Transition to the Nursing profession

By Daniel Brauner, Rachel M. Werner, Tetyana P. Shippee, John Cursio, Hari Sharma, and R. Tamara Konetzka

AGING & HEALTH

Does Nursing Home Compare Reflect Patient Safety In Nursing Homes?

ABSTRACT The past several decades have seen significant policy efforts to improve the quality of care in nursing homes, but the patient safety movement has largely ignored this setting. In this study we compared nursing homes’ performance on several composite quality measures from Nursing Home Compare, the most prominent recent example of a national policy aimed at improving the quality of nursing home care, to their performance on measures of patient safety in nursing homes such as pressure sores, infections, falls, and medication errors. Although Nursing Home Compare captures some aspects of patient safety, we found the relationship to be weak and somewhat inconsistent, leaving consumers who care about patient safety with little guidance. We recommend that Nursing Home Compare be refined to provide a clearer picture of patient safety and quality of life, allowing consumers to weight these domains according to their preferences and priorities.

T he Institute of Medicine (IOM) inspired the quality improvement movement for US nursing homes with its 1986 report titled Improving the Quality of Care in Nursing

Homes.1 The report noted the “shockingly defi- cient” care that people were receiving in many government-licensed nursing homes. This re- port led to the Nursing Home Reform Act of 1987, which mandated extensive regulatory con- trols; regular inspections; and the development of a resident-level assessment, data collection, and care planning system. It wasn’t until 1999 that another groundbreak-

ing report by the IOM, To Err Is Human, ignited the patient safety movement.2 Focusing on the prevention of medical errors in acute care, the patient safety movement inspired a proliferation of attention focused on creating evidence-based methods for improving the safety of care,3,4 along with systems analysts calling for an “integrated framework” to create a universally applicable and coherent approach to quality and patient

safety.5

However, until recently, attention to patient safety in nursing homes remained conspicuously absent. A 2015 review of evidence on patient safety in nursing homes concluded that patient safety outcomes in such facilities have not been well studied and that patient safety measures taken from the hospital setting are unlikely to apply to the nursing home context, with its unique set of resident characteristics and regu- latory environment.6 A related article noted the weak and mixed evidence base on interventions for improving patient safety in nursing homes.7

At the time when patient safety in acute care was gaining momentum, policy makers re- mained focused on improving the quality of care in nursing homes and did not define these efforts as promoting patient safety. Notably, in 2002 the Centers for Medicare and Medicaid Services (CMS) substantially expanded its quality im- provement efforts by launching Nursing Home Compare (NHC), a national effort to publicly report the quality of care in all US nursing

doi: 10.1377/hlthaff.2018.0721 HEALTH AFFAIRS 37, NO. 11 (2018): 1770–1778 ©2018 Project HOPE— The People-to-People Health Foundation, Inc.

Daniel Brauner is an associate professor in the Department of Medicine, University of Chicago, in Illinois.

Rachel M. Werner is a professor of medicine in the Division of General Internal Medicine at the Perelman School of Medicine and a professor of health care management at the Wharton School of Business, both at the University of Pennsylvania, and core faculty at the Center for Health Equity Research and Promotion, Corporal Michael J. Crescenz Veterans Affairs Medical Center, in Philadelphia.

Tetyana P. Shippee is an associate professor in the Division of Health Policy and Management, School of Public Health, University of Minnesota, in Minneapolis.

John Cursio is a research assistant professor in the Department of Public Health Sciences, University of Chicago.

Hari Sharma is an assistant professor in the Department of Health Management and Policy, University of Iowa, in Iowa City. At the time this research was conducted, Sharma was a doctoral student in the Department of Public Health Sciences, University of Chicago.

R. Tamara Konetzka ([email protected]) is a professor in the Department of Public Health Sciences and in the Department of Medicine, University of Chicago.

