Abstract
Background:
The safety and quality of care provided to nursing home residents is a significant concern. Little is unknown whether fostering patient safety culture helps improve the safety and quality of nursing home care.
Methods:
This study determined the associations of nursing home patient safety culture performance, as reported by administrators, directors of nursing, and unit leaders in a large national sample of free-standing nursing homes, with several “Nursing Home Compare” (NHC) performance indicators. We conducted a survey in 2017 using the Agency for Healthcare Research and Quality (AHRQ) Survey on Patient Safety Culture™ (SOPS™) for nursing homes to collect data on 12 core domains of safety culture scores. Survey data were linked to other nursing home files for multivariable regression analyses.
Results:
Overall, 818 out of the 2254 sampled nursing homes had at least one completed survey returned for a response rate of 36%. After adjustment for nursing home, market, and state covariates, every 10 percentage points increase in overall positive response rate for safety culture was associated with 0.56 fewer healthcare deficiencies (p=0.001), 0.74 fewer substantiated complaints (p=0.004), reduced fines by $2285.20 (p=0.059), and 20% increased odds of being designated as 4- or 5-star (vs. 1–3 star) facilities (odds ratio roughly 1.20, p<0.05).
Conclusions:
Efforts to improve nursing home performance in patient safety culture have the potential to improve broad safety and quality of care measures encapsulated in the NHC publication.
Keywords: nursing home, patient safety culture, Nursing Home Compare, quality of care
INTRODUCTION
Nursing homes provide institutional long-term care to frail older Americans, as well as skilled nursing and rehabilitative services to Medicare beneficiaries with recent discharges from acute care hospitals. The safety and quality of care provided to nursing home residents has been a concern for many years, and broad literature has reported safety shortcomings in various clinical areas such as inappropriate medication use and adverse drug events,1–3 high rate of pressure ulcers,4–6 inadequate fall prevention and risk management,5,7,8 and ineffective infection control.9,10
The safety of healthcare in general has received significantly increased attention in the past two decades following the Institute of Medicine (IOM)’s “To Err is Human” report in 1999.11,12 One of the key insights of the IOM report was that most adverse safety events and medical errors reflected problems of the systems rather than of individual healthcare providers. This has shifted the focus of patient safety improvement from blaming individual staff to improving organizational structures and processes to prevent future errors. This new focus started primarily in hospitals and emphasizes support for (rather than blame and punishment of) individual providers, continuous performance improvement, and evidence-based practice. Ongoing hospital-based research shows that these efforts to improve organizational safety culture are associated with better patient outcomes such as reduced hospital mortality rate,13–16 providing empirical support for the critical role of patient safety culture in safer care.
Although substantial efforts have been made to improve patient safety culture in acute-care settings including hospitals, past evidence suggested that patient safety culture might be poorly developed in many nursing homes.5,17–20 Research based on multi-facility surveys further showed that nursing homes varied substantially in staff-reported safety culture scores,17–21 and that these variations could have an impact on care processes and several safety outcomes.8,20,22
Since the Omnibus Budget Reconciliation Act of 1987, federal laws have established stronger quality and safety standards that all nursing facilities eligible for federal funding have to meet.23 To enforce these standards, states surveyors perform annual surveys of nursing homes during which they examine facility operations, review medical records and interview residents and staff. State surveyors issue deficiency citations to nursing homes if they determine that federal or state standards are not met by the facility. Standards and citations cover all aspects of care nursing homes provide, including clinical care, patient safety, quality of life, and resident rights. States may have additional inspections of nursing homes to confirm if previously identified care issues are corrected, or to investigate complaints filed against the facility by residents, their family members, facility staff, or long-term care ombudsmen.24 Although complaints may be filed for a variety of reasons such as abuse, poor care, unsanitary conditions, and dietary problems, recent studies showed that quality of care issues were the primary reasons for substantiated complaints.24,25 Federal and state governments may impose penalties on nursing homes for serious citations, citations that are not corrected for a long time, and substantiated consumer complaints. Penalties are typically fines of nursing homes but can also be manifest less frequently as denial of payments or appointment a temporary manager.23,26,27
These common regulatory enforcement outcomes – deficiency citations, substantiated complaints, and financial penalties – are published and regularly updated on the “Nursing Home Compare” (NHC) website of the Centers for Medicare and Medicaid Services (CMS). As a comprehensive source of quality and safety of care information on the nation’s nursing homes, NHC also publishes a set of quality measures that are derived from the Minimum Data Set (MDS) resident assessments and that capture resident safety of limited scope,28 nurse staffing levels, and 5-star ratings that aggregate alternative sources of performance measures.29 These publications are intended to inform prospective patients of care quality, and to foster quality improvement efforts within nursing homes.30
Although a handful of studies examined the potential impact of nursing home patient safety culture on individual clinical outcomes, no study has examined the impact on broader NHC performance metrics. Motivated by this gap in the literature, this study aimed to determine the associations of patient safety culture, as perceived and reported by administrative or clinical leadership, with the NHC enforcement outcomes and 5-star ratings on multiple domains of care in a large, national sample of nursing homes. Patient safety, or “freedom from accidental or preventable injuries produced by medical care”, is considered as one of the six defining components of quality of care (i.e. safety, effectiveness, patient centeredness, timeliness, efficiency, and equity) and, as such, as being “indistinguishable from the delivery of quality health care” by the National Academy of Medicine (formerly, Institute of Medicine).31 Thus, given the substantial overlap between “patient safety” and “quality of care”, and the indisputable role of patient safety culture in providing safer care, we hypothesize that better reported patient safety culture is associated with better “quality” as measured by the set of NHC performance metrics.
METHODS
Data sources
Primary data on patient safety culture were derived from a survey of nursing home administrative and clinical leaders – administrators, directors of nursing (DONs), and unit leaders – conducted from January to November of 2017. The survey used the AHRQ Survey on Patient Safety Culture™ (SOPS™) for nursing homes to collect data on 12 core domains of safety culture performance (see Table A1 in the Appendix for the list of domains and definitions).32 Each domain was assessed with 3 or 4 items or questions (total of 42 items), and each item used a 5-point Likert scale response category (strongly disagree, disagree, neither, agree, and strongly agree) to measure staff perceptions of safety culture. The AHRQ developed and refined the nursing home SOPS™ instrument based on a previous version for hospitals and evidence-based research and practices that reduce errors and improve clinical outcomes in nursing homes. The AHRQ performed psychometric test for the instrument on more than 3000 staff working in 40 nursing homes nationally, and established its reliability and validity (e.g. Cronbach’s alpha>0.70 for all 12 domains).32 The internal consistency in our own survey sample was also very high (Cronbach’s alpha 0.8–0.9 for all domains).
