| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
a Sinclair School of Nursing, University of Missouri–Columbia
b School of Medicine, University of Missouri–Columbia
c College of Arts and Science, University of Missouri–Columbia
d School of Nursing, University of Wisconsin–Eau Claire
e College of Nursing, University of Iowa, Iowa City
Correspondence: Marilyn J. Rantz, PhD, RN, FAAN, S422 Nursing School Building, University of Missouri, Columbia, MO 65211. E-mail: RantzM{at}health.missouri.edu.
Decision Editor: Eleanor S. McConnell, RN, PhD
 |
Abstract |
|---|
 TOP Abstract Mandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
Key Words: MDS data • Nursing homes • Outcomes
Considerable effort has been devoted to improving quality of care for nursing home residents. Elaborate state and federal systems have been developed to protect the public and assure at least minimal standards of quality (Zimmerman et al. 1995
). Since 1990, federal mandates have directed nursing homes nationwide to conduct quality improvement activities. While quality improvement activities are commonly believed to affect resident outcomes, limited research has supported this premise (Harrington and Carrillo 1999; Sainfort, Ramsay, and Monato 1995). Nonetheless, feedback reports comparing outcomes of one organization to another have become commonplace in quality improvement. To date, they have received limited evaluation (Anderson et al. 1998). To test the benefit of feedback in a quality improvement model, we designed and conducted a randomized controlled trial to determine if (a) simply providing nursing facilities with comparative quality performance information and education about quality improvement would improve clinical practices and subsequently improve resident outcomes, or (b) a stronger intervention, such as expert clinical consultation with nursing facility staff, is needed to improve outcomes.
|
Mandate for Quality Improvement in Nursing Homes |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
; R. L. Kane 1995). Despite those efforts initiated in the 1970s and 1980s, recent media attention and presidential initiatives to address nursing home problems suggest that quality problems persist (Pear 2000).
In 1983, the Institute of Medicine (IOM) began a 2-year study of nursing home quality. The report, Improving the Quality of Care in Nursing Homes (Committee on Nursing Home Regulation, Institute of Medicine 1986), resulted in Congress mandating, in the Omnibus Reconciliation Act of 1987 (OBRA 87), several provisions intended to improve nursing home care. These provisions included developing The Minimum Data Set for Resident Assessment and Care Screening (MDS), mandating routine use of the MDS for all nursing home residents, and requiring that a quality assurance and assessment process be used in all nursing homes to improve the quality of care (McElroy and Herbelin 1989). This standardized resident assessment process was envisioned to improve resident care through the formulation of a resident-specific care plan; to provide nursing home management with resident-level data for monitoring case mix, staffing, and quality of care performance; and to provide regulators with data for case mix, sampling for survey processes, monitoring resident outcomes, and utilization review for Medicare or Medicaid eligibility.
The IOM concluded that "regulation is necessary but not sufficient for high-quality care" (Committee on Nursing Home Regulation, Institute of Medicine 1986, p. 24). The committee further resolved that nursing home staff members need to be well trained, well supervised, and highly motivated to deliver quality services to residents. The committee pointed out that "process measures (of quality) should not be ignored" (p. 55) and that resident outcomes are adversely affected when care delivery processes are overlooked or executed inadequately by staff.
Ten years later, another IOM committee reinforced the importance of staffing in nursing homes and concluded that the "quality of care provided by some nursing facilities still leaves much to be desired" (Committee on the Adequacy of Nurse Staffing in Hospitals and Nursing Homes, Institute of Medicine 1996, p. 140). The committee called for continued research that "could improve both the processes and the outcomes of care" (p. 140).
|
Information Feedback to Improve Quality |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
; Gehlbach et al. 1984; Keller, Chapin, and Soule 1990). Most of the studies, however, have involved changing physician practice patterns or quality improvement strategies in hospitals (Balas et al. 1996; Berwick and Coltin 1986; Horowitz et al. 1996; Myers and Gleicher 1991; Parrino 1989). Other studies show that such feedback can change behavior, improving the quality of care delivered (Frame, Kavolich, and Llewellyn 1984; Hamley et al. 1981). Two comprehensive reviews of feedback of auditing results of practice patterns conclude that performance of health care providers can be generally affected to a small or moderate degree; however, complementary interventions to enhance the effectiveness of audit and feedback have yet to be adequately tested (Thomson O'Brien et al. 2000a, Thomson O'Brien et al. 2000b). Although comprehensive applications in nursing homes have not been conducted, initial success of comparative drug utilization information feedback to nursing homes supports the premise that information feedback using comparative reports may also help change practice behavior in nursing homes (Zimmerman, Collins, Lipowski, and Sainfort 1994).