1770 Health Affairs November 2018 37:11

Aging & Health

homes, thereby informing consumers and in- centing the improvement of quality. NHC was modified to feature a five-star composite rating system for overall quality in late 2008, assigning each nursing home a rating of one to five stars (with more stars indicating higher quality). The notions of health care quality and patient

safety are often conflated. Although they over- lap, patient safety—in the IOM’s sense of preven- tion of harm to patients from medical errors2— may be considered a subdomain of quality that is associated with distinct approaches, processes, and outcomes. The emphasis in patient safety is on identifying errors, determining their cause, and preventing them from happening again, of- ten by addressing gaps or failures in the relevant system of care. In the nursing home setting, res- ident falls are a good example of a typical patient safety target. If a nursing home finds that fall rates are high, a patient safety approach would examine the circumstances that lead to the falls. For example, a fall might result from a loss of balance that may be due to the side effects of medications, environmental hazards, or lack of appropriate staffing to provide oversight and assistance. To prevent future falls, each of these items might be addressed. Ultimately, these ef- forts can result in improved patient safety. A quality improvement perspective, on the

other hand, would consider avoiding adverse events from errors as one of a broader set of goals. Quality improvement efforts also work to enable positive health outcomes beyond im- proving safety. For nursing homes, some of these other goals have been defined as improving or maintaining functional status, treating pain, maintaining weight, avoiding incontinence and catheter use, avoiding depression, avoiding physical restraints and the inappropriate use of antipsychotic medications, and improving vacci- nation rates. There has been little attention placed on differentiating patient safety from oth- er types of quality outcomes in tools designed to measure nursing home quality. Our goal was to examine the relationship be-

tween nursing homes’ performance on standard quality measures overall and on measures spe- cific to patient safety in nursing homes. To do so, we focused on quality measures from NHC, a national, policy-driven public reporting system. Prior research has established that providers monitor their ratings and attempt to improve them,8,9 and NHC scores have generally im- proved over time.10 Yet it is unclear whether nurs- ing homes that score well on NHC also perform well from a patient safety perspective.While NHC was intended first and foremost to help consum- ers choose high-quality nursing homes, arguably it is the extreme examples of poor patient safety

outcomes, sometimes associated with neglect, that consumers fear most when considering nursing home placement. NHC star ratings are based on three domains

of quality: state regulatory health inspections, staffing ratios, and clinical quality. Each facility is assigned a star rating for each domain. The health inspections domain rating is based on the number, scope, and severity of health deficien- cies found at state inspections and the number of repeat visits needed to confirm the correction of deficiencies, all relative to other facilities in the state. One key area of patient safety, medication errors, appears in several potential health defi- ciencies in this domain. The staffing domain rat- ing is based on case-mix-adjusted measures of total nurse hours per resident day and registered nurse hours per resident day, relative to specific national thresholds. The clinical quality domain rating is based on meeting national thresholds across a group of individual outcome measures, some of which are patient safety measures. The overall star rating combines the three do-

mains, using the health inspection star rating as a base and potentially adjusting the rating up or down depending on the staffing and clinical quality star ratings. The health inspections do- main is given thegreatest weight, as itis based on the results of inspections by state surveyors rath- er than facility self-reported data and is therefore considered the most objective. The three individ- ual domain star ratings are reported on NHC in addition to the overall star rating. Details of the process for calculating the star ratings can be found on the NHC website.11

NHC does include some patient safety mea- sures, such as pressure sores and injurious falls, that are often associated in the minds of consum- ers with neglect. However, because NHC mea- sures many factors, its signal in terms of patient safety performance may be weak. How much NHC reflects patient safety and what consumers can learn from NHC scores about patient safety have not been established.

Study Data And Methods Data Source We merged data for the first quar- ter of 2017 from the Nursing Home Compare archives with data from Certification and Survey Provider Enhanced Reporting (CASPER) to cre- ate a nursing home–level data set. The archives, a historical compilation of ratings published on the NHC website, provided us with each home’s overall and domain-specific star ratings and facility-level performance on individual patient safety measures that appear in the quality mea- sures domain. CASPER is a compilation of data collected by surveyors during regular inspec-

November 2018 37:11 Health Affairs 1771

tions of nursing facilities for Medicare and Med- icaid certification. From CASPER, we obtained profit status, payer mix, and chain status, as well as whether the nursing home was cited for medi- cation errors in its most recent prior health inspection—information not published as part of the NHC archives. Methods We examined six key measures

of patient safety in nursing homes: injurious falls, urinary tract infections, and pressure sores among long-stay residents; pressure sores among short-stay residents; and two measures of medication errors. Applying the IOM’s sense of safety as the prevention of harm to patients from medical errors,2 which we took to include acts of omission and commission, we defined the domain of nursing home measures of safety as negative outcomes that were potentially avoid- able. These measures have been identified as safety concerns in the literature as well.6 Al- though the distinction is not always sharp, other types of NHC quality measures (such as pain control, functional status, and vaccination rates) are typically not considered patient safety mea- sures. Each of our selected outcomes is impor- tant in that it potentially leads to hospitalization and subsequent health consequences that are costly in terms of health care use, morbidity, and mortality. Some evidence (though mixed) exists on interventions that could help avoid each of these outcomes.7 Falls could be avoided through education programs, mobility aids, re- ducing polypharmacy, and increased or more consistent staffing levels.12,13 Urinary tract infec- tions could be reduced through minimizing the use of urinary catheters and ensuring their prop- er use, including through better hygiene.14,15

