Care Transitions and Safety

Transitions between care settings (acute to post-acute care to home, long-term care to home, or transition to and from the emergency department) create vulnerability for multiple adverse outcomes. Older, hospitalized patients discharged to skilled nursing facilities (SNFs) are more impaired than those discharged home, and this population may experience an exacerbation of their clinical conditions (e.g., unintentional weight loss, depression, pain) as a result of the hospitalization event. In recent studies, the transition from acute to post-acute care has been shown to be fragmented with incomplete clinical information necessary to provide care safely.^{24, 25} Recent data also suggest that at least some older adults are discharged from the hospital too soon with conditions that could be best treated by remaining in the hospital for a longer period of time²⁶; however, the scope and severity of this problem is unknown.

In short, the probability of adverse events during the transition from hospital to SNFs increases due to the stress of the hospitalization event and the care transition process, a lack of timely, accurate clinical information necessary to care for this complex patient population in the nursing home setting, and the possibility that some patients may not be ready to leave the hospital. For example, adverse events related to medication errors in the SNF cited in the OIG report² may be associated with multiple medications, particularly new medications, being prescribed for older patients discharged from the hospital to this care setting. In a recent study, hospitalized patients discharged to SNFs had an average of 13 medications on their hospital discharge list.²⁷ Thus, SNF providers face the challenge of starting complex new medication regimens with little knowledge of the patient or their medication history upon admission. These same safety issues exist for long-stay residents but to a lesser degree due to fewer care transitions and increased staff knowledge of the resident.

In view of evidence that many hospital readmissions may be avoidable, CMS is currently developing new quality indicators to monitor 30-day readmissions from the SNF as well as successful discharge from SNF to home. In addition, other evidence suggests that many emergency room visits from both the SNF and long-stay nursing home segments may be avoidable, even without hospitalization.²⁸ Thus, separate measures of emergency room visits are also being planned for public reporting and quality monitoring in the nursing home setting.

Rehabilitation Services and Safety

Another primary difference between short- and long-stay residents is the differential focus on rehabilitation services in preparation for discharge home. A major goal of skilled nursing care is to effectively rehabilitate residents to allow discharge to the least restrictive care environment possible which, in most cases, is a return to a community setting (e.g., own home residence or assisted living). Very few long-stay residents are expected to be discharged home or to another community setting. In contrast, 69 percent of SNF residents who were living at home prior to admission return home after their SNF stay.²⁹ Similar to the acute care to post-acute care transition, safety concerns relate to the transition from skilled nursing care to home and include the need for comprehensive care planning; effective, timely communication with outpatient provider(s), scheduled followup visits within less than 30 days of SNF discharge, persons’ ability to safely manage their new medication regimen at home, and an accurate assessment of the need for other support services at home such as transportation, meals and home health).³⁰

Although short-stay residents are generally expected to return to community settings, one recent study reported that 33 percent of short-stay residents discharged home suffered an adverse medication-related event within 45 days of discharge, and only 28 percent of this group remained living at home 90 days after discharge.^{31, 32} Ten percent experienced a hospital readmission within 30 days.²⁹ Based on these findings, new quality indicators are being developed by CMS and others for short-stay residents discharged home related to hospital readmission rates, with the rationale that this measure reflects the quality of the SNF discharge process.

Falls

Multifaceted and individualized falls prevention programs that have demonstrated effectiveness in hospital studies include a focused patient history and physical; educational programs for patients and staff; toileting programs; providing walking aids and making sure they are accessible and used sufficiently; and/or supervised exercise programs, especially those that combine balance, strength and power training. These appear to be most effective when designed and delivered by multidisciplinary teams.^49–53

Medication Errors

Medication review, either by pharmacists during the prescribing phase and/or by patients upon admission to the hospital and/or at discharge, has been associated with reduction in medication errors, including prescribing and dispensing errors; preventable adverse drug events; and adverse drug events (ADEs).^54–60 Clinician engagement, multidisciplinary communication, and the review of medication lists and related risks for risk prevention and patient education, especially with older patients who are at risk due to higher acuity and/or actively take several medications, has also been shown to be effective and increase patient safety in the acute hospital setting.^61–64 It would seem that these types of interventions may be useful in the nursing home setting, particularly among short-stay residents who will be discharged home, and several systematic reviews (GQ3) have addressed medication review in the nursing home.

It is unclear whether technology intensive interventions, which are commonly seen in hospitals, could translate to the nursing home. Some nursing homes have implemented technologies including electronic medication administration systems, with outcomes including reduced errors and identification of key times or situations in which medication errors may occur.^65–68

Health information technology (HIT), such as computerized physician order entry (CPOE) and electronic health records (EHR) and databases that provide decision support (e.g., drug-drug interaction alerts, evidence-based guidelines, dosing alerts, etc.) have been shown to be effective and increase quality and medication safety as well as being more conducive and efficient for medication and chart reviews in the hospital setting.^69–76 It is important to note that hospital resources for intervention implementation included an emphasis on bioinformatics and technological solutions, which may lessen the applicability of many of these interventions in the nursing home setting, or at least limit their generalizability, unless nursing homes also have active and rich informatics systems for health records and management.