Systematic evaluations of individual long-term care organizations suggest that feedback of quality measurement information to staff resulted in better care processes and outcomes (Dennik-Champion, Mareno, and Carlson 1994; Miller and Rantz 1989, Miller and Rantz 1991, Miller and Rantz 1995; Roberts, LeSage, and Radtke-Ellor 1987). One randomized trial provided feedback on quality measurement information to staff in 60 Canadian nursing homes, resulting in performance changes and improvement in quality indicator conditions of hazardous mobility and constipation (Mohide et al. 1988).
|
Clinical Consultation to Improve Quality in Nursing Homes |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
). Educational programming and resident-centered consultation were found to reduce the use of physical restraints in nursing homes without subsequent increases in staffing or resident injury (Ejaz, Folmar, Kaufmann, Rose, and Goldman 1994; Evans et al. 1997; Neufeld et al. 1995, Neufeld et al. 1999; Strumpf, Evans, Wagner, and Patterson 1992; Werner, Koroknay, Braun, and Cohen-Mansfield 1994). Similarly, consultation was shown to reduce falls in nursing homes (Ray et al. 1997). However, some of these studies and others have demonstrated that follow-through by the nursing home staff to the recommendations made during consultation and sustained use of the recommended interventions over time may be difficult to achieve (Ouslander et al. 1995; Schnelle, Newman, White, et al. 1993; Schnelle, Ouslander, Osterweil, and Blumenthal 1993). |
Quality Indicators and the MDS |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
; Shaw and Whelan 1989; Zinn, Aaronson, and Rosko 1993). Accordingly, the Health Care Financing Administration (HCFA) has a basic strategy to develop a system of quality indicators (QIs) across the full range of services paid for by the Medicare and Medicaid programs (Gagel 1995; Jencks 1995).
Mandated by OBRA 87, MDS data are routinely obtained for all nursing home residents nationwide upon admission to all nursing facilities participating in Medicaid and/or Medicare, at times of significant change in condition of the resident, quarterly, and annually. Several authors have recommended using MDS data to facilitate quality improvement in nursing facilities (Schnelle 1997; Schnelle, Ouslander, Osterweil, et al. 1993; Spuck 1992). Data from the MDS are resident-level assessment information that can be aggregated for comparison across units within a nursing home or across nursing homes. As part of the HCFA Multistate Nursing Home Case-Mix and Quality Demonstration Project (NHCMQ), Zimmerman and colleagues at the University of Wisconsin–Madison have developed a series of MDS-based QIs through a systematic process involving extensive interdisciplinary clinical input, empirical testing, and field testing (Ryther, Zimmerman, and Kelly-Powell 1994, Ryther, Zimmerman, and Kelly-Powell 1995; Zimmerman et al. 1995). The most current version includes 30 MDS QIs, measuring such areas as accidents, incontinence, physical function, skin care, cognitive functioning, and behavior (Karon and Zimmerman 1996). Nationally, 24 of the 30 were implemented by HCFA nationwide in 1999 for use in the nursing home survey and certification process. Initial field tests and MDS QI validation studies indicate that they provide valuable information about specific residents, specific nursing homes, and nursing facilities in aggregate (Gagel 1995; Karon and Zimmerman 1996, Karon and Zimmerman 1997; Rantz et al. 1996; Ryther et al. 1994, Ryther et al. 1995; Zimmerman et al. 1995).
Missouri, the state in which this study was conducted, has been collecting MDS data from nursing facilities since the early 1990s. Working cooperatively with the state's major research university, the state survey and certification agency began analyzing MDS QIs with the intent of providing useful facility-level reports, based on MDS data, that would assist facilities to improve quality of care (Rantz et al. 1996; Rantz, Popejoy, Mehr, et al. 1997). Plans for comparative MDS QI feedback reports for nursing home providers began several years before the national plans for MDS QI reports that became available to facilities in March 1999.
|
Methods |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
Feedback Report
Quality performance information was derived from MDS resident assessment data. MDS QIs were calculated using the methods developed in the NHCMQ (CHSRA 1995). A key assumption is that MDS QIs can be used effectively by facility staff to improve resident care, if the MDS QI report is easy to interpret and appropriate clinical consultation and support are provided (Rantz, Petroski, Madsen, Scott, et al. 1997; Rantz et al. 1999, Rantz et al. 2000). The research team designed and field tested such a report format for the state and this study—the Show-Me MDS Quality Indicator Report (Show-Me QI report)." Special features of the Show-Me QI report include five quarterly longitudinal comparisons of MDS QIs in both table and graphic illustration for each nursing home (see Fig. 1). To prepare the report, it was necessary to conduct expert panels to set thresholds to be used in the illustrations to help quality improvement teams target areas of care delivery for further investigation (Rantz, Petroski, Madsen, Scott, et al. 1997; Rantz et al. 2000). Trend lines over time are easy to see and interpret. Comparisons to expert set thresholds are more likely to point to potential clinical problems that can be masked by simple comparisons to statewide averages. A statewide average may be the result of a poor clinical practice that is accepted as the norm in the majority of nursing homes in the state. Some facilities may falsely interpret that they have good quality because they are "average," when the average practice is really indicative of poor clinical care (Rantz, Petroski, Madsen, Scott, et al. 1997, Rantz, Petroski, Madsen, Mehr, et al. 2000).