Pressure sores could be reduced through fre- quent turning of mobility-limited residents and ensuring proper nutrition and hydration.16 Fi- nally, medication errors could be reduced through medication review and improved com- munication across care providers.17,18

The first four measures (injurious falls, uri- nary tract infections, and the two pressure sore measures) were originally based on federally mandated assessments of residents’ condition at regular intervals, the data from which are com- piled into the Minimum Data Set 3.0. The data are reported by the nursing homes and reviewed by nursing home inspectors. Following technical specifications for each measure,19 CMS uses the data to define the cohort at risk. Long-stay residents are defined as those who have been residing in the facility for at least 100 days, and short-stay residents as those with shorter lengths- of-stay, primarily for post–acute care rehabilita- tion. In addition, a resident must be considered at risk for the outcome being measured. For ex-

ample, for the long-stay pressure sore measure, residents at risk are those with mobility limita- tions. Each resident who qualifies for the cohort is then coded as having had the adverse outcome or not, with regression-based risk adjustment incorporated into some measures. The resi- dent-level binary outcomes are aggregated into a facility-level percentage that is averaged over four quarters and reported on NHC. We used these facility-level percentages in our analysis. The two measures of medication errors are

based on deficiency citations from health inspec- tions. Each nursing home in the US that qualifies for Medicare, Medicaid, or both must be in- spected by state surveyors at least once every fifteen months to ensure compliance with regu- lations.When a facility is found to be out of com- pliance, it is cited with a deficiency that needs to be corrected. There are hundreds of possible de- ficiencies. We focused on two related to medica- tion errors, as these errors are prototypical safe- ty failures across health care settings. The first identifies significant medication errors among one or more residents; the second is for a medi- cation error rate of more than 5 percent of all doses prescribed, at any level of significance. These deficiencies are each defined as equal to 1 (cited) or 0 (not cited) at the facility level. Each deficiency is also assigned a score for the scope and severity of the violation. We ignored this information because its use complicated the measure without substantively changing our results. Analysis Our primary goal was to examine

the relationship between nursing homes’ perfor- mance on standard quality measures overall and on measures specific to patient safety in nursing homes. We first tested the Pearson correlation between performance on each patient safety

Our results highlight the differences between patient safety and quality improvement, two related but distinct approaches to improving health care.

Aging & Health

1772 Health Affairs November 2018 37:11

measure and both the NHC overall star rating and the NHC health inspection star rating. Next, we tested the correlation between patient safety measures and NHC ratings graphically by strati- fying nursing homes on overall star level and displaying the average percentage for each mea- sure. The average percentage was calculated as the unweighted average across facilities in each star category of the reported percentage of res- idents who trigger each measure—that is, the average of facility-level percentages.The medica- tion error measures represent the average per- centage of facilities cited for the deficiency. We used analysis-of-variance tests to test for

groupwise differences. If NHC serves as a good marker for patient safety, the means should be highest for one-star homes and lowest for five- star homes, with a monotonic trend in between. The patient safety measures we studied are in- cluded in NHC, so some correlation is to be ex- pected. At the same time, their role in the overall score may be limited. Medication errors are cap- tured only to the extent that nursing homes are cited with a regulatory deficiency for having iso- lated or frequent errors and are just two possible deficiencies among hundreds that feed into a complex formula for the health inspection score. That score and associated star rating do not iden- tify which deficiencies had an impact on the score. One can find the full health inspection results, including specific deficiencies cited, on the NHC website, but these are far less promi- nent than the star ratings and require multiple clicks through the website. The other four pa- tient safety measures we studied (injurious falls, urinary tract infections, and pressure sores among long-stay residents; and pressure sores among short-stay post–acute care residents) are included as separate measures within the clinical quality domain on NHC. Although all four mea- sures are included in the star rating for clinical quality, they constitute a minority of the twenty- four measures (sixteen of which are included in the star rating). Thus, the extent to which patient safety is reflected in NHC depends on whether patient safety variation drives, or is correlated with, overall variation in quality. Our secondary goal was to identify the types of

nursing homes in which patient safety was likely to present the greatest challenges. To do this, we compared rates of patient safety outcomes by key nursing home characteristics: profit status, chain status, and payer mix (whether the facility is dominated by Medicaid, using 90 percent of residents on Medicaid as our threshold).20

