Nursing Home Ratings - The Current Review

Transcription

1 ABSTRACT THE NURSING HOME FIVE STAR RATING: HOW DOES IT COMPARE TO RESIDENT AND FAMILY VIEWS OF CARE? by Anthony R. Williams In 2008, the Centers for Medicare and Medicaid Services (CMS) implemented a five-star rating system of nursing homes in the United States. These star ratings have been widely publicized both by CMS and in the national and state media. Although the components of the star rating system take into account various dimensions of quality, the satisfaction of nursing home residents and their families are not taken into consideration. The current study compares the CMS star rating system to the satisfaction ratings of nursing homes provided by residents and their families throughout the state of Ohio. Findings indicate that the star rating system does not adequately reflect consumer satisfaction and recommend that the star rating system be refined to include a consumer component.

2 THE NURSING HOME FIVE STAR RATING: HOW DOES IT COMPARE TO RESIDENT AND FAMILY VIEWS OF CARE? A Thesis Submitted to the Faculty of Miami University in partial fulfillment of the requirements for the degree of Master of Gerontological Studies Department of Sociology and Gerontology by Anthony R. Williams Miami University Oxford, Ohio 2012 Advisor Robert Applebaum, PhD Reader Jane Straker, PhD Reader Douglas Noe, PhD

3 Table of Contents Chapter 1: Introduction....1 Literature Review...2 Dimensions of quality..2 Role of quality of life and consumer satisfaction...2 Usefulness of quality report card information. 3 Publication of quality report cards and deduced impacts....4 Relevant comparisons of quality dimensions..4 Reliability and validity of Nursing Home Compare...5 Reliability and validity of consumer satisfaction data.7 Focus of Study.8 Chapter 2: Methodology..9 Data Background and development of Ohio s Long-Term Care Consumer Guide Implementation of the 2010 Ohio Nursing Home Family Satisfaction Survey.11 Implementation of the 2009 Ohio Nursing Home Resident Satisfaction Survey..11 Calculation of scores for family and resident satisfaction surveys 12 Background and overview of Nursing Home Compare data.12 Nursing Home Compare data selection.14 Master dataset development Approach to analysis..16 Chapter 3: Results...18 Overall Star Rating in Relation to Consumer Satisfaction Health Inspection Domain in Relation to Consumer Satisfaction.23 Quality Domain in Relation to Consumer Satisfaction..28 Staffing Domain in Relation to Consumer Satisfaction.33 Impact of Consumer Group Summary of Results...38 Chapter 4: Discussion 40 Discrepancies Relative to the Overall Star Rating...42 Discrepancies Relative to the Health Inspection, Quality, and Staffing Domains 42 Family and Resident Views...43 Rationale for Differences and Potential Solutions for CMS Study Limitations and Ideas for Further Research...44 Conclusions...45 References..46 Appendix A Ohio Family Satisfaction Survey Appendix B Ohio Resident Satisfaction Survey..61 ii

4 List of Tables Table 1. Quantity of Nursing Homes Possessing both Satisfaction and NHC data...15 Table 2. Family Satisfaction Means by Overall Star Rating.20 Table 3. Resident Satisfaction Means by Overall Star Rating Table 4. Family Satisfaction Category by Overall Star Rating...21 Table 5. Resident Satisfaction Category by Overall Star Rating...22 Table 6. Family Satisfaction Means by Health Inspection Star Rating.25 Table 7. Resident Satisfaction Means by Health Inspection Star Rating.. 25 Table 8. Family Satisfaction Category by Health Inspection Star Rating.26 Table 9. Resident Satisfaction Category by Health Inspection Star Rating.. 27 Table 10. Family Satisfaction Means by Quality Star Rating Table 11. Resident Satisfaction Means by Quality Star Rating.30 Table 12. Family Satisfaction Category by Quality Star Rating Table 13. Resident Satisfaction Category by Quality Star Rating.32 Table 14. Family Satisfaction Means by Staffing Star Rating...35 Table 15. Resident Satisfaction Means by Staffing Star Rating 35 Table 16. Family Satisfaction Category by Staffing Star Rating...36 Table 17. Resident Satisfaction Category by Staffing Star Rating 37 Table 18. Results Summary...41 iii

5 Acknowledgements I would like to express my sincere gratitude to my committee, the faculty and staff in the Department of Sociology and Gerontology as well as Scripps Gerontology Center, and my fellow students in the department. The support and guidance you have all given me, through both the critical inquiry process and over the past two years, has allowed me to develop into the young professional that I am today. I thank all of you for your encouragement and am very much grateful to have had the opportunity to get to know you all. iv

6 Chapter 1: Introduction In 1998, the Centers for Medicare and Medicaid Services (CMS) launched an online quality report card that was designed to provide consumers with information on every Medicareand Medicaid-certified nursing home in the United States ( (Castle, 2009; Castle, 2010). This report card, called Nursing Home Compare (NHC), originally reported only on basic nursing home characteristics, but has since grown to include an array of quality information (Grabowski & Town, 2011). In 2002, CMS launched a powerful nationwide media campaign consisting of television and newspaper advertisements that promoted NHC as a major source of information on nursing home quality (Grabowski & Town, 2011; Office of Inspector General, 2004). This campaign was greeted with great success, generating about 400,000 visits to the site the month of the campaign launch compared to the fewer than 100,000 visits per month the site was previously generating (Office of Inspector General, 2004). Along with its growth in popularity, NHC has continued to develop throughout the years morphing into its current form. Today, the site publicly rates nursing homes on a five-star scale based on data related to staffing, performance on quality indicators taken from Minimum Data Set (MDS) information, and information gathered through health inspections (Centers for Medicare and Medicaid Services [CMS], 2010a). While these measures include some important elements of nursing home quality, two crucial pieces of information are missing: data on family and resident satisfaction. These data, on nursing home family and resident satisfaction, are not readily available nationwide, as collection requires significant time and fiscal commitments. While the number of states collecting this information is constantly in flux, a review of nursing home report cards conducted in 2009 found only six states publicly reporting family and/or resident satisfaction data; of these, only two states, Rhode Island and Ohio, reported satisfaction data from both groups (Castle, Diesel, & Ferguson, 2010). In Ohio, these data are collected in accordance with Ohio House Bill 403 (2000), which mandates that family and resident satisfaction data be gathered and publicly reported in Ohio s Long-Term Care Consumer Guide ( Little research has compared the ratings reported on NHC and satisfaction data reported by families and residents. The only study previously published examined the link between NHC 1

7 star ratings and family satisfaction in the state of Maryland (Çalıkoğlu et al., 2011). The findings from that study showed somewhat mixed results. The current study will help fill this void in the literature by being the first to compare NHC to state level satisfaction data from both nursing home residents and their families. Ultimately, the implicit assumption that individuals will be more satisfied with nursing homes that rate well on NHC will be put to the test. If the data indicate that NHC ratings do not adequately represent consumer satisfaction, then it may be suggested that a measure of consumer satisfaction should be considered for inclusion in NHC s star rating system. Conversely, if the data suggest that NHC ratings represent consumer satisfaction well, one might question whether the spending by state governments to conduct nursing home consumer satisfaction surveys could be better spent elsewhere. Literature Review Dimensions of quality. Quality in nursing homes is a multidimensional construct and cannot be assessed completely through examination of a single dimension (Castle & Ferguson, 2010; Mukamel & Spector, 2003). Some analysts have divided quality in nursing homes into two categories, hotel services and clinical care (Mukamel & Spector, 2003; Mukamel, et al., 2010). Hotel services, which encompass items such as the friendliness of staff members, environmental conditions, and food quality, are generally straightforward conceptually. Clinical care, on the other hand, can include indicators such as a nursing home s success at preventing pressure ulcers or the quality of the rehabilitative services offered. These can be more difficult to compare between nursing homes because of case-mix differences within and across facilities. Quality in nursing homes can further be broken down using Donabedian s approach to quality through structure, process, and outcome (Castle & Ferguson, 2010; Mukamel & Spector, 2003). Utilizing this framework, structure based indicators include those topics that relate to the organizational and physical aspects of a nursing home such as staffing ratios or the presence of training programs. Process based indicators, such as turning bed bound residents or providing residents with flu shots, deal with the processes that are completed to or for residents. The third, outcome based indicators, examine the desired states that residents would prefer to achieve and are often influenced by both structure and process based indicators. Role of quality of life and consumer satisfaction. Quality indicators for nursing homes have often been subject to the criticism that they are too heavily focused on clinical aspects and 2

