A, The mean out-of-pocket costs per quarter (Q) are shown with associated 95% CIs. The dotted line reflects the date when the Food and Drug Administration mandated that all chlorofluorocarbon (CFC) albuterol inhalers must be withdrawn from the US market. However, phaseout of CFC inhalers occurred gradually in the 2 years prior, as shown in B. B, The percentages of patients filling an albuterol inhaler prescription per quarter are shown with associated 95% CIs. HFA indicates hydrofluoroalkane.
Each point reflects a drug plan that was present in 2004 to 2006 and 2009 to 2010 (32 plans total). The horizontal axis shows each plan’s mean out-of-pocket cost for a 30-day equivalent albuterol inhaler prescription from 2004 to 2006 computed across all individuals in a plan who filled an albuterol prescription during those years. The vertical axis shows the dollar change in the mean out-of-pocket albuterol inhaler cost for a given plan between 2004 to 2006 and 2009 to 2010.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Jena AB, Ho O, Goldman DP, Karaca-Mandic P. The Impact of the US Food and Drug Administration Chlorofluorocarbon Ban on Out-of-pocket Costs and Use of Albuterol Inhalers Among Individuals With Asthma. JAMA Intern Med. 2015;175(7):1171–1179. doi:10.1001/jamainternmed.2015.1665
The US Clean Air Act prohibits use of nonessential ozone-depleting substances. In 2005, the US Food and Drug Administration announced the ban of chlorofluorocarbon (CFC) albuterol inhalers by December 31, 2008. The policy resulted in the controversial replacement of generic CFC inhalers by more expensive, branded hydrofluoroalkane inhalers. The policy’s impact on out-of-pocket costs and utilization of albuterol is unknown.
To study the impact of the US Food and Drug Administration’s CFC ban on out-of-pocket costs and utilization of albuterol inhalers.
Design, Setting, and Participants
Using private insurance data from January 1, 2004, to December 31, 2010, we investigated the effect of the CFC ban on out-of-pocket costs and utilization of albuterol inhalers among individuals with asthma (109 428 adults; 37 281 children), as well as asthma-related hospitalizations, emergency department visits, and outpatient visits. We estimated multivariable models adjusted for age, sex, comorbidities, and mean out-of-pocket costs of albuterol inhalers in an individual’s drug plan. We analyzed whether effects varied between adults vs children and those with persistent vs nonpersistent asthma.
Main Outcomes and Measures
Pharmacy claims for albuterol inhalers, as well as asthma-related hospitalizations, emergency department visits, and outpatient visits.
The mean out-of-pocket albuterol cost rose from $13.60 (95% CI, $13.40-$13.70) per prescription in 2004 to $25.00 (95% CI, $24.80-$25.20) immediately after the 2008 ban. By the end of 2010, costs had lowered to $21.00 (95% CI, $20.80-$21.20) per prescription. Overall albuterol inhaler use steadily declined from 2004 to 2010. Steep declines in use of generic CFC inhalers occurred after the fourth quarter of 2006 and were almost fully offset by increases in use of hydrofluoroalkane inhalers. In multivariable analyses, a $10 increase in out-of-pocket albuterol prescription costs was estimated to lower utilization by 0.92 percentage points (95% CI, −1.39 to −0.44; P < .001) for adults and 0.54 percentage points (95% CI, −0.84 to −0.24; P = .001) for children, with no difference between adults vs children and patients with persistent vs nonpersistent asthma and with no impact on asthma-related hospitalizations, emergency department visits, and outpatient visits.
Conclusions and Relevance
The Federal ban of CFC inhalers led to large relative increases in out-of-pocket albuterol costs among privately insured individuals with asthma and modest declines in utilization. The policy’s impact on individuals without insurance, who faced greater cost increases, is unknown.