Limitations Our approach was subject to sev- eral limitations. First, it is possible that neither the NHC ratings nor the individual patient safety measures reflect true quality or safety. Our goal

was to examine the relationships among types of measures, not to assess their underlying validity. Second, we examined six typical measures of

patient safety that were available in our data, but these measures might not provide a complete picture of patient safety and might not be equally important. Third, the relationships we studied might not

be causal and could be the result of unobservable factors that were correlated with both quality and safety, such as the ability of facility manag- ers. We were mainly interested in whether the NHC ratings acted as a signal for performance in patient safety and did not intend to explore the underlying mechanisms.

Study Results Our national sample of nursing homes in the first quarter of 2017 is summarized in online appendix exhibit 1.21 We analyzed data on 15,652 nursing homes, which includes all nursing homes in the country certified for Medicare, Medicaid, or both. We expected measures of patient safety (all of

which were constructed as adverse outcomes, so that lower rates are better) to have a negative correlation with the Nursing Home Compare overall star rating, in which more stars are better. Exhibit 1 shows that this expected nega- tive relationship was the norm (with some ex- ceptions), but the correlations were quite low. The highest correlation (in absolute value) was for pressure sores among long-stay residents (−0.21), and the lowest was for urinary tract infections among the same population (−0.05).

Exhibit 1

Correlations between patient safety measures and Nursing Home Compare overall and health inspection star ratings in the first quarter of 2017

Nursing Home Compare star rating

Safety measure Overall Health inspection Pressure sores (long-stay residents only) −0.21 −0.13 Pressure sores (short-stay residents only) −0.12 −0.05 Urinary tract infections (long-stay residents only) −0.05 0.04 Injurious falls (long-stay residents only) −0.06 0.02 Medication error rate >5% (all residents) −0.15 −0.17 Significant medication error(s) (all residents) −0.13 −0.14

SOURCE Authors’ analysis of data from Nursing Home Compare and Certification and Survey Provider Enhanced Reporting (CASPER). NOTES The Nursing Home Compare health inspection star rating is for the health inspections domain only, which is one of the three components used to calculate the Nursing Home Compare overall star rating. Long-stay residents are defined as people who have been residing in the nursing home for at least a hundred days. Short-stay residents are defined as people with a shorter length-of-stay, usually for post–acute care purposes. All correlations were significant (p < 0:05). Although we focused our analysis on a single quarter for ease of exposition, similar results were obtained for 2012–16.

November 2018 37:11 Health Affairs 1773

This means that nursing homes with higher star ratings generally had lower rates of both pres- sure sores and urinary tract infections, but these relationships were quite weak. Pearson correla- tion coefficients in general run from 0 to 1 in absolute value, with 0 indicating no relationship between two variables and 1 indicating the stron- gest possible relationship. The correlations be- tween patient safety measures and the star rating for the health inspections domain, which is one of three components used in the overall star rat- ing, were generally even lower. Nursing homes’ performance on patient safety

measures appears to track somewhat with the overall star rating. Notably, for the nondefi- ciency measures (injurious falls, urinary tract infections, and the two pressure sore measures), the differences were most pronounced at the extremes (one-star versus five-star ratings), with little meaningful difference between nursing homes with ratings in the two- to four-star range (exhibit 2). The one exception was pressure sores among long-stay residents, which exhib- ited a more monotonic trend by star rating, with the difference from one star (7.1 percent)

to five stars (4.2 percent) constituting a clinically meaningful magnitude. In contrast to most of the nondeficiency measures, the relationship be- tween the overall star rating and deficiency cita- tions for medication errors was more consistent. The difference between the pattern for medi-

cation-related deficiencies and other patient safety measures is likely due to the fact that defi- ciencies inherently affect the health inspection rating (even if medication-related deficiencies are only two of many), and the health inspection rating is weighted more heavily than the other domains in the overall star rating. To explore this explanation, exhibit 3 displays the relationship between mean patient safety outcomes and the star rating in the health inspection domain. Oth- er than pressure sores, the clinical outcomes had little relationship with the health inspection star rating, with two measures (urinary tract infec- tions and injurious falls) even suggesting a posi- tive relationship. However, medication-related deficiency rates still reflected the expected down- ward slope. Deficiencies for this classic patient safety issue—medication errors—appeared to be better correlated with health inspection ratings