|
). Table 1 is a list of the MDS QIs displayed in the Show-Me QI reports that were used in the intervention in this study.
|
In the two intervention groups there were 17 facilities (9 in Group 1 and 8 in Group 2) that did not attend their training workshop, so they were not able to receive their feedback reports and were excluded from all analyses. Recall that facilities at the time this study was conducted did not have access to MDS QI information without participating in the study. An additional 9 facilities were missing either baseline or follow-up data, so they could not be used in the analysis. After exclusion of these 26 facilities because of failure to attend the intervention workshop or missing data, the analysis is based on 87 facilities: 27 in Group 1, 28 in Group 2, and 32 in Group 3. The resulting groups were of sufficient size for adequate power in planned analyses for treatment effect. Although the sampling unit in this study is the nursing home, it is worth noting that the MDS QI scores of the 87 nursing homes are based on data from 6,381 residents at baseline and 7,385 residents at the one-year postintervention follow-up.
The 87 nursing homes in the analysis are similar to the nursing homes in the remainder of the state except that the proportion of larger nursing homes in the study is higher than the proportion of larger homes in the remainder of the state. It is possible that larger nursing homes began transmitting data sooner than smaller ones in the state and, therefore, were able to volunteer to participate in the study. Table 2 describes the characteristics of study homes by group assignment. Due to random assignment, nursing homes of varying size and ownership participating in the study were distributed relatively evenly among the three groups. Those excluded from analysis reflected the proportions of participating homes' size and ownership.
|
Interventions
An educational program, conducted in regional workshops, was designed for staff from facilities assigned to Groups 1 and 2. The purpose was to teach staff about quality improvement and how to use their Show-Me QI report that they would receive quarterly throughout the study. Content of the workshops included information about MDS QIs, how to initiate quality improvement teams, how to interpret their Show-Me QI report, how to compare themselves to other facilities in the state, and how to implement quality improvement projects targeted at improving resident outcomes measured by the MDS QIs. The staff was encouraged to initiate quality improvement efforts specific to their facility. During the workshops, staff members were given a QI manual prepared by the research team that outlined specific concurrent monitoring plans for each MDS QI to be used in evaluating resident care, as well as a comprehensive reference list of current clinical practice standards for each (Rantz and Popejoy 1998). During the workshop they received a copy of their facility's first quarterly comparative feedback Show-Me QI report. The report included a resident roster that listed residents who met one or more of the definitions of the MDS QIs and could potentially have the clinical problem(s) defined by the indicator. Subsequent quarterly reports were mailed to the administrator and director of nursing in each facility in Groups 1 and 2 who participated in the assigned workshops.
In addition to the educational program and quarterly comparative MDS QI feedback reports, staff from nursing homes assigned to Group 2 were offered access to telephone and/or on-site clinical consultation from a gerontological clinical nurse specialist (GCNS). Use of the GCNS was at the facility's discretion. The overall purpose of the consultation was to assist facilities to interpret their quarterly Show-Me QI report and enable them to make decisions about which clinical issues required further review. Discussions centered on issues related to MDS coding, resident assessment accuracy, and assistance in using the Resident Assessment Instrument (RAI) manual and other RAI reference materials. After coding and assessment issues had been addressed, the GCNS helped facilities identify the clinical problems that were resulting from potentially problematic care practices. In later consultations, assessment of resident problems using RAI resident assessment protocols (RAPs), use of clinical practice guidelines, documentation of care, and care planning were key issues that were discussed.
