Abstract
Research Objective:
Affordable access to medications is important to Medicare enrollees in long-term care (LTC), yet, it is unknown if prescription drug coverage is universal and adequate to meet their high medication needs.
Study Design:
We assessed enrollment in prescription drug coverage, out-of-pocket (OOP) payments and medication use in a nationwide LTC database of prescription-level, resident-level, and facility-level data for the period 2011–2013. Inadequate drug coverage was defined as ≥ 50% medications paid for OOP. Risk-adjusted generalized estimation equations models were estimated to identify predictors of inadequate drug coverage and total prescription fills.
Population Studied:
A nationwide sample of 332,087 Medicare enrollees observed > 100 days in LTC.
Principal Findings:
We found Medicare Part D was the main source of drug coverage (82.4%), followed by private insurance (8.5%), and Veterans Administration (0.2%). No drug coverage could be detected for 8.9% (n = 29,378) who paid for all of their medications OOP or received only temporary drug payment assistance. Inadequate drug coverage was identified in another 2721 persons. LTC Medicare enrollees without drug coverage or who had private insurance received significantly fewer prescriptions than if they had been enrolled in Medicare Part D.
Conclusion:
A substantial proportion of Medicare enrollees in LTC have inadequate or no drug coverage and are receiving less medication than indicated by their health needs.
Policy Implications:
Medicare Part D is an important policy for ensuring affordable access to medications in LTC. However, expansions are needed to increase enrollment and decrease inadequate drug coverage.
Keywords: longterm care, prescription drug coverage, Medicare
Nearly 3.2 million elderly and disabled Americans receive care in long-term care (LTC) facilities each year, with approximately half needing to remain in the institution for an extended period of time until death.1 Medicare enrollees who remain in LTC beyond a short stay (> 100 d) have significant cognitive, functional, and sensory deficits. Less than one third of long-stay LTC residents have intact decision-making capacity and dementia affects 60%.1 For these individuals, LTC facilities are not only where Medicare enrollees receive medical care but also where they make financial decisions such as selecting prescription drug coverage.2
Affordable access to medications is important to Medicare enrollees in LTC who, on average, receive 10 different medications per month.2,3 However, despite the high number of medications, the underuse of effective medications remains a concern.4–8 Early findings suggest there is potential for cost-related underuse of medications in this setting with negative outcomes.3–7
Since 2006, all Medicare enrollees including those in LTC have had access to a federally sponsored drug insurance program (Medicare Part D).8 Medicare enrollees with Medicaid dually-eligibility are auto-enrolled into Part D; however, one third of the LTC population is ineligible for Medicaid so must self-enroll every year and pay monthly premiums to receive the benefit. Part D self-enrollment may be particularly challenging in LTC since Medicare bans facility staff from assisting Part D enrollment.9–11 In addition, it is unknown if drug coverage from other sources is sufficient for meeting the high medication needs of the LTC population.
METHODS
Data Sources
For research purposes, we linked together 5 datasets spanning the period 2011–2013: (1) large all-payer prescription drug dispensing dataset; (2) validated 3.0 Minimum Data Set (MDS); (3) Medicare enrollment files; (4) Medicare Parts A and B data; and (5) Certification and Survey Provider Enhanced Reporting (CASPER).
The prescription drug database contains information on half of all US LTCs. Resident-level files are transmitted electronically to a data repository and deidentified. Data elements include all drugs dispensed regardless of payer [eg, Medicare Part D, out-of-pocket (OOP)], and source of payment. These data are extracted from prescription drug claims so are >99% complete.
The MDS is a federally mandated, 400-item standardized clinical assessment of every resident living in Medicare/Medicaid-certified US LTCs. MDS assessments occur at admission, quarterly intervals, and change in health status. Validity and reliability have been demonstrated with MDS data and the 3.0 version offers improvements in accuracy and validity.12,13
The Medicare data include beneficiary enrollment information including sex, race, age, and Medicaid eligibility. The Medicare Parts A and B claims contain diagnosis information. CASPER is a repository of federally-mandated onsite surveys of all Medicare/Medicaid-certified US LTCs. The LTC surveys are conducted by state survey agencies and must occur at least every 15 months. Validity and reliability have been demonstrated in these data.14,15
Study Population and Observation Period
Our study sampling frame included Medicare-eligible individuals with an admission LTC stay assessment, and residing in a facility receiving pharmacy services from our LTC provider (n = 1,382,954). We excluded individuals observed for <100 days because drug coverage is difficult to determine when medications may be covered in bundled per diem payments. We also excluded LTC stays that did not include at least 1 day in 2012, the reference year for calculating drug coverage, with the assumption that LTC admissions are randomly distributed throughout the year.