Exhibit 2

Nursing homes’ performance on patient safety measures in the first quarter of 2017, by Nursing Home Compare overall star rating

SOURCE Authors’ analysis of data from Nursing Home Compare and Certification and Survey Provider Enhanced Reporting. NOTES The average percentages are calculated as the unweighted average across facilities in each star category of the reported percentage of residents who trigger each measure—that is, the average of facility-level percentages. The medication error measures represent the average percentage of facilities cited for the deficiency. Long-stay residents (LS) and short-stay residents (SS) are defined in the notes to exhibit 1.

Aging & Health

1774 Health Affairs November 2018 37:11

and therefore with overall NHC ratings. Our exploration of heterogeneity by nursing

home type (profit or nonprofit, independent or chain, and high or low Medicaid) revealed no substantial differences in the relationship

between the NHC ratings and patient safety mea- sures, but at least one nursing home character- istic was associated with worse absolute perfor- mance on several of the patient safety measures: For-profit facilities had higher rates of both de-

Exhibit 3

Nursing homes’ performance on patient safety measures in the first quarter of 2017, by Nursing Home Compare health inspection star rating

SOURCE Authors’ analysis of data from Nursing Home Compare and Certification and Survey Provider Enhanced Reporting. NOTES The average percentages are calculated as explained in the notes to exhibit 2. The Nursing Home Compare health inspection star rating is for the health inspections domain only, which is one of the three components used to calculate the Nursing Home Compare overall star rating. Long-stay residents (LS) and short-stay residents (SS) are defined in the notes to exhibit 1.

Exhibit 4

Performance on patient safety measures in the first quarter of 2017, by key nursing home characteristics

Characteristic

For profit Chain High Medicaid

Safety measure Yes No Yes No Yes No All nursing homes

Pressure sores (long-stay residents only) 6.0% 5.0% 5.8% 5.8% 6.2% 5.7% 5.7%

Pressure sores (short-stay residents only) 1.1 1.1 1.0 1.2 1.1 1.1 1.1

Urinary tract infections (long-stay residents only) 4.3 5.1 4.4 4.7 3.6 4.6 4.6

Falls (long-stay residents only) 3.2 3.8 3.3 3.4 2.5 3.4 3.4

Medication error rate >5% (all residents) 15.1 9.2 13.8 13.5 15.1 13.6 13.4

Significant medication error(s) (all residents) 7.3 4.0 6.4 6.0 5.7 6.3 6.3

SOURCE Authors’ analysis of data from Nursing Home Compare and Certification and Survey Provider Enhanced Reporting (CASPER). NOTES High Medicaid is defined as having at least 90 percent of residents with Medicaid as their primary payer. Long-stay residents and short-stay residents are defined in the notes to exhibit 1. The percentages are averages, calculated as explained in the notes to exhibit 2.

November 2018 37:11 Health Affairs 1775

ficiency-based patient safety measures and pres- sure sores among long-stay residents (exhibit 4). As for-profit, chain, and high-Medicaid facilities also tended to score lower on health inspections and on the NHC star ratings, this result for for-profit status was consistent with expecta- tions.22,23 At the same time, rates of urinary tract infections and falls exhibited the opposite pat- tern: For-profit, chain, and high-Medicaid facili- ties reported consistently lower rates than non- profit and independent facilities did.

Discussion Our results highlight the differences between patient safety and quality improvement, two related but distinct approaches to improving health care. In nursing homes, quality improve- ment has been emphasized, with less attention paid to patient safety. Our results reveal that Nursing Home Compare, a key policy initiative aimed at improving quality and empowering consumers to make better choices, does not pro- vide them with much information on which to judge patient safety in nursing homes. Triangu- lating among several sources of data, we found that although there was some correlation be- tween NHC star ratings and patient safety mea- sures, these tended to be weak—and for many safety measures, nonexistent. A rating of one star or five stars seemed to give the most infor- mation about patient safety, with one-star nurs- ing homes having higher rates of adverse safety events and five-star nursing homes having the lowest rates. However, for nursing homes in the middle—those with two, three, or four stars— there was no meaningful difference in adverse safety events between nursing homes with differ- ent star ratings. This lack of correlation makes interpretation more difficult for consumers, as the difference between a one-star and a two-star facility is quite different from the difference be- tween a three-star and a four-star facility. Finally, nursing home characteristics such as profit and chain status and payer mix do not serve as con- sistent proxies for patient safety performance. Nursing homes that scored well on NHC did