Pressure Ulcers

Most individually reported studies regarding pressure ulcers in hospitals assess treatment of ulcers; however, quality indicators for nursing homes are related to the prevention of pressure ulcers. Therefore, a robust body of evidence on wound healing in hospitals provides limited guidance for long-term care beyond care practice guidelines to inform treatment of existing wounds.^77–81

Infection

Catheter-associated UTIs (CAUTIs) may be the most avoidable type of hospital-associated infection (HAI).⁸² A variety of strategies and guidelines for prevention and reduction in catheter-associated infections have been produced^83–89 as well as some AHRQ evidence reports/technology assessments on all healthcare-associated infections. These reports may provide some approaches applicable to use in nursing homes.^{90, 91} Strategies for CAUTI prevention include proper techniques for urinary catheter insertion, which is reinforced by staff training; minimizing urinary catheter usage; using a closed urinary drainage system; avoiding catheter usage for incontinent patients; using external catheters instead of indwelling catheters, if possible; documentation of key information related to urinary catheters; and stop orders or reminders to remove such catheters. These safety issues related to catheter use provide the rationale for the current quality indicator “catheter left in bladder” publicly reported by nursing homes via the CMS website.

Other common infection control approaches in hospitals with applicability to the nursing home include hand washing, environmental cleaning, and antibiotic stewardship. A recent review of antibiotic stewardship interventions in the intensive care unit, which typically seek to optimize appropriate treatment with antibiotics and minimize inappropriate use and antimicrobial resistance, reported reductions in rates of antibiotic use and antimicrobial resistance but no benefits on survival.⁹² A 2013 Cochrane review evaluating clinical trials of interventions for antibiotic stewardship in inpatients similarly reported reductions in antibiotic prescribing and reductions in hospital-acquired Clostridium difficile, Methicillin-resistant Staphylococcus aureus (MRSA), and Enterococcus faecalis infections. Interventions also reduced pneumonia-associated mortality.⁹³ Multiple reviews have reported on the effectiveness of hand hygiene approaches such as chlorhexidine bathing and alcohol rubs at reducing rates of infections or microbial activity,^94–98 but sustained adherence to such approaches is challenging. Reviews of environmental cleaning have discussed the need for thoughtful cleaning practices, especially of high touch sites and sites near the patient, to control potential outbreaks.⁹⁹ Overall, reviews addressing hospital infection control practices commented on a need for multifaceted interventions that combine elements of cleaning, hand hygiene, appropriate control of contacts, and behavioral and system approaches to promote adherence to infection control measures. These elements are equally applicable to nursing home settings, where infection control processes are variable and staff may not be adequately trained in prevention and control measures.^{100, 101}

Finally, in addition to individual studies, another and potentially more informative data source is national data available before and after large-scale payment reform. Specifically, in 2008, Medicare implemented a no-payment policy for specific conditions thought to reflect safety problems with hospital care including new pressure ulcers, falls, and catheter-related infections. Each of these adverse events is also a significant safety issue of concern in the nursing home setting. If scalable interventions are available in the hospital setting, one could reasonably expect that rates should decrease as hospitals strove to implement services to prevent loss of income due to this policy. This should signal the true availability and potential scalability of purported positive intervention effects. In reality, data are mixed or negative on the degree to which improvements were seen after implementation of the no-payment policy in 2008.

Contradictory results reported by studies of pressure ulcer incidence may serve as a good example of issues related to data sources and collection. One retrospective, observational study reported significant reductions but acknowledged that the approach to data collection, which did not necessarily include all diagnosis codes in a given discharge, had high specificity for stage II and IV health care acquired pressure ulcers (HAPUs), but low sensitivity.¹⁰² On the other hand, a quasi-experimental study of 1,381 hospitals participating in the National Database of Nursing Quality Indicators (NDNQI), which relied on trained nurses to assess HAPU prevalence, reported no effect of the payment policy on HAPU incidence. The NDNQI pressure ulcer indicator used in this study is, in fact, that endorsed by the National Quality Forum (NQF).¹⁷

A second study using a non-claims database reported no improvement in injurious falls when measured before and after the introduction of the no payment policy. Data on hospital-acquired infections were similarly mixed with some studies reporting improvements and others reporting no change. National data¹⁰³ suggest substantial decreases since 2008 in many infections but not CAUTI, which have increased by about 6 percent.

Some recent data suggest that this trend is reversing, and the increase in CAUTI is contrary to clearly evidenced reductions in other infections, including central line associated bloodstream infections and hospital-onset MRSA and C. difficile infections. In studies specifically intended to provide data before and after implementation of the policy, conclusions are mixed.^104–107 Indeed, some studies suggest that analyses based on hospital coding data, intended for billing purposes and reported for quality may not provide objective or optimal sources of information.