Study Measures
Prescription Drug Coverage and Generosity
All drug coverage measures were developed from 2012 data to conform to the calendar year enrollment cycle of Part D. Enrollment in Medicare Part D was determined from Medicare administrative records. Drug coverage from other sources was determined from the medication payer sources. In addition, we developed a hierarchy of mutually-exclusive categories of drug coverage: Part D full low-income subsidy (LIS), Part D partial LIS, Part D non-LIS, Veterans Administration (VA), private drug insurance, no coverage and temporary drug payment assistance (temporary). No coverage indicates no Part D enrollment and payer source is limited to only OOP. Temporary drug payment assistance indicates no Part D enrollment and payer sources that include Medicaid, Medicare Parts A or B, facility, or hospice, which are not types of drug insurance.
Drug coverage was further characterized by the generosity of coverage, defined as the proportion of medications paid by each payment source. This measure of generosity has demonstrated reliability and face validity in prior work.16,17 Medications paid OOP as secondary payer represented only 17% of all prescriptions and were weighted by the following scheme: full LIS Part D = 0%, partial LIS Part D = 15%, and non-LIS Part D = 25%, VA = 25%, private = 25%, temporary = 25%. If OOP is the primary payer, then OOP is 100%. Medications with ~3 payment sources were excluded from the analysis, which represented <1% of the total. We defined inadequate drug coverage as ≥50% of medications paid for OOP.
Medication Use
We counted all prescriptions dispensed to our study population in 2012 during the long-stay and summed the total number of fills for each resident.
Demographic Characteristics and Health Status
Age, sex, race, and health status including active diagnosis flags and functional limitations came from the Medicare summary files and the MDS admission assessment. Cognition was measured using the Cognitive Function Scale with demonstrated validity and predictive of medication use.18–20
Risk Adjustment
All of our multivariate analyses were risk-adjusted using the 2012 Rx HCC model.21 The Rx HCC model uses demographic characteristics and diagnoses from Medicare claims to generate a validated score for controlling confounding, and the score is an estimate of the resident’s expected prescription drug costs, relative to the average.21
Long-Term Care Characteristics
Facility characteristics came from CASPER and are associated with LTC quality and prescribing patterns.22
Statistical Analysis
We conducted logistic regression models with generalized estimation equations to account for the clustering of residents within LTCs. Our main outcomes were total prescription drug fills and the likelihood of having inadequate drug coverage, adjusted for both facility characteristics, and resident characteristics.
We also conducted Poisson regressions with generalized estimation equations but found the average medication use exceeded 30 prescriptions and approximated a normal distribution.23
This study was approved by the institutional review board at Northeastern University.
RESULTS
We identified 332,087 long-stay Medicare enrollees in 11,587 nursing homes, which generated 24,839,236 prescription records during the study period of 2012 (Table 1).
TABLE 1.