At the educational workshop, the GCNS offered the nursing homes in Group 2 consultation on site and/or by telephone. After the workshop, the GCNS called each of the 28 nursing homes in Group 2 to ask how they were doing with report interpretations and quality improvement efforts and offered to come for a site visit. Staff from 15 of the nursing homes had one or more on-site visits, as well as telephone calls. Staff from 11 of the nursing homes used telephone consultation only, including conference calls with multiple staff members. Only two homes were not interested in further telephone calls or a site visit. After each telephone call or site visit, homes were encouraged to call the GCNS with further questions. Due to the lengthy travel distances required, phone consultations between visits were encouraged. However, homes that were interested in making practice changes generally desired more site visits. Staff from 10 of the homes in Group 2 were quite receptive to GCNS offers and used on-site consultations to work with groups of their staff several times during the study. These groups typically included the director of nursing, quality assurance coordinator, nurse responsible for MDS completion, other licensed staff, and a few nursing assistants. Staff from the other Group 2 homes (n = 18) seemed interested, but had only one or no on-site consultation and only limited telephone consultation.
Group 3 facilities, the control group, received no information until the end of the study. At that time, they received the same educational program as Groups 1 and 2, the QI study manual, and began receiving their quarterly Show-Me QI report. Additionally, those who were assigned to Groups 1 and 2 who were "no shows" for the intervention workshops were invited to attend these sessions and receive materials and quarterly reports at the end of the study.
Analysis
Outcome measures for this study were selected from the MDS QIs that were included in the feedback report to participating facilities. Thirteen MDS QIs were selected as outcome measures because they are particularly sensitive to clinical intervention by nursing home staff and have sufficient variation in scores to detect changes, as described in a previous study (Rantz et al. 1996). At the facility level, the MDS QIs are calculated as the proportion of residents positive for a particular condition on a particular occasion.
Summary statistics were examined for the outcome measures that included means, standard deviations, and medians (50th percentile). Some of the MDS QI scores have highly skewed (asymmetric) distributions. In such cases the sample median is a more appropriate measure of central tendency than is the sample mean. In cases where the mean and median are strikingly different, the median is generally preferred.
The primary analysis employed logistic regression methods to perform the equivalent of a two-factor analysis of covariance for each MDS QI. The independent variables were Group (three levels) and Time (two levels, baseline and one year) and the interaction of Group and Time; the dependent variables were the MDS QI scores. Because MDS QI scores may be affected by resident case mix, an adjustment for case mix was included as a covariate in each analysis. The case mix variable is the facility average case mix index derived from Version 5.12 of the 44-group RUG-III algorithm using the hierarchical classification method and HCFA case mix index set B0 (Fries et al. 1994; Health Care Financing Administration 1998).
Each regression model included a term for the interaction of Group and Time. In the presence of significant (p 
.05) interaction, further analysis is required because the main effects of Group and Time are not directly interpretable. Statistically significant interaction suggests that the intervention and control groups behaved differently over time, which is what one would expect to see with an effective intervention. Significant interactions were followed by pre–post comparisons to determine which groups changed from their baseline values. Because repeated observations (pre- and postintervention) on the same facility are not independent, the method of generalized estimating equations was used to calculate standard errors.
The primary analysis assumed an intention-to-treat principle in that the analysis is based on the facilities as they were randomized to the three groups. The analysis does not incorporate any measure of the facilities' efforts to utilize the intervention resources beyond attending the training sessions. Particularly, some Group 2 homes made extensive use of the clinical consultant, but others did not draw on this resource. A secondary analysis was performed to examine a subset of Group 2 nursing homes that were intensively involved in the intervention. Using the same methods as in the primary analysis, this intensive intervention group was compared with the control group to detect changes in outcomes from baseline to one year (Group x Time effects). The intent of this secondary analysis was to examine if a more intense intervention might produce any impact on quality. Given the small sample size and the exploratory nature of this analysis, effects were considered to be "suggestive" when p .10. Significant (p .10) interactions were followed by pre–post comparisons to determine which of the groups changed from their baseline values. Line graphs of the group medians were constructed to better appreciate fluctuations over time and possible Group x Time interactions to better understand group quality performance differences. Field notes of all consultations, both on site and telephone, by the GCNS were content analyzed. The numbers of telephone and on-site consultations for each facility were tabulated; the clinical content discussed was categorized as well as the facility staff who participated in the consultation.
|
Results |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
|
Secondary Analysis
To further explore the potential for this type of intervention, a secondary analysis was performed in which a subset of the Group 2 nursing homes that were intensely involved with the intervention (n = 10; 35% of Group 2 nursing homes) were compared to Group 3 (control). These nursing homes utilized on-site and telephone clinical consultation from the GCNS more extensively, that is, more than twice on site and more than twice with telephone consultation. Demographics of ownership and bed size of this subset of Group 2 were reflective of Group 2 and the other study groups; six were 61–120 and four were 120+ bed-size; two were governmental, two were nonprofit, and six were for-profit nursing homes.
