Pressure Ulcers: Using What We Know to Improve Quality of Care

In this issue of the Journal of the American Geriatrics Society, Berlowitz et al. provide incidence data for stage 2 and greater pressure ulcers among residents in more than one hundred nursing homes affiliated with a single, for-profit nursing home chain with facilities in nine southeastern states.1, 2 They have used the Minimum Date Set (MDS) from 1997–1998 (the derivation cohort) to develop a risk-adjustment model for pressure ulcer development and data from 1998 alone (the validation cohort) to evaluate the risk-adjustment model's ability to predict pressure ulcers and to profile nursing home performance related to the development of these lesions. Data demonstrated that 902 of 14,607 residents (6.2%) from the derivation cohort developed a stage 2 or greater pressure ulcer, whereas 608 of 13,457 residents (4.5%) in the validation cohort developed an ulcer. A shorter length of total follow-up can explain the apparent lower cumulative incidence rate in the validation cohort. Subjects in the derivation sample had an average of 2.7 observations per nursing home resident. Because these observations occurred approximately every 3 months, the average length of total follow-up for the derivation cohort was about 8 months. Thus, the 6-month incidence would have been approximately 4.8% in the derivation cohort. The validation cohort had 2.1 observations per resident, so the cumulative pressure ulcer incidence of 4.5% was noted over an average total follow-up of about 6 months and is very similar to that observed in the derivation cohort. The risk-adjustment model developed by Berlowitz et al. is based on pressure ulcer development after an approximately 3-month period of follow-up.1 Each subject-specific MDS assessment was used as a new baseline assessment if a follow-up assessment was available for study subjects. In other words, subjects who remained in the nursing home for 1 year could have had four follow-up MDS assessments, corresponding to 3, 6, 9, and 12 months from the admission assessment. Thus, the 3-month assessment served as the follow-up assessment for the admission assessment and also as the baseline assessment for the 6-month assessment. Similarly, the 6-month assessment served as the baseline for the 9-month assessment, etc. The 3-month cumulative incidence rates observed in the derivation and validation samples analyzing the data in this way were 2.3% and 2.1%, respectively. Previous reports have documented that there is a tremendous degree of variability in pressure ulcer incidence among nursing homes, with 21-month incidence rates for stage 2 and greater ulcers varying from 0% to 31%.3 More than half of persons who develop pressure ulcers experience unrelieved pain, and many suffer complications such as cellulitis, bacteremia, or osteomyelitis and require hospitalization. Available data suggest 26% of stage 2 ulcers and 40% of stage 3 or 4 ulcers remain unhealed even after 6 months of treatment.4 In light of their frequency and poor outcomes, it is not surprising that the occurrence of pressure ulcers have been identified as a marker for quality of care in nursing homes, and this condition has become one of the top-10 most frequently cited deficiencies for long-term care facilities nationally.5 Readers of the Journal should be reminded, however, that most prevalent pressure ulcers occur during acute care hospitalizations,4 and there remains a great need for reducing the risk of pressure ulcers in hospitals as well as in nursing homes. Data-driven quality improvement efforts have resulted in reduction of pressure ulcer incidence in some nursing homes,5 but there has been little evidence that comparative data related to pressure ulcers available from the MDS have been used by nursing homes to foster quality improvement efforts related to pressure ulcer prevention. Failure to use data from the MDS may reflect uncertainty about the quality of reports of pressure ulcer incidence derived from the MDS and questions about the legitimacy of comparing rates among nursing homes because of differences in the residents within nursing homes, that is, the case mix within specific facilities. Some nursing homes may have a large proportion of residents with multiple risk factors for pressure ulcers, whereas other homes may have very few residents who are at risk. Therefore, differences in the pressure ulcer incidence rate could reflect subject-specific differences rather than differences in the care processes within the homes. Berlowitz et al. noted that one nursing home in their validation sample had the third-highest unadjusted rate of pressure ulcer development, but this nursing home had the most severely impaired residents. After risk-adjustment, the nursing home ranked 11th out of 108 nursing homes in terms of pressure ulcer incidence and was no longer defined as an outlier.2 The availability of a risk-adjustment model, such as the one developed and validated by Berlowitz et al. thus may enhance the acceptance and use of data derived