Gerry Oster, PhD

Many patients are now encountering much higher co-pays for generic drugs that have 

been designated “non-preferred” by their insurers, including those recommended as 

first-line treatment in evidence-based guidelines. For some diseases, in fact, many insurers 

have no “preferred” generic medicines, effectively rendering the diseases themselves 

“non-preferred.” Designation of clinically important generic medicines as “non-preferred” 

without ensuring that patients have access to therapeutically equivalent “preferred” 

drugs runs counter to established principles of formulary design, may increase other 

healthcare costs, and ultimately may undermine emerging payment reform initiatives.

The new blockbuster drug sofosbuvir (Sovaldi) 

is offering hope to many patients with hepatitis 

C, but treatment is expensive and many insurers 

are demanding that patients shoulder a large portion 

of the cost. The demand that patients pay a larger share 

of their drug costs, however, is not limited to expensive 

new medicines. In fact, many patients are now facing 

substantially higher co-pays for various generic drugs 

that their insurers have designated “non-preferred,” often 

including those recommended as first-line treatment 

in evidence-based guidelines for hypertension, diabetes, 

epilepsy, schizophrenia, migraine headache, osteoporosis, 

Parkinson’s disease, and human immunodeficiency 

virus (HIV). We are concerned about this relatively recent 

For many years, most insurers had formularies that consisted 

of only 3 tiers: Tier 1 was for generic drugs (lowest 

co-pay), Tier 2 was for branded drugs that were designated 

“preferred” (higher co- pay), and Tier 3 was for “nonpreferred” 

branded drugs (highest co-pay). Generic drugs 

were automatically placed in Tier 1, thereby ensuring that 

patients had access to medically appropriate therapies at 

the lowest possible cost. In these 3-tier plans, all generic 

drugs were de facto “preferred.” Now, however, a number 

of insurers have split their all-generics tier into a bottom 

tier consisting of “preferred” generics, and a second 

tier consisting of “non-preferred” generics, paralleling the 

similar split that one typically finds with branded products. 

Co-pays for generic drugs in the “non-preferred” tier are 

characteristically much higher than those for drugs in the 

To better understand coverage policies in plans with 2 

tiers for generic drugs, we identified several such offerings, 

including both commercial plans and those under the 

Medicare Part D program, via an informal search of the 

Internet. For 6 such plans, we examined coverage policies 

for 10 widely used drugs—all generically available&mdash;

that are recommended as first-line treatment in current 

While 2 of the plans provide access on a “preferred” 

basis to all of the medicines we considered, 1 or more 

of the drugs are “non-preferred” in all of the remaining 

plans. Metformin, for example, is a “non-preferred” drug 

in 1 plan, despite being a first-line treatment for type 2 

diabetes mellitus. Two plans have no “preferred” generic 

anticonvulsant drugs; 3 plans have no “preferred” generic 

antipsychotic medicines; levodopa is designated a 

“non-preferred” agent in 3 plans; 4 plans have no “preferred” 

generic triptans (for migraine headache); and all 

generic antiretrovirals are Tier 2 agents in 4 plans. When 

there are no “preferred” generics from which clinicians 

and patients with particular diseases can choose, it may 

be argued that the diseases themselves effectively are 

We think it is apparent that designation of these generic 

drugs as “non-preferred” was based on cost considerations 

alone; indeed, a number of insurers unabashedly 

refer to their second tiers as consisting of “higher-cost 

generic drugs.” It also is difficult to imagine what criteria 

other than cost could have led to the exclusion of 

highly effective and widely used generic drugs from the 

“preferred” tier. If cost was indeed the reason for designating 

these medications “non-preferred,” it would be 

inconsistent with all accepted principles and standards 

of formulary design and management, including guidelines 

jointly endorsed in 2000 by 7 professional groups, 

representing physicians, pharmacists, business, and other 

It is sometimes argued that patients should have “skin 

in the game” to motivate them to become more prudent 

Articles

The authors comment on the growth of drug plans with tiers for "non-preferred" generics, and argue that most are inconsistent with established principles of formulary design.

Is All "Skin in the Game" Fair Game? The Problem With "Non-Preferred" Generics

To examine the benefit-risk profile of live attenuated influenza vaccine (LAIV) across a range of clinical scenarios in which we varied assumptions regarding both the percentage of children who would receive LAIV in lieu of trivalent inactivated influenza virus (TIV) and the extent of off-label use.

 Model of expected benefits and risks of immunization of young children against influenza.

 We estimated expected numbers of cases of influenza illness (FLU), medically significant wheezing (MSW), and hospitalization in a single influenza season under alternative assumptions regarding use of LAIV in lieu of TIV, based on projections from a large phase III trial.