Characteristics of Study Long-stay Residents and LTC Facilities
| Residents (N = 332,087) | n (%) |
|---|---|
| Age (y) | |
| < 65 | 28,038 (8.4) |
| 65–75 | 50,171 (15.1) |
| 76–85 | 99,047 (29.8) |
| > 85 | 154,831 (46.5) |
| Sex | |
| Male | 98,149 (29.6) |
| Female | 233,566 (70.4) |
| Race/ethnicity | |
| White | 273,615 (82.4) |
| Black | 36,163 (10.9) |
| Hispanic | 14,907 (4.5) |
| Other | 7402 (2.2) |
| Active conditions | |
| Dementia | 152,766 (47.1) |
| Diabetes | 95,905 (29.6) |
| Heart failure | 61,112 (18.8) |
| Stroke | 56,134 (17.3) |
| Alzheimer | 55,924 (17.2) |
| Schizophrenia | 16,094 (5.0) |
| Hip fracture | 13,253 (4.1) |
| Aphasia | 13,017 (4.0) |
| Bipolar disorder | 12,219 (3.8) |
| Activities of Daily Living Score (mean) | 8.5 |
| Cognitive Function Scale (mean) | 2.5 |
| Rx-HCC score (mean) | 1.46 |
| Hospice | 35,220 (10.6) |
| Facilities (n = 11,587) | |
| US Census Region | |
| South | 3932 (33.9) |
| West | 1730 (15.0) |
| Midwest | 3597 (31.0) |
| Northeast | 2243 (19.4) |
| LTC size, number of residents* | |
| < 50 | 1803 (15.5) |
| 50–100 | 5182 (44.7) |
| 101–250 | 4287 (37.0) |
| > 250 | 232 (2.0) |
| Residents on Medicaid* | |
| 0%–24% | 902 (7.8) |
| 25%–49% | 1638 (14.1) |
| 50%–74% | 5843 (50.4) |
| 75%+ | 3052 (26.3) |
| Chain ownership* | |
| Yes | 6577 (56.8) |
| No | 4389 (37.9) |
| For-profit status* | |
| For-profit | 8496 (73.3) |
| Nonprofit | 2534 (21.9) |
| Government | 472 (4.1) |
Not equal to 100% due to missing information on the facility.
LTC indicates long-term care; Rx-HCC, prescription drug hierarchical condition categories.
About 8.4% of the population was under the age of 65 (entitled to the Medicare program through a disabling condition) and 46.5% were 85 years or older. Over two thirds (70.4%) were female and the majority lived in facilities in the South (33.9%) or Midwest (31.0%).
Figure 1 shows the sources of drug coverage for long-stay Medicare enrollees in 2012. We found Part D was the main source of drug coverage (82.4%; 67.0 full LIS, 0.4 partial LIS, and 15.0 non-LIS), followed by private insurance (8.6%), and VA (0.2%). No drug coverage could be detected for 8.9% (n = 2909) who paid for all of their medications OOP or received only temporary drug payment assistance (eg, hospice benefit).
FIGURE 1.
Main sources of drug coverage for long-stay Medicare enrollees in long-term care, 2012. LIS indicates low-income subsidy; VA, Veterans Administration.
We identified inadequate drug coverage in another 2721 persons with some form of drug coverage, which was not subsidized (Table 2). The main mutable predictors of having inadequate drug coverage are the source of drug coverage and facility characteristics. Compared to having Part D, long-stay Medicare enrollees who received only temporary drug payment assistance were 2 times more likely to pay > 50% of their mediations OOP [2.43 odds ratio (OR), 95% confidence interval (CI): 2.23–2.66]. Living in a small facility increased the odds of inadequate drug coverage by 2-fold (2.39 OR, 95% CI: 1.78–33.22), as did the location of the facility in the South (1.83 OR, 95% CI: 1.56–2.15) or Midwest (1.74 OR, 95% CI: 1.49–2.04), relative to in the West. Conversely, living in a facility with high Medicaid participation was protective: 75%+ Medicaid residents (0.54 OR, 95% CI: 0.46–0.65, relative to <25% Medicaid residents.
TABLE 2.
