Directors of nursing (DON) are central to quality of care in nursing homes (NH) because of their role in coordinating and overseeing nursing care. Research is needed to test the association between DON characteristics and quality using large, representative samples of NHs and global measures of quality. One such measure is the quality measure (QM) rating from the Centers for Medicare & Medicaid Services’ Five-Star Quality Rating, which aggregates 10 individual QMs into a single rating.
This study examined whether DON current job tenure or past experience (1) differed across levels of the QM rating, (2) was associated with QM ratings, and (3) was associated with any of the individual 10 QM scores that comprise QM ratings.
Data for a nationally representative sample of 1,174 NHs were obtained from the 2004 National Nursing Home Survey, publicly-reported QMs, and an Area Resource File. Wald tests were used to test differences in mean DON current job tenure and past experience across levels of the QM rating. Multinomial logistic and Poisson regression analyses were used to examine the association between DON current job tenure and past experience and QM ratings and QM scores, respectively, controlling for selected market and organizational characteristics.
NHs with longer DON current job tenure tended to have higher QM ratings. Longer DON current job tenure was associated with higher QM ratings and lower QM scores for several individual QMs, suggesting higher quality. DON past experience did not differ across levels of the QM rating and was neither associated with QM ratings or QM scores.
This study highlights the need for owners and administrators to support DONs as they either transition into the role of the DON for the first time or learn to effectively fulfill their role in a new NH.
Nursing home top management teams, which includes the nursing home (NH) administrator, director of nursing (DON), and medical director, can have an important role in influencing quality in NHs. That is, top managers are responsible for overall oversight of NH care delivery (
Two DON characteristics that existing exploratory studies using small, non-representative samples repeatedly found to be associated with quality of care were current job tenure, which refers to the length of time in the role of DON at a specific NH, and career experience, which refers to the length of time in the role of DON at any NH. Specifically, the studies documented that longer DON current job tenure was associated with lower rates of restraint use (
Longer DON career experience and current job tenure may be linked to higher quality of care because of the “learning curve” when transitioning into the DON role, in general, and when transitioning into the DON role at each unique NH, in particular (
Additional research that builds on existing studies by using a large, representative sample of NHs and utilizes an expanded number of measures of quality of care is needed to inform future policy initiatives and to support future intervention research. Using a large, representative sample of NHs will ensure that the study is adequately powered to detect small effects and will allow conclusions to be drawn about the extent to which the documented linkages between DONs and quality of care would likely be true in the larger population of NHs. By utilizing an expanded number of measures of quality of care, future research would generate knowledge about the types of outcomes that may be influenced by DONs. The Centers for Medicare & Medicaid Services’ (CMS) recently released the new Five-Star Quality Rating System (Five-Star), which assigns NHs ratings of between one star and five stars, with five stars being the highest for the following four categories or domains: staffing levels, certification survey findings, quality measures (QM), and overall quality (
Data for this study were obtained from three sources: 2004 National Nursing Home Survey (NNHS), CMS’ QMs for NHs, and an Area Resource File (
The 2004 NNHS is a nationally representative survey of 1,174 NHs (response rate =81%) administered by the National Center for Health Statistics between August 2004 and January 2005 (
Three quarter averages, corresponding to the data collection period for the 2004 NNHS, of CMS’ publicly-reported QMs were obtained for this study. Specifically, QM data for the third and fourth quarter of 2004 and first quarter of 2005 were used for this study. The QMs are a set of 19 measures derived from resident assessment data that address a range of functioning and health status areas, such as restraint use and mobility decline, reflecting the multidimensional nature of care quality in NHs. Higher values for QMs indicate that a larger percentage of residents are experiencing potentially poor care practices or outcomes (
The ARF is a compilation of data aggregated to the county level from various sources, including the United States Census of Population and Housing, Bureau of Labor Statistics, and CMS. The ARF contains data pertaining to NH market characteristics, including the number of NHs per county and county-level unemployment rates.
Data on DON characteristics were obtained through a series of questions about the DON. Data on two such characteristics--current job tenure and past experience—are included in this study. Current job tenure was the number of years that a DON was employed in that role at the respective NH. Past experience was calculated as the DON’s overall experience, in years, in that role at any NH, less current job tenure.
