Seth A. Seabury, PhD

Pharmaceutical innovation has led to the development of “specialty drugs,” loosely defined as pharmaceuticals produced using advanced biotechnology methods. These products have enabled physicians to more effectively treat diseases like cancer, multiple sclerosis, rheumatoid arthritis, and other conditions that were previously very difficult to manage. Sometimes also referred to as “large molecule” drugs, specialty therapies have attracted attention because their clinical benefi ts are very often accompanied by  higher financial costs compared with their traditional “small molecule” counterparts. With the pressure to curb spending on healthcare, benefit managers have shown increasing interest in monitoring and containing the utilization of specialty drugs.

 For example, some have considered moving coverage of physician-administered injectable products from medical benefit schedules to pharmacy benefits, which can limit benefit generosity and require greater cost sharing by patients.

However, the extent to which containing cost growth in specialty drugs will contribute to aggregate healthcare savings remains unclear. While specialty drugs are often expensive, they are used relatively infrequently except by the sickest of patients. Moreover, even among this very ill population, specialty drug spending remains relatively modest. For example, through 2005, only $1 out of $5 spent on the costliest patients went toward prescription drugs in total, and less than $1 out of $15 went to specialty medicines among this patient group.5 There is also evidence suggesting that patients place a high value on specialty drug therapy and tend to be less responsive to pharmacy benefi t management.

 Moreover, recent evidence suggests that in some cases, the higher immediate cost of therapy with specialty drugs could be offset by lower long-term costs.

More insight is needed into the role that specialty drugs play in healthcare spending, in order to better determine appropriate coverage and cost-sharing policies that balance the goals of cost containment and patient access.

This study characterizes the role of specialty drugutilization in aggregate healthcare spending in the private sector. We used a large database of private sector healthcare claims to compute the share of patients receiving specialty drugs, along with the total share of spending and spending growth accounted for by specialty drugs, from 2000 to 2009.

We used data on medical claims from 2000 to 2009 covering 46 Fortune 500 employers. The data captured all healthcare claims for 11.2 million patient-years for patients under age 65 years who were continuously enrolled for at least 1 full year in all plans offered by these employers, including inpatient, emergency, ambulatory services, and pharmacy benefi t claims. These data have been widely used to study healthcare utilization and spending in a variety of settings, both generally and for specialty drugs.

Specialty drugs were identified differently depending on the method of administration. Physician-administered injectable pharmaceuticals were identified using J-codes on the medical benefit, while those covered under pharmacy benefi t plans were identifi ed using national drug codes (NDCs). A list of the specialty drugs considered is available upon request.

Using these data, we aggregated spending for individual medical claims to create annual spending measures for several different  categories: total spending, medical benefit spending (defi ned as all combined inpatient and outpatient spending), total pharmacy  benefit spending,and spending on specialty drugs. Spending on specialty drugs was broken into pharmacy benefit spending and medical benefit spending. Dollar values were normalized to 2009 dollars using the medical component of the consumer price index (CPI). To evaluate trends, we also computed total spending for each category by year.

Note that in the case of physician-administered injectable products, medical claims did not separate drug spending from the cost of other procedures billed on the same claim. For example, the claim amount might include both the cost of the drug and the amount charged by the physician for administering it. Conservatively, we attributed all spending on any claim with a J-code for a physician-administered specialty drug to the specialty drug itself. As a result, our estimates will overstate spending on specialty drugs themselves and should be interpreted as the total cost of administering them. To understand the clinical utilization of specialty drugs, we identified patients  with chronic disease (eg, cancer) based on the presence of the associated

 International Classification of Disease, Ninth Revision, Clinical Modification 

diagnosis code on at least 1 inpatient or 2 outpatient medical claims in a year.

Because specialty drugs are designed to treat patients with serious illnesses, it is unlikely they would have been widely used in the general population. To better understand the patterns of specialty drug use for at-risk patients, we computed the distribution of total healthcare spending across all individuals in the sample in each year and identified patients in the top 25th, 10th, 5th, and 1st percentiles of spending. We then selected these patients and replicated our analysis of spending in each subsample.Because specialty drugs are most often used by patients with serious conditions who have higher average healthcare utilization, we expected specialty drugs to be more widely used in these high-cost subpopulations.

There is no single defining characteristic that identifies a particular drug as a specialty drug.

 To classify specialty drugs for this study we used a definition developed by IMS Health.

 The IMS definition was based upon 8 objectively measured attributes of drugs, such as the type of condition targeted (chronic or acute), the method of delivery, the need for patient oversight, and cost. These conditions were defined at the molecule level and molecules that met 5 of the 8 attributes were identified as specialty drugs.

Using this definition, there were 215 unique molecules identifi ed as specialty drugs in our data. Of these, there were 178 biologic molecules, 20 blood coagulation factors, and 17 non-biologic therapies. Using the product name and molecule name to link to NDC and J-codes, we identified at least 1 claim for all specialty drugs in our data.

