The economic impact of switching from Synthroid for the treatment of hypothyroidism

Abstract

Aims: To compare hypothyroidism-related costs for patients who continuously used Synthroid and patients who switched from Synthroid to alternative therapies.

Materials and methods: Truven’s Health Analytics MarketScan Commercial Claims and Encounters database from January 1, 2007 to June 30, 2014 was queried for US adults diagnosed with hypothyroidism who initiated Synthroid and adhered to such therapy for at least 6 months. Propensity score matching matched continuous users of Synthroid to patients who switched from Synthroid to alternative levothyroxine agents. Kruskal-Wallis tests assessed differences between the matched cohorts in several categories of costs, including disease-related drug costs, non-drug medical costs, and total direct medical costs.

Results: There were 10,159 individuals included in the study, with 7,991 continuous users of Synthroid and 2,168 switchers. After matching (n = 2,052 for each cohort), continuous use of Synthroid was associated with significantly lower hypothyroidism-related non-drug medical costs ($595 vs $1,023; p = .003) and reduced hypothyroidism-related total medical costs ($757 vs $1,132; p = .010), despite being associated with significantly higher drug costs ($161 vs $109; p < .001). Hypothyroidism-related total medical costs rose as the number of switches of hypothyroidism treatment increased, with continuous users having significantly lower hypothyroidism-related total medical costs ($757) compared with patients who switched twice ($1,179; p = .001) or three or more times ($1,268; p = .004).

Limitations: The analyses focused on continuously insured patients who were adherent to Synthroid for at least 6 months and results may not be generalizable. The reliance on claims data does not allow for clinical examination of hypothyroidism or inclusion of some factors that may be associated with outcomes. The analyses assume that all prescriptions filled are taken as prescribed.

Conclusions: Results indicate that there are significant direct economic healthcare costs associated with switching from Synthroid to alternative levothyroxine therapies, and that these costs increase as patients switch therapies more frequently.

Keywords:

• Continuous use
• hypothyroidism
• switching
• Synthroid
• total medical costs

Introduction

Over the past 30 years, the generic drug market has increased substantially, from 18.6% of all prescriptions filled in the US in 1984¹ to 88% by 2014². The increasing use of generic medications has been associated with considerable total costs savings, which have been estimated at $254 billion in the US for 2014 and $1.68 trillion from 2005–2014². Research has also shown that there are potential large savings associated with switching from brand name to the lowest-priced generic equivalent in both developing countries³ and developed countries outside of the US^4,5. In Europe, where generics are generally more expensive than in the US⁶, increased use of generics has been attributed to both lowering of costs of generics as well as initiatives to increase prescribing of generic medications⁷. In the US, substitution by pharmacists, significantly lower insurance co-payments associated with the use of generic medications, and pre-authorization requirements for branded drugs when generics are available are all factors which have increased the use of generic medications⁸.

The impact of switching medications on patient outcomes is mixed. For instance, research has found that, on average, switching to generic is cost-effective and clinically equivalent, although some patients may be affected by reduced adherence, medication errors, or adverse events^9,10. In contrast, switching medications such as tumor necrosis factor inhibitors among patients with rheumatoid arthritis was associated with a higher disease activity score¹¹. Research on switching among antihypertensive drugs within the same class of medications argues that there may be differences in drug structure, therapeutic actions, drug interactions, and adverse actions that can potentially have detrimental clinical effects¹². Similarly, research has found that switching among statin medications for treatment of high cholesterol was associated with an increased risk of stroke or major cardiovascular events¹³. For patients with hypothyroidism, the American Association of Clinical Endocrinologists (AACE), the Endocrine Society (TES), and the American Thyroid Association (ATA) consider it best practice to maintain patients on the same formulation of thyroid medication throughout treatment^14,15. Reasons given for this recommendation include the uniqueness of specific tablet formulations, as well as the uncertainty regarding the bioequivalence among products¹⁵.

