Prescription Drug Claim Rejections for Nursing Home Residents under Medicare Part D

Abstract

Some nursing facility administrators and physicians have raised concerns about the increased administrative burden of working across multiple Part D drug plans (PDPs), citing the possibility for adverse resident outcomes and financial burden on facilities when PDPs reject drug claims for residents. We examined data on claims in rejected status for one large long-term care pharmacy as of December 2006. There was considerable variation across PDPs in both rejection rates and reasons for rejections.

Medicare Part D represents a substantial departure from how prescription drugs were previously financed and administered in the nursing home setting. Whereas before Part D, Medicaid financed drugs for the almost two-thirds of nursing facility residents dually eligible for Medicaid and Medicare,¹ the new benefit shifted drug coverage for these individuals to Medicare, requiring that they enroll in private Part D plans (PDPs).² Some portion of the remaining third of nursing home residents may have also joined a PDP. Although facilities worked across different coverage types before Part D – including Medicaid, Medicare Part A, and private coverage – the number of plans is almost certainly greater now that dual eligibles are covered by multiple plans instead of a single state Medicaid program.

Some nursing home administrators and physicians have complained that working across multiple PDPs with different formularies and utilization management (UM) processes has increased administrative burdens for facility physicians, nurses, and other staff.³ Concerns have also been raised about the possibility of negative outcomes for residents resulting from gaps in PDP formularies, UM policies like prior authorization, and the complexities of working across multiple plans.⁴ Although 2006 formulary coverage among PDPs serving dual eligibles was fairly comprehensive overall for several drug classes used commonly by nursing home residents, coverage was more limited and prior authorization more common for a minority of plans.⁵

As with other areas of nursing home pharmacy practice, facilities rely heavily on their long-term care pharmacies (LTCPs) to address challenges raised by Part D. LTCPs have become integrally involved in nursing home pharmacy practice, offering specialized services and supplies mandated under federal law.⁶ Through their consultant pharmacists, LTCPs offer facilities comprehensive drug management services and often coordinate related quality assurance and improvement activities, including prospective review of drug regimens to identify potential contraindications.⁷ Although facilities can pay retail pharmacies for such specialized services, LTCPs serve more than 80% of all nursing home beds nationwide.⁸

A new fiscal challenge for nursing homes and LTCPs under Part D is minimizing the use of non-covered medications. Coverage issues are not new, yet negotiating payments and addressing coverage and UM requirements across PDPs introduces greater variation and uncertainty compared with before Part D, when LTCPs received fee-for-service Medicaid payments for dual eligibles.⁹ Although LTCPs – and not nursing homes – now contract with PDPs, nursing homes and their pharmacies share the challenge of minimizing the cost of uncovered drugs. Nursing homes are especially attuned to this issue because the Nursing Home Reform Law of 1987 makes them statutorily responsible for providing medications in the physician's treatment plan regardless of coverage.

Anecdotally, some LTCPs and nursing homes have reported a financial burden associated with PDP claims rejections and wide variation in how “LTC friendly” PDPs are based on coverage of drugs important to the nursing home population and use of prior authorization or step therapy (requirements to fail lower-cost medications before higher-cost drugs are covered).¹⁰ To date, however, there is no knowledge of the types of medications for which claims are commonly rejected, the reasons for rejection, or how claims rejection rates vary across PDPs. To address these questions, we examine data on claims submitted by one large LTCP that were rejected by PDPs.

Data

We obtained data from a large LTCP on its Part D claims in rejected status as of December 2006. The LTCP provided these data under the condition that the LTCP and the specific PDPs with which it contracts not be identified. The data include all claims submitted and subsequently rejected for medications actually dispensed beginning January 1, 2006 that were still in rejected status in December 2006. The data include rejected claims for both nursing facility residents and assisted living residents served by the LTCP, although most were generated by nursing facility residents. The LTCP estimates that approximately 4% of all Part D claims submitted in 2006 were rejected.