In general, hospitals have invested considerable resources in improving safety. Hospital safety has improved in many areas that were not the focus of this review (e.g., surgical adverse events). However, in our brief review for the purposed of this Technical Brief, we could not ascertain that hospital safety has improved for the areas of most relevance to long-term care, with the possible exception of medication errors. Mixed evidence about hospital-based improvements in falls, pressure ulcers and catheter-related infections, even after a “no payment” policy was implemented, suggest that interventions to improve safety that have specific relevance to the nursing home population may not be consistently implemented in the hospital.

Overview of All Systematic Reviews

As a technical brief, the purpose of this report is not to describe outcomes, but to enumerate the number and types of studies available to contribute to an evidence base. We captured this information from existing systematic reviews, supplemented with newly published studies. Studies of interventions intended to improve safety outcomes among the nursing home population may be targeted interventions to change specific outcomes, or more general or multifaceted interventions with multiple outcomes. In this section, we provide an overview of literature on the safety outcomes noted in Table 3. The reviews meeting our inclusion criteria regarding the four safety targets outlined by AHRQ may have addressed multiple outcomes but typically focused on a primary outcome such as falls prevention.

Some reviews also broadly targeted older adults and may only include a small number of studies that were conducted in nursing homes, but also include hospital-based or community based settings. We elected to include these reviews in the Technical Brief to provide a comprehensive picture of the literature potentially relevant to older adults rather than eliminating reviews with mixed care settings. All reviews included nursing home studies. Whenever possible, we focused on data from nursing home or long-term care settings (Appendix C). We also provide an estimate of new comparative studies of interventions published since the ending search date of the reviews addressed in each section and identified in our screening of the primary literature (Appendix D). The purpose of this estimate is to begin to identify areas of research that might warrant an updated or new review. We classified studies as randomized controlled trials (RCTs); prospective intervention studies, which included cohort studies with comparison groups and non-randomized trials; or pre-post studies with comparison groups.

Interventions To Prevent Falls

A wide range of interventions may be associated with falls prevention, including those interventions aimed at toileting, preventing delirium, ensuring appropriate medication use and use of physical restraints. Thus, the reviews included here, which focus specifically on falls prevention interventions represent a subset of potential studies on this topic. Tables 4 and 5 outline the outcomes and intervention components addressed in studies included in the 14 reviews of falls prevention approaches.^{51, 108–120}

Table 4

Outcomes addressed in systematic reviews for falls prevention.

Table 5

Frequently reported components of falls prevention intervention studies.

Several reviews assessed multiple interventions or multimodal approaches.^{51, 108, 111, 112, 114, 117, 120} Among those with low risk of bias, the most recent was published in 2015 and included 13 RCTs published through 2013.¹⁰⁸ Interventions assessed in the review included staff training, written materials, informatics tools for appropriate medication use, vitamin D supplementation, exercise programs, modifications to the environment, management of urinary incontinence and nutrition therapy. Outcomes included numbers of falls, numbers of fallers and numbers of recurrent fallers (individuals who had prior falls). A 2012 Cochrane review included 60 RCTs, also with a diverse set of interventions that were implemented in nursing homes and in hospitals.¹¹¹ This review also examined number of falls and fallers, as well as fractures, complications and economic outcomes. A 2010 review addressing multiple interventions including Vitamin D, exercise, pharmacologic therapies, and hip protectors included 20 RCTs, most assessing Vitamin D or hip protectors.¹¹⁷ A 2007 review of 43 studies assessed multiple interventions including Vitamin D supplementation and hip protectors to prevent fractures in the nursing home.⁵¹

Other reviews specifically examined a single category of intervention such as exercise,^{109, 119} environmental modification,¹¹³ hip protectors,^{115, 116, 118} and vitamin D supplementation.¹¹⁰ The most common outcomes studied in these systematic reviews were the number and rate of fractures and the number and frequency of falls (Table 4).

Table 5 outlines the most frequently evaluated interventions included in unique studies across reviews (individual studies were typically included in multiple reviews). Interventions evaluated in studies typically included multiple components, and we classified components of interventions broadly. We recognize that some elements could be categorized in multiple ways, but the table is intended to present a broad estimate of the approaches addressed in studies. We note also that residents frequently received elements of usual care that are not reflected in the tables reporting components.

Pressure Ulcers

Eight systematic reviews specifically focused on preventing and treating pressure ulcers.^{47, 80, 121–126} One low risk of bias review focused on repositioning and included only three RCTs, with a final search date in 2013.¹²¹ Two low risk of bias reviews conducted by AHRQ EPCs addressed multiple interventions for pressure ulcer prevention¹²⁵ and treatment.¹²⁴ The prevention-focused review included roughly 20 comparative studies (including an intervention and comparison group) in nursing home or long-term care settings. The treatment-focused review included approximately 45 randomized or observational studies in nursing home or long-term care facilities.