Quality measure ratings were calculated for all United States NHs certified to participate in Medicare or Medicaid using the specifications described in the Design for Nursing Home Compare Five-Star Quality Rating System: Technical Users’ Guide (
To calculate QM ratings, NHs were first assigned points based on how they compared to other NHs in the country for non-ADL QMs and other NHs in their state for ADL QMs (see
Each NH was then assigned a QM rating based on how its total score compared to that for other NHs in the country. Specifically, NHs were assigned QM ratings in the following manner: less than the 20th percentile, one star; equal to or greater than the 20th percentile, but lower than the 43.33rd percentile, two stars; greater than or equal to the 43.33rd percentile, but lower than the 66.67th percentile, three stars; greater than or equal to 66.67th percentile, but lower than the 90th percentile, four stars; and greater than or equal to 90th percentile, five stars. No ratings were assigned to NHs with fewer than four long-stay and fewer than two short-stay QMs.
Three quarter averages of QM scores were obtained for the subset of10 QMs that were used to calculate the QM ratings. The QM scores could range from 0 to 100, where 0 indicates that none and 100 indicates that all of the residents experienced the potentially poor care processes or outcomes. Therefore, a QM score at or near to zero is interpreted as suggesting higher quality, whereas a higher QM score indicates lower quality. Three-quarter averages were censored if they were calculated for an average of fewer than 30 residents for long-stay measures or fewer than 20 residents for short-stay measures.
Data on selected organizational characteristics that research has suggested to be associated with quality of care—ownership, chain membership, facility size, occupancy, and Medicaid census—were included in this study (
Two market characteristics that previous research has suggested to be associated with quality of care—market competition and unemployment rates—were included in this study (
To create the analytic file, QM ratings for all U.S. NHs, county-level market characteristics, DON characteristics, and organizational characteristics that could be calculated using publicly available 2004 NNHS data were provided to the Research Data Center (RDC) at the National Center for Health Statistics. The RDC staff then merged that data with restricted 2004 NNHS data and subsequently dropped identifiers and data for NHs that did not participate in the 2004 NNHS. Because this study involved no identifiable patient data, it was determined to be exempt from review by the University of Wisconsin-Madison Institutional Review Board.
This study included 1,012 of the 1,174 NHs surveyed for the 2004 NNHS, approximately 86% of the sample. Forty-two NHs were omitted because they could not be linked with the file containing QM ratings. Another 120 NHs were omitted because they had missing data for study variables. All analysis was conducted using procedures in STATA developed for analysis of complex survey data (
To examine whether DON current job tenure and career experience differed across levels of the QM rating, mean current job tenure and career experience were calculated for each level of the QM rating. Next, pairwise comparisons were tested to determine if mean current job tenure or past experience differed across levels of the QM rating. The
Multinomial logistic regression analysis was used to examine the association between DON current job tenure and past experience and QM ratings, controlling for market competition, unemployment rate, ownership, chain membership, organizational size, occupancy, and Medicaid census. Relative risk ratios and linearized standard errors were calculated for each predictor for NHs with two stars, three stars, four stars, and five stars, compared to NHs with one star. For categorical predictors, for NHs for which the predictor was coded one, relative risk ratios greater than (less than) 1.0 indicated an increased (decreased) relative risk of a NH having a given numbers of stars compared to the risk of having one star. For continuous predictors, for every unit change in the predictor, relative risk ratios greater than (less than) 1.0 indicated an increased (decreased) relative risk of having a given number of stars compared to the risk of having one star.