Relationship Between Drug Coverage and High Out-of-Pocket Prescription Drug Costs (> 50% of Fills Paid Out-of-Pocket)
| 95% CI |
|||||
|---|---|---|---|---|---|
| Parameters | OR | Lower Limit | Upper Limit | P | |
| Drug coverage (reference = Part D) | Private | 0.925 | 0.846 | 1.012 | 0.0889 |
| Temporary Payment Assistance | 2.435 | 2.232 | 2.656 | < 0.0001 | |
| Veterans Administration | 0.132 | 0.031 | 0.556 | 0.0058 | |
| Sex (reference = female) | Male | 1.018 | 0.944 | 1.098 | 0.6439 |
| Race (reference = white) | Black | 0.631 | 0.529 | 0.752 | < 0.0001 |
| Hispanic | 0.387 | 0.269 | 0.557 | < 0.0001 | |
| Other | 0.826 | 0.614 | 1.112 | 0.2076 | |
| Age (reference = > 85) (y) | < 65 | 0.558 | 0.422 | 0.739 | < 0.0001 |
| 65–75 | 0.841 | 0.731 | 0.968 | 0.0157 | |
| 76–85 | 0.856 | 0.790 | 0.928 | 0.0001 | |
| Region (reference = west) | Midwest | 1.744 | 1.492 | 2.039 | < 0.0001 |
| South | 1.832 | 1.558 | 2.153 | < 0.0001 | |
| Northeast | 1.241 | 1.043 | 1.477 | 0.0149 | |
| Size, number of residents (reference = > 250) | < 50 | 2.396 | 1.782 | 3.222 | < 0.0001 |
| 50–100 | 1.712 | 1.293 | 2.266 | 0.0002 | |
| 101–250 | 1.553 | 1.180 | 2.045 | 0.0017 | |
| Residents on Medicaid (reference = 0%–24%) | 25%–49% | 0.708 | 0.599 | 0.837 | < 0.0001 |
| 50%–74% | 0.605 | 0.522 | 0.701 | < 0.0001 | |
| 75%+ | 0.543 | 0.456 | 0.646 | < 0.0001 | |
| Chain ownership (reference = not chain) | Chain | 0.998 | 0.903 | 1.103 | 0.9718 |
| Facility type (reference = nonprofit) | For profit | 0.745 | 0.668 | 0.830 | < 0.0001 |
| Government | 0.764 | 0.590 | 0.989 | 0.0411 | |
| Active chronic conditions | Heart failure | 0.698 | 0.636 | 0.767 | < 0.0001 |
| Diabetes | 0.685 | 0.625 | 0.751 | < 0.0001 | |
| Hip fracture | 0.838 | 0.729 | 0.964 | 0.0134 | |
| Alzheimer | 1.040 | 0.949 | 1.139 | 0.4004 | |
| Aphasia | 0.960 | 0.796 | 1.157 | 0.6643 | |
| Stroke | 0.917 | 0.832 | 1.010 | 0.0787 | |
| Dementia | 1.049 | 0.975 | 1.128 | 0.2027 | |
| Bipolar | 0.618 | 0.439 | 0.872 | 0.0061 | |
| Multiple Sclerosis | 2.090 | 1.456 | 3.000 | < 0.0001 | |
| Schizophrenia | 1.151 | 0.835 | 1.586 | 0.3911 | |
| Activities of Daily Living | Total Activities of Daily Living Score | 0.990 | 0.983 | 0.998 | 0.0171 |
| Cognitive Function Scale | Total Cognitive Function Scale Score | 0.868 | 0.837 | 0.899 | < 0.0001 |
| Risk Adjustment Score | Total Risk Adjustment Score (centered) | 0.384 | 0.311 | 0.474 | < 0.0001 |
| Hospice | Hospice | 0.707 | 0.630 | 0.792 | < 0.0001 |
| Any demographic variable missing | Missing data | 1.166 | 0.945 | 1.439 | 0.1518 |
CI indicates confidence interval; OR, odds ratio.
Table 3 shows that the average number of prescription fills varied substantially by the source of drug coverage. After controlling for health status, we found Medicare enrollees with only temporary drug payment assistance received 43.9 fewer prescriptions, on average, (−43.92, 95% CI: −44.85 to −44.00) than those with Medicare Part D non-LIS. This large difference in medication use is similar to the one found among enrollees without any drug coverage at all (−43.57, 95% CI: −45.28 to −41.86. In comparison, LTC Medicare enrollees with private insurance also received fewer prescriptions that those with Part D non-LIS, although the difference was much smaller (−2.88, 95% CI: −3.72 to −2.04). The average number of prescriptions filled among Medicare enrollees was the same whether in VA or in Part D non-LIS.
TABLE 3.