Two additional reviews (both high risk of bias) assessed multiple interventions.^{123, 126} Three other reviews with high risk of bias largely focused on education and process of care. Table 6 outlines components frequently addressed in studies included in the reviews.

Table 6

Frequently reported components of interventions described in reviews of pressure ulcer prevention and treatment.

Infections, Including Healthcare-Associated Infection, Urinary Tract Infection, and Antibiotic Stewardship

Three reviews addressed infection prevention.^127–129 In one Cochrane review of methods to prevent transmission of MRSA in nursing homes, only one study, a cluster randomized trial of 32 sites (16 matched pairs), was included. The intervention included providing baseline data and infection control education to the staff as well as infection control audits.¹²⁷ Another review evaluated evidence for infection prevention interventions from 24 studies (16 RCTs)¹²⁹ addressing therapeutic or educational interventions including oral hygiene education, antifungal medications, hand sanitizer, vitamin E or other supplements, chlorhexidine bathing, and pneumococcal vaccine across multiple infection sites. The final systematic review addressed the preventive effect of oral hygiene on pneumonia and respiratory tract infection (RTI) in older adults in nursing homes and hospitals. Table 7 lists components frequently included in interventions for infection control in the nursing home.

Table 7

Frequently reported components of interventions described in reviews of infection prevention.

Medication Errors and Adverse Drug Events (e.g., Delirium), Including Inappropriate Medication Use and Polypharmacy

Eleven systematic reviews focused on identifying interventions to prevent medication errors and adverse drug events through approaches to improve prescribing and reduce inappropriate use of drugs. Five reviews focused broadly on optimizing prescribing or medication safety.^{76, 130–134} One addressed antibiotic prescribing specifically,¹³⁵ two evaluated medication reconciliation or communication of medication regimens in care transitions,^{130, 136} and a third evaluated specific effects of medication reviews on mortality and hospitalization.¹³⁷ A final review examined a computer-based approach to identify medications that may contribute to delirium.¹³⁸ To categorize approaches addressed in studies in these reviews further, Table 8 outlines frequently studied components of the interventions.

Table 8

Frequently reported components of interventions described in reviews of medication errors and adverse drug events.

One Cochrane review on polypharmacy included three studies conducted in the nursing home setting.⁷⁶ Most studies evaluated multicomponent interventions (including education and medication review components) intended to promote appropriate medication use. Another review focused on medication reconciliation during transition to and from long-term care.¹³⁰ A second review addressing communication of medication regimens in transitions between hospitals and nursing homes included three studies of medication reconciliation or transfer documentation/summaries.¹³⁶ One review of third party medication reviews in nursing home residents to reduce mortality and hospitalization included seven RCTs and five non-RCTs.¹³⁷ Another review focused specifically on clinical decision support systems to improve medication safety in long-term care settings¹³⁴ and included seven studies (5 RCTs).¹³⁴ One Cochrane review identified two trials that focused on non-pharmacologic delirium prevention approaches.¹³⁸

Across systematic reviews, medication review and clinician or pharmacist education were frequently the focus of interventions. Table 8 provides an estimated count of the intervention components frequently addressed across the nursing home studies included in the reviews. We note that reviews also included studies conducted in hospital or other non-nursing home settings. In addition, we classified components of interventions broadly. We recognize that elements could be categorized in multiple ways, but the table is intended to present a broad assessment of approaches addressed in studies.

Primary Studies Published Since the Publication of the Included Systematic Reviews

We identified an estimated 90 unique new comparative studies evaluating safety-related interventions in the nursing home (Table 9). A reasonable number of RCTs suggests a growing evidence base of potentially high quality studies.

Table 9

Overview of estimated new studies of nursing home safety interventions.

Research Evidence

Perhaps due to mixed results and lack of consensus in the literature, uptake of specific interventions to enhance safety in the nursing home appears to be limited. That said, a fairly large body of primary literature published after the latest systematic reviews addressed in GQ3 is available, and this new literature could inform an update of existing findings for many safety areas in the nursing home care setting, such as the staffing resources necessary for intervention delivery.

Barriers to Uptake

Three primary barriers to uptake appear to be a) a lack of consensus related to the level of adverse events that may be acceptable and thus a target for interventions; b) lack of evidence that Federally-collected quality measures accurately reflect quality in the nursing home setting, and c) lack of implementation data from effective interventions that would support uptake, such as the staffing resources necessary for intervention delivery.

First, limited “natural history” evidence exists to provide expected levels of different safety outcomes absent intervention, given that some degree of decline and associated clinical events will certainly occur in this vulnerable and complex population. It is possible that nursing homes could more confidently adopt evidence-based care practices with realistic targets for achievable outcomes.

Second, the current practice of comparing nursing homes on quality measures to identify variability (e.g., which homes are in top 10^th percentile for falls) assumes that the variability in rates between homes is a true reflection of differences in the quality of care or safety practices within the facility. Empirically, however, the connection between quality measures and facility practices has not been well established for many quality measures and, thus, warrants further consideration.

