital guideline for pressure ulcer care including a visco-elastic foam mattresses [14]. However, their sample was younger than ours and skin assessment was only performed on patients ‘at-risk’, which could mean an underestimation of ulcers due to lack of detection. Several quality improvement reports show examples of success, but often they do not report pressure ulcer grades, reliability of the data collection or patient demographics [10, 13].

Successful implementation of change can be explained by the relationship among evidence, context and facilitation [15]. The organizational context, including leadership and culture, is highlighted as it influences priorities and investments [16]. In the hospital studied, the focus for the last few years has been on two issues: major reorganization of departments and nursing leadership and implementation of the electronic health record. Educational priorities were on facilitating documentation by all professionals in the computerized system, limiting time and energy for substantive focus on pressure ulcer prevention. Another possible explanation for our findings is that improvements in one department were offset in another. A parallel study shows that the incidence of pressure ulcers in our orthopedic ward decreased from 55% in 1997 to 18% in 2006, after stepwise introduction of risk and skin assessment, pressure- reducing mattresses, education and nutritional guidelines [17] (A.-K. Westerlund et al., submitted). Because of the reorganization, it was not possible to compare data at the ward level.

When analyzing the patients with pressure ulcers in detail, it is evident that more than 75% were 70 or older, had acute admissions, were admitted from home, and over 40% had surgery during their hospital stay. These findings are confirmed by others [18, 19]. Patient with pressure ulcers spent more days in the hospital had multiple co-morbidities and a high post-survey death rate. This suggests that patients were frail, lacked biological reserves, required complex medical treatment, and were at risk of pressure ulcers as an iatrogenic consequence of hospitalization [20]. Most (77%) had acute admission, and yet only 41% had admission skin assessment. Very few patients received prevention from the start of their hospitalization.

When care at admission was compared with that on the survey day, more prevention was provided on the survey day. However, it was not possible to identify the timing of mattress use (prior to or following ulcer identification). Over the 4-year period, the mean number of ulcer per patient decreased, but the existing ulcers were slightly deeper. This might be explained by a decrease in planned repositioning and use of chair cushions, although the use of pressure-reducing mattresses significantly increased. These findings show that it is important to emphasize to staff that regular repositioning and heel-protection is needed, even when the patient lies on a pressure-reducing mattress. Findings from this study are consistent with those of De Laat et al. [14], who found that despite implementing a hospital guideline for pressure ulcer prevention, repositioning did not increase. Repositioning is central to prevention [21], yet the issue is how to translate science into practice. A recent systematic review revealed that little is known about how to increase research use in nursing [22].

The Institute for Healthcare Improvement in the United States recommends an ‘all-or-none’ format, meaning that all of the following should be performed: risk assessment, inspect skin daily, moisture management, optimal nutrition, repositioning and use of pressure-relieving surfaces [10]. However, Table 4 shows that patients with pressure ulcers have different risk factors. As expected, our data also show that patients with ulcers received pressure-reducing mattresses and planned repositioning more often that those without ulcers. Individual plans must be tailored to the patient's specific risk factors. Resources for care also need to be considered. Patient and family participation is pivotal in the development and implementation of a prevention plan, yet their role has not been addressed. Studies are needed that document positive outcomes, regardless of whether individual or bundles of interventions are utilized.

The issue of when to initiate prevention remains. Vanderwee et al. [23] found no significant difference in the prevalence of pressure ulcers when prevention was initiated by a Braden score <17 or when a Grade 1 pressure ulcer appeared. Further research is needed to determine what the trigger should be for prevention.

It is important to realize that while the bedside care of turning and positioning patients is primarily a nursing responsibility, pressure ulcer prevention extends beyond nursing and includes the multidisciplinary team. Each profession has a responsibility, e.g. dieticians for assessing nutritional need, physiotherapists for complex mobility issues, physicians for medical issues, etc. [24].

Methodological strengths and limitations

The prevalence methodology used in this study is widely used in Europe and similar to that used in pressure ulcer prevalence studies across the globe. The data collectors were educated in the methodology, and the inter-rater reliability of the pressure ulcer grading was excellent prior to the study. Data were based on the examination of the patient by two registered nurses, which strengthen the validity and reliability of the observations. However, this prevalence study methodology does not provide a way to determine the incidence of pressure ulcers. Thus, effects of prevention cannot be fully known.

The data on prevention on admission were based on retrospective audit of the electronic health records, and it is possible that more prevention was provided than documented. Data from records are limited by the fact that they are self-reports; however, they are also legal documents and are expected to accurately reflect care provided.


Pressure ulcer prevalence did not decrease, despite a comprehensive quality improvement program. Those who developed ulcers were older, at risk of ulcers, incontinent and had longer length of stay. Increased use of pressure-reducing mattresses during the 4-year period reflected the adoption of evidence-based practice [21]. Risk factors for those with more severe ulcers were increased moisture, decreased activity, limited mobility and problems with friction. Major reorganization and implementation of an electronic health record may have negatively influenced the quality improvement program; further research on this is needed.

Data show that special attention is needed to provide prevention to older patients with acute admission. Thus, skin and risk assessment, as well as prevention, should start early in the hospital stay. Increasing the data collection to identify those patients with community-acquired versus hospital-acquired ulcers will strengthen pressure ulcers as an accurate marker of quality of care for hospitalized patients. If possible, data should be reported on ward level for comparison over time.


This study was supported by a grant from Uppsala University Hospital.


We thank all the nurses greatly for participating in the prevalence surveys. We also thank Lisa Jonsson and Ulla-Britt Söderström for auditing the electronic health record and the administrative system, and Åsa Boström for her participation in the design of the audit and in the data collection.


View Abstract