Relationship Between Drug Coverage and Risk-adjusted Prescription Drug Use
| 95% CI |
||||||
|---|---|---|---|---|---|---|
| Parameters | Estimate | Lower Limit | Upper Limit | Z | P | |
| Intercept | Intercept | 27.2867 | 23.2636 | 31.3097 | 13.29 | < 0.0001 |
| Drug coverage (reference = Part D non-LIS) | Private | −2.883 | −3.729 | −2.037 | −6.68 | < 0.0001 |
| Temporary Payment Assistance | −43.928 | −44.853 | −43.004 | −93.17 | < 0.0001 | |
| Veterans Administration | 5.649 | −0.624 | 11.921 | 1.76 | 0.0776 | |
| No coverage | −43.571 | −45.283 | −41.860 | −49.9 | < 0.0001 | |
| Sex (reference = female) | Male | −3.110 | −3.820 | −2.400 | −8.58 | < 0.0001 |
| Race (reference = white) | Black | −4.160 | −5.543 | −2.778 | −5.9 | < 0.0001 |
| Hispanic | −2.710 | −4.822 | −0.599 | −2.52 | 0.0119 | |
| Other | 3.966 | 1.515 | 6.416 | 3.17 | 0.0015 | |
| Age (reference = > 85) | < 65 | −7.249 | −9.468 | −5.031 | −6.41 | < 0.0001 |
| 65–75 | −2.481 | −3.758 | −1.205 | −3.81 | 0.0001 | |
| 76–85 | −0.411 | −1.158 | 0.336 | −1.08 | 0.2812 | |
| Region (reference = west) | Midwest | 8.330 | 6.670 | 9.990 | 9.84 | < 0.0001 |
| South | 7.195 | 5.517 | 8.873 | 8.4 | < 0.0001 | |
| Northeast | 6.360 | 4.565 | 8.155 | 6.94 | < 0.0001 | |
| Size, number of residents (reference = > 250) | < 50 | 7.709 | 4.290 | 11.128 | 4.42 | < 0.0001 |
| 50–100 | 7.155 | 3.984 | 10.325 | 4.42 | < 0.0001 | |
| 101–250 | 6.294 | 3.131 | 9.457 | 3.9 | < 0.0001 | |
| Residents on Medicaid (reference = 0%–24%) | 25%–49% | −4.935 | −7.604 | −2.266 | −3.62 | 0.0003 |
| 50%–74% | −6.810 | −9.167 | −4.453 | −5.66 | < 0.0001 | |
| 75%+ | −9.078 | −11.597 | −6.559 | −7.06 | < 0.0001 | |
| Chain Ownership (reference = not chain) | Chain | 2.489 | 1.282 | 3.695 | 4.04 | < 0.0001 |
| Facility type (reference = nonprofit) | For-profit | −1.957 | −3.433 | −0.480 | −2.6 | 0.0094 |
| Government | −3.953 | −7.421 | −0.485 | −2.23 | 0.0255 | |
| Active chronic conditions | Heart failure | 12.405 | 11.507 | 13.304 | 27.07 | < 0.0001 |
| Diabetes | 11.490 | 10.656 | 12.324 | 27.0 | < 0.0001 | |
| Hip fracture | 2.182 | 0.954 | 3.411 | 3.48 | 0.0005 | |
| Alzheimer | −4.328 | −5.175 | −3.481 | −10.01 | < 0.0001 | |
| Aphasia | 0.368 | −1.303 | 2.039 | 0.43 | 0.6661 | |
| Stroke | 4.520 | 3.624 | 5.416 | 9.88 | < 0.0001 | |
| Dementia | −0.358 | −1.052 | 0.336 | −1.01 | 0.3126 | |
| Bipolar | 6.078 | 3.445 | 8.711 | 4.52 | < 0.0001 | |
| Multiple sclerosis | −5.416 | −9.732 | −1.100 | −2.46 | 0.0139 | |
| Schizophrenia | −3.040 | −6.124 | 0.044 | −1.93 | 0.0534 | |
| Activities of Daily Living | Total Activities of Daily Living Score | 0.369 | 0.296 | 0.441 | 9.99 | < 0.0001 |
| Cognitive Function Scale | Total Cognitive Function Scale Score | 5.992 | 5.632 | 6.351 | 32.66 | < 0.0001 |
| Risk Adjustment Score | Total Risk Adjustment Score (centered) | 47.746 | 45.594 | 49.898 | 43.48 | < 0.0001 |
| Hospice | Hospice | −2.984 | −4.055 | −1.913 | −5.46 | < 0.0001 |
| Any demographic variable missing | Missing data | 3.244 | 0.731 | 5.758 | 2.53 | 0.0114 |
CI indicates confidence interval; OR, odds ratio.