Poisson regression analysis was used to examine the association between DON current job tenure and career experience and scores for each of the 10 QMs that comprise the QM rating, controlling for market competition, unemployment rate, ownership, chain membership, organizational size, occupancy, and Medicaid census. A Poisson distribution, not a negative binomial distribution, was used because the likelihood ratio test of the over-dispersion parameter alpha was not significantly different than zero. In general, ordinary Poisson regression was used, although the zero-inflated model was used to model pain and short-stay pressure ulcers because the Vuong test indicated that the zero-inflated model was superior for those QMs (
Descriptive statistics for all study variables are reported in
Mean DON current job tenure and past experience for each level of the QM rating are reported in
Results from the multinomial logistic regression analysis of the association between DON current job tenure and past experience and QM ratings are reported in
The results of the Poisson regression analyses examining the association between DON current job tenure and past experience and the 10 QM scores that comprise the QM rating, controlling for selected market and organizational characteristics, are reported in
This study adds to the body of research linking DONs to quality of care in NHs by using a nationally representative sample of NHs to demonstrate that longer DON current job tenure, but not past career experience, was associated with higher quality of care. The finding that longer DON current job tenure was associated with higher quality of care was consistent with the findings from other existing, exploratory research (
The association between longer DON current job tenure and higher quality of care may also be explained by DONs being more likely to stay in positions at higher quality NHs. NHs with high DON turnover may experience a “quality meltdown,” wherein persistent DON turnover results in lower quality, which may, in turn, make it difficult to retain well-qualified DONs (
The associations between longer DON current job tenure and higher quality of care—that is, higher QM ratings and lower scores for several individual QMs—were statistically significant, but were of relatively low magnitude. For example, as described previously, for every year increase in DON current job tenure, NHs were 5% more likely to have four stars and 10% more likely to have five stars than one star. If, instead, the unit of measurement had been five year increments, the magnitude of the association would have been more striking: NHs would have been 30% more likely to have four stars and 62% more likely to have five stars than one star. It is noteworthy, however, that the mean DON current job tenure in this sample was only 3.3 years, meaning that many of these NHs may have DONs with too short of current job tenure to experience the potential benefit of longer tenure on quality of care.
Another explanation for the small magnitude of the association between DON current job tenure and quality of care is that accounting for DON current job tenure (and past experience) alone is insufficient to explain differences in QM ratings and the individual 10 QM scores. For example, according to Unruh and Wan’s Systems Framework for Evaluating Nursing Care Quality (2004), NH quality may involve complex associations with contextual, structural, and process components of NH care. This study focused on the association between specific DON characteristics and quality of care, controlling for selected contextual and organizational components. Clearly, opportunities exist for expanding this research to examine additional management characteristics, including management practices (
Another possible explanation for the small magnitude of the association between DON current job tenure and quality of care is that, although DON current job tenure and past experience may have an important influence on quality, the measures used in this study were not sensitive to that influence. For example, a secular trend of performance improvement over time on publicly reported QMs may result in challenges demonstrating performance improvements that could be achieved by DONs with longer current job tenure or past experience. In support of this hypothesis, it is noteworthy that the weighted mean QM scores for the sampled NHs was less than 10% for 5 of 10 QMs—pain, urinary catheters, urinary tract infections, restraint use, and delirium in short-stay residents. Alternatively, some critics of Five-Star have argued that, because the rating system does not subdivide NHs into classes to reflect specialization, the QM ratings may be unduly low for NHs that care for critically ill residents with special needs (
Although not the primary focus of this study, the results of the multivariate analyses presented in Tables
Although this study included data for a large, nationally representative sample of NHs and, therefore, had excellent power to detect small effects and yielded generalizable findings, the use of the 2004 NNHS as a data source for DON characteristics had several limitations. First, beyond the measurement of years of current job tenure and past experience, this study did not include other potentially important information pertaining to DONs because such data were not available. For example, past research has found associations between higher DON turnover and lower quality of care (
Like many other studies in this area, this study used a cross-sectional, correlational design, which limits readers’ abilities to make causal inferences. The 2004 NNHS contained data on DON current job tenure and past experience at a single time point only. Because repeated measures of DON current job tenure were not available, this study could not include a model how changes in DON tenure related to changes in quality of care over time. As a result, readers cannot definitively conclude that it is not actual quality that impacts DON current job tenure. For example, DONs working in low quality NHs may become frustrated and either quit and take a new position as DON in a higher quality NH or leave the profession altogether (
In addition, this study did not control for all factors that may have influenced QM scores and ratings. For example, this study did not control for the potential influence of other top managers, including NHAs and medical directors or other aspects of staffing, such as nursing staffing levels and turnover, on QM ratings and QM scores.