Table 3 displays the summary statistics of the workshop and intensive consultation group (n = 10). Regression results for the secondary analysis are presented in Table 4 . Using the p .10 criterion, the Group x Time interaction was significant in the analysis of the following MDS QIs: MDS QI 2 (falls), MDS QI 3 (behavioral symptoms), MDS QI 27 (little or no activity), MDS QI 29 (pressure ulcers), and MDS QI 29 low risk (pressure ulcers in low-risk residents). For each of these five outcomes, pairwise comparisons revealed that MDS QI scores declined (improved) from pre- to postintervention in the intensive consultation group and remained unchanged in the control group.
|
.10) results include a significant main effect for Time in the analysis of MDS QI 6 (9 or more medications) with nearly uniform increases from baseline in both groups. A similar result was seen in the primary analysis. The analysis for MDS QI 9 (incontinence without a toileting plan) resulted in a highly significant (p = .007) Group effect and a marginally significant (p = .08) Time effect. On this outcome the control and intensive intervention groups were not comparable on their baseline measurements. The baseline median MDS QI score for the control group was 63.3 versus 39.7 for the intensive intervention group. Due to the small sample size the interaction term is not significant (p = .13); however, summary statistics suggest that the intervention group improved from baseline (39.7 vs 23.0) and that the control group's scores were essentially constant (63.3 vs 62.5).
Finally, on MDS QI 26 (physical restraints), there was a significant (p = .06) Group effect. On this MDS QI the control and intervention groups were slightly different from each other at baseline and postintervention, but neither group showed significant changes from their baseline values.
To understand these differences in MDS QIs with suggestive Group x Time interaction results, the median scores for each quarter in the study were used to construct line graphs for Group 1 (workshop only and feedback reports; n = 27), Group 2 intensive consultation (workshop, feedback reports, and intensive on-site GCNS consultation; n = 10), Group 2 limited consultation (workshop, feedback reports, and limited on-site GCNS consultation; n = 18), and control homes (n = 32).
Line graphs in Fig. 2 and Fig. 3 visually reveal trends in improvement in the intensive consultation subgroup of Group 2 that sought the consultation of the GCNS for MDS QI 2 (falls) and MDS QI 29 (pressure ulcers).
|
|
In subsequent site or telephone consultation visits, MDS QI scores were discussed and resident care was explored in depth. When appropriate, specific interventions to try with residents were discussed. Facilities were encouraged to address clinical practice systematically, using the quality monitoring plans provided in the study manual.
Analysis of field notes revealed that staff from most of the 10 nursing homes decided to focus on resident falls and pressure ulcers as their first projects. The GCNS provided the latest clinical information about these topics. All homes were encouraged to use RAIs such as the Risk Assessment for Falls Scale II (RAFS II) (Maas 1991) and the Braden Scale for pressure ulcer risk assessment (Baharestani 1999; Bergstrom, Braden, Boynton, and Brunch 1995; Panel on the Prediction and Prevention of Pressure Ulcers in Adults 1992). Facilities were told how to order and were encouraged to use clinical practice guidelines prepared by the American Medical Directors Association on the topic of falls and urinary incontinence (Falls and Fall Risk Panel 1998; Urinary Incontinence Panel 1996) and the Agency for Health Care Policy and Research guidelines for pressure ulcers and incontinence (Panel on the Prediction and Prevention of Pressure Ulcers in Adults 1992; Urinary Incontinence Guidelines Panel 1992). Facilities were also encouraged to put in place documentation systems for those problems that would allow the clinical staff to identify readily patterns in falls and changes in clinical conditions increasing residents' risk for the development of pressure ulcers.
Analysis of field notes for the 18 nursing homes in Group 2 who decided not to use the consultation of the GCNS revealed a variety of reasons stated for the refusal of a site visit, but short staffing, staff turnover, or other pressing issues were frequently cited. Generally, staff would say they had received their Show-Me QI report, that they were taking care of things themselves, and that they really had no questions or need for the site visit.
|
Discussion |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
When the study was designed, participation was viewed as voluntary. However, inquiries about the consultation seemed to indicate that facilities assigned to Groups 1 and 3 were disappointed that they did not have access to consultation support of the GCNS. Therefore, it was surprising to us that more than half of the nursing homes assigned to Group 2 did not take full advantage of the free consultation offered during telephone follow-up. Reluctance to participate may have been a function of competing priorities for nursing home staff who are busy with moment-to-moment issues of care delivery. Stopping to evaluate clinical practices and design improvements may seem overwhelming or simply not a priority for some. Perhaps the number of homes accepting support would have been higher had the intervention been designed in such a way as to obtain agreement from participants that, if assigned to the clinical consultation group, they would agree to at least quarterly site visits by the GCNS. Our approach of telephone contact and offers for on-site support seemed to be strong enough to involve about a third of the nursing homes in Group 2 intensively. Something stronger is needed to encourage the remaining two thirds.