We randomly sampled 2254 facilities for the survey from a sampling frame of all Medicare and/or Medicaid certified free-standing nursing homes. The sampling frame was obtained from the NHC file, and excluded hospital-based facilities (about 5% of all nursing homes nationally) because they usually have fundamentally different organizational structures, staffing patterns, and strategic aims than other facilities. The survey was mailed to the administrator and the DON of a sampled nursing home in separate preaddressed and stamped return envelopes. DONs were also asked to distribute additional copies of the questionnaire to 1 to 3 nurse managers/unit leaders in their facilities, who were then encouraged to complete the survey independently and return to us in separate envelopes provided. We used several strategies to increase response rate including reminder postcards after initial mailing, 5 follow-up mailings to non-responding facilities, and an alternative electronic version of the survey on REDCap™.
We obtained other nursing home information, such as structural characteristics, deficiency citations, consumer complaints, and 5-star ratings, from the 2018 NHC file. We further used the LTCfocus file maintained by the Brown University to define several more nursing home covariates described below.
Variables
Following the recommendation of the AHRQ32 and also the approach of recent studies,8,20 we first calculated the positive response rate for each safety culture domain; it was done by dividing the number of items in the domain with positive responses by the total number of non-missing items in the domain. For each item, a positive response was defined as a response of agree or strongly agree if the item was positively worded, or a response of disagree or strongly disagree if the item was negatively worded. We then calculated the facility-level positive response rate for each domain (hereafter referred to as facility safety culture score for each domain) as the average of domain-specific positive response rates for all types of staff responses (i.e. administrator, director of nursing, and unit leaders). Finally, we calculated the overall facility safety culture score as the average of the facility safety culture scores for all 12 domains. The key independent variable in analyses was the overall safety culture score with higher value indicating better overall safety culture of the nursing home.
The outcome variables of this study included number of healthcare-related deficiency citations, number of substantiated complaints, total amount of fines paid by the nursing home, and nursing home 5-star ratings. The CMS developed the 5-star ratings to simplify information for consumers by aggregating alternative measures into a rating system of one to five stars, with more stars indicating better quality.33 Specifically, the ratings were developed separately to summarize 3 domains of “quality”: deficiency citations assigned during annual and complaint inspections; clinical outcomes of residents based on Minimum Data Set assessments; and nurse staffing to resident ratios (for registered nurses [RNs] and all nursing staff including RNs, licensed practical nurses, and certified nursing assistants). Overall 5-star ratings were also developed to further aggregate quality information on the three domains.33 This study analyzed the 5-star ratings for overall quality, deficiency citations, and clinical outcomes given our focus on regulatory and patient outcomes.
Additional nursing home covariates included in multivariable analyses were number of beds, total number of residents, ownership type (for-profit, non-for-profit, or government-owned), chain affiliation (yes/no), a case mix index calculated based on the Resource Utilization Groups classification system, percentage of Medicare residents in the nursing home, percentage of Medicaid residents, percentage of racial/ethnic residents. Finally, a county-level measure of market competition for nursing home care was defined based on the Herfindahl–Hirschmann index, which was calculated using nursing home beds. The measure of market competition ranged from zero to one with higher values indicating more competitive markets.
Analysis
All analyses were conducted at the nursing home level. In bivariate analyses we first compared characteristics of responding nursing homes (n=818) to those of nursing homes in the sampling frame nationally (n=14091). Analyses of variance for continuous variables and chi-square tests for categorical variables were used in statistical inference.
In multivariable analyses, we fit separate linear regression models for number of healthcare-related deficiency citations, number of substantiated complaints, and total amount of fines; and separate logistic regression models for three binary outcomes indicating whether the nursing home had a rank of 4 or 5 stars (versus 1 to 3 stars) for overall quality, for deficiency citations, and for risk-adjusted quality measures. These models had overall safety culture score of the nursing home as the independent variable, and adjusted for nursing home covariates, market competition, and a set of dummy variables for states (each for a state or Washington D.C.). All regression models also incorporated random effects of nursing homes and robust variance-covariance estimates to account for the clustering of nursing homes within counties.
To account for possible non-linear association between overall safety culture score and each outcome of interest, we further fit separate sets of models in which reported safety culture score was alternatively specified as a continuous variable; an indicator variable defining whether the score of a nursing home was equal to or greater than 0.849 (the median of scores of all responding nursing homes); or two indicator variables defining tertile groups of nursing homes (scores between 0.898 and 1 as the top tertile group, and scores between 0.785 and 0.898 as the middle tertile group, each compared to the bottom tertile group with scores between 0 and 0.785).
We conducted three sets of sensitivity analyses. First, we repeated the analyses above but with further adjustment for two nursing home covariates in regression models: RN nurse staffing level (hours per resident day) and total nurse staffing level. Second, to determine the robustness of study findings to different types of staff responses (i.e. administrator vs. nursing leaders), we calculated the overall safety culture scores of nursing homes using administrator responses only and then using nursing leader (DON or unit leader) responses only; we then tested the association between safety culture score of each type with the outcome of interest. Third, we fit similar regression models to determine the association of perceived nursing home safety culture for each domain with the outcome of interest; each model had domain-specific score as the independent variable and adjusted for the same set of nursing home, market, and state covariates. Finally, we fit alternative non-linear models (i.e. negative binomial or two-part models) for the three outcomes of number of deficiencies, number of substantiated complaints, and total amount of fines; their results closely resembled those in the linear models and thus were not presented.
RESULTS
We received responses from 529 administrators, 379 DONs, and 539 unit leaders. Overall, 818 out of the 2254 sampled nursing homes had at least one completed survey returned for a response rate of 36%. The overall positive response rate for the 12 safety culture domains and all staff responses had an average of 81.6% and varied considerably over facilities (Table 1 and Figure 1). The positive response rate for each domain had an average ranging from 63.4% (domain 3 – compliance with procedures) to 96.5% (domain 10 – overall perceptions of resident safety), and varied over facilities.
Table 1.