In March 2005, the US Food and Drug Administration (FDA) announced that the production and sale of all chlorofluorocarbon (CFC)-propelled albuterol inhalers would be banned effective December 31, 2008, in the United States.1 This action stemmed from the 1987 Montreal protocol, in which 26 nations assembled and called for a worldwide reduction in use of ozone-depleting substances, such as CFCs, as well as subsequent amendments to the US Clean Air Act that banned use of nonessential products containing CFCs.1,2 Before 2005, the FDA had recommended to the US Environmental Protection Agency that CFC albuterol inhalers should be considered essential medicines and therefore exempt from the Clean Air Act. However, the FDA reversed this policy in 2005 with the growing availability of 2 propelled albuterol inhalers that did not cause ozone depletion.2 The result of this policy was the gradual elimination of generic CFC albuterol inhalers from the US market in the 2 years preceding December 31, 2008, and the availability of only branded hydrofluoroalkane (HFA) albuterol inhalers thereafter.
Although intended to reduce the environmental footprint of CFCs, the FDA ban sparked immediate controversy over how use of albuterol inhalers in patients with asthma and chronic obstructive pulmonary disease (used by these populations for immediate rescue therapy in disease exacerbations) would be impacted by the increase in out-of-pocket expenditures for albuterol inhalers resulting from the elimination of generic formulations.2-5 Initial forecasts estimated that the transition to branded HFA albuterol inhalers could cost individuals up to $30 more per prescription and cost the United States up to $1.2 billion annually.1,6
The impact of the FDA’s CFC albuterol inhaler ban on out-of-pocket costs and utilization of albuterol inhalers has been widely discussed but received little empirical attention.2-5 The effect of this policy is important to consider not only because greater patient cost-sharing has been associated with lower use of health care services,7-13 including patients with asthma,13-15 but also because 4 HFA albuterol inhalers currently exist on the market today, none of which have generic formulations.
Using longitudinal private insurance data from January 1, 2004, to December 31, 2010, we investigated the impact of the FDA ban of CFC albuterol inhalers, which eliminated generic albuterol inhalers, on adults and children with asthma. We estimated the effect of the ban on out-of-pocket albuterol inhaler costs and utilization, medical service utilization (inpatient, emergency department [ED], and outpatient use), and use of other asthma medications, specifically asthma control therapies. We analyzed effects of the ban separately for adults vs children with asthma given evidence that greater cost-sharing may impact adults and children with chronic disease differently.11,13,16-18 Finally, we analyzed whether the effect of the ban on out-of-pocket albuterol costs and albuterol inhaler utilization varied according to the severity of an individual’s asthma.
We assembled data on pharmacy and medical claims from 2004 to 2010 from 77 private health plans offered by 41 geographically diverse, large US employers to their current or retired employees and their dependents. Although these plans were not nationally representative of all US commercial health plans, the geographic distribution of beneficiaries was similar to that of the US population across 4 major census regions. The sample comprised 38.2% from the South, 22.4% from the Midwest, 23.8% from the Northeast, and 16.1% from the West, while the US population in 2010 comprised 37.4% from the South, 22.3% from the Midwest, 18.2% from the Northeast, and 22.7% from the West. Data from these plans have been used in studies11-13,19-21 to analyze the association of pharmacy benefit design with medication use and use of other health care services.
We identified adults and children with asthma using a baseline presence of 2 or more outpatient or inpatient medical claims with an International Classification of Diseases, Ninth Revision, diagnosis code for asthma (code 493.XX).13 Adults were defined as being 18 years or older, and children were defined as being 5 to 17 years old. We excluded individuals with less than 12 months’ follow-up. The data were deidentified and exempt from ethics review at the University of Southern California.
We used national drug codes to identify pharmacy claims for albuterol metered-dose inhalers. Mail-order and 90-day prescriptions were included. We categorized albuterol inhalers as being a generic CFC inhaler, branded CFC inhaler, or branded HFA inhaler. We also identified pharmacy claims for asthma control medications, which included inhaled corticosteroids, long-acting beta agonists, leukotriene receptor antagonists, combined corticosteroid–long-acting beta agonist formulations, methylxanthines, cromolyn sodium, and immunomodulators.
Asthma-related hospitalizations, ED visits, and outpatient visits were identified from medical claims. The primary International Classification of Diseases, Ninth Revision, diagnosis code for asthma (code 493.XX) was used.