Given the complexities associated with switching among levothyroxine (LT4) formulations and the lack of understanding regarding related economic outcomes resulting from switching medications, the present study explored the association between such switches and direct medical costs. This retrospective investigation used records from a large US commercial insurance database to examine differences in disease-related costs between those who continued therapy with Synthroid and those who switched from Synthroid to alternative LT4 formulations. The analysis also examined the association between disease-related medical costs and the number of switches of therapy per patient.

Materials and methods

Truven’s Health Analytics MarketScan Commercial Claims and Encounters (CCAE) database between January 1, 2007 and June 30, 2014 was used for this study. The CCAE database consists of the healthcare records of over 50 million individuals covered by fully- or partially-capitated, fee-for-service health plans from ∼350 payers^16,17. As such, the database provides detailed costs, use, and outcomes data for healthcare services performed in both inpatient and outpatient settings. Medical claims are linked to outpatient prescription drug claims and person-level enrollment information. The data are fully de-identified and Health Insurance Portability and Accountability Act (HIPAA) compliant. Ethics approval was not required for this retrospective database of anonymized data.

This study examined the records of adults age 18 or older who received at least one diagnosis of hypothyroidism (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] of 244.0, 244,1, or 244.9) over the study period. To be included, patients were required to have initiated on Synthroid within 60 days of their initial diagnosis of hypothyroidism, and to be adherent to therapy for the initial 6 months following initiation (i.e. the lead-in period), with adherence defined as receipt of therapy for at least 80% of the lead-in period.

Patients were excluded from the study if they received any diagnosis of congenital hypothyroidism (ICD-9-CM of 243.xx), iodine hypothyroidism (ICD-9-CM of 244.2), iatrogenic hypothyroidism (ICD-9-CM of 244.3), or any other acquired hypothyroidism, including secondary hypothyroidism (ICD-9-CM of 244.8). Also excluded were those who filled prescriptions for multiple LT4 formulations on the same day or were insured in a health maintenance organization or a point-of-service with capitation health plan.

Patients were classified as either switchers or continuous users, based upon their drug treatment patterns following the initial 6-month lead-in time period after initiation on Synthroid. Individuals were categorized as switchers if they changed from Synthroid to another branded or generic LT4 formulation, with a treatment gap of no more than 45 days between the time Synthroid was discontinued and the other levothyroxine therapy was started. For switchers, the index date was the date on which a non-Synthroid LT4 prescription was first filled. In contrast, patients who did not switch from Synthroid were categorized as continuous users, and were assigned an index date 180 days after the initial Synthroid prescription fill date (i.e. first fill of Synthroid during the lead-in phase). Both switchers and continuous users were required to have continuous insurance coverage from 6 months before through 1 year after the index date. Finally, switchers were required to have at least 6 months of stable-dose levothyroxine therapy post-index date, and continuous users were required to have stable-dose continuous use of Synthroid in the 12 months post-index date. Figure 1 illustrates the study design.

Figure 1. Study design.

The primary outcomes of interest in this analysis were hypothyroidism-related healthcare costs. Hypothyroidism-related medical costs (total costs) were identified based upon the receipt of an ICD-9-CM diagnoses of 244.0, 244.1, or 244.9, or the receipt of any LT4 formulation. In addition, total costs were sub-categorized into drug and non-drug medical costs, including inpatient, outpatient, emergency services, and laboratory costs, with all costs adjusted to 2015 dollars using the medical care price index from the Bureau of Labor Statistics¹⁸. To assess the costs associated with switching, patients in the switching cohort were propensity score matched to continuous users using the 1:1 matching without replacement, where the propensity score is based upon patient age, sex, insurance plan type, and baseline comorbidities identified in the Deyo/Charlson Index^19,20. The matched patients in each cohort were followed for 1 year post-index date to assess disease-related healthcare costs. Differences between the continuous users and switchers were examined using Chi-square statistics for categorical variables, t-tests for normally distributed continuous variables, and Kruskal-Wallis tests for non-normally distributed continuous variables, such as costs. All analyses were conducted using SAS, version 9.4 (Cary, NC). A p-value of <.05 was considered statistically significant. Sensitivity analysis was conducted using a sub-set of patients diagnosed with acquired hypothyroidism to investigate whether the type of hypothyroidism may impact differences seen in the costs associated with treatment.