Characteristics of study nursing homes compared to free-standing nursing homes nationally
| Study nursing homes (n=818) | Free-standing Nursing homes nationally (n=14091) | P-value | |
|---|---|---|---|
| Mean±SD or Prevalence (%) | |||
| Safety culture score (positive response rate), % | |||
| Overall (domains 1–12) | 81.6±14.1 | -- | -- |
| Domain 1: teamwork | 86.0±23.2 | -- | -- |
| Domain 2: staffing | 64.5±28.7 | -- | -- |
| Domain 3: compliance with procedures | 63.4±25.9 | -- | -- |
| Domain 4: training & skills | 74.0±28.5 | -- | -- |
| Domain 5: non-punitive response to mistakes | 72.4±25.3 | -- | -- |
| Domain 6: handoffs | 70.8±26.6 | -- | -- |
| Domain 7: feedback & communication about incidents | 96.4±12.0 | -- | -- |
| Domain 8: communication openness | 87.6±21.7 | -- | -- |
| Domain 9: supervisor expectations & actions promoting resident safety | 92.0±19.3 | -- | -- |
| Domain 10: overall perceptions of resident safety | 96.5±13.0 | -- | -- |
| Domain 11: management support for resident safety | 91.5±18.3 | -- | -- |
| Domain 12: organizational learning | 83.6±20.9 | -- | -- |
| Number of healthcare deficiencies | 7.4±6.4 | 7.6±6.8 | 0.314 |
| Number of substantiated complaints | 4.2±7.1 | 4.3±7.6 | 0.766 |
| Total amount of fines, $ x 1000 | 10.8±34.1 | 14.8±56.4 | 0.048 |
| Facilities of 4- or 5-star ratings for overall quality, % | 53.7 | 49.2 | 0.013 |
| Total number of beds | 114.2±52.7 | 107.7±58.9 | 0.002 |
| Total number of residents | 93.7±48.6 | 87.5±52.3 | 0.001 |
| Ownership type, % | 0.000 | ||
| For-profit | 66.8 | 73.0 | |
| Non-for-profit | 25.7 | 21.4 | |
| Government-owned | 7.5 | 5.6 | |
| Chain affiliation | 56.0 | 58.3 | 0.207 |
| Case mix acuity | 1.2±0.1 | 1.2±0.2 | 0.957 |
| Percent of residents funded by Medicare, % | 13.7±10.9 | 14.6±13.0 | 0.060 |
| Percent of residents funded by Medicaid, % | 59.9±20.4 | 59.6±22.3 | 0.685 |
| Percent of racial or ethnic minority residents, % | 17.2±21.1 | 19.3±21.6 | 0.007 |
| Nurse staffing (hrs per resident day) for | |||
| Registered nurses | 0.7±0.3 | 0.6±0.4 | 0.153 |
| Total (RNs+LPNs+CNAs) | 3.8±0.7 | 3.8±0.8 | 0.618 |
| Market competition for nursing home care | 0.8±0.2 | 0.8±0.2 | 0.384 |
RN=registered nurse; LPN=licensed practical nurse; CAN=certified nursing assistant.
Figure 1.
Distribution of overall safety culture score (positive response rate) of study nursing homes.
Note: all items in the AHRQ Survey on Patient Safety Culture for nursing homes use 5-point Likert scales (strongly disagree, disagree, neither, agree, and strongly agree) for responses. We calculated positive response rate for each domain as the percent of positive responses (strongly agree or agree) of all non-missing items in the domain. The figure reports the average rate of positive responses for a nursing home over all 12 domains and staff responses (administrator, director of nursing, and unit leaders).
Compared to free-standing nursing homes in the sampling frame nationally (n=14091), responding nursing homes had, on average, similar number of healthcare deficiencies (Table 1, 7.4 vs. 7.6), number of substantiated complaints (4.2 vs. 4.3), and nurse staffing levels (e.g. 3.8 hours per resident day totally for both groups). However, responding nursing homes were slightly more likely to be larger and non-for-profit facilities, had lower amount of fines ($10.8k vs. $14.8k on average), and were more likely to have overall 4- or 5-star ratings (53.7% vs. 49.2%).
Table 2 shows that in unadjusted analyses, higher overall safety culture score was associated with lower healthcare deficiencies, substantiated complaints, and fines. Multivariable analyses with adjustment for nursing home, market, and state covariates, and using alternative specifications of overall safety culture score (e.g. a continuous variable and categorical variables) largely confirmed these negative associations. For example, in adjusted analyses every 10 percentage points increase in overall safety culture score was associated with 0.56 fewer healthcare deficiencies (p=0.001), 0.74 fewer substantiated complaints (p=0.004), and reduced fines by $2285.20 (p=0.059).
Table 2.
Associations of nursing home safety culture score with “Nursing Home Compare” regulatory outcomes
| Overall safety culture score (average percentage point of positive responses over 12 safety culture domains and administrator, DON, and unit leader responses) | Number of healthcare deficiencies | Number of substantiated complaints | Total amount of fines | |||
|---|---|---|---|---|---|---|
| β-coefficient | P-value | β-coefficient | P-value | β-coefficient | P-value | |
| Continuous variable (percentage point x 10) | ||||||
| Unadjusted a | −0.85 | 0.000 | −1.06 | 0.000 | −2575.86 | 0.017 |
| Adjusted b | −0.56 | 0.001 | −0.74 | 0.004 | −2285.20 | 0.059 |
| Binary groups (≥median vs. <median) | ||||||
| Unadjusted a | −1.27 | 0.003 | −2.05 | 0.000 | −4377.36 | 0.073 |
| Adjusted b | −0.95 | 0.023 | −1.60 | 0.001 | −3419.29 | 0.175 |
| Tertile groups | ||||||
| 2nd vs. 1st tertile group | ||||||
| Unadjusted a | −2.87 | 0.000 | −3.25 | 0.000 | −7928.12 | 0.003 |
| Adjusted b | −1.60 | 0.006 | −1.99 | 0.001 | −5919.83 | 0.030 |
| 3rd vs. 1st tertile group | ||||||
| Unadjusted a | −2.04 | 0.000 | −2.73 | 0.000 | −5307.45 | 0.102 |
| Adjusted b | −1.29 | 0.016 | −1.90 | 0.003 | −1919.64 | 0.599 |
Derived from bivariate linear regression models with nursing home random effects and robust variance-covariance estimates.
Derived from multivariable linear regression models with nursing home random effects and robust variance-covariance estimates, as well as further adjustment for covariates including nursing home bed size, total number of residents, profit status (non-for-profit or government-owned vs. for profit), chain affiliation, a case mix acuity index, percentage of Medicare residents, percentage of Medicaid residents, percentage of racial and ethnic minority residents, market competition for nursing home care, and state dummies.