Third, most intervention studies that have demonstrated positive outcomes do not report either the resource needs or the specific care processes involved in the intervention. This lack of information makes it difficult to achieve the uptake of even successful safety interventions with any fidelity. In the few studies wherein the necessary resources, particularly staffing, have been reported, resources often exceed the typical operating capacity of the facility, suggesting that modifications of effective interventions may be necessary to support uptake in daily care practice in most nursing homes.^139–141

Refine Our Understanding of Safety Within the Nursing Home Context

This report is based on the core measures already identified by AHRQ as safety indicators in the nursing home setting. However, team experts and key informants felt that this small set of four indicators, which were largely informed by hospital practices, did not fully capture the safety issues for nursing home residents.

In reality, there are a number of contributing factors to pressure ulcers, falls, infections and medication errors that may also lead to events that cause injury and adverse events outside of these four safety domains, which are the focus of this report. These conditions and associated adverse events may also be improved with interventions, which suggests that they meet the AHRQ safety definition of a “process or structure that prevents adverse events.” Thus, these domains may provide amenable targets for intervention.

Potential domains that meet these criteria include unintentional weight loss (now a safety quality measure endorsed by the NQF¹⁷); dehydration; decline in ADLs, typically defined as basic tasks of everyday life (eating, bathing, dressing, toileting, transferring to different locations such as chair to bed ); functional independence; fecal and urinary incontinence, including constipation, depressive symptoms, moderate to severe pain, influenza vaccine; pneumococcal vaccine; physical restraints; and catheter left in bladder.^{111, 142–145} It is important to note that these factors may not be possible to mitigate or avoid entirely in the nursing home setting, particular for some segments of the population (e.g., reductions in ADL in severely impaired individuals).

Many of these known factors are commonly referred to in the context of quality of care or quality of life, including those indicators that are tracked via the federally-required Minimum Data Set (MDS), described in Nursing Home Compare, captured through the annual survey and certification process, or identified through complaints. By comparing these sources as well as common litigation targets²² and consulting with key informants, we identified ten conditions (Table 10) that, if not addressed, have substantial potential to result in a safety issue. For example, decline in ADLs could lead to an increased likelihood of falling, and a catheter left in too long could lead to infection. Indeed, increasing support is coalescing around an expanded consideration of safety in this setting, including the recent addition to the NQF safety indicators of specific measures on weight loss and decline in ADLs.¹⁷

Table 10

Common contributing factors to safety events in nursing homes.

Our research team and key informants have proposed that, to best understand and address safety in the nursing home setting writ large, the elements in Table 10 should also be targets for intervention and thus added to future research agendas. A central component of this research would be to identify and explicate any empirical link between these potentially contributing factors and clinical safety outcomes. We recognize that additional factors such as dementia also represent important factors to consider in future research; however, we focused in this brief on potentially preventable issues. Issues such as dementia or physical dependency are also prevalent among nursing home residents and therefore may set the stage for issues described in Table 10 to lead to safety events over time. In sum, future research efforts should consider which measures constitute the body of target outcomes relevant to nursing home safety – both final clinical outcomes and contributing factors. These research efforts should also include understanding what areas or factors are addressed in existing studies and the extent to which these studies offer insights into understanding safety in the nursing home and potential safety-related interventions.

Encourage Implementation Research Specific to Nursing Home Safety

Many studies in this field, to date, have not included standardized, independent measures of safety outcomes, complete information on the care processes thought to be related to those outcomes, details about the staffing resources (number, time, equipment) necessary to provide the interventions, or leadership, regulatory, and payment characteristics that may affect implementation. Because these components are so often lacking in the literature, it is difficult to determine to what extent mixed results in prior systematic reviews for various outcomes are attributable to a lack of an intervention effect or lack of intervention fidelity or lack of resources to support the intervention. It is noteworthy that many of these same implementation issues were discussed in a recent AHRQ-sponsored report on non-pharmacological interventions for behavioral disturbance in long-term care settings.¹⁴⁶ The weaknesses in implementation contribute to the weak/mixed evidence base, a dearth of incentives to change care practice, and questions about staffing resource requirements necessary to improve outcomes as well as what outcomes are even achievable in the nursing home setting. These are inter-related, important issues that should be addressed in future studies, and we provide more specific discussion of some of these implementation research issues in the remainder of this section. In addition, understanding of the sustainability of effects of intervention is lacking; future research should consider how to maximize durability of positive outcomes of interventions.

Develop Consensus Around Common Outcomes

As noted above, specific outcome performance standards (e.g., absolute rather than relative performance) for acceptable quality of care and safety in nursing homes are absent and this likely impedes the uptake of interventions with demonstrated effectiveness. One way to achieve performance standards is to actively develop consensus among experts in the field. Another approach is to encourage the conduct of implementation science to help identify what is achievable under controlled conditions, and, in this context, describe fully the resources and circumstances needed to achieve those outcomes. Finally, better characterization of implementation research also could provide important clinical information about the resident characteristics that may modify the effectiveness of specific interventions.