8 lacking focus on quality of life measures (Castle & Ferguson, 2010). An examination of the current quality measures reported on the NHC website supports this, showing 19 out of the 19 publicly reported measures (10 of which are used to determine the quality domain rating) that deal with resident s physical and clinical needs and zero directly examining aspects of quality of life (CMS, 2010a). While clinical data are undoubtedly important to examine in nursing homes, the fact that residents also often reside in these facilities for long periods of time makes the evaluation of quality of life measures important as well. The examination and utilization of quality of life measures has continued to become more prevalent as time has passed. Proprietary firms, researchers, nursing home trade associations, and government entities at both the state and federal levels have begun developing and implementing nursing home quality of life surveys (Ejaz & Castle, 2007). At the federal level, the introduction of extensive interview-based assessments utilized in the new MDS 3.0 is a move in this direction (Saliba & Buchanan, 2008). Note that it is unclear exactly what role these will play in NHC as they begin being utilized on the site in 2012 (CMS, 2012). Usefulness of quality report card information. Quality report cards are often designed with the intention of helping consumers compare quality across nursing homes, especially among dimensions that may not be easily observable when visiting a facility (Castle & Lowe, 2005; Mukamel & Spector, 2003; Mukamel et al., 2007; Stevenson, 2006). They can serve as valuable tools to consumers when they offer accurate and useful data, provide information on the significance of the measures utilized, and allow for consumers to compare indicators of quality across nursing homes (Castle & Lowe, 2003; Mukamel & Spector, 2003; Stevenson, 2006). In an ideal world, quality report cards would serve as a method of augmenting market competition and, as a result, increase quality across nursing homes; though, it has been suggested that this market driver may not work as well as intended primarily as a result of two factors: 1) noncompetitive markets (i.e., markets with overwhelming demand or a lack of competition); and 2) public payers providing the bulk of the funding for long-term care (Castle & Lowe, 2005; Mukamel et al., 2007). Quality data, both within and outside quality report cards, are also used by some states in their adoption of nursing home pay-for-performance measures. According to a national survey of state Medicaid agencies conducted between 2008 and 2009, eight states including Ohio were found to have pay-for-performance measures in place for nursing homes (Werner et al., 2009). 3

9 Under Ohio Senate Bill 264 (2011), Ohio utilizes pay-for-performance measures for nursing home participating in the Medicaid program. A scoring system is utilized in which nursing homes receive a quality payment for meeting standards including, but not limited to, minimum scores on Ohio s family and resident satisfaction surveys, performing at a certain level on health inspection surveys, and retaining frontline nursing staff. Publication of quality report cards and deduced impacts. Quality report cards in long-term care have taken many forms and have been published by a number of different entities (Harris-Kojetin & Stone, 2007). An assessment of state websites conducted in 2009 found 26 states with publicly available quality report cards (Castle et al., 2011). The contents of these reports varied in terms of format, type, and volume of information provided. Various studies have examined the results of public reporting on nursing homes. One study associated the implementation of public reporting of quality data (namely NHC) with improvements on some quality indicators and not on others (Mukamel et al., 2008). Another suggested that public reporting resulted in an increase in clinical spending at nursing homes relative to spending on hotel services (Mukamel et al., 2010). When looking at the impact of NHC on demand for nursing home services, researchers have found a lack of evidence suggestive of any change in demand (Grabowski & Town, 2011; Stevenson, 2006). From a fiscal point of view, the results of a study by Park et al. (2011) suggested that nursing homes that improved on publicly reporting quality indicators received higher profit margins and revenues after public reporting began, primarily resulting from increased reimbursement by Medicare. Research has also suggested that report card data may be of use to administration teams in nursing homes. Nursing home administrators usually report reviewing the data (Castle, 2005; Mukamel et al., 2007). In one study, 33% of administrators reported that they were utilizing the data in current quality improvement efforts and 51% stated that they planned to utilize the data in the future (Castle, 2005). In another study, 63% of administrators reported investigating their scores, 42% reported reprioritizing existing quality initiatives, and 20% reported being motivated to start new quality initiatives (Mukamel et al., 2007). Relevant comparisons of quality dimensions. A number of studies have looked for relationships among different dimensions of nursing home quality. Two of these, published recently, hold high relevance to the current study, as they examine the relationship between consumer satisfaction and regulatory data. One of these, mentioned earlier, investigated the 4

10 relationship between NHC ratings and Maryland state level family satisfaction data (Çalıkoğlu et al., 2011). Results suggested a relationship between the family satisfaction data and NHC overall rating as well as the health inspection and staffing domains while no relationship was found between family satisfaction and the NHC quality domain. The second examined the relationship between both family and resident satisfaction scores from a small sample of nursing homes as they related to health inspection deficiency citations received by the same facilities. Results showed a moderate positive correlation between both family and resident satisfaction data and the number of deficiency citations received by nursing homes (Tellis-Nayak et al., 2010). Research has also focused heavily on the comparability of family and resident satisfaction data in nursing homes. Prior studies suggest that these two are not interchangeable and, in most cases, have found that families tend to report higher satisfaction than residents on most indicators (Castle, 2006; Gasquet et al., 2003). It should be noted that this directionality does not always exist. In fact, Tellis-Nayak, Shiverick, and Hernandez (2010) found the reverse for the majority of the comparative indicators they examined. A review of previous literature found no studies where family and resident satisfaction data generated identical results. The current study examines both sources of data in relationship to NHC. Reliability and validity of Nursing Home Compare. The validity of the NHC star rating system has been drawn into question by many. Larry Minnix, president and chief executive officer of the industry trade group LeadingAge, described the system as poorly planned, prematurely implemented and ham-handedly rolled out (2008). Bruce Yarwood, president and chief executive officer of the American Health Care Association and National Center for Assisted Living, echoed these comments, stating that the rating system was prematurely implemented without appropriate testing and review and that the rating system is based on the broken survey system, lacks the inclusion of other important quality elements that help consumers make informed decisions, and includes inaccurate data (2008). An overview of the major concerns with each of the three NHC domains that make up the overall score is provided below. The health inspection domain is subject to criticism for a variety of reasons. The largest issue within this domain deals with the absence of reliability and high levels of variability in the survey processes. A 2003 report published by the Office of the Inspector General concluded that 5

11 there was variability in how states went about determining citation practices as a result of inconsistent survey focuses, high staff turnover, unclear guidelines, and the lack of a common review process for draft survey reports (OIG, 2003). More recently, analyses of CMS federal surveys found about 70 percent of state surveys that were compared to CMS federal comparative surveys had missed at least one deficiency citation; the average number of deficiencies missed also varied significantly across states (Governmental Accountability Office [GAO], 2008). As NHC utilizes intra-state data when determining health inspection ratings, it is significant to note that the variation does not stop at the inter-state level. When examining variations in inspection data, a recent study found that different survey districts within states also had variability beyond what could be explained simply by regional variations in quality (Miller & Mor, 2008). CMS does acknowledge that issues exist and indicate that federally conducted validation (comparative and observational) surveys with mandatory follow-ups, appeals and informal dispute resolutions, as well as the new Quality Indicator Survey (QIS) currently being rolled out are tools in place to attempt to increase the validity of the survey process (CMS, 2008). Within the quality domain, the primary concern is that the data are based on the MDS, which is collected and submitted by nursing home staff. Nursing home administrators argue that variations in data coding, unaccounted differences in case-mix, and unusual events impact performance on these quality measures (Mukamel et al., 2007). Research has backed these beliefs, identifying variability between facilities in the way that MDS data are completed as well as the presence of potential issues with the risk adjustment performed on the data (Mor, 2005; Rahman & Applebaum, 2009). Additionally, there is the threat that nursing home staff may purposely doctor the data to make it advantageous to the facility (Rahman & Applebaum, 2009). To counter these concerns, CMS argues that the measures selected for inclusion in the actual ratings determination were selected based on the extent to which the measures were under the facility s control, their reliability and validity, their statistical performance, and their importance (CMS, 2008; CMS, 2010a). Additionally, CMS indicates that some measures patrol themselves; an example CMS has utilized is that indicating a lower rate of residents who developed pressure ulcers would result in lower reimbursement (CMS, 2008). The staffing domain is also subject to the self-report criticism faced by the quality domain. Some suggest that the data utilized, which comes from the Online Survey Certification and Reporting system (OSCAR), may not be completely accurate (PointRight Inc., 2008). There 6

12 has also been concern that the two-week period from which staffing data are obtained may not be representative of staffing that occurs year round (PointRight Inc., 2008). A study conducted by Kash et al. (2007) supported these concerns, finding that OSCAR staffing data typically overreported staffing levels when examined against Medicaid cost report data. To counter these concerns, CMS attempts to identify information that is unreliable through utilization of exclusion criteria to recognize outlier staffing levels, though these may not be successful at detecting all incorrect data (CMS, 2008). CMS is also currently exploring the possibilities of utilizing payroll information from providers as a source for more accurate reporting (Colorado Foundation for Medical Care, 2005). Reliability and validity of consumer satisfaction data. The reliability and validity of consumer satisfaction data, specifically that which is utilized in the current study, should also be taken into consideration. Although the initial development of both the Ohio Nursing Home Family and Resident Satisfaction Surveys included steps to ensure that reliability and validity, some threats still exist. The largest concern with regard to the Ohio Nursing Home Family Satisfaction Survey is the response biases that may have occurred. These biases could be attributed to a number of factors. First, as nursing homes were responsible for the mailing of surveys to family members (Straker et al., 2011), it is unclear if a family member of every resident received a survey in the first place. It is possible that nursing homes may have excluded family members of residents who they believed would express negative opinions of their facilities, or possibly excluded individuals just as a result of errors in distribution. Family members also expressed a number of issues via a survey helpline that likely hindered their participation. These issues included a lack of comfort in discussing their experiences with nursing homes, a lack of willingness to participate due to providing input in previous family satisfaction surveys, concerns with the confidentiality and anonymity of the survey responses, and a lack of time or interest in filling out the survey. Although the survey team tried to reassure family members with regard to some of their concerns (e.g., the confidentiality and anonymity of the data), it is likely that many of these issues resulted in nonresponse by some family members (Straker et al., 2011). With regard to the Ohio Nursing Home Resident Satisfaction Survey, the collection of data from nursing home residents with cognitive impairments, sensory impairments, and health 7