The finding that longer DON current job tenure, but not past experience, was associated with higher quality of care has important implications for NH administrators and owners. In particular, administrators and owners have an important role in supporting DONs as they either transition into the role of the DON for the first time or learn to effectively fulfill their role in a new NH (
Additional research is needed to enhance our ability to draw conclusions about the association between DON characteristics and quality of care in NHs. For example, research is needed that focuses on additional DON characteristics, such as educational preparation and professional certifications; possible non-linear effects between DON characteristics and quality of care; and other important measures of quality of care, including deficiency citations and resident satisfaction. In addition, research is needed that examines how the interplay between DONs and other top managers—including NHAs and medical directors—and other aspects of staffing—including nursing staffing levels and turnover—relates to quality of care. Research is also needed to explore the reason for associations between top level management characteristics and quality of care. This study added to the literature by examining whether DON current job tenure and past experience were associated with quality of care. This study did not, however, seek to explain
By requiring state agencies to collect additional data about DONs, policy makers could greatly facilitate expanding this body of research and provide additional information to consumers for use when making decisions about NHs. One challenge associated with conducting research to examine the link between DON characteristics and quality of care is that there is little nationally available data about DONs. Aside from limited data from CMS’ Online Survey and Certification and Reporting (OSCAR) file and the 2004 NNHS, nationally available data regarding DON characteristics are not widely available. As a result, researchers who wish to study the link between DON characteristics and quality of care must either use data available in the limited data sets or must collect primary data, which can be both time consuming and costly. Further exploratory research, which may include quantitative studies using small, non-generalizable samples as well as descriptive, qualitative studies, could have an important role in informing policy makers which additional data about DONs should be collected.
Melanie R. Krause is employed as a Health Analyst at the U.S. Government Accountability Office (GAO). The views presented in this manuscript are those of the author and are not intended to reflect the views of the GAO.
This work was supported by the Center for Patient-Centered Interventions Predoctoral Traineeship (T32PH10010) from the National Institute for Nursing Research, National Institutes of Health; John A. Hartford Foundation Building Academic Geriatric Nursing Capacity Predoctoral Scholarship (AAN 08-131); Ruth L. Kirschstein National Research Service Award (F31NR010039) from the National Institute for Nursing Research, National Institutes of Health; and a Charles A. Eckburg Scholarship from the University of Wisconsin-Madison, School of Nursing. Additional support during manuscript preparation was received from a Clinical and Translational Science Award (1UL1RR025011) from the National Center for Research Resources, National Institutes of Health.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Descriptive Statistics
| Measure category | Measure | M | % |
|---|---|---|---|
| Director of nursing | Current job tenure | 3.3 (.14) | |
| Past experience | 3.38 (.17) | ||
| Individual QMs | Late loss ADL decline | 16.05 (.23) | |
| Pain | 6.28 (.19) | ||
| Pressure ulcers | 13.5 (.25) | ||
| Urinary catheters | 5.86 (.12) | ||
| Mobility decline | 13.06 (.22) | ||
| UTIs | 9.01 (.17) | ||
| Restraints | 7.41 (.27) | ||
| Short-stay delirium | 3.24 (.15) | ||
| Short-stay pain | 22.15 (.42) | ||
| Short-stay pressure ulcers | 18.62 (.32) | ||
| Market | Market competition | .2 (.01) | |
| Unemployment rate | 5.74 (.06) | ||
| QM rating | One star | 18.78 (1.25) | |
| Two stars | 23.11 (1.33) | ||
| Three stars | 24.97 (1.37) | ||
| Four stars | 23.03 (1.34) | ||
| Five stars | 10.11 (.94) | ||
| Organizational | Ownership (not-for-profit) | 62.9 (1.51) | |
| Chain membership (non-chain) | 55.47 (1.55) | ||
| Facility size ( ≥ 100 beds) | 50.61 (.69) | ||
| Occupancy | |||
| Low ( < 80%) | 23.02 (1.29) | ||
| Moderate ( ≥ 80%, but ≤ 94%) | 41.32 (1.56) | ||
| High (≥ 95%) | 35.66 (1.54) | ||
| Medicaid Census | |||
| Low (< 60%) | 36.49 (1.51) | ||
| Moderate (≥ 60%, but ≤ 79%) | 42.32 (1.53) | ||
| High (≥ 80%) | 21.19 (1.29) |
Weighted mean.
Weighted percentage.
Mean
| QM rating | Current job tenure | Past experience |
|---|---|---|
| One star | 2.68 (.27) | 3.54 (.39) |
| Two stars | 2.82 (.27) | 3.7 (.38) |
| Three stars | 3.25 (.24) | 3.34 (.37) |
| Four stars | 3.69 (.31) | 2.94 (.34) |
| Five stars | 4.8 (.63) | 3.44 (.51) |
Weighted means are presented.