Note: estimates with p-value<0.05 are highlighted in bold.
Table 3 further shows that in both unadjusted and adjusted analyses, improved overall safety culture score was associated with increased odds of facilities being ranked as 4- or 5-star for overall quality, survey deficiencies, and risk-adjusted quality measures. In adjusted analyses, every 10 percentage points increase in overall safety culture score was associated with 20% increased odds of facilities having 4- or 5-star ratings (odds ratio [OR]=1.23, p=0.003 for overall ratings; OR=1.18, p=0.019 for deficiency ratings; and OR=1.19, p=0.012 for quality measure ratings).
Table 3.
Associations of nursing home safety culture score with “Nursing Home Compare” 5-star ratings (4 or 5 stars vs. 1–3 stars)
| Overall safety culture score (average percentage point of positive responses over 12 safety culture domains and administrator, DON, and unit leader responses) | Overall ratings | Ratings for survey deficiencies | Ratings for quality measures | |||
|---|---|---|---|---|---|---|
| Odds ratio | P-value | Odds ratio | P-value | Odds ratio | P-value | |
| Continuous variable (percentage point x 10) | ||||||
| Unadjusted a | 1.24 | 0.000 | 1.16 | 0.005 | 1.14 | 0.027 |
| Adjusted b | 1.23 | 0.003 | 1.18 | 0.019 | 1.19 | 0.012 |
| Binary groups (≥median vs. <median) | ||||||
| Unadjusted a | 1.65 | 0.002 | 1.29 | 0.069 | 1.24 | 0.224 |
| Adjusted b | 1.82 | 0.001 | 1.47 | 0.022 | 1.36 | 0.127 |
| Tertile groups | ||||||
| 2nd vs. 1st tertile group | ||||||
| Unadjusted a | 1.87 | 0.001 | 1.76 | 0.006 | 2.11 | 0.000 |
| Adjusted b | 1.88 | 0.006 | 1.64 | 0.039 | 2.75 | 0.000 |
| 3rd vs. 1st tertile group | ||||||
| Unadjusted a | 1.68 | 0.006 | 1.37 | 0.077 | 1.47 | 0.057 |
| Adjusted b | 1.69 | 0.017 | 1.42 | 0.105 | 1.75 | 0.024 |
Derived from bivariate logistic regression models with nursing home random effects and robust variance-covariance estimates.
Derived from multivariable logistic regression models with nursing home random effects and robust variance-covariance estimates, as well as further adjustment for covariates including nursing home bed size, total number of residents, profit status (non-for-profit or government-owned vs. for profit), chain affiliation, a case mix acuity index, percentage of Medicare residents, percentage of Medicaid residents, percentage of racial and ethnic minority residents, market competition for nursing home care, and state dummies.
Note: estimates with p-value<0.05 are highlighted in bold.
In sensitivity analyses when we repeated the analyses in Tables 2 and 3 but with further adjustment for RN and total nurse staffing levels in regressions models, the results were closely similar and thus are not presented. The robustness of the results in Tables 2 and 3 was also fairly confirmed by sensitivity analyses in which the overall safety culture score was calculated based on administrator responses only or DON/unit leader responses only (see Tables A2 and A3 in the Appendix which show qualitatively similar estimates between the two types of responses). Lastly, higher scores of many of the individual safety culture domains were associated with reduced deficiencies, substantiated complaints, and fines (Table A4 in the Appendix), as well as with increased odds of being 4- or 5-star facilities (Table A5 in the Appendix), although the magnitude of associations for each safety culture domain seemed to be smaller than that of the associations for the overall domain score shown in Tables 2 and 3.
DISCUSSION
This study surveyed a large, national sample of nursing homes and demonstrated that patient safety culture as perceived by facility leaders varied substantially across facilities. Better patient safety culture score predicted better NHC performance indicators including reduced deficiency citations for healthcare, fewer substantiated complaints, lower amount of fines paid by nursing home to the CMS for quality and safety issues, and increased odds of being designated as 4- or 5-star facilities. These findings persisted after multivariable adjustment for nursing home, market, and state covariates, and were robust to alternative ways of computing safety culture scores.
Only several studies empirically tested the impact of staff-reported patient safety culture on nursing home quality and outcomes. In a national survey of top managers (administrators and DONs) in 3557 facilities based on the AHRQ nursing home SOPS™, Thomas and colleagues8 found that higher overall safety culture score was significantly associated with lower prevalence of restraint use, confirming the finding of another study,22 and with lower fall rate in nursing homes.8 A more recent AHRQ initiative designed to reduce catheter-associated urinary tract infections (CAUTI) and other healthcare-associated infections in nursing homes reported that patient safety culture improved over time in nursing homes, but that improved safety culture was not significantly associated with reduced CAUTI rate.20
Our study extended this limited literature by examining the potential impact of perceived patient safety culture on a set of performance metrics published on the NHC website. Together, these publications capture nursing home’s potential care issues identified during annual state inspections (e.g. measured by deficiency citations and amount of fines), resident and family complaints about care delivery, and patient clinical outcomes. These measures are publicly disseminated to inform consumer choice of nursing homes, and are being used as essential measures in state pay-for-performance programs34 as well as in more recent healthcare reform efforts such as Medicare accountable care organizations.35 The magnitude of the estimated associations between perceived safety culture and NHC performance metrics is relatively strong and bears important implications for care practices. For example, compared to facilities with lower than median patient safety culture scores, facilities in the higher-than-median group on average had 1 less deficiency, 1.6 less substantiated complaints, and 80% higher odds of having 4 or 5 stars in overall rating (Tables 2 and 3). This suggests a substantial potential that nursing home efforts to foster an organizational culture of safe practices can improve multi-dimensional performance in safety, effectiveness, and patient-centered care that are broadly targeted by current quality improvement and payment reform initiatives. Furthermore, our findings showed that the associations of overall safety culture score with focused performance metrics (Tables 2 and 3) are stronger and more consistent than those of individual safety culture domains (Tables A4 and A5 in the Appendix). This further suggests that nursing home’s focus on a particular aspect of patient safety culture (e.g. non-punitive response to mistakes) would not be very effective in improving safety of care; instead, efforts to improve multiple aspects of safety culture and thus overall organizational performance may be a more effective approach to improving overall safety and outcomes of care.