exhibit a more consistent relationship between the star ratings and two important measures of patient safety—rates of pressure sores among long-stay residents and citations for medication errors—but again, these correlations were small. Furthermore, a high star rating says little about therates of such outcomes as fallsand infections. In fact, we found that those facilities with the highest star ratings in the health inspections domain (arguably the most objective domain we studied) reported more falls and urinary tract infections. These measures may simply present

different clinical challenges. However, another possible explanation for this seemingly paradox- ical finding is that superior facilities may be better able to identify and track falls and urinary tract infections, which suggests an ascertain- ment bias that exposes an inherent problem in reporting systems such as NHC. Overall, our findings suggest that NHC does

not adequately reflect patient safety in nursing homes, even though it contains some patient safety measures. Accordingly, consumers who prioritize patient safety might not find NHC a useful tool with which to assess the performance of nursing homes. This conclusion suggests one straightforward and plausible solution: Nursing home performance on patient safety should be emphasized more in NHC. Specifically, to rectify the lack of focus on patient safety in NHC, we recommend that patientsafety measures be iden- tified as a separate subset within the clinical quality measures. Precedent exists for this type of grouping of measures in some state report cards.24 The development and inclusion of a mea- sure of medication errors would be a helpful addition to current patient safety measures, which could be as simple as explicitly reporting medication-related deficiencies as a new mea- sure. Despite the importance of medication er- rors to patient safety, NHC captures this safety failure only as one deficiency among many used to calculate the health inspection star rating. Raising the visibility of patient safety measures should serve not only to inform consumers but also to provide a stronger incentive for nursing homes to improve on these measures. More broadly, our findings suggest that the

patient safety movement has not been well inte-

To rectify the lack of focus on patient safety, we recommend that patient safety measures be identified as a separate subset within the clinical quality measures.

Aging & Health

1776 Health Affairs November 2018 37:11

grated into the nursing home setting. This is a missed opportunity, but it may also reflect the challenges of patient safety in nursing homes. These include the challenges of providing appro- priate care to older people with long-term care needs and the potential incompatibility between patient safety models and the nursing home setting, as the models were designed largely in the context of acute care. Not only do nursing homes serve particularly complex and vulnera- ble residents, they are also residents’ homes. Thus, care providers must balance patient safety needs with residents’ autonomy and freedom. Recent criticisms of patient safety models sug- gest that many aspects of medical care are too complex to be corrected using the models, which are based on industrial and occupational safety methods.25 This criticism may be best exempli- fied by the lack of correlation we found between nursing home star ratings and rates of falls. An outcome such as a fall represents a much more complex event, the result of myriad combina- tions of problems—for example, cognitive im- pairment, arthritis, and lack of balance. In addi- tion, other positive factors such as becoming more mobile with therapy and gaining an in- creasing degree of autonomy are both associated with an increasing possibility of falling, com- pared with being confined to one’s bed. Achiev- ing this possibility with its increased risk of falling may represent a major improvement in quality and greater resident autonomy at the same time it heralds an adverse event. In the quest to improve patient safety in nurs-

ing homes, the need to reconcile competing pri- orities for long-stay residents will present an ongoing challenge in NHC and other contexts. However, the answer to this challenge is not to

deemphasize important patient safety concerns. Rather, NHC needs to do a better job of repre- senting the competing concerns of long-stay residents by incorporating measures related to quality of life, such as resident and caregiver satisfaction. Although this would be inherently challenging because these measures are not con- veniently derived from administrative data, it is feasible. Indeed, some states have made substan- tial progress in measuring and reporting resi- dent and caregiver satisfaction.26–28 A public re- porting system that included both patient safety concerns and quality of life would be the most straightforward way to represent these two po- tentially conflicting goals in NHC, while allow- ing consumers to weight the domains according to their preferences and priorities. The challenges of patient safety in nursing

homes will remain embedded in the much broader challenges of providing appropriate care to elderly people with long-term care needs. At the same time, the predominant focus in nurs- ing homes is improving selected aspects of the quality of care, with an emphasis on nursing home star ratings.While tweaking NHC to more completely measure and report patient safety metrics is important, there remains the more central challenge of addressing the tensions and potential incompatibilities that exist between patient safety, quality of care, and quality of life in nursing homes. Within the limited scope of what a public reporting system can do, NHC should strive to achieve a more comprehensive picture that reflects all important domains of quality—including patient safety and quality of life, both of which are currently underrepre- sented. ▪

The authors are grateful for funding for this work from the Agency for Healthcare Research and Quality (Grant No. R01HS024967).