13 conditions has in previous years been identified as a threat to the reliability and validity of nursing home data collected from residents (IOM, 2001). While this may be true in some instances, researchers have found that it is feasible to collect reliable and valid data directly from cognitively impaired nursing home residents (Applebaum et al., 2006; Kane et al., 2004; Sangl et al., 2007; Straker et al., 2007; Uman et al., 2000; Wheatley et al., 2007). The Ohio Nursing Home Resident Satisfaction Survey took many steps to promote the usability of the survey by a broad array of individuals. These steps included, but were not limited to, the utilization of faceto-face interviews, extensive screening and training of interviewers, and the use of large print cards that could be utilized to remind residents of response categories (VR, 2010). Additionally, interviews were stopped if residents were unable to answer multiple questions (VR, 2010). A second concern worth noting is that residents may have felt uncomfortable speaking negatively about their nursing homes (Applebaum et al., 2006). To try to promote the comfort of residents in expressing their true opinions, each satisfaction survey was conducted in a private location selected by the resident. All residents were also able to refuse to participate and could discontinue the interview at any time if they chose to do so (VR, 2010). Focus of Study As NHC has become a popular resource for information on nursing homes, the current study takes advantage of the family and resident satisfaction data available in the state of Ohio to examine how the views of each consumer group compare to the star ratings reported on NHC. If the ratings of families and residents are comparable to NHC, then we can assume the CMS star rating system covers the views of consumers. However, if the ratings of families and residents are not comparable to NHC, we might want to consider how the views of consumers can be better integrated into the NHC system. 8

14 Chapter 2: Methodology The current study brings into question whether the star ratings provided on NHC accurately reflect consumers satisfaction with nursing homes. In order to address this inquiry, a comparative approach was applied utilizing secondary data from NHC and Ohio s Nursing Home Resident and Family Satisfaction Surveys. Two research questions served as the foundation for this study: 1. How do the star ratings on NHC compare to residents subjective ratings of satisfaction with their nursing home? 2. How do the star ratings on NHC compare to families subjective ratings of satisfaction with the nursing homes in which family members are located? The first question, focusing on examining the relationship between the NHC star ratings and resident satisfaction, is novel to the literature. Although this direct relationship has not yet been addressed, one would assume that residents in facilities with higher star ratings in any given NHC domain would have higher resident satisfaction scores relative to other nursing homes in the same domain. One way to attempt to capture some insight into what results may exist is by examining the research conducted utilizing the underlying factors for each domain. With regard to the health inspection domain, prior research on how health inspection data relates to resident satisfaction suggests that higher NHC health inspection ratings would result in higher levels of resident satisfaction (Tellis-Nayak et al., 2010). Within the NHC staffing domain, prior research utilizing staffing data from the OSCAR database (same database utilized by NHC to obtain staffing data) found that nursing homes reporting higher staffing levels on OSCAR have had significantly higher levels of resident satisfaction (Lucas et al., 2007). When examining the quality domain, no prior research could be identified that examined the relationships between the set of measures utilized for the NHC quality domain and resident satisfaction data. One would assume that higher ratings in the quality domain would also be associated with higher resident satisfaction scores, though an examination of how NHC quality measures reflects on family satisfaction data has not supported this postulation (Çalıkoğlu et al., 2011). From these assumptions and research on the primary underlying data utilized to calculate the ratings, it can be assumed that nursing homes with higher NHC overall ratings would have, on average, higher resident satisfaction scores. 9

15 The second question, examining the relationship between NHC and family satisfaction, has been tested once previously in the literature. The previous study found that family satisfaction data trended in agreement with NHC overall rating as well as the health inspection and staffing domains. No clear pattern was found when examining how the NHC quality domain related to the family satisfaction data (Çalıkoğlu et al., 2011). Looking at the underlying factors here as well, research could only be identified examining the relationship between health inspection performance and family satisfaction; findings suggested that better performance on health inspections results in better family satisfaction scores (Tellis-Nayak, 2010). Data Several secondary data sources were utilized in the current study. These included overall family and resident satisfaction scores that were collected for inclusion in Ohio s Long-Term Care Consumer Guide, as well as two sets of archived data that were previously published on NHC. An overview of each of these datasets is provided below. Background and development of Ohio s Long-Term Care Consumer Guide. Under Ohio House Bill 403 (2000), the Ohio Department of Aging was required to develop the Ohio Long-term Care Consumer Guide. The legislation also required the development of distinct, but complementary, surveys for examining the satisfaction of nursing home residents and their families. Researchers were originally given a six to seven month time frame to develop, ensure the reliability and validity of, and pre-test the surveys. Currently, the surveys are being administered to each group, family members and residents, in alternating years (Ejaz et al., 2003; Ejaz et al., 2007; Straker et al., 2007) The development of both surveys was guided by literature on nursing home satisfaction, previously collected nursing home satisfaction data from various entities, and an advisory council consisting of a broad array of stakeholders. Both surveys were pre-tested utilizing cognitive interview methods on nursing home residents, as researchers believed the core set of questions and response categories for both surveys should be based on resident input. Following initial modifications based on nursing home resident input, the family satisfaction instrument was further tested on a small group of family members. After initial testing, each survey was administered to its respective group in larger scale tests. Following administration of the surveys to larger groups, the reliability and validity of the instruments were tested by examining testretest and inter-rater reliability among the resident sample as well as by conducting follow-up 10

16 interviews with family members who completed the survey. Researchers also conducted additional cognitive interviews with residents to examine response categories. Based on the findings of these evaluations, factor analyses conducted on the data, and investigator judgment, final versions of each survey were developed for statewide utilization (Ejaz et al., 2003; Ejaz et al., 2007; Straker et al., 2007). Implementation of the 2010 Ohio Nursing Home Family Satisfaction Survey. The 2010 Ohio Nursing Home Family Satisfaction survey consisted of 17 background questions and 54 close-ended satisfaction questions from 11 domains. The domains included were activities, administration, admissions, choices, direct care and nursing staff, environment, laundry, meals and dining, social services, therapy, and general questions. The response categories ranged from yes, always to no, never and all options were presented concurrently. The survey instrument utilized for the family survey has been included as appendix A. All surveys were returned to Scripps Gerontology Center and facilities had no knowledge of who completed surveys. In all, surveys were returned from family members of residents in 931 (97%) of the 961 nursing homes provided with survey packets. The total number of family surveys returned to Scripps Gerontology Center was 29,876; this is estimated to be about a 44.8% response rate, though the exact percentage is unknown as not all nursing homes reported the number of surveys mailed (Straker et al., 2011). Implementation of the 2009 Ohio Nursing Home Resident Satisfaction Survey. The 2009 Ohio Nursing Home Resident Satisfaction Survey consisted of one screening question and 50 close-ended satisfaction questions. These questions encompassed ten domains including activities, administration, choice, direct care and nurse assistants, environment, laundry, meals and dining, social services, therapy, and overall/general satisfaction. The response categories for these items were branched, wherein residents first answered yes or no to questions then, if they did not specify based on their knowledge of the answer categories, were asked to elaborate by choosing always or sometimes if yes was answered at first or hardly ever or never if no was answered at first. The survey instrument utilized for the resident survey has been included as appendix B. Surveys were administered to consenting residents in a private location through face-toface interviews conducted by interviewers recruited and trained by Vital Research staff, who were contracted to conduct the survey. The quantity of interviews conducted in each facility 11

17 varied. Researchers selected quantities of residents at each facility based upon the number of residents eligible to participate and a 10% margin of error that was mandated by the Ohio Department of Aging. Residents were randomly selected from a census provided to Vital Research two weeks prior to the interview team s visit. Residents who were in isolation or whose legal guardian declined participation on the behalf of the resident were not included in the sampling process. In all, 955 out of 957 facilities eligible for participation across the state of Ohio participated in the survey. Within these facilities, a total of 28,412 interviews were requested with residents. Of these, 23,462 residents were willing to participate, yielding a resident participation rate of 83%. Of the 23,462 residents who were willing, interviews were able to be completed with 22,822. Started, but incomplete, interviews resulted from resident inability to respond (n = 347), resident refusal to continue (n = 137), fatigue (n = 72), and other reasons (n = 84) (Vital Research, 2010). Calculation of scores for family and resident satisfaction surveys. The scores for the satisfaction data on both the family and resident satisfaction surveys were calculated in an identical manner. For each satisfaction survey question, answers were recoded on a 101-point scale as follows: yes, always = 100; yes, sometimes = 67; no, hardly ever = 33; and no, never = 0. Responses of don t know/not applicable and blanks were recoded as missing. Domain scores were calculated as the mean from all questions within a domain; these were only calculated if the party, resident or family member, completed all or all but one of the questions in the domain. A composite (overall) satisfaction score was computed for each individual as the mean on all the domain scores calculated for each resident. Data at each level (question, domain, composite) were averaged across all individuals at each facility and this determined each facility s scores (Straker et al., 2011; Vital Research, 2010; Wheatley et al., 2007). Background and overview of Nursing Home Compare data. NHC was launched in 1998 by CMS to provide consumers with information on each Medicare- and Medicaid-certified nursing home in the United States (Castle, 2009; Castle, 2010). Initially, the report only provided information on deficiency citations, facility characteristics, and resident characteristics (Castle, 2009; Stevenson, 2006). In 2000, staffing information was also added to NHC (Castle, 2009). Following a pilot study in six states, including Ohio, quality measures were added 12