from MDS to guide and encourage quality improvement efforts within nursing homes. However, many other barriers to data-driven quality improvement efforts related to pressure ulcers, and other important clinical issues such as fall prevention, end-of-life care, medication use, and nutrition, must be overcome to facilitate improving the quality of care for nursing home residents. Although current nursing home regulations mandate a quality assurance structure, the infrastructure for true data-driven quality improvement efforts has not been developed. Key elements of such an infrastructure include the need not only for comparative risk-adjusted outcome data, but also for comparative data related to care processes such as the frequency of using a pressure-reducing mattress or documenting a repositioning schedule for nursing home residents at risk for pressure ulcers. The availability of comparative data alone will not lead to improved quality of care. Nursing home personnel need to be able to interpret and use such data to identify root causes for variations in facility outcomes and care processes. They then need to be able to test changes in care processes within their facilities that will improve outcomes for their residents. Unfortunately, although the nursing home industry may be one of the most regulated industries in the nation, the current approach to regulation and certification of nursing homes has tended to create an environment of blame and litigation, rather than a culture of using data available from the MDS or other sources to identify opportunities for improvement, to make changes in care processes, and to document the effectiveness of such changes. More than 4 years ago, the Health Care Financing Administration committed to using its peer review organization (PRO) contractors to systematically promote improved performance on quality measures identified for common clinical conditions among fee-for-service Medicare beneficiaries cared for in both outpatient and inpatient settings.6 This national program emphasizes a voluntary, collaborative, and nonpunitive educational strategy. Available evidence suggests that such an educational strategy can result in improved processes and outcomes of care for Medicare beneficiaries.7, 8 Many of the PRO contractors are beginning to work with nursing homes to test whether such a collaborative approach can be used to foster quality improvement within nursing homes. We can hope that such collaborations will prove successful and will be expanded to assist in the development of a culture and an infrastructure that will support continuous quality improvement throughout the nation's nursing homes. The risk-adjustment model developed by Berlowitz et al. identifies resident-specific risk factors for pressure ulcers. Their model suggests that male gender, poor transfer performance, being bedfast, altered urinary continence (either incontinence or the presence of an indwelling urinary catheter), deteriorating cognitive status, new admissions or a change in health status prompting a nonroutine MDS assessment, decreased body mass index, a stage 1 pressure ulcer or a history of a resolved pressure ulcer, and edema all are significantly associated with the development of stage 2 or greater pressure ulcers.1, 2 Similar subject-specific risk factors all have been identified in other prospective cohort studies.4 Inclusion of such well-documented subject-specific factors in the risk adjustment model permitted Berlowitz et al. to stratify nursing home residents into 10 different categories of risk, with observed 3-month pressure ulcer incidence rates ranging from 0.2% in the lowest-risk category to 7.3% in the highest-risk category in the derivation cohort.1 Similar results were noted in the validation cohort.2 It has previously been reported that newly admitted hospitalized patients whose activity is limited to bed or chair can be stratified into four risk categories based on five pressure ulcer risk factors: nonblanchable erythema of intact skin; lymphopenia; inability to reposition without assistance; decreased body weight; and dry skin. The 3-week incidence of hospital-acquired pressure ulcers for patients with none, one, two, or three or more of these factors was 0%, 11.4%, 39.6%, and 67.9%, respectively.9 These data suggest that knowledge of subject-specific pressure ulcer risk factors can enable clinicians to target prevention efforts to those who need them most. Pressure ulcers remain an all-too-common problem among nursing home residents and hospitalized patients with activity limitation. Facility- and subject-specific data can be used to improve care processes and reduce the incidence of pressure ulcers. Healthcare professionals, hospitals, nursing homes, accreditation and certification bodies, and the government need to work together to encourage innovative ways of encouraging greater use of comparative risk-adjusted outcome and care process data to improve the quality of preventive care for older persons at risk for pressure ulcers and other common conditions.