Our data did not include formulary information on the specific copayments charged by health plans for various asthma medications. Based on prior studies,11-13,19-21 we therefore computed the mean out-of-pocket cost for albuterol inhalers for each health plan in a given calendar quarter (ie, at the plan quarter level). Although plan benefit designs (eg, branded vs generic copayments and tiered formularies) typically vary by year, generic CFC albuterol inhalers were gradually phased out by US health plans during 2007-2008, raising out-of-pocket costs in each plan midway through the plan year as only branded formulations became available in each plan. We computed the mean out-of-pocket albuterol inhaler costs at the plan quarter level by summing the total out-of-pocket expenditures for albuterol inhalers in a given plan quarter (including generic and branded CFC inhalers and branded HFA inhalers) and dividing by the number of 30-day equivalent prescriptions. The mean out-of-pocket costs for asthma control medications were computed similarly at the plan quarter level.
Our primary outcome measure was an indicator for whether an individual filled a prescription for an albuterol inhaler in a given quarter. We also considered as an outcome whether an individual filled a prescription for an asthma control medication in a given quarter. Increases in cost-sharing for albuterol inhalers could drive patients to substitute toward control therapies to better manage symptoms and reduce the demand for albuterol inhalers. Alternatively, increases in albuterol inhaler costs could lead patients to reduce use of control therapies if patients prioritize asthma medication spending on albuterol. In addition to these drug utilization measures, we also analyzed the impact of the FDA’s CFC ban on asthma-related hospitalizations, ED visits, and outpatient visits both separately and combined.
We began with a descriptive analysis of trends in the mean out-of-pocket albuterol inhaler costs and overall utilization by quarter. Within each quarter, we averaged across plans the mean out-of-pocket cost for albuterol inhalers weighted by the number of individuals with asthma in each plan. Before 2008, the mean out-of-pocket costs included costs of branded or generic CFCs and branded HFAs, and only branded HFAs were included after 2008. Within each quarter, we next computed the percentage of individuals with asthma in a given quarter who filled a prescription for an albuterol inhaler, averaged across all plans.
To estimate the association between changes in out-of-pocket albuterol inhaler costs and albuterol inhaler utilization, we next estimated a multivariable linear probability model in which the dependent variable was a binary variable for whether an individual filled a prescription for an albuterol inhaler in a given quarter. The primary independent variable was the mean out-of-pocket cost for albuterol inhalers in an individual’s plan in a given quarter (continuous variable). We estimated 2 models. The first model adjusted for individual characteristics, including the following: age, sex, number of quarters in the sample, quarter of year indicators to account for seasonal differences in utilization, zip code–level socioeconomic information from the 2010 census (percentage population that was rural, percentage of nonwhite race/ethnicity, and median family income), and annual indicators for several comorbid conditions (hypertension, congestive heart failure, diabetes mellitus, hyperlipidemia, coronary artery disease, peripheral vascular disease, allergic rhinitis, depression, and anxiety).13 To account for the possibility that unobserved individual characteristics are correlated with both utilization and mean out-of-pocket albuterol costs (eg, by an individual’s choice of drug plan), we estimated a second model that adjusted for age, annual indicators for comorbid conditions, quarter of year indicators, and individual fixed effects. Individual fixed effects account for observed and unobserved time-invariant individual characteristics that may be associated with participation in a given drug plan and albuterol inhaler utilization. We estimated these models separately for adults and children. Models that included children did not include the full set of comorbid adult conditions and instead included only an indicator for allergic rhinitis.13 We estimated linear probability models (rather than logistic models) given known issues of model convergence when the number of fixed effects is large.22,23
We also examined whether the effect of out-of-pocket albuterol inhaler costs on utilization varied according to an individual’s level of asthma severity. Although we did not have data on symptoms or pulmonary function tests that would allow us to clinically categorize individuals as having intermittent vs persistent asthma,24 we were able to define individuals as having persistent asthma based on Healthcare Effectiveness Data and Information Set criteria. These criteria identify individuals with persistent asthma as those having 4 or more prescriptions filled for asthma medications, 1 or more acute inpatient discharges or ED visits with a primary diagnosis of asthma, or 4 or more outpatient visits with any diagnosis of asthma plus 2 or more prescriptions filled for asthma medications.25
Because higher out-of-pocket albuterol costs resulting from the FDA’s CFC ban could impact utilization of asthma control medications as well, we also estimated the association between out-of-pocket albuterol inhaler costs and asthma control medication use. Specifically, we estimated multivariable linear probability models in which the dependent variable was a binary variable of whether an individual filled a prescription for an asthma control medication in a given quarter, and the primary independent variable was the mean out-of-pocket cost for albuterol inhalers in an individual’s plan in a given quarter. In addition to similar adjustments used in the albuterol inhaler utilization models, we also adjusted for the mean out-of-pocket costs for asthma control medications at the plan quarter level.
