Examining Black-White Disparities Among Medicare Beneficiaries in Assisted Living Settings in 2014

ObjectiveAssisted living (AL) provides housing and personal care to residents who need assistance with daily activities. Few studies have examined black-white disparities in larger (25 + beds) ALs; therefore, little is known about black residents, their prior residential settings, and how they compare to whites in AL. We examined racial differences among a national cohort of AL residents and how the racial variation among AL Medicare Fee-For-Service (FFS) beneficiaries compared to differences among community-dwelling and nursing home cohorts.Study designRetrospective cohort study.ParticipantsWe included (1) a prevalence sample of 442,018 white and black Medicare beneficiaries residing in large AL settings, (2) an incidence sample of new residents (n = 94,741), and (3) 10% random samples of Medicare FFS community-dwelling and nursing home beneficiaries in 2014.MeasuresThe Medicare Master Summary Beneficiary File was used to identify AL residents and provided demographic, entitlement, chronic condition, and health care utilization information. We used the American Community Survey and prior ZIP code tabulation areas of residents to examine differences in prior neighborhoods. Medicare claims and the Minimum Data Set yielded samples of Medicare FFS community-dwelling older adults and nursing home residents.ResultsBlacks were disproportionately represented in AL, younger, more likely to be Medicaid eligible, had higher levels of acuity, and more often lived in ALs with fewer whites and more duals. New black residents entered AL with higher rates of acute care hospitalizations and skilled nursing facility utilization. Across the 3 cohorts, blacks had higher rates of dual-eligibility.ConclusionsBlack-white differences observed among AL residents indicate a need for future work to examine how disparities manifest in differences in care received and residents’ outcomes, as well as the pathways to AL. More research is needed to understand the implications of inequities in AL as they relate to quality and experiences of residents.     

Hospital Proximity and Emergency Department Use among Assisted Living Residents

ObjectivesTo examine the relationship between AL communities' distance to the nearest hospital and residents’ rates of emergency department (ED) use. We hypothesize that when access to an ED is more convenient, as measured by a shorter distance, assisted living (AL)-to-ED transfers are more common, particularly for nonemergent conditions.DesignRetrospective cohort study, where the main exposure of interest was the distance between each AL and the nearest hospital.Setting and Participants2018-2019 Medicare claims were used to identify fee-for-service Medicare beneficiaries aged ≥55 years residing in AL communities.MethodsThe primary outcome of interest was ED visit rates, classified into those that resulted in an inpatient hospital admission and those that did not (ie, ED treat-and-release visits). ED treat-and-release visits were further classified, based on the NYU ED Algorithm, as (1) nonemergent; (2) emergent, primary care treatable; (3) emergent, not primary care treatable; and (4) injury-related. Linear regression models adjusting for resident characteristics and hospital referral region fixed effects were used to estimate the relationship between distance to the nearest hospital and AL resident ED use rates.ResultsAmong 540,944 resident-years from 16,514 AL communities, the median distance to the nearest hospital was 2.5 miles. After adjustment, a doubling of distance to the nearest hospital was associated with 43.5 fewer ED treat-and-release visits per 1000 resident years (95% CI −53.1, −33.7) and no significant difference in the rate of ED visits resulting in an inpatient admission. Among ED treat-and-release visits, a doubling of distance was associated with a 3.0% (95% CI −4.1, −1.9) decline in visits classified as nonemergent, and a 1.6% (95% CI −2.4%, −0.8%) decline in visits classified as emergent, not primary care treatable.Conclusions and ImplicationsDistance to the nearest hospital is an important predictor of ED use rates among AL residents, particularly for visits that are potentially avoidable. AL facilities may rely on nearby EDs to provide nonemergent primary care to residents, potentially placing residents at risk of iatrogenic events and generating wasteful Medicare spending.     