Results

Table 1 presents unadjusted descriptive statistics both before and after matching. Prior to matching, the analyses revealed several differences between the two cohorts. Specifically, the continuous users were more likely to reside in the Northeast (13.4% vs 10.2%) or Western (12.1% vs 11.3%) regions of the country, and were less likely to live in the North Central (26.9% vs 28.7%) or Southern (46.5% vs 49.4%) regions, compared to the switchers. Continuous users were also less likely to be insured via a preferred provider organization (75.8% vs 76.3%) or point of service insurance (11.1% vs 12.8%) as compared to switchers. Matching resulted in a successful match rate of 94.6%, with a final matched sample of 2,052 individuals who continuously used Synthroid in the 1 year post-period and 2,052 individuals who switched between Synthroid and an alternative LT4. As expected, after propensity score matching, the continuous users and switchers shared a number of common demographic and clinical characteristics. For instance, there was no difference in baseline comorbidities between the cohorts. In addition, after matching, both patient groups had a mean age of 48.9 years; and a female majority (81.3%), and they were more likely to reside in the South (48.8% or 49.7%) or Northeast (10.0 or 12.4%) regions of the US. In addition, most patients in both cohorts were insured via a preferred provider organization (79.1%) and were diagnosed with acquired hypothyroidism (84.8%) (ICD-9-CM of 244.9). Table 1 shows that, among the matched cohort, the majority of patients who switched therapies did so only once (51.6%), while 36.5% changed medications two times, and 11.9% switched three or more times during the 1-year follow-up period. Among the patients who switched only once, 94.9% switched to generic LT4, while, among patients who switched multiple times, 76.9% switched back to Synthroid. Figure 2 compares hypothyroidism-related costs in the post-period between the patient cohorts. As Figure 2 illustrates, hypothyroidism-related drug costs were significantly higher for patients who continuously used Synthroid ($161 vs $109; p < .001), while non-drug medical costs were significantly lower for this cohort ($595 vs $1,023; p = .003). Furthermore, total hypothyroidism-related medical costs (drug and non-drug) were found to be significantly lower for continuous users compared to switchers ($757 vs $1,132; p = .010). While Figure 2 compares costs between the cohorts, Figure 3 examines differences in total hypothyroidism-related costs based upon the number of switches. These results illustrate that total disease-related costs increased as patients switched more frequently. Furthermore, compared to patients who maintained therapy on Synthroid (n = 2,052, $757) patients who switched twice (n = 748, $1,178; p = .001) or three or more times (n = 245, $1,268; p = .004) had significantly higher hypothyroidism-related total medical costs.

Figure 2. Post-period hypothyroidism-related costs for matched cohort of continuous users of Synthroid and switchers from Synthroid to alternative levothyroxine therapies.

Figure 3. Post-period hypothyroidism-related total medical costs for matched cohort by number of switches.

Table 1. Patient characteristics—pre- and post-matching.