18 nationally to the report card in 2002 (Castle, 2009). These measures have changed over time and currently 19 quality measures are reported (Castle, 2009; Castle, 2010). Through consultation with an expert panel consisting of individuals from academia, patient advocacy groups, and nursing home provider groups, CMS developed and implemented a significant change to the NHC system in December of 2008 (CMS, 2008). This change was the implementation of a five-star rating system on NHC. These star ratings serve as the primary NHC data utilized for this study. This system includes an overall star rating developed from ratings in three domains based on health inspections, quality measures, and staffing (CMS, 2010a). Each rating is based on a five point scale with one representing much below average and five representing much above average (CMS, 2010a). The rating for the health inspection domain for each nursing home is calculated based on the severity, scope, and number of deficiencies obtained from health inspection surveys (including complaint surveys) over the preceding three years and data are compared separately for each state. The top 10% of nursing homes in each state in this domain receive a rating of five, the middle 70% receive ratings of four, three, or two (in equal proportions), and the lowest 20% receive a rating of one. In the quality domain, a rating is obtained based on each nursing home s performance on 10 MDS-based measures (note that these are a subset of the full list of quality measures reported on NHC). A rating for the quality domain is based on the national distribution of performance (as opposed to state-by-state) utilizing the same percentile distributions as the health inspections rating. A rating for the staffing domain for each nursing home is calculated based on RN hours per resident day and total staffing hours (RN, LPN, and nurse aide) per resident day. Both of these are case-mix adjusted and each of the two measures is weighted equally in calculating the score for staffing. The rating distribution is based on a mixture of percentile rankings relative to free-standing facilities nationwide and, additionally, whether a nursing home meets CMS identified staffing thresholds (CMS, 2010a). The overall rating on NHC for each facility is based on the ratings for the three domains discussed above. To obtain the overall rating for a nursing home, one would first begin with the rating that the facility received in the health inspection domain; this serves as the base score. Then, if the nursing home received four or five in the staffing domain, the base score is increased by one. Alternatively, if the nursing home received a score of one in the staffing domain, the base score is decreased by one. An additional increase by one to the base score occurs if the 13

19 nursing home received a five star rating in the quality domain or, conversely, a decrease by one occurs if the nursing home received a one in the quality domain. For all other actions not mentioned, no adjustment to the base rating occurs. This altered rating serves as the nursing home s overall rating. The exceptions to this rule are as follows: 1) nursing homes that receive a one in the health inspections domain can only be upgraded by, at most, one; and 2) nursing homes designated as special focus facilities can have an overall rating no higher than three (CMS, 2010a). Nursing Home Compare data selection. The current study utilizes NHC data that were archived on two separate dates. NHC generally archives data four times per year, limiting the specific data that could be selected for utilization. For comparison to the 2010 family satisfaction data, archived data from December 20, 2010 were utilized. The rationale behind selecting this data was that only three data sets were archived in 2010, and two of the three had little or no overlap with the actual dates of family satisfaction survey administration; meaning that the majority of the health inspection data were collected prior to the families being surveyed. Additionally, although the data set did not contain the greatest number of health inspections, the differences in the number of health inspections between each archived 2010 dataset were miniscule. For comparison to the 2009 resident satisfaction data, NHC data that were archived on December 23, 2009 was utilized. The rationale behind selecting this data is that it was based on the greatest number of health inspections conducted and, additionally, had the greatest overlap period with the period of time that the resident satisfaction survey data were being collected. Master dataset development. Utilizing unique identifiers assigned to nursing homes to link nursing home information from the different data sources (Medicare Provider Numbers, Facility Identification Numbers, and State Identification Numbers), facility data from each source were aggregated into a master file on SPSS 19 statistical software. Following compilation of the data sources, data were cleaned to develop the dataset utilized in the current study. The pre-cleaned dataset consisted of 973 nursing homes containing data from at least one of the four primary datasets utilized for analyses in the current study (2010 family satisfaction, 2009 resident satisfaction, 2010 NHC, or 2009 NHC). From this dataset, 14 nursing homes were removed due to not having NHC data for neither of the years; it is likely that these facilities did 14

20 not participate in Medicare or Medicaid. Thirty-five additional facilities were hospital-based and also dropped from the sample. In a third step, five more facilities containing only one year of NHC data and no other survey data from other sources were also excluded, as there would be no ability to make comparisons with these. In a final step, one additional nursing home, which possessed only 2010 family satisfaction data and 2009 NHC data, was also removed. After cleaning, the final dataset consisted of 918 nursing homes. Of these, 880 (95.9%) possessed data from all four primary datasets. 892 (97.2%) possessed both an overall family satisfaction score and a corresponding NHC overall star rating. 907 (98.9%) possessed both an overall resident satisfaction score and a corresponding NHC overall star rating. Additional data regarding the quantity possessing satisfaction data by each type of star rating can be found in Table 1. Table 1 Quantity of Nursing Homes Possessing both Satisfaction and NHC data Possess Overall Satisfaction Score Corresponding NHC Data Overall Star Health Inspection Star Quality Measures Star Staffing Star Family Data Resident Data Note. Data from 2010 and 2009 Nursing Home Compare, 2010 Ohio Nursing Home Family Satisfaction Surveys, and 2009 Ohio Nursing Home Resident Satisfaction Surveys. 15

21 Approach to Analysis Information from each component of the NHC dataset was analyzed separately. The first set of analyses utilized the 2010 overall family satisfaction scores and the 2010 NHC overall and domain star ratings. The second set of analyses utilized the 2009 overall resident satisfaction scores and the 2009 NHC overall and domain star ratings. Two phases of data analyses were conducted. The first phase utilized an absolute approach to examining the data; in this phase, the data were kept on their original scales. In the second phase, a relative approach was utilized in which NHC star ratings and overall consumer satisfaction scores were compared on relative scales. In the first phase of analyses, the data were analyzed utilizing a series of one-way ANOVAs with Scheffé post-hoc tests. Utilizing this method, mean overall satisfaction scores for nursing homes at each star rating level (one star through five stars) of each NHC rating were compared. Results allowed for examination of mean satisfaction scores by star rating level and also for identification of any significant differences in satisfaction scores between each possible pair of levels (e.g., one star versus five stars; one star versus four stars; etc.) within each NHC rating. In the second phase of analyses, overall consumer satisfaction data were converted to five-point ordinal scales ranging from very low satisfaction to very high satisfaction. Utilizing this method, it was possible to compare both NHC and satisfaction data on similar metrics. The first comparison, for instance, was conducted between overall family satisfaction scores and NHC overall star ratings from First, any nursing homes that did not possess both an overall family satisfaction score and a NHC overall star rating were removed from the population. Then, the frequency distribution of each NHC star rating among the population was obtained. The same frequency distribution was then applied to the nursing homes based on their overall family satisfaction score. The 180 lowest performing nursing homes (quantity of 180 equal to the number of nursing homes receiving one star in the NHC overall domain) in terms of overall family satisfaction were categorized as having very low satisfaction. The 98 highest performing nursing homes (quantity of 98 equal to the number of nursing homes receiving five stars on the NHC overall domain) in terms of overall family satisfaction were categorized as having very high satisfaction. Data were distributed similarly in the middle three categories. 16

22 This same methodological approach was then applied to all other possible pairs relating consumer satisfaction data to each NHC domain. The data were then analyzed utilizing crosstabulations. By utilizing cross-tabulations, it was possible to examine how performance on NHC compared to performance on consumer satisfaction. A mean score for each satisfaction category was also calculated. 17

23 Chapter 3: Results In this section, the data from each of the NHC star ratings is compared to consumer satisfaction data. The first set of analyses in each domain allow for examination of the mean satisfaction scores of nursing homes broken down by the star ratings received within each domain. By doing this, it is possible to see if, on average, nursing homes rated higher on the NHC star rating system have higher consumer satisfaction scores. In the second set of analyses, nursing home consumer satisfaction ratings are distributed to nursing homes utilizing the distribution of NHC star ratings within each domain. Overall Star Rating in Relation to Consumer Satisfaction Mean comparisons for family and resident satisfaction data across levels of star ratings within the NHC overall domain are located in Tables 2 and 3. The tables show both mean scores at each level, as well as provide an overview of where significant differences existed between the mean satisfaction scores at different levels. Significant differences were identified in the tables at p <.05 and p <.01. Mean satisfaction scores for both family and resident data tended to increase as the number of stars received in the NHC overall domain increased. Family satisfaction scores ranged from (one star) to (five stars). Resident satisfaction scores had slightly less spread, ranging from (one star) to (five stars). Within the family data, examination of whether significant differences existed in mean satisfaction scores revealed that not all mean satisfaction scores varied significantly from neighboring scores (e.g., one star may not have possessed a significantly different mean satisfaction score than two stars, etc.). Within the family data, the mean satisfaction score at one star was significantly less than the mean satisfaction scores at three, four, and five stars, but not significantly different from the mean satisfaction score at two stars. Similarly, the mean satisfaction score at the five stars was significantly greater than the mean satisfaction scores at three, two, and one stars, but not significantly different from the mean satisfaction score at four stars. Within the other mean satisfaction ratings, significant differences were found in the following relationships: the mean satisfaction score at two stars was significantly less than the mean satisfaction scores at three, four, and five stars; the mean satisfaction score at three stars was significantly greater than the mean satisfaction scores at one and two stars and significantly 18