DISCUSSION
This study found that drug coverage is not universal in the LTC and is often inadequate to meet the high demand for medications in this setting. No drug coverage could be detected for 29,378 long-stay Medical enrollees who paid for all of their medications OOP or received only temporary drug payment assistance (eg, hospice benefit). Inadequate drug coverage was identified in another 2721 long-stay Medicare enrollees who had drug coverage but still paid for more than half of their prescriptions OOP. Adding both of these groups together, we estimate nearly 10% of long-stay Medicare enrollees in LTC have insufficient drug coverage.
The main factor in having insufficient drug coverage was not enrolling in the Medicare Part D prescription drug program. Most of our population participated in Part D, however, nearly 20% did not, which placed them at high risk for paying most or all of their medications OOP. Individuals with temporary drug payment assistance were especially at-risk. Only participation in the VA prescription drug benefit provided comparable coverage to Part D, although the VA benefit has restrictive eligibility requirements.
Last, as a result of these large discrepancies in sufficient drug coverage, we found large differences in medication use that could not be explained by differences in health status. Medicare enrollees in LTC who did not have reliable drug coverage received far less medication, on average than enrollees with reliable drug coverage. These differences remained even after controlling for underlying medical conditions, suggesting potential underuse of necessary medications and potential economic barriers.
Limitations
This study was designed to reduce threats to validity but nonetheless has several limitations. First, we lacked complete clinical information, so cannot determine the appropriateness of prescribing. Second, MDS and claims-based measures may underreport outcomes (diseases, limited functioning). However, our main measures of payer sources and medication use should be unbiased and accurate. Third, temporary drug payment assistance may have masked some drug coverage, although our assessment of Part D came from enrollment records. Fourth, our estimates are nationwide but not nationally representative, limiting generalizability. A comparison of our study LTC population to the US LTC population in the 2010 US OSCAR data shows substantial overlap (eg, 61.9% vs. 66.4% female; 66.4% vs. 71.4% aged 75+; 54.5% vs. 66.0% Medicaid-eligible). Fifth, our data span the period 2011–2013 so it may be outdated. However, the Medicare Part D program has been relatively stable since this period.24 The Affordable Care Act included provisions to decrease gaps in coverage by 2020, however, OOP costs have generally increased for Part D enrollees without LISs since 2013. Thus, our findings may be a conservative underestimation of the current adequacy of prescription drug coverage in nursing homes. Sixth, our prescription dispensing dataset does not contain cost information (Appendix, Supplemental Digital Content 1, http://links.lww.com/MLR/B954).
CONCLUSIONS
The Medicare Part D prescription drug program was implemented to improve access to medications for all Medicare enrollees. Achievement of these goals has been mixed in LTC. On the one hand, Part D has become the main payer of medications in this setting and largely protects residents from burdensome drug costs. Yet, ~10% of long-stay Medicare enrollees in LTC do not enroll in Part D and do not have sufficient drug coverage. The consequences are inequitable access to necessary medications.
Supplementary Material
ACKNOWLEDGMENT
The authors thank Lauren Bigger, MPH, for her support in the preparation of this manuscript.
Supported by NIH grants 1 R21 AG049269-01 and R01 AG046341-01A1.
Footnotes
The authors declare no conflict of interest.
Supplemental Digital Content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website, www.lww-medicalcare.com.