Exhibit 1 shows rejections at the medication-level and Exhibit 2 at the PDP-level. Once an initially-rejected claim is paid, it is no longer considered in “rejected status.” For example, if a PDP rejected a Seroquel claim because of incomplete medical history documentation, but the claim was later paid when the required documentation was provided, the claim would not appear in the data. In contrast, if a PDP rejected a claim for Lexapro because the plan does not cover Lexapro, and the physician subsequently switched the patient to a covered antidepressant, the original Lexapro claim would appear. Thus, these data do not reflect all claims rejections during 2006, only claims that have not been paid as of December 2006. As such, these data capture the current financial burden of unpaid claims for the LTCP at a point in time.

Results

At the end of 2006, 507,875 claims were in rejected status for this LTCP.¹¹ Exhibit 1 presents the reasons for rejection for all these claims and for the twenty drugs with the largest number of rejections.

Approximately 27% of “rejected status” claims were rejected because the drug was not covered by the plan's formulary. Another 25% were rejected because of UM requirements, including prior authorization that was not granted (7%) or inadequate/incomplete clinical history documentation (18%). Nine percent were rejected because a refill claim was submitted too early (i.e., the plan did not think the prescription should be depleted). The remaining 39% were rejected for a variety of administrative or other coverage reasons, including that plan limitations had been exceeded, the claim was already paid or was too old, or the claim's enrollee identification number was invalid.

The twenty drugs with the largest number of claims in rejected status include many brand-name medications (e.g., Risperdal) and some frequently prescribed generic medications (e.g., lisinopril/Prinivil). Exhibit 1 highlights how reasons for rejections vary across drugs. For example, some rejections primarily stem from lack of coverage (e.g., Lexapro), others from not meeting prior authorization requirements (e.g., Namenda), and still others from having incomplete medical history documentation (e.g., Seroquel).

Exhibit 2 presents PDP rejection ratios and the proportion of rejections made for different reasons across sixteen PDPs that contract with the LTCP. The rejection ratios convey the relative portion of all claims rejected by each plan. For example, Plan B's rejection ratio implies that, relative to its claims volume, its share of rejected claims is 0.39 times the average rate. This share varies by more than a factor of ten across the sixteen plans, from 0.23 to 3.54. Given the LTCP's estimate of a 4% overall Part D rejection rate in 2006, we estimate that the corresponding rejection rates vary roughly between 1−14% of claims across plans.¹² The proportion of claims rejections made for specific reasons varies considerably across PDPs, likely reflecting differences in preferred methods for cost controls.

Discussion

Among claims in rejected status for one large LTCP at the end of Part D's first year, approximately one-quarter were rejected because the drug was not covered, one-quarter due to coverage requirements for the medication (e.g., prior authorization, medical history documentation), and most of the remainder for general administrative reasons. Although one might have expected only expensive brand-name drugs to be rejected frequently, some of the drugs with the greatest volume of rejections were relatively inexpensive generics. While the high volume of rejections for these generics is likely driven by their high frequency of use among elderly nursing home residents, high rejection rates for relatively inexpensive medications are surprising. The most common reason for rejection of the generics was incomplete medical history documentation, suggesting that rejections could stem from quality concerns (e.g., ensuring a particular medication is clinically appropriate). As might be expected, some drugs of contested efficacy (e.g., Alzheimers’ medications Namenda and Aricept) and drugs that may be used inappropriately (e.g., Procrit, used to treat anemia) were also among those with the greatest volume of rejections, with rejections typically due to prior authorization requirements and incomplete medical history documentation.

The clinical impact of limited coverage, extensive prior authorization or stepped therapy requirements depends on several factors, including prevalence of use, available alternatives, and the efficacy of specific medications. For instance, if a drug has clinically-superior alternatives, low coverage levels may be appropriate and cause minimal burden clinically. Similarly, prior authorization might add a valuable safeguard when prescribing could be questionable or inappropriate, either due to controversy about efficacy or concerns about risks or side effects. However, coverage restrictions and prior authorization may also create obstacles to receiving appropriate, clinically-important medications. Although interviews of key stakeholders did not report adverse clinical outcomes for residents, empirical study of administrative and clinical data is needed to understand impacts on resident outcomes.¹³

PDPs vary considerably in rejection rates and reasons for rejections. The variation across PDPs in rejection rates due to lack of coverage, insufficient medical history documentation, and prior authorization requirements suggests that formulary coverage and UM procedures vary considerably across plans, as do general administrative procedures. This variation no doubt adds to the administrative burden of facility and LTCP staff. These data also suggest that some plans reject a much lower proportion of claims than others, resulting in variable administrative burden across PDPs.