Relative Risk Ratios from the Logistic Regression Model
| Two stars | Three stars | Four stars | Five stars | |
|---|---|---|---|---|
| Market characteristics | ||||
| Market competition | 0.62 (.26) | 0.79 (.3) | 0.69 (.27) | 0.31 (.19) |
| Unemployment | 0.97 (.05) | 0.95 (.04) | 0.97 (.04) | 0.93 (.06) |
| Organizational characteristics | ||||
| For-profit : not-for-profit | 0.81 (.19) | 0.73 (.17) | 0.92 (.22) | 1.29 (.4) |
| Chain : non-chain | 1.1 (.24) | 0.85 (.19) | 1.02 (.23) | 0.73 (.21) |
| ≥ 100 beds : < 100 beds | 1.97 (.41) | 1.65 (.35) | 1.16 (.24) | 0.57 (.16) |
| Moderate : low occupancy | 1.11 (.27) | 1.95 (.49) | 1.88 (.48) | 1.92 (.64) |
| High : low occupancy | 1.29 (.35) | 2.25 (.62) | 2.44 (.67) | 2.84 (.96) |
| Moderate : low Medicaid | 1.49 (.35) | 1.77 (.42) | 2.17 (.51) | 1.8 (.57) |
| High : low Medicaid | 1.77 (.53) | 2.01 (.59) | 3.11 (.92) | 5.88 |
| DON characteristics (in years) | ||||
| Current job tenure | 1.01 (.002) | 1.03 (.002) | 1.05 (.002) | 1.1(.002) |
| Past experience | 1.0 (.002) | 0.99 (.002) | 0.98 (.002) | 1.01 (.002) |
p < .05.
p < .01.
p <.001
Incidence Rate Ratios from the Logistic Regression Models
| Late loss | Pain | Pressure | Urinary | Mobility | UTI | Restraints | Short- | Short- | Short- | |
|---|---|---|---|---|---|---|---|---|---|---|
| n = 968 | n = 974 | n = 949 | n = 974 | n = 961 | n = 974 | n = 975 | n = 877 | n = 890 | n = 877 | |
| Market | ||||||||||
| Market | 1.16 | 1.18 (.12) | .83 (.07) | 1.41 | 1.02 (.06) | 1.23 | .85 (.11) | 1.8 | 1.0 (.08) | .79 (.07) |
| Unemployment | .99 (.006) | 1.0 (.01) | 1.02 (.01) | 1.0 (.01) | 1.0 (.01) | 1.0 (.01) | 1.02 (.01) | .99 (.02) | 1.0 (.01) | 1.03 |
| Organizational | ||||||||||
| For-profit : not- | 1.02 (.03) | 1.08 (.07) | 1.07 (.05) | 1.06 (.06) | .99 (.04) | 1.07 (.05) | 1.11 (.1) | .9 (.08) | .86 | .94 (.04) |
| Chain : non- | 1.09 | .97 (.06) | 1.0 (.04) | .98 (.05) | 1.04 (.04) | .98 (.04) | .95 (.07) | .88 (.08) | 1.0 (.04) | .94 (.04) |
| ≥100 beds : | 1.02 (.03) | .87 (.05) | 1.13 | 1.05 (.05) | .97 (.03) | 1.0 (.04) | 1.18 | .77 | .87 | 1.08 |
| Moderate : low | 1.02 (.04) | .78 | .99 (.05) | 1.01 (.05) | 1.03 (.05) | 1.05 (.05) | .83 (.07) | .97 (.11) | .82 | .91 |
| High : low | .96 (.04) | .78 | .88 | .89 (.05) | 1.02 (.05) | 1.02 (.06) | .83 (.08) | .92 (.11) | .77 | .97 (.05) |
| Moderate : low | .95 (.03) | 1.01 (.07) | 1.01 (.04) | .86 | .89 | .9 (.04) | 1.02 (.09) | 1.16 (.11) | .89 | .94 (.04) |
| High : low | .86 (.04) | 1.06 (.09) | 1.11 (.07) | .78 | .75 | .72 | 1.25 | .93 (.12) | .8 | 1.04 (.05) |
| DON characteristics | ||||||||||
| Current job | 0.99 | 0.99 | 0.99 | 1.0 | 0.99 | 0.99 | 0.98 (.14) | 0.98 | 1.0 | 1.0 |
| Past experience | 1.0 | 1.01 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.01 | 1.0 | 1.0 |
| Chi-Square | F (11, | F (11, | F (11, | F (11, | F (11, | F (11, | F (11, | F (11, | F (11, | F (11, |
p < .05.
p < .01.
p < .001.