REFERENCES
- 1.Centers for Medicare & Medicaid Services. Nursing Home Data Compendium 2010. Rockville, MD: Centers for Medicare & Medicaid Services; 2012. [Google Scholar]
- 2.Avorn J, Gurwitz JH. Drug use in the nursing home. Ann Intern Med. 1995;123:195–204. [DOI] [PubMed] [Google Scholar]
- 3.Briesacher BA, Soumerai SB, Gurwitz JH. Nursing home residents and enrollment in Medicare Part D. J Am Geriatr Soc. 2009;57:1902–1907. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Parikh S, Shrank WH, Mogun H, et al. Statin utilization in nursing home patients after cardiac hospitalization. J Gen Intern Med. 2010;25:1293–1299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Levy CR, Radcliff TA, Williams ET, et al. Acute myocardial infarction in nursing home residents: adherence to treatment guidelines reduces mortality, but why is adherence so low? J Am Med Dir Assoc. 2009;10: 56–61. [DOI] [PubMed] [Google Scholar]
- 6.Gaboda D, Lucas J, Siegel M, et al. No longer undertreated? Depression diagnosis and antidepressant therapy in elderly long-stay nursing home residents, 1999 to 2007. J Am Geriatr Soc. 2011;59:673–680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hanlon JT, Handler SM, Castle NG. Antidepressant prescribing in US nursing homes between 1996 and 2006 and its relationship to staffing patterns and use of other psychotropic medications. J Am Med Dir Assoc. 2010;11:320–324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Allsworth JE, Toppa R, Palin NC, et al. Racial and ethnic disparities in the pharmacologic management of diabetes mellitus among long-term care facility residents. Ethn Dis. 2005;15:205–212. [PubMed] [Google Scholar]
- 9.Briesacher BA, Soumerai SB, Field TS, et al. Medicare part D’s exclusion of benzodiazepines and fracture risk in nursing homes. Arch Intern Med. 2010;170:693–698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Huskamp HA, Stevenson DG, O’Malley AJ, et al. Medicare Part D plan generosity and medication use among dual-eligible nursing home residents. Med Care. 2013;51:894–900. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Stevenson DG, O’Malley AJ, Dusetzina SB, et al. Effect of Part D coverage restrictions for antidepressants, antipsychotics, and cholinesterase inhibitors on related nursing home resident outcomes. J Am Geriatr Soc. 2014;62: 1666–1674. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ryan C, O’ Mahony D, Kenney J, et al. Potentially inappropriate prescribing in older residents in Irish nursing homes. Age Ageing. 2013;42: 116–120. [DOI] [PubMed] [Google Scholar]
- 13.Centers for Medicare & Medicaid Services, Lower Medicare Part D costs than expected in 2009 in Press Releases, US Department of Health and Human Services; 2008. [Google Scholar]
- 14.Stevenson DG, Huskamp HA, Newhouse JP. Medicare part D and the nursing home setting. Gerontologist. 2008;48:432–441. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Stevenson DG, Huskamp HA, Keating NL, et al. Medicare Part D and nursing home residents. J Am Geriatr Soc. 2007;55:1115–1125. [DOI] [PubMed] [Google Scholar]
- 16.Huskamp HA, Oakman TS, Stevenson DG. Medicare Part D and its impact in the nursing home sector: an update. Washington, DC: MedPAC; 2010. [Google Scholar]
- 17.Gambassi G, Landi F, Peng L, et al. Validity of diagnostic and drug data in standardized nursing home resident assessments: potential for geriatric pharmacoepidemiology. SAGE Study Group. Systematic Assessment of Geriatric drug use via Epidemiology. Med Care. 1998;36:167–179. [DOI] [PubMed] [Google Scholar]
- 18.Mor V, Angelelli J, Jones R, et al. Inter-rater reliability of nursing home quality indicators in the US. BMC Health Serv Res. 2003;3:20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Bostick JE, Rantz MJ, Flesner MK, et al. Systematic review of studies of staffing and quality in nursing homes. J Am Med Dir Assoc. 2006;7: 366–376. [DOI] [PubMed] [Google Scholar]
- 20.Zhang X, Grabowski DC. Nursing home staffing and quality under the nursing home reform act. Gerontologist. 2004;44:13–23. [DOI] [PubMed] [Google Scholar]
- 21.Artz MB, Hadsall RS, Schondelmeyer SW. Impact of generosity level of outpatient prescription drug coverage on prescription drug events and expenditure among older persons. Am J Public Health. 2002;92: 1257–1263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Doshi JA, Brandt N, Stuart B. The impact of drug coverage on COX-2 inhibitor use in Medicare. Health Aff (Millwood). 2004;Suppl Web Exclusives: W4–105. [DOI] [PubMed] [Google Scholar]
- 23.Thomas KS, Dosa D, Wysocki A, et al. The Minimum Data Set 3.0 Cognitive Function Scale. Med Care. 2017;55:e68–e72. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Cubanski J, Neuman T, Damico A. Coverage gap: trends, recent changes, and what’s ahead. Kaiser Family Foundation; Available at: https://www.kff.org/medicare/issue-brief/closing-the-medicare-part-d-coverage-gap-trends-recent-changes-and-whats-ahead. Accessed August 21, 2018. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.