Specific research questions include:

• What are the maximum achievable outcomes (e.g., fall or pressure ulcer incidence rate reduction) when specific care processes thought to be related to the outcomes are implemented with high fidelity?
• What are the resources required to consistently implement the intervention to all residents in need within a given facility (e.g., number of staff, training, equipment)?
• What resident characteristics modify intervention effectiveness such that clinically meaningful criteria can be used to best target interventions, especially in the context of limited staffing resources?

Empirically Assess the Role of Performance Monitoring Approaches in Nursing Home Settings

Outside of the Federal and State inspection process and recent targeted chart review protocols used in OIG studies,² most research to date has relied solely on self-reported information from nursing facilities. Because discrepancies have been noted between self-reported and externally collected data, validation research as well as direct observation studies would be informative.^{2, 147–149}

In particular, questions that have arisen about the accuracy of self-reported staffing information that is part of the current 5-Star reporting systems, staffing data submitted to CMS will soon be based on nursing home payroll data in lieu of facility self-report. As an example of changes in practice that may improve available data, CMS has now instructed survey staff to increase their audits during the survey process.^{150, 151}

Research questions include:

• What auditing approaches are most efficacious and effective more broadly for verifying the accuracy of adverse event reporting in nursing homes?
• Does the effectiveness of such auditing differ based on the frequency of the audits?
• What are the most accurate information sources to identify safety issues in the nursing home setting?

Rigorously Study the Role of Staffing Models, Levels, and Types of Staff in Achieving Safe Nursing Home Environments

In contrast to hospitals and hospital care, nursing homes rely on many non-licensed personnel (e.g., nurse aides) who are responsible for the majority of labor-intensive and non-clinical ADL care (e.g., helping residents get in/out of bed, dressing, bathing, feeding, toileting, walking assistance) delivered to residents over a longer period of time. Also in contrast to hospitals, nursing homes rely more on Licensed Practical Nurses or Licensed Vocational Nurses (LPNs/LVNs) as opposed to Registered Nurses (RNs), which may have implications for medication errors as well as overall clinical management. A mixed body of research has explicitly explored the association of staffing levels and quality of care issues that may contribute to safety, with most studies reporting positive outcomes and others not demonstrating any effect. Systematic and narrative reviews have identified substantial variation in the study methods and measurement of staffing interventions. Studies overall were considered methodologically flawed in the reviews that reported study quality, and reviews generally commented on variation in methods to measure staffing.^152–156

The largest and most standardized database on this topic is perhaps the five-star nursing home rating system, which catalogues staffing, survey deficiency and quality outcome data for all nursing home facilities in the U.S. on a routine basis using a standardized reporting format. Data from these reports also suggest positive associations between total staffing (i.e., licensed nurses plus nurse aides), survey deficiencies, 30-day hospital readmissions from SNF and successful community discharge from SNF.^157–159 However, the mechanism through which staffing may be affecting care quality in these five star reports is unclear.

Future studies should be used to determine the effects of specific staffing models on care processes related to safety outcomes. Concurrently, effectiveness studies should report details about staffing that can be used to assess this potential modifier of effectiveness.^139–141 An increased focus on a more rigorous application of implementation science in the evaluation of nursing home interventions could provide a basis for understanding the role of staffing models in the future and would, in fact, reduce the need for minimum staffing standards in nursing homes by providing more nuanced information about models and staffing types that affect outcomes. Many nursing homes nationwide currently report total staffing levels that are consistent with expert consensus recommendations (Table 1). However, staffing levels and types of staff still vary significantly among facilities, with little evidence to suggest that any particular model is optimal for improving quality and safety.^{152, 160} A potential reason for the current variability in staffing levels is the absence of reliable and empirically established models describing an optimal level and mix of staffing resources based on resident acuity.

Research questions include:

• What are the staff-time requirements and type(s) of staff associated with interventions known to prevent adverse events?
• Given staff time requirements and staffing skill mix, are any staffing models more efficient and effective than others in practice for preventing adverse events?
• How can labor resource data be converted into an information system useful for planning the number and skill mix of staff necessary to prevent adverse events?

Better Understand What Works for Staff Training and Management and Organizational Culture

At a very practical level, methods of training and managing staff should be studied and reported, including operational research and rigorously conducted qualitative research to understand contextual factors with the nursing home setting and staff perspectives. It is conceivable that managers may not be cognizant of safety lapses in their facilities or the best approaches to training staff, documenting care, or implementing programs to improve safety. Recent studies from multiple hospital systems have shown that both nurses and patients report frequent omitted care, particularly care related to pressure ulcer and fall prevention (e.g. repositioning and mobility assistance). Hospital nurse staffing levels have been shown to be the primary predictor of care omissions and there is no reason to believe that this is not also a potential issue in nursing homes^{161, 162} Rigorous evaluations of both staff training and management models would provide needed information for broader implementation. Potential examples include web-based approaches that would be continually available to new staff. Programs that do exist, including the Quality Assessment and Performance Improvement (QAPI) framework, should be rigorously evaluated.