The NHC metrics encapsulate important patient safety outcomes. Healthcare deficiency citations, for example, may be issued due to significant safety issues identified during on-site inspections such as widespread medication errors.28 Complaints filed by consumers, staff, or long-term care ombudsmen, and financial penalties made to nursing homes may also reflect similar safety shortcomings. Furthermore, the five-star ratings on risk-adjusted quality measures are derived from a set of 16 MDS- or claims-based measures, several of which are common safety outcomes (e.g. pressure ulcer rate and rate of injurious falls) that may be amenable to improved patient safety culture.8,22 Thus, the findings of this study, when not targeting specific safety deficiencies or outcomes, provide evidence supporting the expectation that overall improved safety of care can be achieved in nursing homes as a result of improved organizational safety culture.
Although patient safety is an important component of quality of care, the NHC performance metrics are designed to capture the broader, multidimensional construct of nursing home quality for the purposes of informed consumer choices of facilities and market-driven incentives for quality improvement. Many of these measures (e.g. those of depressive symptom management or improved physical function) are not directly linked to safety of care, at least in the traditional sense. Nevertheless, it is plausible that efforts to improve patient safety culture in a facility would be focused on multiple nursing home process of care, staff knowledge and training, and team practices, which may overlap with other organizational-level quality improvement efforts.36 Thus, improved safety culture may positively influence broad aspects of care and resident outcomes that may or may not fall exactly under the purview of care safety.
We acknowledge potential limitations of this study. First, although this study documented associations between patient safety culture and nursing home quality measures, we could not confirm causal relationships given the cross-sectional design of this study. In addition, although our regression analyses adjusted for an intensive set of facility, market, and state covariates, we could not totally rule out the possibility that part of the associations were confounded by unmeasured factors. However, the sensitivity analyses confirmed the robustness of key findings of this study. Finally, response bias may exist in this voluntary survey with a 36% response rate; specifically, previous studies showed that staff in the same facility might perceive safety culture differently and it was possible staff over-state their facilities’ safety culture performance.18 However, our results are relatively robust to responses from administrators vs. nursing leaders, and the large number of participants help minimize this concern.
In conclusion, better patient safety culture as reported by facility administrative and clinical leaders was associated with better NHC performance metrics that were alternatively defined to incorporate broad perspectives of quality and care provision in nursing homes. Efforts to improve nursing home performance in safety culture thus have the potential to improve both safety of care and other aspects of care quality that are central to patient welfare.
Acknowledgment:
This study is funded by the Agency for Healthcare Research and Quality (AHRQ) under grant R01HS024923. The views expressed in this article are those of the authors and do not necessarily represent the view of the AHRQ. The authors thank Mr. Joseph Duckett for his contributions to the conduct of the survey.
APPENDIX
Table A1.
Safety culture domains in the Nursing Home Survey on Patient Safety Culture (SOPS™) of the Agency for Healthcare Research and Quality (AHRQ).a
| There are 12 safety culture domains in the Nursing Home SOPS™, which measure the extent to which… Staff treat each other with respect, support one another, and feel like they are part of a team (Domain 1: Teamwork); There are enough staff to handle the workload, meet residents’ needs during shift changes, and keep residents safe, because there is not much staff turnover (Domain 2: Staffing); Staff follow standard procedures to care for residents and do not use shortcuts to get their work done faster (Domain 3: Compliance With Procedures); Staff get the training they need, have enough training on how to handle difficult residents, and understand the training they get in the nursing home (Domain 4: Training & Skills); Staff are not blamed when a resident is harmed, are treated fairly when they make mistakes, and feel safe reporting their mistakes (Domain 5: Nonpunitive Response to Mistakes) Staff are told what they need to know before taking care of a resident or when a resident’s care plan changes, and have all the information they need when residents are transferred from the hospital (Domain 6: Handoffs); Staff discuss ways to keep residents safe, tell someone if they see something that might harm a resident, and talk about ways to keep incidents from happening again (Domain 7: Feedback & Communication About Incidents); Staff speak up about problems, and their ideas and suggestions are valued (Domain 8: Communication Openness); Supervisors listen to staff ideas and suggestions about resident safety, praise staff who follow the right procedures, and pay attention to safety problems (Domain 9: Supervisor Expectations & Actions Promoting Resident Safety); Residents are well cared for and safe (Domain 10: Overall Perceptions of Resident Safety); Nursing home management provides a work climate that promotes resident safety and shows that resident safety is a top priority (Domain 11: Management Support for Resident Safety); And There is a learning culture that facilitates making changes to improve resident safety and evaluates changes for effectiveness (Domain 12: Organizational Learning). |
Excerpted from Sorra J, Gray L, Famolaro T, et al. Nursing Home Survey on Patient Safety Culture. (Prepared by Westat, under Contract No. HHSA290201300003C). AHRQ Publication No. 18–0038-EF (Replaces 08(09)-0060, 15(16)-0052-EF). Rockville, MD: Agency for Healthcare Research and Quality; July 2018. http://www.ahrq.gov/professionals/quality-patient-safety/patientsafetyculture/nursing-home/index.html
Table A2.
Associations of nursing home safety culture performance with “Nursing Home Compare” regulatory outcomes (sensitivity analyses on administrator responses versus DON/unit leader responses).
| Overall safety culture score (average percentage point of positive responses over 12 safety culture domains); Continuous variable (percentage point x 10) |
Number of healthcare deficiencies | Number of substantiated complaints | Total amount of fines | |||
|---|---|---|---|---|---|---|
| β-coefficient | P-value | β-coefficient | P-value | β-coefficient | P-value | |
| Administrator responses only (n=529 facilities) | ||||||
| Unadjusted a | −0.62 | 0.004 | −0.80 | 0.000 | −620.72 | 0.592 |
| Adjusted b | −0.50 | 0.002 | −0.57 | 0.015 | −340.80 | 0.771 |
| DON or unit leader responses only (n=443 facilities) | ||||||
| Unadjusted a | −0.96 | 0.000 | −1.18 | 0.001 | −4449.45 | 0.005 |
| Adjusted b | −0.64 | 0.022 | −0.89 | 0.012 | −4362.57 | 0.016 |
Derived from bivariate linear regression models with nursing home random effects and robust variance-covariance estimates.
Derived from multivariable linear regression models with nursing home random effects and robust variance-covariance estimates, as well as further adjustment for covariates including nursing home bed size, total number of residents, profit status (non-for-profit or government-owned vs. for profit), chain affiliation, a case mix acuity index, percentage of Medicare residents, percentage of Medicaid residents, percentage of racial and ethnic minority residents, market competition for nursing home care, and state dummies.