Implementing Large-Scale Data-Driven Quality Improvement in Assisted Living

ObjectivesDevelop and evaluate the implementation of a proposed model for large-scale data-driven quality improvement in assisted living.DesignWe conducted a mixed-methods evaluation of the implementation of a large-scale data-driven quality improvement collaborative of Wisconsin assisted living communities (ALCs).Setting and ParticipantsThe model has been voluntarily implemented by 810 Wisconsin-licensed ALCs serving >20,000 residents.MethodsThe model was codesigned iteratively 2009-2012 by a public-private multistakeholder advisory group. Using system usage statistics and project records, we evaluated implementation outcomes: appropriateness, acceptability, adoption, feasibility, fidelity, penetration, and sustainability.ResultsImplementation for ≥1 quarter was feasible for 92% of the 810 ALCs that enrolled. The model has been deemed appropriate and acceptable by public-private stakeholders representing residents, providers, regulators, and payers, and appropriateness for ALCs serving different populations has been iteratively improved through targeted workgroups. The model is currently adopted in Wisconsin by 31% of the 1573 ALCs in provider associations. Among adopters, 88% on average implemented the model with fidelity to key membership rules per quarter. The model achieved demographic and institutional penetration by currently reaching 24% of Wisconsin ALC residents and by leveraging initial grant funding to become integrated in Wisconsin's annual Medicaid budget and being central to Wisconsin's incentive program to managed care organizations. Model implementation for 8 years has been sustained by member enrollment for nearly 4 years on average, with 71% of members enrolled >2 years and sustained early adopters representing 37% that have been enrolled >5 years.Conclusions and ImplicationsThis is the first implementation study of large-scale data-driven quality improvement in assisted living, despite its demonstrated value in other health care sectors. The article proposes a model with core components and implementation strategies drawing on a decade-long public-private collaboration. The implementation study findings establish a promising path and future directions for wider implementation.     

Reforming Access: Trends in Medicaid Enrollment for New Medicare Beneficiaries, 2008–2011

Objective

To evaluate whether aligning the Part D low‐income subsidy and Medicaid program enrollment pathways in 2010 increased Medicaid participation among new Medicare beneficiaries.

Data Sources

Medicare enrollment records for years 2007–2011.

Study Design

We used a multinomial logistic model with state fixed effects to examine the annual change in limited and full Medicaid enrollment among new Medicare beneficiaries for 2 years before and after the reforms (2008–2011).

Data Extraction Methods

We identified new Medicare beneficiaries in the years 2008–2011 and their participation in Medicaid based on Medicare enrollment records.

Principal Findings

The percentage of beneficiaries enrolling in limited Medicaid at the start of Medicare coverage increased in 2010 by 0.3 percentage points for individuals aging into Medicare and by 1.3 percentage points for those qualifying due to disability (p < .001). There was no significant difference in the size of enrollment increases between states with and without concurrent limited Medicaid eligibility expansions.

Conclusions

Our findings suggest that streamlining financial assistance programs may improve Medicare beneficiaries’ access to benefits.     

The Effect of Health Plan Characteristics on Medicare+Choice Enrollment

Objective. To provide national estimates of the effect of out-of-pocket premiums and benefits on Medicare beneficiaries' choice among managed care health plans.
Data Sources/Study Setting. The data represent the population of all Medicare+Choice (M+C) plans offered to Medicare beneficiaries in the United States in 1999.
Study Design. The dependent variable is the log of the ratio of the market share of the j th health plan to the lowest cost plan in the beneficiary's county of residence. The explanatory variables are measures of premiums and benefits in the j th health plan relative to the premiums and benefits in the lowest cost plan.
Data Collection Methods. The data are from the 1999 Medicare Compare database, and M+C enrollment data from the Centers for Medicare and Medicaid Services (CMS).
Principal Findings. A $10 increase in an M+C plan's out-of-pocket premium, relative to its competitors, is associated with a decrease of four percentage points in the j th plan's market share (i.e., from 25 to 21 percent), holding the premiums of competing plans constant.
Conclusions. Although our price elasticity estimates are low, the market share losses associated with small changes in a health plan's premium, relative to its competitors, may be sufficient to discipline premiums in a competitive market. Bidding behavior by plans in the Medicare Competitive Pricing Demonstration supports this conclusion.     

Assisted Living Communities During Hurricane Irma: The Decision to Evacuate or Shelter in Place and Resident Acuity

ObjectivesTo investigate the factors associated with whether assisted living communities (ALCs) in Florida evacuated or sheltered in place for Hurricane Irma in 2017, focusing on license type as a proxy for acuity of care.DesignCross-sectional study using data collected by the state through its emergency reporting system and a post-hurricane survey.Setting and ParticipantsAnalyses included all 3112 ALCs in the emergency reporting system. A subset of 1880 that completed the survey provided supplementary data.Methodsχ² tests were used to examine differences between ALC characteristics (license type, size, payment, profit status, rural location, geographical region, and being under an evacuation order) and whether they evacuated. Logistic regression was used to test associations between characteristics and evacuation status.ResultsOf 3112 ALCs, 560 evacuated and 2552 sheltered in place. Bivariate analysis found significant associations between evacuation status and evacuation order, license type (mental health care), payment, and region. In the adjusted analysis, medium and larger ALCs were 43% (P < .001) and 53% (P < .001) less likely to evacuate than ALCs with fewer than 25 beds. Compared with ALCs in the Southeast, nearly every region was more likely to evacuate, with the highest likelihood in the Central West (odds ratio 1.76, 95% confidence interval 1.35‒2.30). ALCs under an evacuation order were 8 times more likely to evacuate (P < .001). We found no relationship between evacuation status and having a license to provide higher care.Conclusions and ImplicationsPrior research highlighting harm associated with evacuation has led to recommendations that long-term care facilities carefully consider resident impairment in evacuation decision-making. Evidence that small ALCs are more likely to evacuate and that having a higher-care license is not associated with evacuation likelihood shows research is needed to understand how ALCs weigh resident risks in decisions to evacuate or shelter in place.     