 Assuming no use of LAIV in nonindicated children (aged <24 months and those with history of recurrent wheezing or asthma), and 50% use in lieu of TIV among children in the indicated population, there would be 2099 fewer FLU cases per 100,000 children aged 12 to 59 months, and no change in MSW or hospitalization. If LAIV also were used in lieu of TIV among 20% of children aged 12 to 23 months and 20% of children aged 24 to 59 months with a history of recurrent wheezing or asthma, there would be a further reduction of 397 FLU cases and 12 hospitalizations per 100,000 children aged 12 to 59 months, but 51 additional MSW cases.

 Our study suggests that even if LAIV were sometimes used inadvertently in clinical practice in young children for whom it is not indicated, the overall balance of expected benefits and risks would remain favorable.

	Live-attenuated influenza vaccine (LAIV) is not recommended for use in children aged 24 to 59 months with a history of recurrent wheezing or asthma, or approved for use in children aged 12 to 23 months. In clinical practice, however, some of these children may inadvertently receive LAIV.

		If LAIV were used inadvertently in lieu of TIV among 20% of vaccinees aged 12 to 23 months, and 20% of vaccinees aged 24 to 59 months with a history of recurrent wheezing or asthma, the number of cases of influenza and hospitalization would decline below levels otherwise expected, but the number of cases of medically significant wheezing would increase. 

		Even if LAIV were sometimes used inadvertently in clinical practice among young children for whom it is not indicated, the overall balance of expected benefits and risks would remain favorable.

The burden of influenza is substantial in young children. During the 2003-2004 influenza season, the influenza-attributable hospitalization rate for children younger than 2 years was 3 per 1000, or approximately the same as that of persons aged 65 years and older, for whom annual vaccination has long been recommended. 

 Young children (ie, those younger than 5 years) also have very high rates of outpatient visits for influenza—10 to 250 times their rate of hospitalization.

 In light of this burden, the Advisory Committee on Immunization Practices of the Centers for Disease Control and Prevention (CDC) recommended routine annual vaccination of all children aged 6 to 23 months with trivalent  inactivated influenza vaccine (TIV), beginning in the 2004-2005 winter season.

 This recommendation was expanded in 2006 to include children aged 24 to 59 months, and again in 2008 to include children aged 5 to 18 years.

Although the safety of TIV in young children has been demonstrated in several studies, its efficacy in this age group appears to be less than that in older children and adults.

A  review of randomized trials reported a pooled efficacy estimate of 63% for children younger than 9 years,

 as contrasted with the generally accepted efficacy range of 70% to 90%  among healthy adults younger than 65 years.

 In 2003, a live attenuated influenza vaccine (LAIV) was approved for use in healthy persons aged 5 to 49 years. Although data  available at the time of approval indicated that LAIV was efficacious against influenza in children younger than 5 years, approval for this age group was withheld at that time due to uncertainty regarding safety.

To shed additional light on the comparative efficacy and safety of LAIV versus TIV in young children, a large phase III clinical trial was undertaken in 8475 children aged 6 to 59 months.

 Although children randomized to LAIV had a significantly lower rate of culture-confirmed influenza compared with those receiving TIV (55% relative efficacy), and risks of medically significant wheezing (MSW) and hospitalization were comparable for the 2 vaccines in selected (especially older) children, potential safety concerns were noted. In  particular, among children aged 6 to 11 months, LAIV was associated with a higher risk of MSW in the 42-day postvaccination period (3.8% vs 2.1% for TIV, 

 = .08). A higher risk of  hospitalization—mostly for common pediatric (eg, respiratory, gastrointestinal) disorder&mdash;also was noted for LAIV among children aged 6 to 11 months over the period beginning with the first vaccine dose and ending 180 days after the last vaccination (6.1% vs 2.6% for TIV, 

 = .002). Children aged 12 to 23, 24 to 35, and 36 to 47 months with a history of wheezing or asthma who received LAIV also had nominally (but not significantly) higher risks of hospitalization.

In 2007, the US Food and Drug Administration (FDA) expanded its approval of LAIV to include children aged 24 to 59 months.

 Approval came with several warnings and  precautions, however, including one against its use in children younger than 24 months (because of increased risks of hospitalization and wheezing) and another against its use in those with asthma or recurrent wheezing. Notwithstanding the product labeling, LAIV may be used inadvertently in clinical practice in children for whom it is not currently indicated. Although it is unknown whether—and to what extent&mdash;use of LAIV may deviate from labeled indications and recommendations, off-label use of vaccines among children does not appear to be uncommon. For example, in a recent study of rotavirus immunization in Philadelphia, Pennsylvania, 23% of children who received rotavirus vaccine were reported to have been given their first dose when they were outside the recommended age range.

 We therefore undertook a study to examine the potential benefits and risks associated with approval of LAIV for use in children aged 24 to 59 months, considering the potential impact of possible off-label use.