Note: estimates with p-value<0.05 are highlighted in bold.
Table A3.
Association of nursing home safety culture performance with “Nursing Home Compare” 5-star ratings (4 or 5 stars vs. 1–3 stars; sensitivity analyses on administrator responses versus DON/unit leader responses)
| Overall safety culture score (average percentage point of positive responses over 12 safety culture domains); Continuous variable (percentage point x 10) |
Overall ratings | Ratings for survey deficiencies | Ratings for quality measures | |||
|---|---|---|---|---|---|---|
| Odds ratio | P-value | Odds ratio | P-value | Odds ratio | P-value | |
| Administrator responses only (n=529 facilities) | ||||||
| Unadjusted a | 1.26 | 0.000 | 1.18 | 0.037 | 1.11 | 0.162 |
| Adjusted b | 1.29 | 0.005 | 1.27 | 0.012 | 1.09 | 0.948 |
| DON or unit leader responses only (n=443 facilities) | ||||||
| Unadjusted a | 1.26 | 0.003 | 1.14 | 0.061 | 1.17 | 0.051 |
| Adjusted b | 1.27 | 0.313 | 1.15 | 0.186 | 1.24 | 0.025 |
Derived from bivariate logistic regression models with nursing home random effects and robust variance-covariance estimates.
Derived from multivariable logistic regression models with nursing home random effects and robust variance-covariance estimates, as well as further adjustment for covariates including nursing home bed size, total number of residents, profit status (non-for-profit or government-owned vs. for profit), chain affiliation, a case mix acuity index, percentage of Medicare residents, percentage of Medicaid residents, percentage of racial and ethnic minority residents, market competition for nursing home care, and state dummies.
Note: estimates with p-value<0.05 are highlighted in bold.
Table A4.
Association of each nursing home safety culture domain performance with “Nursing Home Compare” regulatory outcomes
| Domain-specific average score (average percentage point of positive responses to a safety culture domain over administrator, DON, and unit leader responses; continuous variable x 10) | Number of healthcare deficiencies | Number of substantiated complaints | Total amount of fines | |||
|---|---|---|---|---|---|---|
| β-coefficient | P-value | β-coefficient | P-value | β-coefficient | P-value | |
| Domain 1: teamwork | ||||||
| Unadjusted a | −0.36 | 0.004 | −0.53 | 0.000 | −1687.95 | 0.006 |
| Adjusted b | −0.30 | 0.004 | −0.42 | 0.001 | −1609.32 | 0.020 |
| Domain 2: staffing | ||||||
| Unadjusted a | −0.32 | 0.001 | −0.43 | 0.000 | −832.12 | 0.068 |
| Adjusted b | −0.21 | 0.011 | −0.33 | 0.002 | −577.60 | 0.242 |
| Domain 3: compliance with procedures | ||||||
| Unadjusted a | −0.30 | 0.000 | −0.53 | 0.000 | −775.37 | 0.180 |
| Adjusted b | −0.21 | 0.007 | −0.33 | 0.000 | −535.69 | 0.361 |
| Domain 4: training & skills | ||||||
| Unadjusted a | −0.27 | 0.002 | −0.40 | 0.000 | −1138.63 | 0.039 |
| Adjusted b | −0.13 | 0.075 | −0.28 | 0.019 | −741.83 | 0.192 |
| Domain 5: non-punitive response to mistakes | ||||||
| Unadjusted a | −0.33 | 0.000 | −0.35 | 0.001 | −819.32 | 0.111 |
| Adjusted b | −0.20 | 0.035 | −0.17 | 0.132 | −875.11 | 0.104 |
| Domain 6: handoffs | ||||||
| Unadjusted a | −0.32 | 0.000 | −0.39 | 0.001 | −931.85 | 0.047 |
| Adjusted b | −0.21 | 0.013 | −0.24 | 0.037 | −694.47 | 0.160 |
| Domain 7: feedback & communication about incidents | ||||||
| Unadjusted a | −0.73 | 0.004 | −0.75 | 0.045 | −1443.20 | 0.316 |
| Adjusted b | −0.63 | 0.010 | −0.65 | 0.091 | −1892.66 | 0.289 |
| Domain 8: communication openness | ||||||
| Unadjusted a | −0.40 | 0.000 | −0.49 | 0.003 | −1270.34 | 0.097 |
| Adjusted b | −0.28 | 0.016 | −0.36 | 0.033 | −1281.24 | 0.108 |
| Domain 9: supervisor expectations & actions promoting resident safety | ||||||
| Unadjusted a | −0.20 | 0.185 | 0.05 | 0.699 | −371.85 | 0.577 |
| Adjusted b | −0.02 | 0.893 | 0.15 | 0.157 | −100.10 | 0.879 |
| Domain 10: overall perceptions of resident safety | ||||||
| Unadjusted a | −0.77 | 0.000 | −0.89 | 0.006 | −2175.22 | 0.104 |
| Adjusted b | −0.52 | 0.010 | −0.63 | 0.036 | −2349.57 | 0.066 |
| Domain 11: management support for resident safety | ||||||
| Unadjusted a | −0.35 | 0.024 | −0.37 | 0.048 | −602.48 | 0.448 |
| Adjusted b | −0.31 | 0.028 | −0.39 | 0.027 | −687.31 | 0.453 |
| Domain 12: organizational learning | ||||||
| Unadjusted a | −0.35 | 0.004 | −0.31 | 0.057 | −1084.45 | 0.104 |
| Adjusted b | −0.22 | 0.070 | −0.21 | 0.210 | −1186.36 | 0.124 |
Derived from bivariate linear regression models with nursing home random effects and robust variance-covariance estimates.
Derived from multivariable linear regression models with nursing home random effects and robust variance-covariance estimates, as well as further adjustment for covariates including nursing home bed size, total number of residents, profit status (non-for-profit or government-owned vs. for profit), chain affiliation, a case mix acuity index, percentage of Medicare residents, percentage of Medicaid residents, percentage of racial and ethnic minority residents, market competition for nursing home care, and state dummies.
Note: estimates with p-value<0.05 are highlighted in bold.
Table A5.