Finally, we analyzed whether increases in albuterol costs resulting from the CFC ban were associated with increases in asthma-related hospitalizations, ED visits, and outpatient visits. For each outcome (and a composite measure of use of any of these services), we estimated multivariable linear probability models in which the dependent variable was a binary outcome for whether a given service was utilized in a quarter and the primary independent variable was the mean out-of-pocket albuterol inhaler cost in an individual’s plan in a given quarter. Models were estimated with and without individual fixed effects.
To make estimated coefficients on out-of-pocket albuterol cost more interpretable in each model, we reported effects of a simulated $10 increase in out-of-pocket costs on the various outcomes studied. Statistical software (STATA, version 13.1; StataCorp LP) was used for analyses. The 95% CI around reported means reflects 0.025 in each tail or P ≤ .05. Model standard errors were clustered at the plan level.
Our sample included 109 428 adults (1 339 002 person-quarters) and 37 281 children (457 438 person-quarters). Persistent asthma was common among adults (19 850 [18.1%]) and children (6037 [16.2%]) (Table 1). Consistent with epidemiological data on the opposite sex predominance of asthma among adults and children,26 most children with asthma were male (21 635 [58.0%]), whereas males were a minority among adults (38 323 [35.0%]). Most adults and children filled at least 1 prescription for an albuterol inhaler (69 673 adults [63.7%] and 25 917 children [69.5%]) or for asthma control therapy (63 450 adults [58.0%] and 20 178 children [54.1%]). Among adults, 10 405 (9.5%) were hospitalized for asthma at least once during the study period, and 7579 (6.9%) had at least 1 asthma-related ED visit. Among children, 1429 (3.8%) were hospitalized for asthma at least once, and 3147 (8.4%) had at least 1 asthma-related ED visit.
The mean out-of-pocket costs for albuterol inhalers were stable at $13.60 (95% CI, $13.40-$13.70) per prescription from the first quarter of 2004 to the fourth quarter of 2006 and rose during the first quarter of 2007 to $19.90 (95% CI, $19.70-$20.10) as CFC inhalers began to be phased out. The mean out-of-pocket costs reached their high of $25.00 (95% CI, $24.80-$25.20) in the first quarter of 2009 after the ban went into effect and declined to $21.00 (95% CI, $20.80-$21.20) by the fourth quarter of 2010 (Figure 1A). The mean out-of-pocket costs fell slightly across quarters within each year, reflecting a larger proportion of individuals meeting out-of-pocket deductibles throughout the year.
Utilization of generic CFC albuterol inhalers began to decline well before the December 31, 2008, FDA-mandated date of removal of CFC inhalers from the US market (Figure 1B). For example, utilization of CFC inhalers declined slightly from the first quarter of 2004 to the fourth quarter of 2006 and sharply declined after the first quarter of 2007, with 14.4% of individuals filling prescriptions for a generic CFC albuterol inhaler in the fourth quarter of 2006 compared with 10.2% in the first quarter of 2007 and zero prescriptions after the fourth quarter of 2008. Utilization of branded HFA albuterol inhalers increased concurrently with the reduction in generic CFC inhalers. Overall, total albuterol inhaler utilization exhibited a slight decline from the first quarter of 2004 to the fourth quarter of 2006, but there was no further decline in that trend after the first quarter of 2007, suggesting that FDA-mandated reductions in CFC inhalers were offset by increases in use of HFA inhalers.
There was significant variation across plans in out-of-pocket albuterol inhaler costs before the FDA’s CFC ban and after, suggestive of differences in formulary placement of CFC and HFA inhalers across plans (Figure 2). Plans varied considerably in the extent to which the mean out-of-pocket albuterol inhaler costs rose after phaseout of generic CFC inhalers by the fourth quarter of 2008. For example, the mean out-of-pocket albuterol inhaler costs in some plans increased by more than $20 between 2004 to 2006 and 2009 to 2010. Almost half of all plans had increases in the mean out-of-pocket albuterol inhaler costs exceeding $10.