We developed a model to characterize expected benefits and risks of immunization of young children (aged 12-59 months) against influenza in a given season. Children aged 6 to 11 months were not considered in the model because they are at least 1 full year removed from the lower age limit reflected in the revised indication for LAIV, and thus were assumed highly unlikely to receive LAIV in clinical practice. In the model, we conservatively assumed that children would receive either LAIV or TIV, and, therefore, that LAIV would be used exclusively in lieu of TIV during that influenza season. For children who were vaccine-naïve, we allowed vaccination to be either partial (1 dose only) or complete (2 doses). To reflect differential benefits and risks, we stratified the model population into 4 subgroups based on age (12-23 months vs 24-59 months) and risk profile (with or without a history of recurrent wheezing or asthma). Outcomes of interest included expected numbers of cases of influenza illness and MSW, and hospitalizations. Risk of these events was assumed to depend on age, risk status, vaccine received, and whether vaccination was partial or complete. We estimated our model using data from a large phase III clinical trial of LAIV versus TIV in young children and other sources, as described below. We examined expected benefits and risks under various scenarios regarding the extent of use of LAIV among children in the above-defined subgroups. The time frame of interest was a single influenza season.

Estimation of model parameters was based primarily on data from the safety population (all children who received at least 1 dose of study vaccine with any follow-up data) in the phase III clinical trial (

). The design of this randomized, double-blind, active-controlled trial has been described in detail elsewhere.

 Briefly, 8475 children aged 6 to 59 months (primarily from the United States, Europe, and the Middle East) were randomized 1:1 to receive either LAIV or TIV before the beginning of the 2004 influenza season. Randomization was stratified by age at first dose, previous influenza vaccination, history of recurrent wheezing (prospectively defined as >3 medically attended wheezing illnesses), and country of residence. Exclusion criteria included a history of hypersensitivity to any component of LAIV or TIV, immunosuppressive condition, medically diagnosed/treated wheezing within 42 days before study enrollment, history of severe asthma (as assessed by the investigator), acute febrile illness within 3 days before enrollment, and recent use of aspirin or salicylate-containing products. Children not previously vaccinated against influenza received 2 doses of study vaccine; all others received 1 dose. Each dose of LAIV and TIV contained the three 2004-2005 influenza virus strains recommended by the FDA.

The proportions of the model population in the 4 age and risk subgroups (ie, vaccinated children aged 12-23 and 24-59 months, respectively, with and without a history of recurrerent wheezing/asthma) were estimated based on the size of each subgroup in the total US population and corresponding vaccine coverage levels. The size of each subgroup was estimated by multiplying the number of children aged 12 to 23 months and 24 to 59 months, respectively, in the United States (based on 2005 US Census data)

 and the age-specific percentage of children with a history of recurrent wheezing or asthma (based on unpublished data from the phase III trial, MedImmune, 2007). Because the age distribution of children in the phase III trial was different from that in the United States (ie, children aged 6-23 months were overrepresented in the trial), percentages from the trial  were standardized to the US population. Vaccine coverage levels were assumed to be the same in all subgroups, based on the recent Advisory Committee on Immunization Practices recommendation of annual influenza vaccination for older children and a recent survey suggesting comparable coverage levels among children aged 12 to 23 months and those aged 24 to 59 months with asthma, groups for whom annual vaccination was recommended at the time.

 Among vaccine-naïve children, the proportion receiving only 1  dose was estimated based on data from the National Immunization Survey, the National Health Interview Survey, and other sources.

Due to the greater likelihood of prior vaccination, fewer older children were estimated to be vaccine-naïve.

Influenza illness was defined, per the phase III clinical trial protocol, as body temperature >100°F oral or equivalent plus cough, sore throat, or runny nose/nasal congestion occurring on the same day or consecutive days, in association with a positive nasal swab culture for wild-type influenza (modified CDC-ILI [culture-confirmed influenza-like illness, as defined by the CDC, caused by any strain]). Rates for children fully vaccinated with TIV and LAIV were based on data from the phase III trial, with age standardization to the US population, and included disease caused by both matched (ie, included in the vaccine) and unmatched strains occurring up to 180 days after the last dose of vaccine. For children  receiving only partial vaccination with LAIV, the rate of modified CDC-ILI was conservatively assumed to be 28% higher than the risk with full vaccination, based on results of another phase III study.

 For partial vaccination with TIV, the rate of modified CDCILI was assumed to be 50% higher than that for full vaccination with TIV.

Medically significant wheezing was defined as wheezing documented by a healthcare provider and accompanied by (1) tachypnea, retractions, or dyspnea; (2) O2 saturation <95%; or (3) a new prescription for daily bronchodilators.

Rates of MSW for children aged 12 to 23 months who were assumed to be vaccinated with LAIV (full or partial) were based on age-adjusted data from the phase III trial; all such events through 84 days after first dose of vaccine were considered (which corresponds to the prespecified maximum follow-up period for evaluating the occurrence of MSW in the trial—ie, 42 days after receipt of last dose of vaccine, which for vaccine-naïve children was specified to occur 28-42 days after receipt of first dose). Rates of MSW for children of this age assumed to receive TIV were based on trial data and were assumed to be the same for full and partial vaccination. For children aged 24 to 59 months, rates of MSW in the phase III trial were, in fact, slightly lower for LAIV than for TIV, but they were conservatively assumed to be the same in our analysis, irrespective of risk status.