Association of each nursing home safety culture domain performance with “Nursing Home Compare” 5-star ratings (4 or 5 stars vs. 1–3 stars)
| Domain-specific average score (average percentage point of positive responses to a safety culture domain over administrator, DON, and unit leader responses; continuous variable x 10) | Overall ratings | Ratings for survey deficiencies | Ratings for quality measures | |||
|---|---|---|---|---|---|---|
| Odds ratio | P-value | Odds ratio | P-value | Odds ratio | P-value | |
| Domain 1: teamwork | ||||||
| Unadjusted a | 1.12 | 0.001 | 1.05 | 0.136 | 1.06 | 0.117 |
| Adjusted b | 1.12 | 0.007 | 1.04 | 0.337 | 1.10 | 0.014 |
| Domain 2: staffing | ||||||
| Unadjusted a | 1.06 | 0.034 | 1.06 | 0.043 | 1.01 | 0.663 |
| Adjusted b | 1.09 | 0.008 | 1.08 | 0.016 | 1.06 | 0.093 |
| Domain 3: compliance with procedures | ||||||
| Unadjusted a | 1.08 | 0.009 | 1.08 | 0.011 | 1.05 | 0.152 |
| Adjusted b | 1.09 | 0.007 | 1.10 | 0.006 | 1.03 | 0.416 |
| Domain 4: training & skills | ||||||
| Unadjusted a | 1.07 | 0.007 | 1.05 | 0.068 | 1.06 | 0.059 |
| Adjusted b | 1.06 | 0.050 | 1.06 | 0.037 | 1.08 | 0.050 |
| Domain 5: non-punitive response to mistakes | ||||||
| Unadjusted a | 1.05 | 0.093 | 1.03 | 0.339 | 0.99 | 0.755 |
| Adjusted b | 1.02 | 0.550 | 1.00 | 0.919 | 1.00 | 0.930 |
| Domain 6: handoffs | ||||||
| Unadjusted a | 1.11 | 0.001 | 1.07 | 0.015 | 1.06 | 0.043 |
| Adjusted b | 1.10 | 0.005 | 1.08 | 0.011 | 1.07 | 0.035 |
| Domain 7: feedback & communication about incidents | ||||||
| Unadjusted a | 1.20 | 0.003 | 1.11 | 0.163 | 1.10 | 0.092 |
| Adjusted b | 1.23 | 0.009 | 1.17 | 0.132 | 1.11 | 0.120 |
| Domain 8: communication openness | ||||||
| Unadjusted a | 1.13 | 0.000 | 1.05 | 0.117 | 1.12 | 0.001 |
| Adjusted b | 1.13 | 0.007 | 1.06 | 0.118 | 1.12 | 0.005 |
| Domain 9: supervisor expectations & actions promoting resident safety | ||||||
| Unadjusted a | 1.02 | 0.580 | 1.00 | 0.958 | 1.04 | 0.375 |
| Adjusted b | 0.98 | 0.646 | 0.95 | 0.351 | 1.03 | 0.520 |
| Domain 10: overall perceptions of resident safety | ||||||
| Unadjusted a | 1.18 | 0.007 | 1.31 | 0.015 | 1.06 | 0.282 |
| Adjusted b | 1.19 | 0.029 | 1.31 | 0.035 | 1.13 | 0.052 |
| Domain 11: management support for resident safety | ||||||
| Unadjusted a | 1.08 | 0.036 | 1.04 | 0.336 | 1.06 | 0.179 |
| Adjusted b | 1.08 | 0.080 | 1.04 | 0.476 | 1.04 | 0.320 |
| Domain 12: organizational learning | ||||||
| Unadjusted a | 1.10 | 0.007 | 1.10 | 0.016 | 1.09 | 0.016 |
| Adjusted b | 1.09 | 0.035 | 1.08 | 0.078 | 1.11 | 0.005 |
Derived from bivariate logistic regression models with nursing home random effects and robust variance-covariance estimates.
Derived from multivariable logistic regression models with nursing home random effects and robust variance-covariance estimates, as well as further adjustment for covariates including nursing home bed size, total number of residents, profit status (non-for-profit or government-owned vs. for profit), chain affiliation, a case mix acuity index, percentage of Medicare residents, percentage of Medicaid residents, percentage of racial and ethnic minority residents, market competition for nursing home care, and state dummies.
Note: estimates with p-value<0.05 are highlighted in bold.
Footnotes
Conflicts of Interests: no conflict of interest for any author.
REFERENCE
- 1.Briesacher BA, Limcangco MR, Simoni-Wastila L, et al. The quality of antipsychotic drug prescribing in nursing homes. Arch Intern Med 2005;165(11):1280–1285. [DOI] [PubMed] [Google Scholar]
- 2.Chen Y, Briesacher BA, Field TS, Tjia J, Lau DT, Gurwitz JH. Unexplained variation across US nursing homes in antipsychotic prescribing rates. Arch Intern Med 2010;170(1):89–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Handler SM, Wright RM, Ruby CM, Hanlon JT. Epidemiology of medication-related adverse events in nursing homes. The American journal of geriatric pharmacotherapy 2006;4(3):264–272. [DOI] [PubMed] [Google Scholar]
- 4.Berlowitz DR, Wilking SV. Pressure ulcers in the nursing home. In: Rubenstein LZ, Wieland D, eds. Improving Care in the Nursing Home: Comprehensive Reviews of Clinical Research Newbury Park, CA: Sage; 1993. [Google Scholar]
- 5.Gruneir A, Mor V. Nursing home safety: current issues and barriers to improvement. Annu Rev Public Health 2008;29:369–382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Li Y, Yin J, Cai X, Temkin-Greener J, Mukamel DB. Association of race and sites of care with pressure ulcers in high-risk nursing home residents. JAMA 2011;306(2):179–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rapp K, Lamb SE, Erhardt-Beer L, et al. Effect of a statewide fall prevention program on incidence of femoral fractures in residents of long-term care facilities. J Am Geriatr Soc 2010;58(1):70–75. [DOI] [PubMed] [Google Scholar]
- 8.Thomas KS, Hyer K, Castle NG, Branch LG, Andel R, Weech-Maldonado R. Patient safety culture and the association with safe resident care in nursing homes. Gerontologist 2012;52(6):802–811. [DOI] [PubMed] [Google Scholar]
- 9.Eriksen HM, Koch AM, Elstrom P, Nilsen RM, Harthug S, Aavitsland P. Healthcare-associated infection among residents of long-term care facilities: a cohort and nested case-control study. The Journal of hospital infection 2007;65(4):334–340. [DOI] [PubMed] [Google Scholar]
- 10.Ye Z, Mukamel DB, Huang SS, Li Y, Temkin-Greener H. Healthcare-associated pathogens and nursing home policies and practices: results from a national survey. Infect Control Hosp Epidemiol 2015;36(7):759–766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Institute of Medicine Committee on the Quality of Health Care. To err is human: building a safer health care system Washington, DC: America National Academy Press; 1999. [Google Scholar]