24 less than the mean satisfaction score at five stars; and the mean satisfaction score at four stars was significantly greater than the mean satisfaction scores at one and two stars. The data on resident satisfaction showed much weaker levels of significance between values. Within the data, mean satisfaction score at one star was significantly less than the mean satisfaction scores at three, four, and five stars. No other significant differences emerged. Categorical comparisons between consumer satisfaction categories and the NHC overall ratings are displayed in Tables 4 and 5. The tables display the percentage of relative rankings for family and resident satisfaction based on the NHC overall rating assigned to each home. The categories that one might expect to correspond to the greatest extent (e.g., very low satisfaction and one star ) are indicated by boxes within each table. The tables also display the mean scores for each satisfaction category. Within the family data, of the nursing homes that received one star on the NHC overall rating, 41.1% were also categorized as having very low satisfaction. Of those that received one star, 63.3% were categorized as having either very low satisfaction or low satisfaction. In contrast, 20.0% that received one star on the NHC overall rating were categorized as having high or very high satisfaction. Of the nursing homes that received five stars on the NHC overall rating, 22.4% were also categorized as having very high satisfaction. Of those that received five stars, 59.1% were categorized as having either very high satisfaction or high satisfaction. In contrast, 40.8% that received five stars on the NHC overall rating were categorized as having moderate, low, or very low satisfaction. Within the resident data, of the nursing homes that received one star on the NHC overall rating, 36.2% were also categorized as having very low satisfaction. Of those that received one star, 59.9% were categorized as having either very low satisfaction or low satisfaction. In contrast, 19.3% that received one star on the NHC overall rating were categorized as having high or very high satisfaction. Of the nursing homes that received five stars on the NHC overall rating, 26.3% were also categorized as having very high satisfaction. Of those that received five stars, 46% were categorized as having either very high satisfaction or high satisfaction. In contrast, 53.9% receiving five stars in the NHC overall rating were categorized as having moderate, low, or very low satisfaction. 19

25 Table 2 Family Satisfaction Means by Overall Star Rating Overall Star Rating Family Satisfaction Mean P-Value between Family Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star **.000**.000** 2 stars **.000**.000** 3 stars **.003** * 4 stars **.000** stars **.000**.011* Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. * p <.05, ** p <.01 Table 3 Resident Satisfaction Means by Overall Star Rating Overall Star Rating Resident Satisfaction Mean P-Value between Resident Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star **.000**.000** 2 stars stars ** stars ** stars ** Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. * p <.05, ** p <.01 20

26 Table 4 Family Satisfaction Category by Overall Star Rating Family Satisfaction Category Family Satisfaction Category Mean Overall Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 27.6% 11.9% 11.0% 5.1% Low Satisfaction % 25.5% 27.3% 15.9% 14.3% Moderate Satisfaction % 22.9% 21.1% 25.6% 21.4% High Satisfaction % 17.7% 30.9% 31.7% 36.7% Very High Satisfaction % 6.3% 8.8% 15.9% 22.4% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 21

27 Table 5 Resident Satisfaction Category by Overall Star Rating Resident Satisfaction Category Resident Satisfaction Category Mean Overall Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 24.1% 21.8% 20.6% 19.7% Low Satisfaction % 27.1% 22.3% 17.1% 19.7% Moderate Satisfaction % 22.2% 22.8% 24.1% 14.5% High Satisfaction % 22.7% 23.4% 27.6% 19.7% Very High Satisfaction % 3.9% 9.6% 10.6% 26.3% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 22

28 Health Inspection Domain in Relation to Consumer Satisfaction Tables 6 and 7 provide mean comparisons for family and resident satisfaction data across levels of star ratings within the NHC health inspection domain. Similar to Tables 2 and 3, these tables show mean satisfaction scores at each star rating level, as well as provide an overview of where significant differences existed between the mean satisfaction scores at different levels. Significant differences are identified in these tables at p <.05 and p <.01 levels. Mean satisfaction scores for both family and resident data tended to increase as the number of stars received in the NHC health inspection domain increased. Family satisfaction scores ranged from (one star) to (five stars). Resident satisfaction scores had slightly less spread, ranging from (one star) to (five stars). Within the family data, examination of whether significant differences existed in mean satisfaction scores revealed that not all mean satisfaction scores varied significantly from neighboring scores. The mean satisfaction score at one star was significantly lower than the mean satisfaction scores at all other ratings. The mean satisfaction score at the five stars, on the other hand, was significantly higher than the mean satisfaction scores at only one and two stars, but was not significantly different from the mean satisfaction scores at four or three stars. Within the other mean satisfaction ratings, significant differences were found in the following relationships: the mean satisfaction score at two stars was significantly higher than the mean satisfaction score at one star and significantly lower than the mean satisfaction scores at four and five stars; the mean satisfaction score at three stars was significantly higher than the mean satisfaction score at one star; and the mean satisfaction score at four stars was significantly higher than the mean satisfaction scores at one and two stars. The data on resident satisfaction showed slightly weaker levels of significance between values. The mean satisfaction score at one star was significantly lower than the mean satisfaction scores at three, four, and five stars. The mean satisfaction score at the five stars, alternatively, was significantly higher than the mean satisfaction scores at only one and two stars, but was not significantly different from the mean satisfaction scores at four or three stars. Within the other mean satisfaction scores, significant differences were found in the following relationships: the mean satisfaction score at two stars was significantly lower than the mean satisfaction scores at four and five stars; the mean satisfaction score at three stars was 23

29 significantly higher than the mean satisfaction score at one star; and the mean satisfaction score at four stars was significantly higher than the mean satisfaction scores at one and two stars. Categorical comparisons between the consumer satisfaction categories and the NHC health inspection domain are presented in Tables 8 and 9. Within the family data, of the nursing homes that received one star in the NHC health inspection domain, 42.7% were also categorized as having very low satisfaction. Of those that received one star, 68.4% were categorized as having either very low satisfaction or low satisfaction. In contrast, 16.9% that received one star in the NHC health inspection domain were categorized as having high or very high satisfaction. Of the nursing homes that received five stars in the NHC health inspection domain, 18.1% were also categorized as having very high satisfaction. Of those that received five stars, 55.4% were categorized as having either very high satisfaction or high satisfaction. In contrast, 44.5% that received five stars in the NHC health inspection domain were categorized as having moderate, low, or very low satisfaction. The mean scores for the satisfaction categories within the family satisfaction categorical comparisons showed a noticeable change between each category. Within the resident data, of the nursing homes that received one star in the NHC health inspection domain, 35.1% were also categorized as having very low satisfaction. Of those that received one star, 58.3% were categorized as having either very low satisfaction or low satisfaction. In contrast, 22.1% that received one star in the NHC health inspection domain were categorized as having high or very high satisfaction. Of the nursing homes that received five stars in the NHC health inspection domain, 22.8% were also categorized as having very high satisfaction. Of those that received five stars, 51.9% were categorized as having either very high satisfaction or high satisfaction. In contrast, 48.2% that received five stars in the NHC health inspection domain were categorized as having moderate, low, or very low satisfaction. The mean scores for the satisfaction categories within the resident satisfaction categorical comparisons showed a noticeable change between each category. 24

30 Table 6 Family Satisfaction Means by Health Inspection Star Rating Health Inspection Star Rating Family Satisfaction Mean P-Value between Family Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star **.000**.000**.000** 2 stars ** *.000** 3 stars ** stars **.022* stars **.000** Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. * p <.05, ** p <.01 Table 7 Resident Satisfaction Means by Health Inspection Star Rating Health Inspection Star Rating Resident Satisfaction Mean P-Value between Resident Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star **.000**.000** 2 stars *.018* 3 stars ** stars **.013* stars **.018* Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. * p <.05, ** p <.01 25

31 Table 8 Family Satisfaction Category by Health Inspection Star Rating Family Satisfaction Category Family Satisfaction Category Mean Health Inspection Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 20.7% 10.9% 12.3% 6.0% Low Satisfaction % 27.9% 22.4% 22.4% 12.0% Moderate Satisfaction % 22.1% 26.4% 24.1% 26.5% High Satisfaction % 25.5% 29.9% 29.4% 37.3% Very High Satisfaction % 3.8% 10.4% 11.8% 18.1% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 26

32 Table 9 Resident Satisfaction Category by Health Inspection Star Rating Resident Satisfaction Category Resident Satisfaction Category Mean Health Inspection Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 20.7% 16.2% 13.4% 15.2% Low Satisfaction % 27.6% 25.7% 20.4% 20.3% Moderate Satisfaction % 26.3% 27.1% 23.1% 12.7% High Satisfaction % 21.7% 23.3% 29.6% 29.1% Very High Satisfaction % 3.7% 7.6% 13.4% 22.8% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 27

33 Quality Domain in Relation to Consumer Satisfaction Mean comparisons for family and resident satisfaction data across levels of star ratings within the NHC quality domain are located in Tables 10 and 11. Similar to earlier tables, these tables show mean scores at each level, as well as provide an overview of where significant differences existed between the mean satisfaction scores at different levels. Significant differences discussed here refer to those at p <.05 and p <.01. Neither mean satisfaction scores for family nor resident data showed any clear pattern as the quantity of stars received in the NHC quality domain changes. The ranges of both were miniscule, with mean satisfaction scores for family satisfaction ranging from (one star) to (five stars) and mean satisfaction scores for resident satisfaction ranging from (five stars) to (two stars). No significant differences were present between any of the pairs of mean satisfaction scores; this held true for both the family and resident data. Tables 12 and 13 display the categorical comparisons between the consumer satisfaction categories and the NHC quality domain. Within the family data, of the nursing homes that received one star in the NHC quality domain, 19.3% were also categorized as having very low satisfaction. Of those that received one star, 46.7% were categorized as having either very low satisfaction or low satisfaction. In contrast, 24.4% that received one star in the NHC quality domain were categorized as having high or very high satisfaction. Of the nursing homes that received five stars in the NHC quality domain, 17.3% were also categorized as having very high satisfaction. Of those that received five stars, 36% were categorized as having either very high satisfaction or high satisfaction. In contrast, 64.0% that received five stars in the NHC quality domain were categorized as having moderate, low, or very low satisfaction. The mean scores for the satisfaction categories within the family satisfaction categorical comparisons showed a noticeable change between each category. Within the resident data, of the nursing homes that received one star in the NHC quality domain, 23.4% were also categorized as having very low satisfaction. Of those that received one star, 46.8% were categorized as having either very low satisfaction or low satisfaction. In contrast, 23.8% that received one star in the NHC quality domain were categorized as having high or very high satisfaction. Of the nursing homes that received five stars in the NHC quality domain, 10.2% were also categorized as having very high satisfaction. Of those that received five stars, only 16.3% were categorized as having either very high satisfaction or high 28