NOTES

1 Institute of Medicine. Improving the quality of care in nursing homes. Washington (DC): National Acade- mies Press; 1986.

2 Corrigan JM, Kohn LT, Donaldson MS editors. To err is human: building a safer health system. Washington (DC): National Acade- mies Press; 1999.

3 Leape LL, Berwick DM, Bates DW. What practices will most improve safety? Evidence-based medicine meets patient safety. JAMA. 2002; 288(4):501–7.

4 Shojania KG, Duncan BW, McDonald KM, Wachter RM. Safe

but sound: patient safety meets evi- dence-based medicine. JAMA. 2002; 288(4):508–13.

5 Runciman WB, Williamson JA, Deakin A, Benveniste KA, Bannon K, Hibbert PD. An integrated frame- work for safety, quality, and risk management: an information and incident management system based on a universal patient safety classi- fication. Qual Saf Health Care. 2006; 15(Suppl 1):i82–90.

6 Simmons S, Schnelle J, Slagle J, Sathe NA, Stevenson D, Carlo M, et al. (Vanderbilt Evidence-based Practice Center, Nashville, TN).

Resident safety practices in nursing home settings [Internet]. Rockville (MD): Agency for Healthcare Re- search and Quality; 2016 May [cited 2018 Sep 28]. (Technical Brief No. 24). Available from: https:// effectivehealthcare.ahrq.gov/sites/ default/files/pdf/nursing-home- safety_technical-brief.pdf

7 Simmons SF, Schnelle JF, Sathe NA, Slagle JM, Stevenson DG, Carlo ME, et al. Defining safety in the nursing home setting: implications for future research. J Am Med Dir Assoc. 2016; 17(6):473–81.

8 Mukamel DB, Spector WD, Zinn JS,

November 2018 37:11 Health Affairs 1777

Huang L, Weimer DL, Dozier A. Nursing homes’ response to the Nursing Home Compare report card. J Gerontol B Psychol Sci Soc Sci. 2007;62(4):S218–25.

9 Perraillon MC, Brauner DJ, Konetzka RT. Nursing home re- sponse to Nursing Home Compare: the provider perspective. Med Care Res Rev. 2017 Aug 1. [Epub ahead of print].

10 Abt Associates. Nursing Home Compare five-star quality rating system: year five report [public ver- sion]: final report [Internet]. Balti- more (MD): Centers for Medicare and Medicaid Services; 2014 Jun 16 [cited 2018 Sep 28]. Available from: https://www.cms.gov/ Medicare/Provider-Enrollment- and-Certification/Certificationand Complianc/Downloads/NHC- Year-Five-Report.pdf

11 Medicare.gov. About Nursing Home Compare data: technical details [In- ternet]. Baltimore (MD): Centers for Medicare and Medicaid Services; [cited 2018 Sep 28]. Available from: https://www.medicare.gov/Nursing HomeCompare/Data/About.html# technicalDetails

12 Vlaeyen E, Coussement J, Leysens G, Van der Elst E, Delbaere K, Cambier D, et al. Characteristics and effec- tiveness of fall prevention programs in nursing homes: a systematic re- view and meta-analysis of random- ized controlled trials. J Am Geriatr Soc. 2015;63(2):211–21.

13 Cameron ID, Gillespie LD, Robertson MC, Murray GR, Hill KD, Cumming RG, et al. Interventions for preventing falls in older people in

care facilities and hospitals. Cochrane Database Syst Rev. 2012; 12:CD005465.

14 Washington EA. Instillation of 3% hydrogen peroxide or distilled vine- gar in urethral catheter drainage bag to decrease catheter-associated bac- teriuria. Biol Res Nurs. 2001;3(2): 78–87.

15 Fendler EJ, Ali Y, Hammond BS, Lyons MK, Kelley MB, Vowell NA. The impact of alcohol hand sanitizer use on infection rates in an extended care facility. Am J Infect Control. 2002;30(4):226–33.