Patient Safety Culture in Assisted Living: Staff Perceptions and Association with State Regulations

ObjectivesTo examine perceptions of patient safety culture (PSC) among assisted living (AL) administrators and direct care workers (DCWs), and their associations with state regulations.DesignWe conducted a survey using the PSC instrument developed by the Agency for Healthcare Research & Quality. Secondary data on ALs and residents were derived from the Medicare Master Beneficiary Summary Files. Other data sources were the Area Health Resource Files, a previously compiled national AL directory, and the US census. Data on state AL regulations were available from a prior study.Setting and ParticipantsParticipants included administrators and DCWs working in assisted living communities serving Medicare beneficiary residents.MethodsWe employed exploratory factor analysis, examined Pearson correlations, and obtained standardized Cronbach alphas to test the PSC instrument. We estimated linear regression models with the dependent variable being the proportion of positive PSC assessments, for each PSC domain, with SEs clustered at the AL level.ResultsSurveys were completed by 714 administrators and DCWs in 257 ALs. The PSC instrument tested reliable and valid for AL communities. Administrators’ and DCWs’ perceptions of PSC differed significantly across almost all domains. A 1-unit increase in state regulatory specificity for DCW staffing was associated with a 4.13–percentage point (P < .05) increase in the PSC staffing domain. Associations with regulatory specificity in staff training were also found for other PSC domains.Conclusions and ImplicationsPSC is an important metric for assessing organizational performance. DCWs have significantly worse perceptions of PSC than do administrators, suggesting it is crucial to understand the source of these differing perceptions. Because state regulations relate to PSC, achieving a comprehensive focus on patient safety in AL may require regulatory action, particularly increasing specificity with regard to staffing and training.     

Defining the Assisted Living Sector in Canada: An Environmental Scan

ObjectivesIn this study, we (1) identify the terms used to describe the assisted living sector and the legislation governing operation in all Canadian provinces and territories; (2) identify the cost estimates associated with residency in these homes; and (3) quantify the growth of the sector.DesignEnvironmental scan.Setting and ParticipantsInternet searches of Canadian provincial and territorial government websites and professional associations were conducted in 2021 to retrieve publicly accessible sources related to the assisted living sector.MethodsWe synthesized data that identified the terms used to describe the sector in all provinces and territories, the legislation governing operation, financing, median fees per month for care, and growth of the sector from 2012 to 2020. Counts and proportions were calculated for some extracted variables. All data were narratively synthesized.ResultsThe terms used to describe the assisted living sector varied across Canada. The terms “assisted living,” “retirement homes,” and “supportive living” were prevalent. Ontario was the only province to regulate the sector through an independent, not-for-profit organization. Ontario, British Columbia, and Alberta had some of the highest median fees for room, board, and care per month (range: $1873 to $6726). The licensed assisted living sector in Ontario doubled in size (768 in 2020 vs 383 in 2012), and there was a threefold increase in the number of corporate-owned chain assisted living facilities (465 in 2020 vs 142 in 2012).Conclusions and ImplicationsThe rapid growth of the assisted living sector that is primarily financed through out-of-pocket payments may indicate a rise in a two-tier system of housing and health care for older adults. Policymakers need better mechanisms, such as standardized reporting systems and assessments, to understand the needs of older adults who reside in assisted living facilities and inform the need for sector regulation and oversight.     

