And the most prescribed drug is……

Medicare Itemizes Its $103 Billion Drug Bill

April 30, 2015

The federal government popped the cap off drug spending on Thursday, detailing doctor-by-doctor and drug-by-drug how Medicare and its beneficiaries spent $103 billion on pharmaceuticals in 2013.

The data show that 14 drugs cost the federal government and Medicare beneficiaries more than $1 billion each, accounting for nearly a quarter of Medicare prescription drug spending in 2013. Most of those drugs are used to treat chronic conditions that plague the elderly, including diabetes, depression, high cholesterol and blood pressure, dementia and asthma.

pill money 570The brand drug Nexium, used to treat heartburn, acid reflux and related stomach ailments, cost the most: $2.5 billion for 1.5 million Medicare patients, who filled 8 million prescriptions and refills. The total cost included what was paid by Medicare, beneficiaries and third party groups such as supplemental health plans. The cost covered not just the drug ingredients but also sales tax and dispensing fees. It did not, however, include sometimes substantial manufacturer rebates, and the drug makers’ trade group warned that omission distorted the actual cost.

The most frequently prescribed drug was lisinopril, a generic used to treat high blood pressure and help patients survive after heart attacks. The drug was prescribed or refilled nearly 37 million times by more than 7 million Medicare beneficiaries at a cost of $307 million.

Federal officials said they hoped that disseminating the data would lead to new revelations about the prescribing patterns of doctors and for particular drugs. The database identifies doctors by name.

Niall Brennan, the chief data officer for the Centers for Medicare & Medicaid Services, said agency analysts have been examining the data for several years but that “the data is larger and diverse enough that other outside folks may develop insights that we have missed.”

Dan Mendelson, the CEO of Avalere, a Washington, D.C., consulting firm, said the data could provide patients with new questions about their prescription history when they visit their physician. “It’s really important to stimulate conversations that get patients more actively engaged in their care,” he said.

However, he noted that some doctors may not take kindly to a more inquisitive patient and longer conversations. “In the shorter term, I think it will irk some physicians,” he said.

The database tracked 3,450 different drugs prescribed by a million doctors, nurse practitioners, medical students, dentists and other providers.

Medicare Drugs 1

The most expensive drug per prescription was Carbaglu, a man-made enzyme used to treat people with high ammonia levels in the blood caused by a rare disorder, according to a Kaiser Health News analysis of the data. The drug was dispensed only 24 times, but at nearly $60,000 per claim it cost the government $1.4 million.

Among drugs dispensed to at least 10,000 beneficiaries, the most expensive was Revlimid, KHN found. It is used for some cancer patients. Dispensed for 24,637 patients, Revlimid cost $8,778 per claim. That totaled more than $1.3 billion.

Drug prescribing varied considerably among states, KHN found. Rhode Island and Nebraska had the most claims per Medicare beneficiary, averaging 4.6 per patient. Delaware had the lowest number, with the average number of claims per beneficiary at 3.3.

The CMS data is likely to be used in conjunction with other datasets the government has previously released, including what procedures individual doctors billed to Medicare and how much those cost. Analysts are also sure to look for relationships between drugs commonly prescribed by doctors and another Medicare database showing payments physicians received from drug companies for research, gifts, speaking fees, meals or travel.

Medicare Drugs by cost

Pro Publica, a nonprofit news site, obtained similar data for earlier years through the Freedom of Information Act and has already published analyses.

The Pharmaceutical Research and Manufacturers of America called the data misleading. “Significant price negotiation exists in Part D and results in rebates of as high as 20 to 30 percent for branded medicines,” the association’s president, John Castellani, said in a written statement. “These savings are not reflected in the data. Rebates have been a significant factor in keeping Part D program costs hundreds of billions of dollars below original estimates, while still offering beneficiaries steady premiums and a robust choice of plans.”

The American Medical Association also cautioned that the data could be misinterpreted.

“The data does not account for varying strengths or dosage levels of the medications or varying patient needs,” the association said in a written statement. “For example, a physician could prescribe a low dose of a medication and at a later time need to prescribe another, stronger dosage for the same patient if the low dose isn’t meeting their need or if the patient has an adverse react.”

The government noted that the top 10 most commonly prescribed drugs were generic and the 10 most expensive drugs were all brand name. The finding is not surprising since some brand name drugs are protected from competition by their patents.

An analysis Medicare released with the data found that in some parts of the country brand drugs were dispensed much more frequently than generics. Doctors in the western part of the country, including Washington, Oregon, Idaho and Nevada, and parts of in the Midwest leaned heavily toward generics, which tended to be dispensed between 78 percent and 81 percent of times. Brand drugs were favored in much of Texas and Alaska, where generics were dispensed in between 65 percent and 75 percent of cases.