The significant improvement in MDS QI 27 (little or no activity) for both intervention Groups 1 and 2 is likely due to heightened awareness about accurately coding the MDS items used in this indicator. The importance of accurate coding of the MDS items was reinforced in the workshops and teaching materials for the intervention groups. Alternatively, it is possible that more activities were planned and carried out for residents, and that coding changes reflect the increase in activity.
Travel distance is an issue for on-site consultation. For consultation to be effective, travel distances must be reasonable so that the consulting staff can make appointments, travel to the nursing homes, and have adequate time for discussion with staff and on-site observation assistance with clinical problems. If at all possible, consulting staff located within regions of a state would be beneficial to a study such as this. Some appointments were frustrating because situations would occur that prevented the scheduled site visit at the last moment, after the consultant had traveled 2 or more hours to meet with staff. Finding consultation staff close to the area would reduce travel time and provide more options for scheduling site visits.
The changes in fall and pressure ulcers scores for the nursing homes that used intensive consultation is most likely related to several things. There are clear standards of practice on both of these issues. Both problems are sensitive to interventions at the resident level. For example, often simple discussions with the GCNS about different approaches enabled facilities to make changes in interventions on plans of care that reduced fall rates. The GCNS encouraged all nursing homes to use RAIs for falls and for pressure ulcer development. Facilities were encouraged to use clinical practice guidelines about fall and pressure ulcer prevention and treatment. Staff could grasp the clinical changes needed for better management of these clinical problems.
The increase in scores for all groups for MDS QI 6 (9 or more medications) has some possible clinical explanations. This MDS QI was discussed in facilities where it was high (indicating a problem). The increase over time may indicate an increase in resident acuity. It may also be a reflection of changes in practice guidelines that have occurred in the last 2 years that now recommend multidrug regimens for some conditions. For example, managing severe congestive heart failure often routinely now includes several medications (Bottorff 2001; Feldman 2001). Apparently, some broad practice changes or increases in acuity are affecting nursing homes across the state and were detected across all groups in the study.
The overall message of this study is clear: To effect improvement in resident outcomes, simply providing comparative performance feedback is not enough. There may be some exceptional nursing homes that can independently put a team together to examine and interpret comparative performance feedback reports such as the one used in this study or the one now available for every nursing home in the country from the federal MDS data system. There may be some exceptional places with teams that can plan quality improvement data collection, interpret results, and plan actions to improve their clinical practices. However, the results suggest that active clinical consultation support in the context of a ready environment is needed to help staff in most nursing homes conduct quality improvement activities that will effect improvement in resident outcomes. This is consistent with the findings of two recent comprehensive research reviews that found that performance can be affected to a small or moderate degree with feedback, whereas other interventions to increase the effectiveness of feedback have yet to be adequately tested (Thomson O'Brien et al. 2000a; Thomson O'Brien et al. 2000b). Similarly, Solberg, Brekke, Fazio, and colleagues (2000) concluded that multiple strategies are needed to successfully change health care provider practice patterns and influence them to incorporate clinical guidelines. Although these and other studies are not nursing home-based, it appears that enhancing feedback interventions with additional strategies may improve effectiveness and actually facilitate a positive change in clinical practice by health care providers.
It also appears that while we can generate a myriad of quality indicator information for teams to examine, they can only focus on one or two areas for improvement at a time. For those who are experienced in quality improvement, this will come as no surprise. There is a limit to the time and energy of staff that can be harnessed to implement and sustain change. Selecting a limited number of topics for further examination, collecting data about current care practices, interpreting the data collected, planning actions, educating staff about necessary changes, and following up to see that the changes in practice actually happened as planned, takes time. The problems of staff turnover and too few staff to participate in a quality improvement team also interfere with the number of areas that can be addressed, changed, and sustained as an accepted clinical practice.
Nursing homes participated in the study for a full year (four quarters of feedback reports) with the quarter before the study as baseline. We anticipated that staff from participating facilities would need the first quarter to select topics and begin their data collection to examine problem areas further. Action plans could be implemented in the next quarter. Because residents are assessed every quarter using the MDS instrument, we anticipated that changes in their outcomes could be detected in the reassessment processes during the third and fourth quarters. This timeline appears to have worked for those nursing homes that did embrace the quality improvement process. For the others in the study, perhaps a longer period of time is needed for staff to implement changes in practice and to detect changes in resident outcomes in quarterly assessments.