Hospitalization was defined as an inpatient admission for any reason lasting 24 hours or more. Hospitalization rates for fully vaccinated children were based on age-standardized trial data (through 180 days after the last dose). Rates for partially vaccinated children aged 12 to 23 months were assumed to be higher than those for fully vaccinated children in this age group, based on higher modified CDC-ILI rates and the conditional probability of influenza-related hospitalization.

 Rates for children aged 24 to 59 months receiving TIV and LAIV were again conservatively assumed to be the same, despite a slightly lower rate for LAIV in the trial.

We first conducted analyses to compare expected benefits and risks in each of the 4 subgroups of children assumed to receive either LAIV or TIV: (1) aged 24 to 59 months without a history of recurrent wheezing or asthma (ie, those not covered by warnings and precautions); (2) aged 24 to 59 months with a history of recurrent wheezing or asthma; (3) aged 12 to 23 months without a history of recurrent wheezing or asthma; and (4) aged 12 to 23 months with a history of recurrent wheezing or asthma. For the first group (ie, those not covered by warnings and precautions), substitution of LAIV for TIV was allowed to vary from 25% to 100%. In analyses focusing on the other 3 groups, substitution of LAIV for TIV was varied from 10% to 30% (lower rates were assumed, due to the warnings and precautions).

We then pooled results from these separate analyses, weighting estimates for each subgroup by its respective proportion of the total vaccinated population aged 12 to 59 months, to yield expected changes in the outcomes of interest at various assumed levels of use of LAIV. In these pooled analyses, rates of use of LAIV in children covered by warnings and precautions were assumed never to exceed LAIV coverage rates among children aged 24 to 59 months without a history of recurrent wheezing or asthma (ie, those not covered by the warnings and precautions). We expressed our results on a standardized basis, per 100,000 children vaccinated.

Use of LAIV in lieu of TIV among children aged 24 to 59 months without a history of recurrent wheezing or asthma would result in 650 fewer cases of influenza illness per 100,000 vaccinated in this group for every 10% increase in share (

). Thus, for example, if 50% of these children received LAIV rather than TIV, cases of influenza illness would be expected to decrease by 3248 per 100,000 vaccinated in this subgroup. Similar reductions in the number of cases of influenza illness (per 100,000 vaccinated) would be expected with 10% increases in share in the other 3 groups: 704 fewer cases among children aged 24 to 59 months with a history of recurrent wheezing or asthma (

), 479 fewer cases among children aged 12 to 23 months without a history of recurrent wheezing or asthma (

), and 958 fewer cases among children aged 12 to 23 months with a history of recurrent wheezing or asthma (

Change in Expected Cases of Medically Significant Wheezing and Hospitalization

In the 2 groups of children aged 12 to 23 months, however, there would be offsetting increases in MSW: an additional 84 cases per 100,000 vaccinated among those without a history of recurrent wheezing or asthma, and an additional 454 cases per 100,000 vaccinated among those with such a history. Changes in rates of hospitalization also would be expected among children aged 12 to 23 months: a decrease of 40 admissions per 100,000 vaccinated children without a history of recurrent wheezing or asthma, but an additional 293 admissions per 100,000 vaccinated children with such a history. (Because rates of MSW and hospitalization [ie, the values for corresponding model inputs set forth in the Table] were assumed to be the same for LAIV and TIV among all children aged 24 to 59 months, projections do not differ by the assumed use of LAIV in lieu of TIV.)

With aggregation across the 4 age and risk subgroups based on population weights, if 50% of vaccinated children in the indicated population receive LAIV in lieu of TIV—and assuming no use among children aged 12 to 23 months or those with a history of recurrent wheezing or asthma&mdash;the rate of influenza illness among vaccinated children aged 12 to 59 months would decline by 2099 cases per 100,000 (

), while rates of MSW and hospitalization would remain unchanged (

, respectively). If LAIV also was used in lieu of TIV by 20% of vaccinated children age 12 to 23 months and 20% of those with a history of recurrent wheezing or asthma, the expected rate of influenza illness  among vaccinated children aged 12 to 59 months would decline by an additional 397 cases per 100,000, and the expected rate of hospitalization would decline by 12 cases per 100,000. Such use, however, would result in an additional 51 cases of MSW per 100,000 vaccinated.

Use of live attenuated influenza vaccine in young children has a favorable benefit-risk profile.

Benefits and Risks of Live Attenuated Influenza Vaccine in Young Children

To compare healthcare utilization and costs in the year preceding and following initial diagnosis of fibromyalgia (FM).