Federal officials also calculated how prescription patterns varied among medical specialties. Family practice doctors prescribed the most drugs, followed by internal medicine doctors. Among the biggest medical specialties, psychiatrists prescribed the most expensive drugs, averaging $104 for a prescription or refill. While hematologists and oncologists were not among the top prescribers, their drugs averaged $550 per claim. The average cost of all prescriptions or refills was $75.

The data does not present a complete picture of physician prescribing. Most notably, it includes only those drugs for 36 million beneficiaries that were billed to Medicare’s Part D program, which make up 68 percent of all the people on Medicare. It does not reflect the prescriptions doctors wrote for privately insured patients or those on other government programs such as Medicaid. It also reveals nothing about the quality of these treatments or what kind of patients each doctor saw. The data also omit drugs administered in doctors’ offices and billed to Medicare’s Part B program.

To ensure that people could not identify beneficiaries, Medicare omitted prescriptions that were based on 10 or fewer claims per drug. That excluded 13 percent of claims.

Answer to CC #14

Case Challenge #14 presented a 46 year old female presenting with a persistent cough for 10 years, no significant exposures, no smoking history, no travel, and an otherwise negative ENT, allergy, GI, speech, etc. work-up. PFTs reveal obstructive pattern, but standard COPD medications do not prove to be helpful. CT reveals bilateral nodular disease and mosaicism. Biopsy of the lung reveals neuroendocrine cell hyperplasia which stain for carcinoid.

What is the best initial management strategy?

Serial CT Scans! The patient has DIPNECH (Diffuse idiopathic pulmonary neuroendocrine cell hyperplasia)

  • Definition: Hyperplasia of bronchial neuroendocrine cells in airway epithelium with aggregates of cells (“tumorlets”) extending beyond the basement membrane
    • first described in 1992 (Aguayo et al, NEJM)
    • typically female nonsmokers, mean age 45 years
    • absence of other pulmonary processes or environmental exposures to explain symptoms
  • Presentation:
    • years of cough, +/- wheeze and dyspnea
    • PFTs: irreversible airflow obstruction & a third of patients had restriction
    • CT: multiple nodules and air-trapping due to bronchiolitis. Moasaicism.
  • Management: “…longterm follow-up and treatment remains incomplete…” (Am J Resp Crit Care 2011)
    • COPD management: BA +/- inhaled or systemic steroids for symptomatic relief
    • Serial CTs to watch for development of carcinoid tumors
      • “Given its only minor risk of invasive metastatic spread, serial imaging for timely detection of progressing lesions and optimizing of concomitant obstructive pulmonary disease are the preferred treatment approaches.” (Dtsch Med Wochenschr. 2014 Jun;139(23):1245-8. doi: 10.1055/s-0034-1370073. Epub 2014 May 27.)
    • There is some evidence that somatostatin analogues are helpful, but long-term safety and efficacy data are lacking

Case challenge #14 (aka the Parkland Files!) will be posted next week!    

Grand Rounds Review: Severe Alcohol Withdrawal

This morning, Dr. Matthew Leveno, director of the Parkland Medical ICU and youngest winner of the housestaff teaching award, gave an incredible grand rounds presentation about the history, pathophysiology, diagnosis, and management of severe alcohol withdrawal. Take a look below for some important information. 


Alcohol abuse is a common problem globally, and it is estimated to result in 2.5 million deaths annually. Of the drugs of abuse, alcohol is the most common, with an estimated 18.3 million individuals dependent on it in the United States. Alcohol abuse has a prevalence of 22.4% in a hospitalised general medical population. In one analysis, alcohol-related admissions accounted for 9% of admissions to a population of mixed medical intensive care unit (ICU) and surgical ICU patients; in addition these patients accounted for 13% of total ICU costs. One population with a particularly high rate of alcohol abuse are trauma patients, with estimates of prevalence ranging from 31% to 70% across centers.

Alcohol-related complications in the ICU affect nearly every organ system (Table 1). Alcohol abuse in patients is associated with increased length of stay, outpatient pneumonia and an almost three times higher incidence of healthcare-associated infections.

Clinical Manifestations and Diagnosis

The gold standard for the diagnosis of alcohol withdrawal syndrome (AWS) is the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. It requires that a patient’s alcohol usage is heavy and prolonged, there is a cessation in alcohol intake and also that there is no other general condition that better accounts for the diagnosis.