Characteristic	Unmatched					Matched
	Continuous Synthroid users (n = 7,991)		Switchers from Synthroid (n = 2,168)		p value	Continuous Synthroid users (n = 2,052)		Switchers from Synthroid (n = 2,052)		p value
	Mean/n	SD/%	Mean/n	SD/%		Mean/n	SD/%	Mean/n	SD/%
Age (mean; SD)	48.8	10.7	48.7	10.3	.825	48.9	10.3	48.9	10.3	1.000
Sex (n; %)					.472					1.000
Female	6,381	79.9%	1,716	79.2%		1,669	81.3%	1,669	81.3%
Male	1,610	20.2%	452	20.9%		383	18.7%	383	18.7%
Region (n; %)					<.001					<.001
North Central	2,149	26.9%	622	28.7%		487	23.7%	586	28.6%
Northeast	1,070	13.4%	220	10.2%		254	12.4%	205	10.0%
South	3,712	46.5%	1,071	49.4%		1,002	48.8%	1,019	49.7%
West	968	12.1%	245	11.3%		298	14.5%	232	11.3%
Unknown	92	1.2%	10	0.5%		11	0.5%	10	0.5%
Plan Type (n; %)					.010					1.000
Consumer Directed Health Plan	434	5.6%	113	5.3%		92	4.6%	92	4.6%
Comprehensive	299	3.8%	65	3.1%		49	2.4%	49	2.4%
Exclusive Provider Organization	135	1.7%	28	1.3%		21	1.0%	21	1.0%
High Deductible Health Plan	155	2.0%	25	1.2%		20	1.0%	20	1.0%
Point of Service	865	11.1%	272	12.8%		241	11.9%	241	11.9%
Preferred Provider Organization	5,905	75.8%	1,621	76.3%		1,601	79.1%	1,601	79.1%
Baseline Comorbidities (mean; SD)	0.41	0.7	0.38	0.7	.115	0.34	0.6	0.34	0.6	1.000
Number of switches					<.001					<.001
Zero	7,991	0.0%	0	0.0%		2,052	100%	0	0%
One	0	0.0%	1,111	51.2%		0	0%	1,059	51.6%
Two	0	0.0%	800	36.9%		0	0%	748	36.5%
Three	0	0.0%	257	11.9%		0	0%	245	11.9%
Acquired Hypothyroidism (n; %)	7,317	91.6%	2,007	92.6%	.249	1,741	84.8%	1,741	84.8%	1.000

Figure 4 shows the results of the sensitivity analysis examining hypothyroidism-related medical costs for the matched sub-set of patients diagnosed with acquired hypothyroidism (n = 3,482). In general, the results were not sensitive to an examination of this sub-group. As with the larger cohort of all hypothyroidism patients, continuous use of Synthroid, compared to switching LT4 formulations, was associated with significantly higher hypothyroidism-related drug costs ($156 vs $106; p < .001) but significantly lower hypothyroidism-related non-drug medical costs ($616 vs $1,076; p = .006) and total medical costs ($771 vs $1,182; p = .015). Similarly, as Figure 5 illustrates, post-period hypothyroidism-related costs increased as the number of switches increased. Patients who continuously used Synthroid over the 1 year post-period were found to have significantly lower costs (n = 1,741, $771) compared to patients who switched twice (n = 629, $1,252; p = .038). However, in this sub-group there was no statistically significant difference in total hypothyroidism-related costs between continuous users and those who switched LT4 therapies three or more times (n = 205, $771 vs $1,352; p = .114).

Figure 4. Post-period hypothyroidism-related costs for matched cohort of acquired hypothyroid patients.

Figure 5. Post-period hypothyroidism-related total costs by number of switches for matched cohort of acquired hypothyroidism patients.

Discussion

To examine the differences in disease-related costs between continuous users of Synthroid and patients who switched from Synthroid to an alternative LT4, this study focused on a relatively stable sample of individuals who were adherent to Synthroid therapy for at least 6 months. Even among such patients, 21.3% switched LT4 formulations in the 1 year post-period, despite AACE/TES/ATA guidelines that recommend against such a practice¹⁴. While there may be medically valid reasons for patients to switch medications, in this study nearly half (48.8%) of patients who switched did so multiple times and, of these, 76.9% switched back to Synthroid. Among patients who switched only one time, 94.9% switched to a generic LT4. These findings suggest that Synthroid may not have been consistently dispensed as written or that switching from Synthroid to an alternative LT4 formulation may not have been successful for many patients.