34 satisfaction. In contrast, 83.7% that received five stars in the NHC quality domain were categorized as having moderate, low, or very low satisfaction. The mean scores for the satisfaction categories within the resident satisfaction categorical comparisons showed a noticeable change between each category. 29

35 Table 10 Family Satisfaction Means by Quality Star Rating Quality Star Rating Family Satisfaction Mean P-Value between Family Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star stars stars stars stars Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. * p <.05, ** p <.01 Table 11 Resident Satisfaction Means by Quality Star Rating Quality Star Rating Resident Satisfaction Mean P-Value between Resident Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star stars stars stars stars Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. * p <.05, ** p <.01 30

36 Table 12 Family Satisfaction Category by Quality Star Rating Family Satisfaction Category Family Satisfaction Category Mean Quality Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 12.9% 15.3% 14.2% 17.3% Low Satisfaction % 26.7% 28.4% 26.7% 24.0% Moderate Satisfaction % 25.4% 22.8% 21.8% 22.7% High Satisfaction % 27.1% 27.0% 28.9% 18.7% Very High Satisfaction % 7.9% 6.5% 8.4% 17.3% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 31

37 Table 13 Resident Satisfaction Category by Quality Star Rating Resident Satisfaction Category Resident Satisfaction Category Means Quality Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 21.3% 20.1% 29.3% 38.8% Low Satisfaction % 27.7% 28.3% 27.2% 14.3% Moderate Satisfaction % 20.4% 25.1% 20.1% 30.6% High Satisfaction % 25.5% 21.5% 18.5% 6.1% Very High Satisfaction % 5.1% 5.0% 4.9% 10.2% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 32

38 Staffing Domain in Relation to Consumer Satisfaction Tables 14 and 15 provide an overview of mean comparisons between family and resident satisfaction data across levels of star ratings within the NHC staffing domain. Similar to earlier tables, these tables show mean satisfaction scores at each star rating level, as well as provide an overview of where significant differences existed between the mean satisfaction scores at different levels. Significant differences are identified here as well at p <.05 and p <.01. Mean satisfaction scores for family satisfaction tended to increase as the number of stars received in the NHC staffing domain increased. Mean satisfaction scores for resident satisfaction also showed a gradual increase, with the exception of a decrease between one and two stars. Family satisfaction scores ranged from (one star) to (five stars). Resident satisfaction scores had slightly less spread, ranging from (two stars) to (five stars). Within the family data, examination of whether significant differences existed in mean satisfaction scores revealed limited significant differences between mean satisfaction scores. The mean satisfaction score at one star was significantly lower than the mean satisfaction scores at three, four, and five stars, but not significantly different from the mean satisfaction score at two stars. The mean satisfaction score at the five star rating, on the other hand, was only significantly greater than the mean satisfaction score at one star. Within the other mean satisfaction ratings, significant differences were found in the following relationships: the mean satisfaction score at two stars was significantly less than the mean satisfaction score at four stars; the mean satisfaction score at three stars was significantly greater than the mean satisfaction score at one star; and the mean satisfaction score at four stars was significantly greater than the mean satisfaction scores at one and two stars. The data on resident satisfaction showed even weaker levels of significance between values. The only significant difference was the mean satisfaction score at two stars being significantly less than the mean satisfaction score at four stars. Tables 16 and 17 display the categorical comparisons between consumer satisfaction categories and the NHC staffing domain. Within the family data, of the nursing homes that received one star in the NHC staffing domain, 34.7% were also categorized as having very low satisfaction. Of those receiving one star, 56.5% were categorized as having either very low satisfaction or low satisfaction. In contrast, 20.6% that received one star in the NHC staffing domain were categorized as having high or very high satisfaction. Of the nursing homes that 33

39 received five stars in the NHC staffing domain, 16.1% were also categorized as having very high satisfaction. Of those that received five stars, 51.6% were categorized as having either very high satisfaction or high satisfaction. In contrast, 48.4% that received five stars in the NHC staffing domain were categorized as having moderate, low, or very low satisfaction. The mean scores for the satisfaction categories within the family satisfaction categorical comparisons showed a noticeable change between each category. Within the resident data, of the nursing homes that received one star in the NHC staffing domain, 29.2% were also categorized as having very low satisfaction. Of those that received one star, 51.8% were categorized as having either very low satisfaction or low satisfaction. In contrast, 27.4% that received one star in the NHC staffing domain were categorized as having high or very high satisfaction. Of the nursing homes that received five stars in the NHC staffing domain, 22.2% were also categorized as having very high satisfaction. Of those that received five stars, 44.4% were categorized as having either very high satisfaction or high satisfaction. In contrast, 55.6% that received five stars in the NHC staffing domain were categorized as having moderate, low, or very low satisfaction. The mean scores for the satisfaction categories within the resident satisfaction categorical comparisons showed a noticeable change between each category. 34

40 Table 14 Family Satisfaction Means by Staffing Star Rating Staffing Star Rating Family Satisfaction Mean P-Value between Family Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star **.000**.001** 2 stars ** stars ** stars **.005** stars ** Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. * p <.05, ** p <.01 Table 15 Resident Satisfaction Means by Staffing Star Rating Staffing Star Rating Resident Satisfaction Mean P-Value between Resident Satisfaction Means 1 star 2 stars 3 stars 4 stars 5 stars 1 star stars * stars stars * stars Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. * p <.05, ** p <.01 35

41 Table 16 Family Satisfaction Category by Staffing Star Rating Family Satisfaction Category Family Satisfaction Category Means Staffing Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 20.6% 16.2% 13.3% 9.7% Low Satisfaction % 20.0% 19.6% 15.5% 22.6% Moderate Satisfaction % 31.5% 21.1% 21.0% 16.1% High Satisfaction % 24.8% 41.7% 46.6% 35.5% Very High Satisfaction % 3.0% 1.5% 3.6% 16.1% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2010 Nursing Home Compare and 2010 Ohio Nursing Home Family Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 36

42 Table 17 Resident Satisfaction Category by Staffing Star Rating Resident Satisfaction Category Resident Satisfaction Category Mean Staffing Star Rating 1 star 2 stars 3 stars 4 stars 5 stars Very Low Satisfaction % 31.3% 21.9% 21.1% 16.7% Low Satisfaction % 22.2% 19.4% 15.6% 22.2% Moderate Satisfaction % 19.9% 23.9% 24.7% 16.7% High Satisfaction % 26.1% 33.3% 35.6% 22.2% Very High Satisfaction % 0.6% 1.5% 2.9% 22.2% Total Percentage 100.0% 100.0% 100.0% 100.0% 100.0% Number with Star Rating Note. Data from 2009 Nursing Home Compare and 2009 Ohio Nursing Home Resident Satisfaction Surveys. Totals may not sum up to exactly 100% due to rounding. 37

43 Impact of Consumer Group Two trends emerged when comparing the results of each consumer group to the overall star ratings. The first trend was that mean satisfaction scores for family members at any star rating within any given domain were, in almost all comparisons, higher than the counterpart mean satisfaction scores for residents at the same star rating in the same domain. The second trend was that the levels of disagreement among residents were higher than the levels of disagreement among family members. In instances where this was not the case, the two groups had similar (within 0.7%) levels of disagreement. Summary of Results When mean satisfaction scores were examined for the NHC overall rating, nursing homes that received a higher rating on NHC, on average, had higher satisfaction scores than those nursing homes with fewer stars. While this may be true, the differences between the mean satisfaction scores at each star rating were often not significant. The categorical comparisons between the NHC overall rating and consumer satisfaction scores also showed discrepancies that were present. About one out of every five nursing homes that received one star had high or very high family and resident satisfaction scores. About two out of every five nursing homes that received five stars on the NHC overall rating had moderate, low, or very low family satisfaction and more than half had having moderate, low, or very low resident satisfaction. Similar trends were found in the examinations of the NHC health inspection and staffing domains. Generally, the mean satisfaction scores of families and residents tended to increase as the number of stars received in the NHC domains increased (one exception was present in the comparison between resident satisfaction and staffing). In many of the cases, there were no significant differences between mean satisfaction scores. When looking at the health inspection domain, about one out of every six nursing homes that received one star in the health inspection domain had high or very high family satisfaction and slightly more than one out of every five nursing homes rated as one star in the health inspection domain had high or very high resident satisfaction. Conversely, just under half of the nursing homes that received five stars in the health inspection domain had moderate, low, or very low family and resident satisfaction. When comparing to the staffing domain, about one out of every five nursing homes that received one star had high or very high family and slightly more than one out of every four nursing homes rated as one star had high or very high resident 38