Because these data reflect claims for dispensed medications that were in rejected status as of December 2006, they do not capture claims that were initially rejected and then resolved. This could explain why rejections due to prior authorization requirements seem relatively low (6%) compared to reports by facility staff, physicians, and LTCPs.¹⁴ We are unable to assess the proportion of patients with claims in rejected status that were subsequently switched to another medication, and we have no data on pre-Part D rejection rates. We also cannot assess the relative magnitude of administrative burden involved with each PDP's administrative processes; we measure only whether the processes resulted in claims in rejected status, and thus presumably required at least some level of administrative and/or clinical attention by the facility and/or LTCP. These data do, however, provide an important indication of the clinical, administrative, and financial burden for nursing homes and LTCPs resulting from Part D claims rejections and information on variability across plans in rejection rates.

Rejection rates may drop in 2007 and beyond as LTCPs and nursing home staff become more familiar with coverage rules and administrative practices of the PDPs that serve their residents or if use of electronic prescribing systems that identify coverage issues at the point of prescribing become more widespread. In 2006 most PDPs had limited experience serving nursing home residents. As a result, plans may have been experimenting with strategies for UM and coverage for this population. Some PDPs may change policies given their experience in Part D's first year, potentially resulting in continued complexity and burden for nursing homes at least for the near future. Also, as risk corridors for PDPs widen over the next few years, plans may adopt stricter UM and formulary coverage for drugs used commonly by sicker beneficiaries, including nursing home residents, to avoid disproportionately enrolling higher-cost enrollees if the risk adjustment methodology does not adequately account for differences in financial risk across beneficiaries.¹⁵ Alternatively, as PDPs gain experience with this population, they may be able to streamline administrative processes.

Given the considerable variation across PDPs in claims rejection rates, nursing facilities and LTCPs might benefit from steering their residents to a subset of plans with more generous coverage and less stringent UM for drugs used commonly by residents net of any effect on premiums. CMS regulations, however, prohibit facilities from directing residents to particular plans, although facilities can provide residents with educational materials about coverage differences across plans.¹⁶ These restrictions are intended to minimize nursing home providers’ ability to steer residents in ways financially beneficial to the home, although they could also increase the likelihood that some residents are enrolled in plans that fail to cover their medications. Some facility administrators have advocated loosening these restrictions to allow advising residents and their families on plan choice, while others would like other entities (e.g., State Health Insurance Assistance Programs) to be able to assist residents.¹⁷

Alternately, an approach referred to as intelligent random assignment has been used by some state pharmacy assistance programs (e.g., Maine) to assign individuals to particular PDPs.¹⁸ Unlike the initial random assignment of dual eligibles to PDPs with premiums below regional benchmarks, intelligent random assignment may consider beneficiaries’ current medication use. Such an approach applied to the assignment of dual eligibles could result in a better match of plan and beneficiary medication use, possibly reducing administrative burden and costs associated with rejected claims for nursing homes. However, such an approach could result in higher Part D costs if pure random assignment resulted in switching some patients to lower-cost medications or if it caused premiums to rise. The approach could also lead to greater plan concerns about adverse selection and perhaps more restrictive medication coverage overall.

Other ways to reduce administrative burden associated with working across multiple PDPs for nursing home residents include two potential options recently discussed by MedPAC.¹⁹ Both would involve “carving out” the nursing home sector from the rest of Part D. Under the first approach, PDPs would compete for a contract to serve all nursing home residents in a given region, so a facility would be subject to a single set of coverage and UM processes for residents covered by Part D. Under the second, Medicare would reimburse LTCPs directly for medications dispensed, so residents would no longer be enrolled in a PDP. MedPAC examined these two options along with a third (continuing the current approach of contracting with multiple PDPs) in its June 2007 report but made no recommendations on this issue.

It may be premature to consider dramatic changes to the structure of Part D for nursing home residents in the absence of data on its impact on utilization, outcomes, and quality of care for this population. Because nursing home residents are a particularly vulnerable population, continued research and monitoring of the benefit's impacts are needed to ensure Part D works well to meet the needs of this group.