COVID-19 Recommendations for Assisted Living: Implications for the Future

ObjectivesAssisted living (AL) emerged over 2 decades ago as a preferred residential care option for older adults who require supportive care; however, as resident acuity increased, concern has been expressed whether AL sufficiently addresses health care needs. COVID-19 amplified those concerns, and an examination of recommendations to manage COVID-19 may shed light on the future of AL. This review summarizes recommendations from 6 key organizations related to preparation for and response to COVID-19 in AL in relation to resident health and quality of life; compares recommendations for AL with those for nursing homes (NHs); and assesses implications for the future of AL.DesignNonsystematic review involving search of gray literature.Setting and ParticipantsRecommendations from key governmental bodies and professional societies regarding COVID-19 in AL, long-term care facilities (LTCFs) in general, and NHs.MeasuresWe collected, categorized, and summarized these recommendations as they pertained to quality of life and health care.ResultsMany recommendations for AL and NHs were similar, but differences provided insight into ways the pandemic was recognized and challenged AL communities in particular: recommending more flexible visitation and group activities for AL, providing screening by AL staff or an outside provider, and suggesting that AL staff access resources to facilitate advance care planning discussions. Recommendations were that AL integrate health care into offered services, including working with consulting clinicians who know both the residents and the LTC community.Conclusions and ImplicationsLong-term care providers and policy makers have recognized the need to modify current long-term care options. Because COVID-19 recommendations suggest AL communities would benefit from the services and expertise of social workers, licensed nurses, and physicians, it may accelerate the integration and closer coordination of psychosocial and medical care into AL. Future research should investigate different models of integrated, interdisciplinary health care in AL.     

End-of-Life Care Transitions in Assisted Living: Associations With State Staffing and Training Regulations

ObjectiveWe examined the frequency and categories of end-of-life care transitions among assisted living community decedents and their associations with state staffing and training regulations.DesignCohort study.Setting and ParticipantsMedicare beneficiaries who resided in assisted living facilities and had validated death dates in 2018–2019 (N = 113,662).MethodsWe used Medicare claims and assessment data for a cohort of assisted living decedents. Generalized linear models were used to examine the associations between state staffing and training requirements and end-of-life care transitions. The frequency of end-of-life care transitions was the outcome of interest. State staffing and training regulations were the key covariates. We controlled for individual, assisted living, and area-level characteristics.ResultsEnd-of-life care transitions were observed among 34.89% of our study sample in the last 30 days before death, and among 17.25% in the last 7 days. Higher frequency of care transitions in the last 7 days of life was associated with higher regulatory specificity of licensed [incidence risk ratio (IRR) = 1.08; P = .002] and direct care worker staffing (IRR = 1.22; P < .0001). Greater regulatory specificity of direct care worker training (IRR = 0.75; P < .0001) was associated with fewer transitions. Similar associations were found for direct care worker staffing (IRR = 1.15; P < .0001) and training (IRR = 0.79; P < .001) and transitions within 30 days of death.Conclusions and ImplicationsThere were significant variations in the number of care transitions across states. The frequency of end-of-life care transitions among assisted living decedents during the last 7 or 30 days of life was associated with state regulatory specificity for staffing and staff training. State governments and assisted living administrators may wish to set more explicit guidelines for assisted living staffing and training to help improve end-of-life quality of care.     

Self‐Reported Cancer Screening Among Elderly Medicare Beneficiaries: A Rural‐Urban Comparison

Purpose: We examined the rural‐urban disparity of screening for breast cancer and colorectal cancer (CRC) among the elder Medicare beneficiaries and assessed rurality's independent impact on receipt of screening. Methods: Using 2005 Medicare Current Beneficiary Survey, we applied weighted logistic regression to estimate the overall rural‐urban disparity and rurality's independent impact on cancer screening, controlling for patient, and area factors. Results: From urban, large rural, small rural, and isolated rural areas, the rates for mammogram last year were 53%, 52%, 45%, and 44%, respectively. They were 56%, 50%, 48%, and 43% for CRC screening, respectively. After controlling for patient and area level characteristics, rurality is significantly associated with CRC screening, but not mammogram. Conclusions: We found rural‐urban disparities for both mammogram and CRC screenings. Patient and area factors totally eliminated the rural‐urban disparity for mammogram but not CRC screening. Health promotions to improve cancer screening should focus more on small and isolated rural areas.     