Using a large US health insurance claims database, we identified all persons with newly diagnosed FM (

International Classification of Diseases,

diagnosis code 729.1) between January 1, 2003, and December 31, 2005 ("FM patients"). Each patient's first-noted claim with a diagnosis of FM was designated the "index date," and all pharmacy, outpatient, and inpatient claims were compiled over the 12-month periods preceding and following this date ("prediagnosis" and "postdiagnosis," respectively). Patients with incomplete pre- or postdiagnosis data were excluded. Healthcare utilization and costs were compared between the 2 periods.

A total of 1803 patients met all study inclusion criteria; mean (SD) age was 50.4 (9.4) years; 91% were women. Comorbidities were common, including arthritis (21% of study subjects), back pain (20%), and painful neuropathic disorders (16%). The percentage of study subjects receiving various pain-related medications increased from pre- to postdiagnosis, including opioids (51.3% vs 55.9%), antiepileptics (22.6% vs 28.6%), and tricyclic antidepressants (15.5% vs 21.2%) (all 

<.01). Mean total healthcare costs also increased by $1725 between these periods (mean [95% confidence interval]: $9324 [$8655, $10,092] vs $11,049 [$10,245, $11,973], respectively; 

Patients with FM are often seen for other medical problems prior to initial diagnosis. Levels of healthcare utilization and costs are high during both the pre- and postdiagnosis periods.

Fibromyalgia (FM) is a widespread disorder of unknown etiology that affects an estimated 2% to 4% of the general population,

 and more than 5% of patients in general medical practice.

 Women are approximately 9 times more likely to develop FM than men.

Symptoms-predominantly widespread musculoskeletal pain-typically appear between the ages of 20 and 55 years. Other symptoms are often present, including sleep disturbance, fatigue, morning stiffness, paresthesias, headaches, and exercise intolerance. The symptoms of FM can be prolonged and debilitating, and evidence of complete remission is rare.

Information on the etiology of FM is limited. Symptoms and pathophysiologic findings suggest several levels of dysfunction, including peripheral nociceptive input, dorsal horn of the spinal cord, and brain processing. The underlying cause of FM may be some form of immune dysfunction, either triggered or exacerbated by infectious conditions and/or traumatic life events.

 FM also has been linked to abnormal central sensitization,

 pathologies at the level of the muscles,

 Family members of FM patients are themselves more likely to have FM, suggestive of a genetic predisposition toward the disease.

Although the epidemiology of FM has been widely studied, comparatively little is known about the economic burden of the disease. In a prior study based on US healthcare claims data, we reported that total healthcare costs were significantly higher among FM patients than those of similar age and sex without evidence of this disease ($9573 vs $3921 for comparison group; 

<.01)10; the prevalence of various comorbidities also was significantly higher among FM patients. More recently, and also using US healthcare claims data, White and colleagues examined healthcare utilization and costs in patients with "established" FM (n = 13,588) and newly diagnosed disease (n = 27,947), respectively.

 Among the latter, the authors reported that mean annual healthcare costs increased by $2099 between the pre- and postdiagnosis periods.

 study are of interest, the specificity of the algorithm that they used to identify patients with FM (ie, 2 or more encounters with a diagnosis of FM over a 7-year period) was probably quite low. In the present study, we examine healthcare utilization and costs in the periods immediately preceding and following initial diagnosis of FM, using a case-ascertainment algorithm that we believe has greater specificity than that employed in earlier research. Also in contrast to prior research, we include an examination of absenteeism and short-term disability (ie, indirect costs) during these 2 periods.

Data were obtained from the Medstat MarketScan Health and Productivity Management Database. The database is comprised of information from enrollment files and medical and outpatient pharmacy claims from a variety of private insurers providing healthcare coverage in the United States. The database is a subset of the entire MarketScan Database, which consists of approximately 28 million covered lives enrolled in about 80 health plans from 8 states in the United States. All patient identifiers in the database have been fully encrypted (ie, patients are fully deidentified), and the database is fully compliant with the Health Insurance Portability and Accountability Act of 1996.

Age, sex, and dates of medical benefit eligibility and claims availability are available for all patients. Data available for each outpatient pharmacy claim ("prescription") include the drug dispensed (in National Drug Code format), the dispensing date, and the quantity and number of therapy-days dispensed. Data available for each facility or professional service claim ("medical claim") include dates of service and 

) diagnosis codes. Facility claims (for selected plans) also include 

procedure codes and discharge disposition. Professional service claims also include procedure codes (in 

and Healthcare Common Procedure Coding System formats). All claims include charged and paid (ie, reimbursed) amounts, including patient deductibles, copayments, and/or coinsurance. For employed enrollees ("primary beneficiaries"), the database also includes information on short-term disability (eg, case days, total payments), as well as dates and times of work absence and reasons thereof (eg, vacation, disability, sick). Using a unique encrypted identifier, all patient-level data can be arrayed chronologically to provide a longitudinal profile of all medical and pharmacy services used by each plan member. The database for this study encompassed the period January 1, 2003, to December 31, 2005.