An intervention that currently lacks rigorous evaluation is the use of point of care documentation systems, which are commercially available, that may provide timely methods to identify care frequency and care omissions. A second solution that does not require technology is the use of standardized observational tools by managers to intermittently monitor daily care processes related to adverse events. Recent studies have demonstrated that standardized observations can be used by both nursing home staff and federal surveyors to monitor nutritional care quality and resident-centered care practices^{147, 163–168} but both technological and non-technological approaches to management improvement need to be rigorously evaluated in future studies.

Further, understanding of how management and leadership practices and work processes affect care and the safety-related climate of nursing homes is also needed, and some literature from the hospital setting may guide efforts in this area.^169–171

Research questions include:

• What is the effectiveness of various staff training and management models for improving staff ability to provide optimal care consistently?
• What is the role of direct observation in multicomponent approaches to staff management of clinical care?
• What is the effectiveness of point-of-care documentation systems for reducing adverse events?
• What is the role of care process information (collected either through technology or standardized observations) for improving staff performance?
• What are the costs of implementing new care process documentation systems or the staff training and management models based on the data generated by these systems?
• How do leadership style, management practices, and work organization affect care and safety?
• How do organizational and management factors contribute to a culture of safety?

Understand the Effects of Care Omissions on Safety

A number of researchers have suggested that the basis for safety issues may be identifiable omissions of care, and thus that omissions of care may be a focus of study. Care omissions can be defined as (1) care documented in a resident’s medical record but not actually provided by staff; and, (2) the presence of a clinical condition not identified by staff and thus not reflected in the care plan and/or treatment decisions. Finally, prolonged delays in care delivery wherein care is provided but not in a timely manner may occur (e.g., delayed incontinence care or repositioning).

Care omissions may reflect a discrepancy between medical record documentation and direct observations of care delivery in the nursing home setting for incontinence care (e.g., toileting assistance and changing), feeding assistance during meals, nutritional supplement and snack delivery between meals, repositioning and mobility exercise.¹⁴⁸ Separate studies have shown that many clinical conditions go unrecognized and untreated by staff including depression, moderate to severe pain, inadequate food and fluid intake and unintentional weight loss.^{147, 172–175} Similarly, other clinical conditions may be recognized by staff but not adequately addressed in resident care plans (e.g., incontinence and a toileting program). A recent Government Accounting Office (GAO) report indicated that fall risk, nutritional problems and incontinence are commonly experienced by nursing home residents but are not addressed in care plans.^{78, 159} Prolonged and/or excessive omissions of care may increase a resident’s risk for harm and subsequently be deemed as neglect/abuse in litigation (e.g., recurrent episodes of inadequate feeding assistance can lead to weight loss/dehydration).

Research questions include:

• How can omissions of care be captured and measured?
• How do omissions of care contribute to adverse events?
• How do omissions of care for risk factors related to pressure ulcers and falls (e.g., exercise and incontinence care) affect these outcomes?

Rigorously Evaluate Person-Centered Care

Key Informants added one additional concept to key safety and quality issues, that of person-centered care as a way of supporting an environment less susceptible to safety lapses. A central tenet of person-centered care is staff compliance with residents’ preferences and/or the ability of residents to make choices about aspects of their daily care and lives. Person-centered care approaches may have a potential impact on resident wellbeing and quality of life, and anecdotal evidence suggests that some approaches (e.g., Eden Alternative, Greenhouses) may produce both clinical and quality of life benefits. However, the research literature is fairly limited, and most often does not explicitly describe the key components of person-centered care in a way that is measurable, and thus replicable, by other facilities.

Given potential trade-offs between personal freedom (a common tenet of these approaches) and safety, good evaluations are needed to better understand the role and optimal implementation of person-centered care. Conversely, studies to evaluate clinical interventions to improve resident health status and other clinical outcomes also should consider the potential risks and benefits related to residents’ quality of life and wellbeing. Future studies also should define the specific daily care processes related to person-centered care and objectively measure associated outcomes to allow such models to be replicated in other facilities. Understanding the impact of the physical structure of the home, such as providing a more homelike environment, is also lacking. Research in this area may also include rigorously conducted qualitative research to better inform our understanding of contextual factors related to of care as well as staff, family, and resident perspectives.

Research questions include:

• How does daily care differ between facilities based on person-centered care models and other models (e.g. are residents offered more choices in their daily lives?)
• What are the key, measurable daily care processes that define the person-centered care model, and how can these be replicated more broadly?
• How do daily care differences relate to adverse events?
• How does the physical environment of the nursing home affect care outcomes and safety of care?