AWS has four clinical stages: (1) autonomic hyperactivity, (2) hallucinations, (3) neuronal excitation and (4) delirium tremens (Table 2). Patients generally start the withdrawal process at 5 h, with hallucinations at 24 h, and delirium at 48 h; it is rare for this to persist for more than 120 h.


AWS is the result of a disruption of the delicate neurochemical balance that is controlled via inhibitory and excitatory neurotransmitters. The principal inhibitory neurotransmitter is gamma aminobutyric acid (GABA), which exerts its effect on the GABA-A neuroreceptor. A principal excitatory transmitter is glutamate, which affects the N-methyl-D-aspartate neuroreceptor. With chronic alcohol exposure, the brain has a tolerance to the effects of the alcohol due to down-regulation of the GABA-A receptor over time. This down-regulation may occur by modification of the GABA-A receptor in the alpha 1 subunit to make the receptor less susceptible to the effects of alcohol exposure.

Pharmacological treatment

The severity of the symptoms of AWS should direct the appropriate pharmacotherapeutic interventions. The patient’s comorbidities, other active diagnoses as well as exposure to any other drug of abuse should also be factored into the development of their treatment plan.


Benzodiazepines have historically been the mainstay pharmacologic intervention of AWS; they are generally considered to be the ‘gold standard’ treatment. It has been shown that sedative-hypnotic agents such as benzodiazepines, in comparison with other agents, reduce mortality and control the symptoms of AWS. All benzodiazepines have the same mechanism of action on the GABA receptor. Several agents have been used for AWS including chlordiazepoxide, lorazepam, valium, oxazepam and midazolam. Lorazepam is suggested as the benzodiazepine of choice for AWS due to its intermediate half-life, which balances a smooth withdrawal, with the potential for delayed metabolism in those with impaired hepatic function such as geriatric or cirrhotic patients.

Benzodiazepines were traditionally administered to AWS patients in a fixed dose regimen. There has now been over two decades of experience accumulated with the use of on demand or ‘symptom-triggered’ dosing of benzodiazepines for AWS treatment. This method of symptom-triggered dosing relies on the Clinical Institute Withdrawal Assessment for Alcohol [CIWA-A or CIWA-Ar (revised)]. In studies, the symptom-triggered dosing method results in both a decrease in the amount of benzodiazepines administered and a shortened duration of withdrawal symptom. While the symptom-triggered approach has these advantages, there is quite limited experience of the use of this approach in critical care settings, and it has not shown the same benefit across all studies.

Adjunctive agents

The alpha-2-agonist, clonidine, has traditionally been used to blunt the sympathomimetic effects of AWS. This has been done outside critical care settings. While intravenous clonidine is available in Europe, it is not currently available for use in the United States. This has resulted in intensivists to turn to dexmedetomidine, a drug derived from clonidine. Dexmedetomidine is not FDA-approved for AWS, but rather for procedural conscious sedation and sedation for mechanical ventilation


Beta-blockers have been used as an adjunctive agent in AWS. Given the sympathetic outflow associated with autonomic hyperactivity, betablockers are a direct antagonist. This medication can be administered either orally or intravenously, and it serves to normalise tachycardia and hypertension in non-agitated patients that are otherwise comfortable. In a randomised trial by Gottlieb, atenolol in patients with AWS served to make a more rapid resolution of their vital sign abnormalities and clinical signs such as tremor. Betablockers serve an important role as part of a multimodal pharmacological plan, but they should never be used without a GABA agent.


Haloperidol is a phenothiazine that is commonly prescribed in ICUs for acute agitation. It has the benefit of haemodynamic neutrality, and the possible complications of an elevation in the QTc interval and tardive dyskinesia. While haloperidol is an adjunctive agent in AWS setting, it is particularly useful for the symptoms related to delirium. Most frequently, it is used in patients with underlying psychiatric disorders. 


AWS continues to challenge clinicians in critical care settings. Keys to good outcomes in this area include early recognition of the disorder and rapid implementation of appropriate pharmacologic treatment. The range of symptoms represents a spectrum; the pharmacologic strategy needs to match the severity that the patient is experiencing. While some patients have a good therapeutic response to a single benzodiazepine agent, more severe cases may require a multimodality therapy. The current protocol used at our institution is presented in Table 3. With a stepwise protocol-driven plan, intubation and mechanical ventilation can be avoided except in the more severe cases, contributing to better outcomes in terms of length of stay and VAP. 

Modified from: DeMuro JP. Alcohol withdrawal syndromes in the critically ill. OA Alcohol 2013 Feb 01;1(1):1. Under the terms of the Creative Commons Attribution License (CC-BY).