- 12.Bates DW, Singh H. Two Decades Since To Err Is Human: An Assessment Of Progress And Emerging Priorities In Patient Safety. Health Aff (Millwood) 2018;37(11):1736–1743. [DOI] [PubMed] [Google Scholar]
- 13.DiCuccio MH. The Relationship Between Patient Safety Culture and Patient Outcomes: A Systematic Review. Journal of patient safety 2015;11(3):35–42. [DOI] [PubMed] [Google Scholar]
- 14.Huang DT, Clermont G, Kong L, et al. Intensive care unit safety culture and outcomes: a US multicenter study. Int J Qual Health Care 2010;22(3):151–161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mardon RE, Khanna K, Sorra J, Dyer N, Famolaro T. Exploring relationships between hospital patient safety culture and adverse events. Journal of patient safety 2010;6(4):226–232. [DOI] [PubMed] [Google Scholar]
- 16.Vogus TJ, Sutcliffe KM. The impact of safety organizing, trusted leadership, and care pathways on reported medication errors in hospital nursing units. Med Care 2007;45(10):997–1002. [DOI] [PubMed] [Google Scholar]
- 17.Bonner AF, Castle NG, Perera S, Handler SM. Patient Safety Culture: A Review of the Nursing Home Literature and Recommendations for Practice. The annals of long-term care : the official journal of the American Medical Directors Association 2008;16(3):18–22. [PMC free article] [PubMed] [Google Scholar]
- 18.Castle NG, Wagner LM, Ferguson JC, Handler SM. Safety culture of nursing homes: opinions of top managers. Health Care Manage Rev 2011;36(2):175–187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Scott-Cawiezell J, Vogelsmeier A, McKenney C, Rantz M, Hicks L, Zellmer D. Moving from a culture of blame to a culture of safety in the nursing home setting. Nursing forum 2006;41(3):133–140. [DOI] [PubMed] [Google Scholar]
- 20.Smith SN, Greene MT, Mody L, Banaszak-Holl J, Petersen LD, Meddings J. Evaluation of the association between Nursing Home Survey on Patient Safety culture (NHSOPS) measures and catheter-associated urinary tract infections: results of a national collaborative. BMJ Qual Saf 2018;27(6):464–473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Castle NG, Handler S, Engberg J, Sonon K. Nursing home administrators’ opinions of the resident safety culture in nursing homes. Health Care Manage Rev 2007;32(1):66–76. [DOI] [PubMed] [Google Scholar]
- 22.Bonner AF, Castle NG, Men A, Handler SM. Certified nursing assistants’ perceptions of nursing home patient safety culture: is there a relationship to clinical outcomes? J Am Med Dir Assoc 2009;10(1):11–20. [DOI] [PubMed] [Google Scholar]
- 23.Harrington C, Carrillo H. The regulation and enforcement of federal nursing home standards, 1991–1997. Med Care Res Rev 1999;56(4):471–494. [DOI] [PubMed] [Google Scholar]
- 24.Troyer JL, Sause WL. Complaints against nursing homes: comparing two sources of complaint information and predictors of complaints. Gerontologist 2011;51(4):516–529. [DOI] [PubMed] [Google Scholar]
- 25.Hansen KE, Hyer K, Holup AA, Smith KM, Small BJ. Analyses of Complaints, Investigations of Allegations, and Deficiency Citations in United States Nursing Homes. . Med Care Res Rev 2017: epub ahead of print. [DOI] [PubMed] [Google Scholar]
- 26.Harrington C, Mullan JT, Carrillo H. State nursing home enforcement systems. J Health Polit Policy Law 2004;29(1):43–73. [DOI] [PubMed] [Google Scholar]
- 27.Harrington C, Stockton J, Hoopers S. The Effects of Regulation and Litigation on a Large For-Profit Nursing Home Chain. J Health Polit Policy Law 2014. [DOI] [PubMed]
- 28.Brauner D, Werner RM, Shippee TP, Cursio J, Sharma H, Konetzka RT. Does Nursing Home Compare Reflect Patient Safety In Nursing Homes? Health Aff (Millwood) 2018;37(11):1770–1778. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Temkin-Greener H, Campbell L, Cai X, Hasselberg MJ, Li Y. Are Post-Acute Patients with Behavioral Health Disorders Admitted to Lower-Quality Nursing Homes? Am J Geriatr Psychiatry 2018;26(6):643–654. [DOI] [PubMed] [Google Scholar]
- 30.Harrington C, O’Meara J, Kitchener M, Simon LP, Schnelle JF. Designing a report card for nursing facilities: what information is needed and why. Gerontologist 2003;43 Spec No 2:47–57. [DOI] [PubMed] [Google Scholar]
- 31.IOM. Institute of Medicine, Committee on Data Standards for Patient Safety. Patient Safety: Achieving a New Standard for Care THE NATIONAL ACADEMIES PRESS, Washington, D.C. 2004. [PubMed] [Google Scholar]
- 32.On-line resource available at https://www.ahrq.gov/sops/surveys/nursing-home/index.html Accessed September 27, 2018.
- 33.CMS. Centers for Medicare and Medicaid Services Design for Nursing Home Compare Five-Star Quality Rating System, Technical users’ guide. April 2018. Available at https://www.cms.gov/Medicare/Provider-Enrollment-and-Certification/CertificationandComplianc/downloads/usersguide.pdf.
- 34.Arling G, Cooke V, Lewis T, Perkins A, Grabowski DC, Abrahamson K. Minnesota’s provider-initiated approach yields care quality gains at participating nursing homes. Health Aff (Millwood) 2013;32(9):1631–1638. [DOI] [PubMed] [Google Scholar]
- 35.Lage DE, Rusinak D, Carr D, Grabowski DC, Ackerly DC. Creating a network of high-quality skilled nursing facilities: preliminary data on the postacute care quality improvement experiences of an accountable care organization. J Am Geriatr Soc 2015;63(4):804–808. [DOI] [PubMed] [Google Scholar]
- 36.Carayon P, Wooldridge A, Hose BZ, Salwei M, Benneyan J. Challenges And Opportunities For Improving Patient Safety Through Human Factors And Systems Engineering. Health Aff (Millwood) 2018;37(11):1862–1869. [DOI] [PMC free article] [PubMed] [Google Scholar]