The study sample consisted of all persons with 2 or more paid medical claims with a diagnosis of FM (

diagnosis code 729.1) at least 90 days apart between January 1, 2003, and December 31, 2005, at least 1 of which had to be rendered by a rheumatologist. The date of the first such claim was designated the "index date." We excluded all patients: (1) aged less than 18 years as of their index date; (2) with less than 12 months of complete data preceding and following their index date ("prediagnosis" and "postdiagnosis," respectively); (3) enrolled in a Medicaid program; or (4) aged 65 years or more if enrolled in a Medicare supplemental plan (due to possibly incomplete claims histories). All pharmacy, professional service, and facility claims were then compiled for all remaining subjects over 24 months (ie, 12 months preceding their index date to 12 months following this date).

We examined the prevalence of selected comorbidities during the prediagnosis period. Patients were deemed to have any of the conditions of interest if, during the 12-month prediagnosis period, they had 2 or more outpatient claims on different days, or any hospitalizations, with a corresponding diagnosis code or prescription (eg, diabetes was defined based on presence of either 

diagnosis codes specific to diabetes or prescriptions for diabetes medications).

We examined the use of both pain-related and non-pain-related medications during the pre- and postdiagnosis periods. Medications were designated as "pain-related" based on their classification as analgesics or adjuvant medications in the World Health Organization's "analgesic ladder." Although this algorithm was developed initially for the treatment of cancer pain,

 published literature has supported the use of these medications for the treatment of neuropathic pain in patients with and without cancer.

Pain-related medications accordingly were defined to consist of the following: (1) antiepileptic drugs; (2) benzodiazepines; (3) corticosteroids; (4) cyclooxygenase (COX)-2 inhibitors and other prescription nonsteroidal anti-inflammatory drugs (NSAIDs); (5) muscle relaxants; (6) sedatives/hypnotics; (7) opioids (both shortand long-acting); (8) antidepressants (including tricyclic antidepressants, monoamine oxidase inhibitors, selective serotonin reuptake inhibitors) and other selected agents (eg, trazodone, bupropion, venlafaxine, duloxetine, buspirone); (9) antimigraine agents; and (10) miscellaneous agents (injectable agents [eg, bupivacaine], topical analgesics [eg, lidocaine]). All other medications were designated as non-pain-related and were stratified into various categories (eg, angiotensin-converting enzyme inhibitors, antibiotics). Use was examined in terms of the percentage of patients receiving each type of medication.

Use of healthcare services during the pre- and postdiagnosis periods was examined in terms of the following: (1) physician office visits, (2) other outpatient office visits, (3) emergency department (ED) visits, and (4) hospitalizations. Use was examined in terms of the percentage of patients receiving each service, as well as the number of times each service was received; for hospitalizations, we also examined length of stay (LOS).

Healthcare costs were tallied in terms of (1) pharmacotherapy, (2) physician office visits, (3) other outpatient office visits, (4) E D visits, (5) inpatient care, and (6) all other services. Reimbursed amounts (including patient liability, such as copays) were used in all analyses of healthcare costs. No attempt was made to attribute care specifically to the treatment of FM, given inherent difficulties of attribution in electronic healthcare claims databases.

Absenteeism and short-term disability were examined for all study subjects designated in the database as "primary beneficiaries" (productivity information is limited to these persons). For absenteeism, we tallied the number of hours and days absent from work; for short-term disability, we ascertained the total number of days absent from work and total disability payments. We also estimated the number of primary beneficiaries with at least 1 instance of absenteeism and short-term disability, respectively. Total payments for short-term disability also were estimated.

The statistical significance of differences in the measures of interest between the pre- and postdiagnosis periods was ascertained using either paired 

tests or Wilcoxon signed-rank tests (for continuous measures), or McNemar's and Bowker's tests for categorical variables, as appropriate. All analyses were conducted using PC-SAS Proprietary Software, Release 8.4 (SAS Institute Inc, Cary, NC).

There were 120,685 patients in the study database with 2 or more medical encounters with a diagnosis of FM at least 90 days apart. After applying all inclusion/exclusion criteria, the total number of study subjects was 1803; the most common reason for exclusion was no visit to a rheumatologist resulting in a diagnosis of FM.