This view is supported by the fact that correcting inaccurate MDS assessments takes time. When staff in nursing homes first came together to examine their Show-Me QI report, much time was devoted to explaining the definitions of the MDS QIs and clarifying MDS coding. Because most nursing homes hire a nurse RAI coordinator to be responsible for timely and accurate completion of the MDS, turnover in this position is devastating to the accuracy of MDS coding. Many teams in the nursing homes participating in the study were confronted with staff turnover in this key position that compounded their efforts to correct coding errors so that they could more accurately evaluate MDS QI scores. Because residents are reassessed with a version of the MDS every quarter, it takes one quarter to see the changes of the reassessment in the next MDS QI report. It was not unusual for some teams to work for two or three quarters to correct coding errors, especially if there was staff turnover or the interdisciplinary care planning team was not functioning well. Some teams seemed to never get as far as we had hoped they would into the quality improvement process that focused on clinical care delivery changes. They seemed to be mired in the MDS assessment process and coding issues.
Most of the participating nursing facilities did not have well-developed quality improvement programs with systems to support implementing changes needed in care delivery. While staff seem to be able to alter care for short periods of time for some residents, there seems to be little systematic change that would broadly improve quality of care throughout the facility. It is difficult to convince staff to use continuous quality improvement principles. Most nursing homes do not use specific teams to address problems, nor do they report accomplishments. Many facilities continue to only use the quality assurance measures found in the OBRA regulations. In others, there is a crisis management approach, and problems are not addressed until they are so severe that they cannot be ignored. These findings may be related to the small numbers of professional staff who work in nursing homes. There may simply not be enough professional staff to have the critical mass needed to commit time and energy to quality improvement methods. Alternatively, it may be a function of leadership not embracing quality improvement as a way to improve care and services to residents. Nursing homes that did have continuous quality improvement systems in place were often part of larger health care systems that have ongoing support from a quality improvement expert. We noticed that large and complex facilities also are more likely to have well-organized quality improvement processes. Those homes are structured in such a way that there are multiple nurses responsible for the RAI process, as well as a quality manager on staff to support care delivery improvements.
While it would seem that simply educating staff about quality improvement and how to implement quality improvement programs should improve resident outcomes, it is probably much more complex. Findings from a recent quality improvement study in primary care clinics found no effect from quality improvement training, consultation, and networking to help the teams of staff develop and implement prevention services (Solberg, Kottke, Brekke, et al. 2000). Similarly, Goldberg and colleagues 1998 found in a randomized clinical trial that quality improvement teams were generally ineffective in improving guideline compliance and primary care clinical outcomes of hypertension and depression. Quality improvement strategies that actually affect resident outcomes in a positive way apparently involve more than education about quality improvement methods. It is likely that the context of care—with its myriad factors such as leadership, performance expectations, organizational culture, staff mix, and others—will impact the success (or failure) of quality improvement efforts. Clinical consultation with a GCNS does appear to be effective and capable of improving resident outcomes. Our results of improved resident outcomes in the nursing homes that sought additional support from the GCNS are strikingly similar to Ryden and colleagues 2000. In that study, weekly consultation of 10 hours of a GCNS did significantly improve outcomes of pressure ulcer development, incontinence, and aggressive behavior. Although more evaluation of effectiveness is clearly needed, it appears that ongoing GCNS consultation may be an important strategy to influence and improve clinical care and subsequently improve resident outcomes in nursing homes.
Limitations of this study are that we conducted it in a single state, so regional variations were not addressed. When we implemented the intervention we did not provide feedback reports to the nursing homes that failed to attend the required workshops. Therefore, those facilities were excluded from the analysis, so we could not use them in a complete intention to treat analysis that some may favor in a field study such as this. The subgroup analysis needs careful interpretation because the subgroup that used intensive consultation was a self-selected group. However, the results suggest that more intensive interventions may be effective to impact resident outcomes. Using feedback to support quality improvement is a complex intervention that may require substantially more effort. Further research is needed to explore the amount of consultation support as well as the organizational context of care that is necessary for a quality improvement feedback intervention to positively affect resident outcomes. Another potential limitation of this study is that our outcome measures rely on the accuracy of MDS data. Although reliabilities of MDS data are reported as good, particularly for those areas used as outcomes in this study (Casten, Lawton, Parmelee, and Kleban 1998; Hawes et al. 1995; Morris et al. 1997; Phillips, Chu, Morris, and Hawes 1993), one must always be concerned about data accuracy when using data collected for clinical research purposes.