44 satisfaction. Conversely, about half of those nursing homes that received five stars on staffing had moderate, low, or very low family and resident satisfaction. The quality domain was the least in line with the satisfaction data. The mean satisfaction scores among the star ratings showed no real trends. No significant differences were found between any of the mean satisfaction scores in neither the family nor the resident satisfaction data. The categorical comparisons further supported this. About a quarter of the nursing homes that received one star in the quality domain were had high or very high family and resident satisfaction. Potentially even more surprising, about two out of every three nursing homes rated five stars in the quality domain had moderate, low, or very low family satisfaction and over eight out of every 10 nursing homes that received five stars in the quality domain had moderate, low, or very low resident satisfaction. Two additional findings emerged from the data analyses that identified impacts of each consumer group. In most cases, mean satisfaction scores for family members at any star rating within any given domain were higher than the counterpart mean satisfaction scores for residents at the same star rating in the same domain. The data also showed that the levels of disagreement among residents were similar or greater than than the levels of disagreement among family members. Overall, the results of this study show that inconsistencies are present between the NHC star rating system and the consumer satisfaction scores of both families and residents. These findings hold significant implications for nursing home policy and practice. In the next and final chapter, additional discussion of these findings along with the implications they hold is presented. 39

45 Chapter 4: Discussion The current study explores how the NHC star rating system compares to family and resident satisfaction with nursing homes. A summary of the study s findings in Table 18 displays the information from the mean comparisons as well as the categorical comparisons. Within the mean comparisons section of the table, information on the ranges of mean satisfaction scores by star ratings, visible trends (i.e., whether or not mean satisfaction scores increased as star ratings increased), and detection at the extremes of each comparison (i.e., whether or not one star was significantly different than five stars) are provided. Within the categorical comparisons section of the table, levels of disagreement (operationalized as one star nursing homes possessing high or very high consumer satisfaction or five star nursing homes possessing moderate, low, or very low consumer satisfaction) among one star and five star facilities are presented along with the ranges of the mean scores of each satisfaction category. Findings suggest that NHC is not a valid substitute for direct examination of consumer satisfaction. In this chapter, these findings are elaborated on and their implications for the NHC star rating system are explored. 40

46 Table 18 Results Summary Mean Comparisons Categorical Comparisons Range of Mean Satisfaction Scores by Star Rating Level of Level of Trend Visible? Detection at Extremes? Disagreement Disagreement among 1-star NHs: among 5-star NHs: Range of Mean Satisfaction Scores by Category Overall Star vs. Family Satisfaction (1 star) (5 stars) Yes Yes 20.0% 40.8% Overall Star vs. Resident Satisfaction (1 star) (5 stars) Yes Yes 19.3% 53.9% Health Insp. Star vs. Family Satisfaction (1 star) (5 stars) Yes Yes 16.9% 44.5% Health Insp. Star vs. Resident Satisfaction (1 star) (5 stars) Yes Yes 22.1% 48.2% Quality Star vs. Family Satisfaction (1 star) (5 stars) No* No 24.4% 64.0% Quality Star vs. Resident Satisfaction (5 stars) (2 stars) No No 23.8% 83.7% Staffing Star vs. Family Satisfaction (1 star) (5 stars) Yes Yes 20.6% 48.4% Staffing Star vs. Resident Satisfaction (2 stars) (5 stars) Yes* No 27.4% 55.6% *Note that if the rating at 2 stars was excluded, each of these would exhibit visible linear trends; neither exhibit perfect linearity. Quality Star vs. Family Satisfaction was labeled as not having a meaningful linear trend due to the limited range. Staffing Star vs. Resident Satisfaction was labeled as having a meaningful linear trend due to the larger range. 41

47 Discrepancies Relative to the Overall Star Rating The overall star rating on NHC receives the most attention from the media and consumers. In some years CMS has actually posted results in major newspapers across the nation. Given how the star rating is used, the link to consumer satisfaction ratings is a critical question to both CMS and to consumers. While it was promising to find that nursing homes receiving higher ratings on NHC tended to have higher consumer satisfaction scores, the limited relationship between consumer satisfaction scores and the NHC star ratings is cause for concern. For example, although the mean satisfaction score for those nursing homes receiving one star in the NHC overall domain was significantly lower than the mean satisfaction score for those nursing homes receiving five stars, the categorical comparisons between the NHC overall ratings and the consumer satisfaction categories demonstrated the inconsistencies that exist. Many nursing homes that received five stars on the NHC overall rating had moderate to very low consumer satisfaction (40.8% when compared to family satisfaction and 53.9% when compared to resident satisfaction) and many nursing homes that received one star on the NHC overall rating had high to very high consumer satisfaction (20.0% when compared to family satisfaction and 19.3% when compared to resident satisfaction). With such high rates of inconsistency between the NHC overall ratings and consumer satisfaction scores, it is difficult to say that consumer s views of nursing homes are adequately reflected in the NHC overall rating. Discrepancies Relative to the Health Inspection, Quality, and Staffing Domains The three NHC domains that make up the NHC overall rating vary in how well they reflect consumer satisfaction ratings. The health inspection and staffing domains were at least generally similar to consumer ratings; that is, satisfaction scores were often higher, but not significantly, for nursing homes receiving more stars within each of these two domains. Within the health inspection and staffing domains, family mean satisfaction scores at one star were significantly lower than mean satisfaction scores at five stars. Within the health inspection domain, the mean resident satisfaction score at one star was also significantly lower than mean satisfaction score at five stars. This suggests that, to some extent, the ratings in these domains are able to detect consumer satisfaction at the extremes. Those scenarios where significant differences between one star mean satisfaction scores and five star mean satisfaction scores were not present require further examination. It is possible that the resident mean satisfaction scores in the staffing domain are not significantly different 42

48 due to the varying perceptions of staffing between families and residents. While families may observe the quantity of staff in a facility and thus make a similar rating as that on NHC, residents likely focus more on the quality of care they are receiving, regardless of the quantity of staff present. The quality domain, which poorly reflected satisfaction among both family and residents, may vary due to the different nature of the dimensions of quality being assessed (e.g., clinical care versus quality of life). All three of these domains demonstrated a great deal of inconsistency in the categorical comparisons. Within the health inspection and staffing domains, the levels of disagreement ranged from 16.9% to 55.6%. The level of agreement within the quality domain was even higher, ranging from 23.8% to 83.7%. At these levels, inferring how satisfied consumers might be with nursing homes based on NHC star ratings is not viable. Family and Resident Views The differentiation between family and resident satisfaction scores is likely due to the differing perspectives of each consumer group. While residents may provide subjective views of their experiences (e.g., perceived quality of staff), family members may be looking to more objective measures (e.g., quantity of staff) to establish their views. It is possible to see this phenomenon occurring by examining the levels of disagreement within the categorical comparisons. As family members in most cases had lower levels of disagreement with the NHC ratings, it is possible to assume that family members are basing their opinions more along the lines of those dimensions of quality captured by the NHC system than are their resident counterparts. This finding suggests that although furthering efforts to obtain input from both families and residents is important, future efforts might be more valuable if they focus on obtaining input from residents rather than families. Rationale for Differences and Potential Solutions for CMS There are a number of possible reasons why the ratings provided on NHC do not accurately reflect consumer satisfaction. First and foremost, the NHC star rating system is not designed to directly assess consumer satisfaction. Rather, it is focused on assessing other dimensions of quality. Although one might argue that performance on health inspections, MDS quality measures, and staffing may be valid indicators of consumer satisfaction, this study shows that they are by no means a substitute. 43

49 Additionally, performance on the measures utilized in the calculations of the NHC ratings may not adequately reflect year-round nursing home practice. In addition to those topics discussed in the literature review, a number of factors can account for the discrepancies between NHC and the consumer satisfaction ratings. For instance, nursing homes generally know the time frame when state surveyors will arrive. They can prepare for this, add extra staffing, and correct issues that may be faulty during other times throughout the year. As a result, family and residents may be assessing nursing homes based on their long-term experiences with facilities whereas NHC is only assessing nursing homes based on data collected at limited periods in time. As such, nursing home consumer satisfaction may have the ability to offer a more holistic view of consumers experiences within nursing homes when compared to the information offered on NHC. At a minimum, the complete exclusion of consumer satisfaction data from NHC is cause for concern. In order to serve as a more comprehensive and reliable resource for consumers, CMS can work to better integrate the thoughts and opinions of consumers into its NHC star rating system. If large scale surveying, such as that conducted in Ohio, is not a viable option for nationwide implementation, residents should at least be provided the opportunity to provide some input into how their nursing homes are rated. Such information could be obtained by introducing a satisfaction component to the state survey process wherein, at the very least, surveyors could ask a core set of satisfaction questions to a sample of the residents in each nursing home. Study Limitations and Ideas for Further Research Multiple limitations should be noted with regard to the current study. First, as this study utilized satisfaction data from only Ohio nursing homes, the findings may not represent what is true in other states. A second limitation is that not all nursing homes across the state of Ohio were utilized. It is possible that those nursing homes that were not utilized may have differed from those utilized in the analyses. Further research should be conducted to examine whether similar findings exist in other states that possess consumer satisfaction data as well as on future Ohio consumer satisfaction data as it becomes available. Additional research should also explore what underlying structure, process, and outcome factors result in optimal performance on both NHC and consumer 44

50 satisfaction metrics. By obtaining this information, nursing homes will be better able to focus their quality improvement initiatives. Conclusions In closing, although NHC can serve as one resource for obtaining information on nursing homes, it fails to adequately capture the actual satisfaction levels of nursing home consumers. While consumers may desire information such as that available on NHC, ultimately many want to know whether or not they will have satisfactory experiences with nursing homes. By failing to include this information in NHC, CMS is missing a large opportunity to increase the usefulness of the NHC database. With small changes, such as introducing simple satisfaction questions to the nursing home survey process, CMS can greatly improve the rigor of the information it offers on NHC to those who are in search of a nursing home. 45