Study Approaches to Managing Polypharmacy

Polypharmacy is common in both hospitalized older patients and nursing home residents and can be associated with a number of adverse events and other poor clinical outcomes. We do not know, however, to what extent it can be improved for this medically complex population, while managing challenging clinical conditions. For example, older adults discharged from the hospital to post-acute care (SNFs) have an average of more than 13 medications and new medications are prescribed during their hospital stay.^{27, 176} This high number of medications per patient may increase the probability of adverse medication-related events and also is related to multiple geriatric syndromes associated with safety outcomes (e.g., falls, urinary incontinence, weight loss, delirium, depression).^{27, 177–180} Literature on medication-related adverse events alludes to some of these issues. However, evidence that medications can be safely reduced for this frail older population or if improved health outcomes related to safety can be achieved with medication reductions is lacking for older hospitalized patients discharged to SNF but also for those discharged to home.⁷⁶ Future research should evaluate interventions related to polypharmacy and medication reductions and should assess appropriate medication management to optimally balance reducing unnecessary prescriptions while also effectively managing clinical needs.

Research questions include:

• Is there evidence that polypharmacy is associated with adverse events in the nursing home population?
• What interventions, including technology-based approaches, may safely reduce the number of prescribed medications for hospitalized older adults discharged to SNF and subsequently to the nursing home or home demonstrate promise?

Establish What Lessons Can Be Learned From Hospital Safety for Older Patients

The assumption that effective hospital safety interventions are transferrable to the nursing home setting is untested, as noted in this report. Furthermore, analyses of hospital discharge records highlight a lack of documentation for problems related to safety and experienced by geriatric patients in the hospital.^{25, 27, 31} A separate comprehensive literature review of hospital-based safety practices specifically as they relate to older patients could identify aspects of hospital care and the discharge process that warrant improvement.

Research questions include:

• What is the evidence that hospital- based interventions to improve safety are transferrable to the nursing home population?
• What barriers to generalizability exist?
• What modifiers of effectiveness exist in the nursing home setting that are the same or different than those in the hospital?

Population Shifts and Clinical Challenges

Our Key Informants suggested that several shifts in the target population are occurring rapidly and require that safety interventions and related research adapt as part of future efforts to improve safety outcomes. These include increases in the psychiatric needs of nursing home residents, individuals with HIV-AIDS living longer lives and moving to nursing home care, and the care of aging prisoners. Perhaps most significantly, a greater proportion of older adults who are higher functioning with fewer care needs are moving into assisted-living facilities (ALFs), rather than nursing homes, which changes the population still moving into nursing homes to be of significantly higher acuity. Thus, the nursing home population is becoming more medically complex with higher care needs. As this shift occurs, the dominant paradigm may shift to palliative care, which has the potential to affect definitions of target safety outcomes as well. Also inherent in this shift is a need to focus increasingly on educating families and residents to make informed treatment decisions such that a resident’s life expectancy and quality of life are strongly considered.

ALFs and Dementia Care Within ALFs

ALFs are not only the fastest growing segment of older adult congregate living but ALFs also house residents with multiple ADL and cognitive impairments.^{181, 182} Some State-level regulations govern ALF staffing, but these vary by State and are less restrictive than those for nursing homes. In particular, the significant growth in dementia care services within ALFs makes this segment of the ALF population similar to those with dementia in nursing homes.^{182, 183} This similarity suggests that safety issues for those with dementia in the ALF care setting may be comparable. One of the biggest challenges in ALFs is the lack of standardized quality or safety data; thus, the extent of care quality and/or safety problems in this care setting is largely unknown, with only a few studies examining ALF care quality.^182–185 Future research in this area is needed for multiple reasons.

First, the number of ALFs is growing with an estimated 36,000 facilities serving over one million older adults nationwide.¹⁸¹ A recent nationwide survey of 31,100 ALFs revealed dementia as one of the most prevalent chronic conditions.¹⁸² At least partially due to the prevalence of dementia, 74 percent of ALF residents require caregiver assistance with one or more ADLs, such as bathing (72 percent), dressing (52 percent), and toileting (36 percent).¹⁸² Moreover, a longitudinal study showed that ALF residents and long-stay nursing home residents both experienced significant and comparable decline in their ability to independently perform ADLs.¹⁸⁴ Functional decline is a quality indicator for both short- and long-stay nursing home residents, and evidence suggests that optimal care can prevent decline.¹⁴¹ Thus, safety issues related to functional decline may be similar in both the nursing home and ALF care settings. Because the ALF industry began as a hospitality industry, it is also likely that measures related to person-centered care and quality of life are also equally applicable across settings.

One Key Informant who represented the ALF industry noted that resident acuity is increasing in this population, and current ALF staffing, both in terms of number and skill set, may be inadequate to meet future needs. Some of the safety concerns raised by Key Informants included medication errors, at least partially due to the skill set of the staff responsible for medication management (e.g., use of medication aides as opposed to licensed nurses); falls; and accurate assessments of clinical conditions (e.g., delirium, dehydration, depression) in the absence of licensed nurses with this skill set to support timely treatment.