Change in any organization is difficult. Nursing homes are no exception. Researchers working with nursing home staff to improve resident continence have repeatedly found that it is extremely difficult to maintain toileting programs, even those that are well designed and found to be effective (Ouslander et al. 1995; Schnelle, Newman, White et al. 1993; Schnelle, Ouslander, Osterweil et al. 1993; Specht, Bergquist, and Frantz 1995). Follow-through on recommendations for fall reduction has met with the same difficulty (Ray et al. 1997). Follow-through to implement and sustain change is necessary for quality improvement. In nursing homes where there is clear administrative support and expectation that care innovations be planned and effectively implemented, changes in practice occur (Levine, Marchello, and Totolos 1995; Rantz and Miller 1989; Specht et al. 1995; Specht and Lyons in press). If we are to implement true quality improvement programs in nursing homes throughout the country, there must be commitment from leadership within each facility that quality improvement is important and encouragement for staff to participate in quality improvement activities.
For states designing statewide strategies to encourage quality improvement in nursing homes using MDS QIs, the message is clear. Nursing facilities need more than feedback reports to improve resident outcomes. Clinical expertise is essential. Quality improvement and team development expertise is essential as is administrative support and commitment to excellence in clinical practice. Somehow, these ingredients must come together for clinical practice changes to be implemented and sustained that will improve resident outcomes. Clinical consultation provided by an advanced practice gerontological nurse appears to be an effective strategy that can be used, given administrative encouragement to use the consultation.
Practice Concepts
The Forum
Book Reviews
|
Acknowledgments |
|---|
We acknowledge the contribution of other University of Missouri–Columbia MDS and Nursing Home Quality Research Team: Brad Chancellor and Ken Lobenstein, ITS Research and Support Development Group; Steve Miller, Data Support staff. The members of the MU MDS and Nursing Home Quality Research Team gratefully acknowledge the support of the Missouri Division of Aging staff; they are truly committed to helping homes embrace quality improvement. Opinions are those of the authors and do not represent the Missouri Division of Aging or the Health Care Financing Administration.
Received for publication November 28, 2000. Accepted for publication April 17, 2001.
|
References |
|---|
|
TOPAbstractMandate for Quality Improvement...Information Feedback to Improve...Clinical Consultation to Improve...Quality Indicators and the...MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. O. Kjos, G. Botten, and T. I. Romoren Quality improvement in a publicly provided long-term care system: the case of Norway Int. J. Qual. Health Care, December 1, 2008; 20(6): 433 - 438. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. A. Miller and V. Mor Balancing Regulatory Controls and Incentives: Toward Smarter and More Transparent Oversight in Long-Term Care Journal of Health Politics Policy and Law, January 1, 2008; 33(2): 249 - 279. [Abstract] [PDF]
|
|
|
|
 |
|
 G. Arling, R. L. Kane, T. Lewis, and C. Mueller Future Development of Nursing Home Quality Indicators Gerontologist, April 1, 2005; 45(2): 147 - 156. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Lucas, T. Avi-Itzhak, J. P. Robinson, C. G. Morris, M. J. Koren, and S. C. Reinhard Continuous Quality Improvement as an Innovation: Which Nursing Facilities Adopt It? Gerontologist, February 1, 2005; 45(1): 68 - 77. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. R. Jones, R. Fink, G. Pepper, E. Hutt, C. P. Vojir, J. Scott, L. Clark, and K. Mellis Improving Nursing Home Staff Knowledge and Attitudes About Pain Gerontologist, August 1, 2004; 44(4): 469 - 478. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Rantz, L. Hicks, V. Grando, G. F. Petroski, R. W. Madsen, D. R. Mehr, V. Conn, M. Zwygart-Staffacher, J. Scott, M. Flesner, et al. Nursing Home Quality, Cost, Staffing, and Staff Mix Gerontologist, February 1, 2004; 44(1): 24 - 38. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Rantz, L. Hicks, G. F. Petroski, R. W. Madsen, D. R. Mehr, V. Conn, M. Zwygart-Staffacher, and M. Maas Stability and Sensitivity of Nursing Home Quality Indicators J. Gerontol. A Biol. Sci. Med. Sci., January 1, 2004; 59(1): M79 - 82. [Full Text] [PDF]
|
|
|
|
|
|
T. J. Lowe, J. A. Lucas, N. G. Castle, J. P. Robinson, and S. Crystal Consumer Satisfaction in Long-Term Care: State Initiatives in Nursing Homes and Assisted Living Facilities Gerontologist, December 1, 2003; 43(6): 883 - 896. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Rantz, A. Vogelsmeier, P. Manion, D. Minner, B. Markway, V. Conn, M. A. Aud, and D. R. Mehr Statewide Strategy to Improve Quality of Care in Nursing Facilities Gerontologist, April 1, 2003; 43(2): 248 - 258. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
