Statistics from Altmetric.com
Commentary on: OpenUrl
Implications for practice and research
Pressure ulcers (PUs) are common and often occur in older persons and those with multiple comorbidities.
Patients with hospital-acquired PUs (HAPUs) have longer hospital stays and higher rates of morbidity and mortality; therefore, prevention and early detection are important priorities for nursing care.
Interpretation of epidemiological studies of PUs is difficult due to variances in data collection methods and PU classification employed, making cross comparisons between clinical sites challenging.
Development of HAPUs is considered an adverse event.1 Thus, the incidence of HAPU is commonly used as a quality marker for healthcare facilities. One of the challenges in interpreting data lies in the lack of homogeneity in how figures are collated, for example, variability in sample sizes, data collection and PU grading systems employed. For institutions, the ability to benchmark against others is therefore limited, compounding the challenges in determining success in prevention strategies. It is with this background that the current study was undertaken, the aim of which was to determine PU incidence across the USA.
Data were collected, retrospectively, for hospital discharges for a 2-year period (January 2006 to December 2007). The records of 51 842 Medicare fee-for-services in patient discharges, across the 50 states, were analysed for documented presence of PU on admission (prevalence) and the development of new PUs during hospital stay (incidence). Demographic data and the presence of comorbidities associated with PU risk were also recorded. Diagnosis of a PU was based on the nurse and physician documentation within the medical record. Inter-rater reliability of data collection was established at 90%, between the data abstractors and the principal investigator, before commencement of the study. Data were analysed using descriptive and bivariate analysis, with the focus to compare the differences in characteristics and outcomes for those who did and did not develop a HAPU.
The nationwide HAPU incidence was 4.5% (n=2313/51 842); however, there was a variance between states (3.2%, SE=1.2%). PU prevalence on admission was 5.8% (n=2999/51 842) and of these 16.7% (n=502/2999) developed a new PU during hospital stay. The majority of PUs occurred in individuals aged 75–85 years, who were non-white and had higher rates of comorbidities (p≤0.05). There was a statistically significant association between the development of a HAPU and inpatient mortality, and mortality within 30 days postdischarge (p≤0.001; OR 2.81, 95% CI 2.44 to 3.23; p≤0.001; OR 1.69, 95% CI 1.67 to 1.77, respectively). Furthermore, those with a HAPU had a statistically significant longer length of stay compared to those without a HAPU (11.6±10.1 days vs 4.9±5.2 days, p<0.001).
Understanding the epidemiology of PUs is important in order to drive forward targeted interventions that will help reduce prevalence and incidence. It is recognised that benchmarking using cross comparisons between different facilities is inherently challenging, often due to the poor consistency in data collection methods. Lyder and colleagues have addressed some of the methodological weaknesses of previous incidence studies by using medical record data for a large cohort of US Medicare patients. Incidence, associated risk factors and clinical outcomes are clearly identified from the medical record; however, since recorded incidence of PUs often differs from the incidence identified on actual patient examination, HAPU may still have been under-reported. Indeed, Lyder and colleagues acknowledge that the grades of PUs could not be identified due to poor documentation of the clinical staff within the hospital records.
A further point of consideration is the use of retrospective data collection to predict risk factors. A prospective design is considered more appropriate, particularly as consistency in data collection is essential.2 Finally, it is important to report the incidence of PUs within the context of risk status of the population of interest, in tandem with the availability and use of preventative strategies.3 It is only when all this information is made available can the significance of the problem be truly interpreted.
Competing interests None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.