The mean (SD) age of the study subjects was 50.4 (9.4) years; 90.5% were women (

). Numerous comorbidities and encounters for various physical complaints were noted during the prediagnosis period, including arthritis (21.2% of study subjects), back pain (20.2%), hypertension (17.5%), respiratory infections (17.4%), painful neuropathic disorders (15.8%), hyperlipidemia (14.1%), and cervical pain (13.8%); 11.4% had depression, 4.1% had sleep disorders, and 3.5% had anxiety. Fifty-two percent of patients had diagnoses categorized as "symptoms, signs, and ill-defined conditions," including abdominal pain (12.6%), chest pain (9.5%), and fatigue (9.3%). In addition, 46.3% of study subjects had diagnoses of "other body/joint pain," including "pain in soft tissues of limb" (

diagnosis code 729.5) (16.7%), "pain in joint, site unspecified" (719.40) (13.3%), "pain in joint, multiple sites" (719.49) (11.9%), "pain in joint, lower leg" (719.46) (9.3%), "pain in joint, pelvic region, and thigh" (719.45) (7.9%), and "pain in joint, shoulder region" (719.41) (6.6%) (data not presented).

Seventy-nine percent of patients received pain-related medications during the prediagnosis period (

)-most commonly, NSAIDs (including COX-2 inhibitors) (54.1% of study subjects), opioids (predominantly short-acting agents) (51.3%), and antidepressants (53.0%); 24.1% received benzodiazepines, and 22.6% received antiepileptics. Receipt of a number of pain-related medications was somewhat higher following diagnosis of FM (82.8% vs 78.6% during the prediagnosis period; 

<.01)-especially, opioids (predominantly short-acting agents) (55.9% vs 51.3%), antiepileptics (28.6% vs 22.6%), antidepressants (61.7% vs 53.0%), muscle relaxants (46.3% vs 37.0%), and sedatives/hypnotics (22.3% vs 18.9%) (all 

<.01). Use of NSAIDs, benzodiazepines, and other agents was substantially unchanged, as was that of non-pain-related medications (

Relative to prediagnosis, FM patients averaged 3.9 more office visits during the postdiagnosis period (mean [SD]: 23.6 [17.0] vs 19.7 [15.9]; 

). The percentage of patients with 1 or more office visits increased during the 12 months preceding diagnosis, from 58.2% in the first month to 68.9% during the last; it decreased slightly following diagnosis, from 76.6% during the first month to 71.2% in the last. While the mean number of ED visits was unchanged between the 2 periods (1.3 [1.8] prediagnosis vs 1.5 [2.7] postdiagnosis; 

= .30), the percentage of patients with 1 or more ED visits was highest around the time of initial FM diagnosis. A similar phenomenon was noted with respect to hospitalization.

Patterns of Healthcare Utilization and Cost in Patients With Newly Diagnosed Fibromyalgia

: To compare healthcare utilization and costs in patients with end-stage renal disease (ESRD) beginning peritoneal dialysis (PD) or hemodialysis (HD). 

: Using a US health insurance database, we identified all patients with ESRD who began dialysis between January 1, 2004, and December 31, 2006. Patients were designated as PD patients or as HD patients based on first-noted treatment. Patients with less than 6 months of pretreatment data and those with less than 12 months of data following initiation of dialysis (“pretreatment” and “follow-up,” respectively) were dropped from the study sample. The PD patients were matched to HD patients using propensity scoring to control for differences in pretreatment characteristics. Healthcare utilization and costs were then compared over 12 months between propensity-matched PD patients and HD patients using paired 

 tests and Wilcoxon signed rank tests for continuous variables and using Bowker and McNemar tests for categorical variables, as appropriate.

: A total of 463 patients met all study entrance criteria; 56 (12%) began treatment with PD, and 407 (88%) began treatment with HD. Fifty PD patients could be propensity matched to an equal number of HD patients. The HD patients were more than twice as likely as matched PD patients to be hospitalized over the subsequent 12 months (hazard ratio, 2.17; 95% confidence interval, 1.34-3.51; 

 <.01). Their median healthcare costs over the 12-month follow-up period were $43,510 higher ($173,507 vs $129,997 for PD patients, 

: Among patients with ESRD, PD patients are less likely than HD patients to be hospitalized in the year following initiation of dialysis. They also have significantly lower total healthcare costs.

Using a large US healthcare claims database, we compared healthcare utilization and costs in patients beginning renal replacement therapy with peritoneal dialysis (PD) versus hemodialysis (HD), matching PD patients to HD patients based on pretreatment characteristics.

The PD patients were significantly less likely to be hospitalized during the year following initiation of dialysis.

The PD patients also had significantly lower total healthcare costs over this period.

Hemodialysis (HD) and peritoneal dialysis (PD) are the main dialysis modalities for patients with end-stage renal disease (ESRD). Hemodialysis is typically performed 3 times weekly at a dialysis center, with each treatment taking 3 to 5 hours

; nocturnal HD and short daily home HD are also available.

 In contrast, PD uses the lining of the abdomen (the peritoneal membrane) instead of a dialyzer to filter the blood. The abdomen is filled with dialysis solution (a combination of minerals and sugar designed to draw wastes and excess fluids from the body into the solution) and is then drained several hours later (a process known as “exchange”). There are 3 different types of PD: continuous ambulatory PD (CAPD), automated PD (APD), and combination CAPD and APD.