Changes in Long-Term Care Markets: Assisted Living Supply and the Prevalence of Low-Care Residents in Nursing Homes

ObjectivesTo assess the effect of changes in assisted living (AL) capacity within a market on prevalence of residents with low care needs in nursing homes.DesignRetrospective, longitudinal analysis of nursing home markets.Setting and participantsTwelve thousand two hundred fifity-one nursing homes in operation during 2007 and 2014.MeasurementsWe analyzed the percentage of residents in a nursing home who qualified as low-care. For each nursing home, we constructed a market consisting of AL communities, Medicare beneficiaries, and competing nursing homes within a 15-mile radius. We estimated the effect of change in AL beds on prevalence of low-care residents using multivariate linear models with year and nursing home fixed effects.ResultsThe supply of AL beds increased by an average 258 beds per nursing home market (standard deviation = 591) during the study period. The prevalence of low-care residents decreased from an average of 13.0% (median 10.5%) to 12.2% (median 9.5%). In adjusted models, a 100-bed increase in AL supply was associated with a decrease in low-care residents of 0.041 percentage points (P = .026), controlling for changes in market and nursing home characteristics, county demographics, and year and nursing home fixed effects. In markets with a high percentage of its Medicare beneficiaries (≥14%) dual eligible for Medicaid, the effect of AL is stronger, with a 0.066–percentage point decrease per 100 AL beds (P = .026) vs a 0.016–percentage point decrease in low-duals markets (P = .48).Conclusions and implicationsOur analysis suggests that some of the growth in AL capacity serves as a substitute for nursing homes for patients with low care needs. Furthermore, the effects are concentrated in markets with an above-average proportion of beneficiaries with dual Medicaid eligibility.     

Online Customer Reviews of Assisted Living Communities: Association with Community,County, and State Factors

ObjectivesOnline reviews provided by users of assisted living communities may offer a unique source of heretofore unexamined data. We explored online reviews as a possible source of information about these communities and examined the association between the reviews and aspects of state regulations, while controlling for assisted living, county, and state market-level factors.DesignCross-sectional, observational study.Setting and ParticipantsSample included 149,265 reviews for 8828 communities.MethodsPrimary (eg, state regulations) and secondary (eg, Medicare Beneficiary Summary Files) data were used. County-level factors were derived from the Area Health Resource Files, and state-level factors from the integrated Public Use Microdata series. Information on state regulations was obtained from a previously compiled regulatory dataset. Average assisted living rating score, calculated as the mean of posted online reviews, was the outcome of interest, with a higher score indicating a more positive review. We used word cloud to visualize how often words appeared in 1-star and 5-star reviews. Logistic regression models were used to determine the association between online rating and a set of community, county, and state variables. Models were weighted by the number of reviews per assisted living bed.ResultsOverall, 76% of communities had online reviews. We found lower odds of positive reviews in communities with greater proportions of Medicare/Medicaid residents [odds ratio (OR) = 0.986; P < .001], whereas communities located in micropolitan areas (compared with urban), and those in states with more direct care worker hours (per week per bed) had greater odds of high rating (OR = 1.722; P < .001 and OR = 1.018, P < .05, respectively).Conclusions and ImplicationsOnline reviews are increasingly common, including in long-term care. These reviews are a promising source of information about important aspects of satisfaction, particularly in care settings that lack a public reporting infrastructure. We found several significant associations between online ratings and community-level factors, suggesting these reviews may be a valuable source of information to consumers and policy makers.     

An Algorithm for the Use of Medicare Claims Data to Identify Women with Incident Breast Cancer

Objective. To develop and validate a clinically informed algorithm that uses solely Medicare claims to identify, with a high positive predictive value, incident breast cancer cases.
Data Source. Population-based Surveillance, Epidemiology, and End Results (SEER) Tumor Registry data linked to Medicare claims, and Medicare claims from a 5 percent random sample of beneficiaries in SEER areas.
Study Design. An algorithm was developed using claims from 1995 breast cancer patients from the SEER-Medicare database, as well as 1995 claims from Medicare control subjects. The algorithm was validated on claims from breast cancer subjects and controls from 1994. The algorithm development process used both clinical insight and logistic regression methods.
Data Extraction. Training set: Claims from 7,700 SEER-Medicare breast cancer subjects diagnosed in 1995, and 124,884 controls. Validation set: Claims from 7,607 SEER-Medicare breast cancer subjects diagnosed in 1994, and 120,317 controls.
Principal Findings. A four-step prediction algorithm was developed and validated. It has a positive predictive value of 89 to 93 percent, and a sensitivity of 80 percent for identifying incident breast cancer. The sensitivity is 82–87 percent for stage I or II, and lower for other stages. The sensitivity is 82–83 percent for women who underwent either breast-conserving surgery or mastectomy, and is similar across geographic sites. A cohort identified with this algorithm will have 89–93 percent incident breast cancer cases, 1.5–6 percent cancer-free cases, and 4–5 percent prevalent breast cancer cases.
Conclusions. This algorithm has better performance characteristics than previously proposed algorithms. The ability to examine national patterns of breast cancer care using Medicare claims data would open new avenues for the assessment of quality of care.