16 Sullivan N, Schoelles KM. Prevent- ing in-facility pressure ulcers as a patient safety strategy: a systematic review. Ann Intern Med. 2013; 158(5 Pt 2):410–6.

17 Marasinghe KM. Computerised clinical decision support systems to improve medication safety in long-term care homes: a systematic review. BMJ Open. 2015;5(5): e006539.

18 Fleming A, Browne J, Byrne S. The effect of interventions to reduce po- tentially inappropriate antibiotic prescribing in long-term care facili- ties: a systematic review of rando- mised controlled trials. Drugs Aging. 2013;30(6):401–8.

19 RTI International. MDS 3.0 quality measures: user’s manual (v10.0 03- 25-2015): effective April 1, 2016 [Internet]. Baltimore (MD): Centers for Medicare and Medicaid Services; 2016 Apr [cited 2018 Sep 28]. Available from: https://www.cms .gov/Medicare/Quality-Initiatives- Patient-Assessment-Instruments/ NursingHomeQualityInits/

Downloads/MDS-30-QM- Users-Manual-V10.pdf

20 Mor V, Zinn J, Angelelli J, Teno JM, Miller SC. Driven to tiers: socioeco- nomic and racial disparities in the quality of nursing home care. Mil- bank Q. 2004;82(2):227–56.

21 To access the appendix, click on the Details tab of the article online.

22 Harrington C, O’Meara J, Kitchener M, Simon LP, Schnelle JF. Designing a report card for nursing facilities: what information is needed and why. Gerontologist. 2003;43(Spec No 2): 47–57.

23 Konetzka RT, Gray ZS. The role of socioeconomic status in nursing home quality ratings. Seniors Hous Care J. 2017;25(1):3–14.

24 Arling G, Job C, Cooke V. Medicaid nursing home pay for performance: where do we stand? Gerontologist. 2009;49(5):587–95.

25 Braithwaite J,Wears RL, Hollnagel E. Resilient health care: turning patient safety on its head. Int J Qual Health Care. 2015;27(5):418–20.

26 Shippee TP, Henning-Smith C, Gaugler JE, Held R, Kane RL. Family satisfaction with nursing home care. Res Aging. 2017;39(3):418–42.

27 Shippee TP, Henning-Smith C, Kane RL, Lewis T. Resident- and facility- level predictors of quality of life in long-term care. Gerontologist. 2015; 55(4):643–55.

28 Williams A, Straker JK, Applebaum R. The nursing home five star rating: how does it compare to resident and family views of care? Gerontologist. 2016;56(2):234–42.

Aging & Health

1778 Health Affairs November 2018 37:11

Brauner D, Werner RM, Shippee TP, Cursio J, Sharma H, Konetzka RT. Does nursing home compare reflect patient safety in nursing homes? Health Aff (Millwood). 2018;37(11).

Appendix Exhibit 1: Summary of the Sample

Number of Nursing Homes 15,652

Size and Occupancy, mean (SD)

Total residents 86.1 (54.0)

Certified beds 106.2 (60.9)

Occupancy, % 80.6 (16.1)

Ownership

For profit 10,855 (69.4%)

Government 1,114 (7.1%)

Nonprofit 3,683 (23.5%)

Certification

Medicaid only 382 (2.4%)

Medicare only 758 (4.8%)

Medicaid and Medicaid 14,512 (92.7%)

% Medicare residents, mean (SD) 14.6 (15.0)

% Medicaid residents, mean (SD) 58.5 (24.8)

Hospital-based

No 14,862 (94.9%)

Yes 790 (5.0%)

Nursing Home Compare Overall Star

Rating

1 (lowest quality) 2,540 (16.2%)

2 2,954 (18.9%)

3 2,843 (18.2%)

4 3,537 (22.6%)

5 (highest quality) 3,647 (23.3%)

Patient Safety Measures, mean percent

of residents (SD)

Pressure Sores (long stay) 5.7 (4.4)

Pressure Sores (short stay) 1.1 (1.7)

Urinary tract infections (long stay) 4.6 (4.3)

Falls (long stay) 3.4 (2.8)

Medication error deficiency citations

Error rate > 5%, number (%) cited 2,106 (13.4)

Significant medication error(s),

number (%) cited

984 (6.3)

SOURCE: Authors’ analysis of Nursing Home Compare and CASPER data.

Reproduced with permission of copyright owner. Further reproduction prohibited without permission.