 In CAPD, patients undergo the exchange process usually 4 to 5 times during a 24-hour period; no machine is required. In APD, the patient uses an automated cycler to perform 3 to 5 exchanges during the night while sleeping (the abdomen remains filled with dialysis solution throughout the day).

In the United States, the cost of dialysis is largely borne by the Medicare ESRD system, which accepts all patients previously enrolled in Medicare on initiation of dialysis (principally, persons ≥65 years) and those otherwise not eligible for Medicare benefits after they have received a minimum of 3 months of dialysis (for these latter patients, there is an additional 30-month “coordination of benefits” period during which Medicare remains the secondary payer, while the private insurer is the primary payer).

 Persons 65 years or older who are still working or who have a spouse who is still working also may have their costs borne (in part or in full) by private health insurers. It has been estimated that 25% of all patients with ESRD beginning HD and 37% of all such patients beginning PD are privately insured.

There is a wealth of information about healthcare utilization and costs among patients with ESRD who are insured through the Medicare program. Comparatively little is known about the use and cost of healthcare services among patients with ESRD who are privately insured and, in particular,  those beginning treatment with PD versus HD.

Data were obtained from the Phar-Metrics Patient-Centric Database, which is composed of facility, profes-sional service, and retail (ie, outpatient) pharmacy claims from more than 85 US health plans (PharMetrics, Watertown, MA). The plans provide healthcare coverage to approximately 14 million persons annually throughout the United States (35% in the Midwest, 21% in the Northeast, 31% in the South, and 13% in the West). All patient identifiers in the database are fully encrypted, and the database is fully compliant with the Health Insurance Portability and Accountability Act of 1996.

Information available for each facility and professional service claim includes the date and place of service, diagnoses (in 

International Classification of Diseases, Ninth Revision, Clinical Modification

 [selected plans only] and Health Care Financing Administration Common Procedural Coding System formats), provider specialty, and charged and paid amounts. Data available for each retail pharmacy claim include the drug dispensed (in National Drug Code format), dispensing date, and quantity dispensed and number of days of therapy supplied (selected plans only). All claims include a charged amount; the database also provides the paid amounts (ie, total reimbursed, including patient deductible, copayment, and coinsurance).

Selected demographic and eligibility information is also available, including age, sex, geographic region, coverage type, and the dates of insurance coverage. All patient-level data can be arrayed chronologically to provide a detailed longitudinal profile of all medical and pharmacy services used by each insured person. Because this study was retrospective in nature, used completely anonymized data, and did not involve patient contact, institutional review board approval was neither required nor sought.

Using the PharMetrics database, we identified all patients with 1 or more medical encounters for PD or HD between January 1, 2004, and December 31, 2006 (“study period”), irrespective of whether they had any claims with a diagnosis of renal failure (

 diagnosis codes 403.X1, 404.X2, 404.X3, 585, 585.X, and 586) (additional criteria, listed herein, were used to exclude patients receiving dialysis for reasons other than ESRD). For each such patient, we then identified the first-noted claim for dialysis (either PD or HD) during the study period; the date of this claim was designated the “index date,” and patients were stratified into 2 groups (“PD patients” or “HD patients”) based on the treatment received on this date. All patients were required to be continuously enrolled for 6 months before their index date (“pretreatment”) and for 12 months after this

date (“follow-up”). Identification of patients who received PD versus HD was based on algorithms developed by us (

). Patients with claims for both PD and HD on their index date were excluded, as were patients enrolled in a Medicaid program and those 65 years or older who were enrolled in Medicare supplemental or capitated plans (because of incomplete claims histories). Additional exclusion criteria were taken from prior studies

 that identified patients receiving dialysis based on electronic claims databases and included the following: (1) any claims encounters with dialysis-related codes (ie, diagnostic, procedural, or equipment) during the pretreatment period, (2) less than 3 months of continuous enrollment following the index date, (3) evidence of initiation of dialysis for reasons other than ESRD (eg, because of trauma), and (4) patients who underwent renal transplantation during the first month of follow-up.

Patients with end-stage renal disease who began peritoneal dialysis had lower 1-year hospitalization rates and lower total healthcare costs than those who began therapy with hemodialysis.

Gerry Oster, PhD

Articles

Is All "Skin in the Game" Fair Game? The Problem With "Non-Preferred" Generics

Benefits and Risks of Live Attenuated Influenza Vaccine in Young Children

Patterns of Healthcare Utilization and Cost in Patients With Newly Diagnosed Fibromyalgia

Cost Comparison of Peritoneal Dialysis Versus Hemodialysis in End-Stage Renal Disease

Childhood Vaccination Against Pneumococcal Otitis Media and Pneumonia: An Analysis of Benefits and Costs

The Clinical and Economic Burden of Obesity in a Managed Care Setting

Patterns of Antidepressant Use and Their Relation to Costs of Care