Table 2 (top) lists the estimated effect of a $10 increase in out-of-pocket albuterol inhaler costs on albuterol inhaler utilization (approximately equal to the mean increase in out-of-pocket albuterol inhaler costs between the fourth quarter of 2006 and the first quarter of 2009) according to asthma severity and for statistical models that adjusted for individual patient characteristics or included individual fixed effects. Overall, in the first statistical model, a $10 increase in out-of-pocket albuterol inhaler cost was estimated to lower utilization by 0.92 percentage points (95% CI, −1.39 to −0.44 percentage points; P < .001) among adults and by 0.54 percentage points (95% CI, −0.84 to −0.24 percentage points; P = .001) among children. Given observed quarterly albuterol inhaler utilization rates among adults and children of approximately 20% in the year before the CFC ban, this finding implies relative declines in albuterol inhaler utilization of approximately 5% and 3%, respectively. Similar effect sizes were noted in models with individual fixed effects. Effect sizes were similar in magnitude for patients with persistent vs nonpersistent asthma.
Table 2 (bottom) lists the estimated effect of a $10 increase in out-of-pocket albuterol inhaler costs on utilization of asthma control medications according to asthma severity. Among adults and children, out-of-pocket albuterol inhaler costs were not statistically significantly associated with utilization of asthma control medications overall or among patients with persistent vs nonpersistent asthma.
Increased out-of-pocket albuterol inhaler costs were not statistically significantly associated with changes in asthma-related hospitalizations, ED visits, outpatient visits, or a composite measure of these services. These results are summarized in Table 3.
In 2005, the FDA announced that all CFC-propelled albuterol inhalers would be banned from the US market by December 31, 2008, in accord with amendments to the Clean Air Act. The policy created immediate concerns that the elimination of generic albuterol inhalers from the market and the availability of only branded HFA inhalers would not only increase out-of-pocket costs to patients but also reduce utilization of these important treatments.2-5 Despite these concerns and the fact that generic albuterol inhalers still do not exist in the United States today, no empirical evidence exists on the impact of this important FDA policy on albuterol out-of-pocket costs and utilization to date.
We found that the FDA ban led to a gradual phaseout of CFC albuterol inhalers from commercial health plans from 2006 onward, culminating with their elimination by the fourth quarter of 2008. During this time, the mean out-of-pocket costs for albuterol inhalers rose from $14 per prescription, where they had been priced from 2004 to 2006, to almost $25 immediately after the ban. By 2010, out-of-pocket costs had fallen to $21 per prescription. Plans varied considerably not only in their out-of-pocket albuterol costs before the CFC ban but also in the magnitude to which out-of-pocket albuterol costs increased as generics were phased out. Overall, albuterol inhaler utilization fell slightly as a result of these price increases without any differential effect for patients with persistent vs nonpersistent asthma. Asthma control medication utilization and asthma-related hospitalizations, ED visits, and outpatient visits were also not affected. Therefore, the main impact of this policy on individuals with asthma who were privately insured was a large relative increase in out-of-pocket burden without substantial decreases in utilization or increases in asthma-related medical services.
Stratospheric ozone reductions have been implicated in higher rates of melanoma and nonmelanoma skin cancers, photokeratitis, and cataracts.27-29 Although the intended effect of the FDA’s CFC ban was to reduce these health risks by eliminating CFC-based products, our findings suggest that the policy also led to a modest decline in albuterol inhaler utilization among the privately insured and to large relative increases in out-of-pocket costs. Larger declines in use may also have occurred among the uninsured, for whom we did not have data. Although a full evaluation of the FDA policy would balance these 2 effects, our findings suggest that ensuring affordable access to HFA inhalers through public subsidies or price regulation may have been rational policy considerations.