51 References 2010 Family Survey FAQs. (2010). Retrieved October 24, 2011 from Applebaum, R. A., Uman, G. C., & Straker, J. K. (2006). Capturing the voices of consumers in long-term care: If you ask them they will tell. In S. R. Kunkel & V. Wellen (Eds.), Consumer Voice and Choice in Long-Term Care (pp ). New York: Springer Publishing Company. Çalıkoğlu, Ş., Christmyer, C. S., & Kozlowski, B. U. (2011). My eyes, your eyes the relationship between CMS five-star rating of nursing homes and family rating of experience of care in Maryland. Journal of Health Care Quality, DOI: /j x Castle, N. G. (2005). Nursing home administrators opinions of nursing home compare web site. The Gerontologist, 45(3), Castle, N. G. (2006). Family members as proxies for satisfaction with nursing home care. Joint Commission Journal on Quality and Patient Safety, 32(8), Castle, N. G. (2009). The nursing home compare report card: consumers' use and understanding. Journal of Aging and Social Policy, 21(2), DOI: / Castle, N. G., Diesel, J., Ferguson, J. C. (2011). The evolution of nursing home report cards. Journal of Applied Gerontology, 30(6): Castle, N. G. & Ferguson, J. C. (2010). What is nursing home quality and how is it measured? The Gerontologist, 50(4), Castle, N. G., & Lowe, T. J. (2005). Report cards and nursing homes. The Gerontologist, 45(1), Centers for Medicare and Medicaid Services (December 2008). Questions and answers: Improving the Nursing Home Compare web site: The five-star nursing home quality rating system: Volume 1. Retrieved from /CertificationandComplianc/Downloads/QsandAsFinal.pdf Centers for Medicare and Medicaid Services. (December 23, 2009). Nursing Home Compare Archived Dataset [Archived Dataset]. Retrieved from /qualityinitiativesgeninfo/amcd/itemdetail.asp?filtertype=none&filterbydid=- 99&sortByDID=1&sortOrder= ascending&itemid=cms &intnumperpage=10 46

52 Centers for Medicare and Medicaid Services (July 2010a). Design for nursing home compare five-star quality rating system Technical users guide. Retrieved October 18, 2011, from Centers for Medicare and Medicaid Services. (December 20, 2010b). Nursing Home Compare Archived Dataset [Archived Dataset]. Retrieved from /qualityinitiativesgeninfo/amcd/itemdetail.asp?filtertype=none&filterbydid=- 99&sortByDID=1&sortOrder= ascending&itemid=cms &intnumperpage=10 Centers for Medicare and Medicaid Services. (2012). Nursing Home Compare. Retrieved January 10, 2012 from Colorado Foundation for Medical Care. (2005). Development of staffing quality measures: Phase 1: Final report Donabedian, A. (1985). Twenty years of research on the quality of medical care: Evaluation and the Health Professions, 8, Ejaz, F. K., & Castle, N. G. (2007). Resident satisfaction with long-term care services. Journal of Aging and Social Policy, 19(2), 1-8. Ejaz, F. K., Straker, J. K., Fox, K., & Swami, S. (2003). Developing a satisfaction survey for families of Ohio's nursing home residents. The Gerontologist, 43(4), Gasquet, I., Dehe, S., Gaudebout, P., Falissard, B. (2003). Regular visitors are not good substitutes for assessment of elderly patient satisfaction with nursing home care and services. Journal of Gerontology: Medical Sciences, 58A(11), Government Accountability Office. (May 2008). Federal monitoring surveys demonstrate continued understatement of serious care problems and CMS oversight weaknesses (Publication No. GAO ). Retrieved from Grabowski, D. C., & Town, R. J. (2011). Does information matter? Competition, quality, and the impact of nursing home report cards. Health Services Research, DOI: /j x H.B. 403, 123rd Gen. Assem., Reg. Sess. (OH. 2000). Harris-Kojetin, L. D., & Stone, R. I. (2007). The role of consumer satisfaction in ensuring quality long-term care. Journal of Aging and Social Policy, 19(2),

53 Institute of Medicine (2001). Improving the Quality of Long-Term Care. National Academy Press: Washington, DC. Kash, B., Hawes, C. and Phillips, C.D. (2007) Differences in nursing home staffing measures obtained from OSCAR data and Medicaid cost reports. The Gerontologist, 47: Lucas, J. A., Levin, C. A., Lowe, T. J., Robertson, B., Akincigil, A., Sambamoorthi, U.,... Crystal, S. (2007). The relationship between organizational factors and resident satisfaction with nursing home care and life. Journal of Aging and Social Policy, 19(2): Miller, E. A., & Mor, V.(2008). Balancing regulatory controls and incentives: Toward smarter and more transparent oversight in long-term care. Journal of Health Politics, Policy and Law, 33(2), Minnix, L. (2008). AAHSA Statement on CMS five-star rating system for nursing homes [press release]. Retrieved from Mukamel, D. B., & Spector, W. D. (2003). Quality report cards and nursing home quality. The Gerontologist, 43(Special Issue II), Mukamel, D. B., Spector, W. D., Zinn, J. S., Huang, L., Weimer, D. L., & Dozier, A. (2007). Nursing homes' response to the nursing home compare report card. Journal of Gerontology: Social Sciences, 62B(4), S218-S225. Mukamel, D. B., Spector, W. D., Zinn, J., Weimer, D. L., & Ahn, R. (2010). Changes in clinical and hotel expenditures following publication of the nursing home compare report card. Medical Care, 48(10), Mukamel, D. B., Weimer, D. L., Spector, W. D., Ladd, H., & Zinn, J. S. (2008). Publication of quality report cards and trends in reported quality measures in nursing homes. Health Research and Educational Trust, 43(4), Office of Inspector General. (2003). Nursing home deficiency trends and survey and certification process consistency (OIG publication No. OEI ). Washington, DC: Office of Inspector General. Retrieved from Office of Inspector General. (2004). Inspections results on nursing home compare: Completeness and accuracy (OIG publication No. OEI ). Washington, DC: 48

54 Office of the Inspector General. Retrieved from pdf Ohio Department of Aging. (2011). LTCCG. Retrieved October 18, 2011, from Park, J., Konetzka, R., & Werner, R. M. (2011). Performing Well on Nursing Home Report Cards: Does It Pay Off? Health Services Research, 46(2), doi: /j x PointRight Inc. (2008). Review of CMS five-star nursing home rating system [PowerPoint Slides]. Retrieved from /Documents/Review%20of%20CMS%205%20Star%20Rating%20System.pdf Rahman, A. N., & Applebaum, R. A. (2009). The nursing home minimum data set assessment instrument: Manifest functions and unintended consequences - Past, present, and future. The Gerontologist, 49(6), doi: /geront/gnp066 S.B. 264, 129 th Gen. Assem., Reg. Sess. (OH. 2011) Saliba, D., & Buchanan, J. (2008). Development & validation of a revised nursing home assessment tool: MDS Sangl, J., Buchanan, J., Cosenza, C., Bernard, S., Keller, S., Mitchell, N., Brown, J., Castle, N., Sekscenski, E., & Larwood, D. (2007). The development of a CAHPS instrument for nursing home residents (NHCAHPS). Journal of Aging & Social Policy, 19(2), Scripps Gerontology Center. (2010). Ohio Nursing Home Family Satisfaction Data. Unpublished Raw Data. Stevenson, D. G. (2006). Is a public reporting approach appropriate for nursing home care? Journal of Health Politics, Policy, and Law, 31(4), Straker, J.K., Chow, K., Mwangi, S., & Reddecliff, L. (2011). Implementation of the 2010 Ohio nursing home family satisfaction survey. Oxford, OH. Straker, J. K., Ejaz, F. K., McCarthy, C., & Jones, J. A. (2007). Developing and testing a satisfaction survey for nursing home residents. Journal of Aging and Social Policy, 19(2), Tellis-Nayak, V. V., Shiverick, B. N., & Hernandez, M. (2010). Where Allies Part Ways and Strangers Converge: Nursing Home Performance in the Eyes of Residents, Families, and State Surveyors. Seniors Housing & Care Journal, 18(1),

55 Uman, G. C., Hocevar, D., Urman, H. N., Young, R., Hirsch, M., & Kohler, S. (2000). Satisfaction surveys with the cognitively impaired. In J. Cohen-Mansfield, F. K. Ejaz, & P. Werner (Eds.), Satisfaction Surveys in Long-Term Care (pp ). New York: Springer Publishing Company. Vital Research. (2009). Ohio Nursing Home Resident Satisfaction Data. Unpublished Raw Data. Vital Research. (2010). Implementation of the 2009 long-term care resident satisfaction survey. Los Angeles, CA. Werner, R. M., Tamara Konetzka, R., & Liang, K. (2009). State adoption of nursing home payfor-performance. Medical Care Research and Review, 67(3), Wheatley, M.V., Choi, K., Hirsch, M., Walley, J.E., Lee, C. S., Urman, H. N., & Uman, G. C. (2007). Implementation and results of the statewide Ohio nursing home resident satisfaction survey. Journal of Aging & Social Policy, 19(2), Yarwood, B. (2008). Rankings are flawed. USA Today. 50

56 Appendix A Ohio Nursing Home Family Satisfaction Survey 51

57 52

58 53

59 54

60 55

61 56

62 57

63 58

64 59

65 60

66 Appendix B Ohio Nursing Home Resident Satisfaction Survey 61

67 62

68 63