In our data, average annual healthcare spending was $4076 per patient (

). Of this total, the majority of spending (78%) was attributed to medical costs ($3175). Specialty drugs accounted for 2.7% of spending ($112), largely due to infrequency of use. For example, only 1.6% of patients in the sample (178,161) used specialty drugs. About 75% of spending on pharmacy drugs was paid for through the pharmacy benefit side. Overall, spending on specialty drugs accounted for 9.3% of all pharmacy spending.

As expected, a larger share of patients received specialty drugs in the high-cost subsamples. Among patients in the top quartile of the total healthcare spending distribution, about 6.7% had at least 1 specialty drug claim. Among those in the very top percentile of this distribution, the rate was 27.3%. Moreover, specialty drugs accounted for larger shares of pharmacy spending in these high-cost groups, rising to as much as 28.8% of pharmacyspending for patients in the top 4 percentile of cost. Even so, the share of specialty drugs in total spending was relatively modest in the high-spending groups, never exceeding 5.6%. The majority of spending (90%) for the highest cost patients was due to medical services unrelated to specialty drugs.

 analyzes various categories of spending for all patients and those with at least 1 specialty drug claim, and reports statistics for groups at different points in the spending distribution. Average annual spending was higher for patients taking specialty drugs compared with the average patient ($28,863 compared with $4076). This is consistent with the hypothesis that specialty drug users are generally sicker than average. This trend was true for all subpopulations studied. For patients with at least 1 specialty drug claim in a year, the average annual spending on specialty drugs was $7095, a little less than one-fourth of their total spending. For patients in the high-cost sub groups, specialty drugs accounted for about the same percentage or less of total spending. For  patients in the top percentile who received specialty drugs, the average cost of specialty drugs was $21,576, or 14.6% of their total spending. Thus, for patients taking specialty drugs, even in the highest cost subpopulations, medical spending unrelated to specialty drugs accounted for the majority of total healthcare costs.

Approximately 90% of all spending on specialty drugs was for patients with cancer, Crohn’s disease, kidney disease, multiple sclerosis, psoriasis, and rheumatoid arthritis. Spending ranged from 28% for cancer to 4% for Crohn’s disease.

We also compared how the use of specialty drugs changed over the study period. We found that over the 10-year period, the number of patients taking specialty drugs and their contribution to spending grew. In 2000 specialty drugs accounted for 1.0% of total healthcare spending; in 2009, specialty drugs accounted for 3.7% of total healthcare spending.

From 2000 to 2009, 1.6% of all patients used a specialty drug, accounting for 2.8% of total spending and 9.3% of spending in the pharmacy benefit. For patients in the highest percentile of total spending, specialty drugs accounted for 4.4% of total spending and 28.8% of pharmacy benefi t spending. The share of spending on specialty drugs grew from 2000 to 2009, but remained modest, rising from 1.0% to 3.7% of total healthcare costs. At current growth rates, we would expect the share of spending on specialty drugs to be approximately 6.7% by 2020. In the context of healthcare spending growth rates, these shares are relatively modest. Given recent annual increases in Medicare spending of nearly 8%,

 and similar recent annual increases in private health insurance premiums,

 even complete elimination of specialty drugs would stave off less than a year’s worth of healthcare cost growth.

Our study has limitations worth noting. First, while the database used in our study is national in the sense that it includes patients in all regions of the country, it is not designed to be a representative sample of the US population. Nevertheless, these data have been used extensively in past work to study healthcare spending in many contexts and appear to closely mimic national trends. Second, our data could not separate spending on physician-administered drugs from other outpatient spending on the medical claim. For this reason, our estimate of spending on specialty drugs in medical benefit claims could have overstated the share of aggregate healthcare spending on specialty drugs, though by how much is unclear. Our analysis also focused only on the direct cost of specialty drugs, but there could be indirect costs we did not capture. For example, if specialty drugs were effective at extending life for diseases that would otherwise be fatal (such as cancer), this could lead to more downstream medical costs simply because the patients would be living longer. Additionally, our analysis did not attempt to capture any of the indirect benefits that accrue from the use of specialty drugs, such as increased quality of life or productivity gains. Our study focused only on the cost and use of therapy, and said nothing about the effi cacy of treatment or the value to patients. Ultimately, allowing access to expensive therapies could benefit society if those costs were associated with high value. Similarly, even if a treatment contributes little to total cost, if it produces no value for patients eliminating it could still be effi cient. Focusing solely on cost ignores this important aspect of the policy debate.

In the search for ways to lower healthcare costs, attention naturally shifts to higher-cost treatment and therapies. This has led to increased efforts to employ methods to contain the utilization of specialty drugs as a tool for cutting costs. However, while specialty drugs may be a high-cost option for a select group of patients, the fact that their use is relatively uncommon suggests that even if spending on specialty drugs were to fall signifi cantly, it would have no more than a modest impact on aggregate healthcare spending in the private sector.

Articles

This study characterizes the role of specialty drug utilization in aggregate private sector healthcare spending in the United States from 2000 to 2009.