Our study closely relates to previous investigations demonstrating that greater patient cost sharing is associated with lower use of health care services,7-13 including in patients with asthma.13-15 Several of these studies use variation in out-of-pocket costs across health plans to identify the effect of out-of-pocket costs on utilization, and fewer studies rely on changes over time in out-of-pocket costs within plans. An advantage of our study compared with prior studies is that it evaluates the impact of out-of-pocket medication costs on utilization by relying on a plausibly exogenous natural experiment in which out-of-pocket albuterol costs rose as a result of FDA policy, rather than other unobserved factors that may lead health plans to change the generosity of their drug coverage.
We estimated that a $10 increase in out-of-pocket albuterol inhaler costs (approximately equal to the mean increase in out-of-pocket albuterol inhaler costs between the fourth quarter of 2006 and the December 31, 2008, FDA ban) lowered utilization by approximately 5%. Although others have found larger declines in drug utilization with similar increases in out-of-pocket costs,12 this result may occur because other studies analyze an entire range of drugs rather than a single drug. Moreover, patients may be reluctant to reduce use of medications that quickly relieve symptoms (eg, albuterol) despite large out-of-pocket cost increases.13
Our study had several limitations. First, we analyzed the impact of the FDA’s CFC albuterol ban on out-of-pocket costs of individuals with private insurance. Although several health plans in our data had large increases in out-of-pocket albuterol costs after the CFC ban, presumably reflective of placement of HFA inhalers on higher formulary tiers, declines in albuterol inhaler utilization may have been greater for those without insurance who faced substantially higher prices for albuterol inhalers as only branded HFA inhalers became available. Prior research has demonstrated high rates of medication nonadherence among the uninsured,30-33 suggesting that albuterol inhaler utilization among uninsured patients with asthma may have been more adversely affected than observed in our privately insured population. Second, although we adjusted for zip code–level socioeconomic variables, such as the median household income, we did not have specific data on family income, which may further influence the impact of albuterol cost-sharing increases on utilization.15 Third, employers sometimes offer their employees multiple health plans from which to choose, which may confound our estimated associations between plan out-of-pocket albuterol costs and medication use if individuals select health plans based on unobserved characteristics that are correlated with utilization. However, selection bias across drug benefit plans should be limited because most employers in our study offered a single employer-specific drug plan regardless of the various medical plans they offered.13 Moreover, our findings were unaffected by the inclusion of individual fixed effects in our statistical model, which account for unobserved time-invariant individual characteristics that are correlated with drug plan choice and albuterol inhaler utilization. Fourth, while we found that increases in out-of-pocket albuterol costs were not associated with lower use of other asthma medications, it is possible that essential nonasthma medications were reduced. Fifth, changes in asthma morbidity in the overall United States occurred during the study period, which may have affected trends in overall albuterol inhaler utilization. Our analysis addressed this issue by studying effects on albuterol inhaler utilization of changes in out-of-pocket albuterol costs within plans, therefore accounting for national trends in asthma morbidity and albuterol inhaler utilization. Sixth, while large and geographically diverse, our data were not nationally representative.
In summary, we found that the FDA ban on albuterol CFC inhalers was associated with large relative increases in out-of-pocket inhaler costs and slight declines in albuterol inhaler utilization among privately insured individuals with asthma. The impact of the FDA policy on individuals without insurance who faced greater increases in out-of-pocket costs warrants further exploration.
Accepted for Publication: February 14, 2015.
Corresponding Author: Anupam B. Jena, MD, PhD, Department of Health Care Policy, Harvard Medical School, 180 Longwood Ave, Boston, MA 02115 (firstname.lastname@example.org).
Published Online: May 11, 2015. doi:10.1001/jamainternmed.2015.1665.
Author Contributions: Dr Jena had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Jena, Goldman, Karaca-Mandic.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Jena, Karaca-Mandic.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Jena, Ho, Karaca-Mandic.
Obtained funding: Jena, Goldman, Karaca-Mandic.
Administrative, technical, or material support: Karaca-Mandic.
Study supervision: Jena, Goldman, Karaca-Mandic.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was funded by Health Early Independence Award 1DP5OD017897-01 from the National Institutes of Health (Dr Jena), by grants 5P01AG033559 (Dr Goldman) and 5K01AG036740 (Dr. Karaca-Mandic) from the National Institute on Aging, and by a University of Minnesota Academic Health Center Faculty Development Grant (Dr Karaca-Mandic).
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.