The matched cohort analysis revealed that hypothyroidism-related drug costs were significantly lower among patients who switched from Synthroid to an alternative LT4 formulation, most likely due to the lower costs associated with the use of generic drugs, which are ∼25% of the price of branded LT4²¹. However, hypothyroidism-related non-drug medical costs and hypothyroidism-related total medical costs were significantly greater for patients who switched therapies. Furthermore, patients who changed LT4 therapies multiple times had increased costs: for patients who switched three or more times in the post-period, the hypothyroidism-related total medical costs were 40.3% higher compared to these costs for continuous users. These findings are consistent with previous research which has reported that switching from branded to generic LT4 was associated with larger all-cause total healthcare costs, despite lower prescription drug costs, and that multiple switches were linked to greater costs²². This finding is also consistent with research which reported that switching from branded to generic formulations can increase costs for patients with epilepsy²³.

In addition to recommending that practitioners maintain patients on the same brand of thyroid drug, guidelines advise that those who do change therapies should have their thyroid-stimulating hormone (TSH) values monitored within 6 weeks¹⁴. This re-testing may account, at least partially, for the increased hypothyroidism-related costs associated with switching found in this study. Research has shown that 17.6% of patients taking thyroid medication are sub-clinically hypothyroid²⁴, a condition associated with an increased risk of coronary heart disease, heart failure, and fractures^25–27. Given the narrow therapeutic index for LT4^28,29, as well as concerns regarding the bioequivalence of thyroxine preparations¹⁴, switching may also increase medical costs due to complications associated with non-optimal treatment of hypothyroidism. Furthermore, the increased monitoring associated with switching LT4 formulations that is recommended in clinical guidelines³⁰ may also increase medical costs. Research on other disease states has found that switching medications in general may be associated with poorer patient outcomes^11–13. Notably, given that the majority of individuals in the present study switched to generics, evidence also suggests that switching from branded to generic formulations is associated with reduced medication adherence, increases in adverse events, and worsening of patient symptoms^31–33. These factors discussed above are all consistent with, and may in part explain, our findings of higher hypothyroidism-related medical costs associated with switching.

The present findings must be interpreted in the context of the limitations of the study. First, the analyses focused on a well-insured cohort of patients who were identified as being adherent to Synthroid for at least 6 months. As such, the results may not be generalizable to all patients. Second, the use of diagnostic codes may not be as rigorous as formal assessments, such as TSH laboratory results, for the identification of patients with hypothyroidism. Third, the use of claims data precluded the analyses from directly controlling for factors that may impact treatment but were not recorded in the claims database, including factors such as patient behavior, race, socioeconomic class, or severity of hypothyroidism. Fourth, while claims data provides information regarding when patients fill prescriptions, they are unable to provide insight into whether the medication was taken. Finally, observational studies such as the claims analyses presented herein are generally, by design, unable to determine any causality between switches in medications and costs, and our results, therefore, present associations only. Despite these limitations, a strength of the claims database was that it enabled adjustment for potential confounders using a matched cohort of patients whose demographics were consistent with the general distribution of age²⁴ and sex³⁴ associated with hypothyroidism.

Conclusions

This study examined hypothyroidism-related costs in a nationwide sample of patients who initiated therapy on Synthroid and were generally adherent to their treatment for at least 6 months. In this analysis, patients who continued on Synthroid after the initial 6 months of lead-in had significantly lower hypothyroidism-related total medical costs over the subsequent 12-month follow-up period relative to patients who switched LT4 therapy during that same follow-up timeframe. Furthermore, hypothyroidism-related costs increased as the number of switches in therapy increased. In support of current guidelines that recommend against such switches in medication for therapeutic reasons, the results of this study illustrate that there is also a significant economic burden associated with switching from Synthroid therapy.

Transparency

Declaration of funding

Support for this study was provided by AbbVie, Inc. AbbVie participated in the study design, data collection, analysis, interpretation of data, writing, reviewing, and approving the publication. AbbVie also funded the editorial services provided by HealthMetrics Outcomes Research. The data was not collected as part of a contract with a company marketing the original product.

Declaration of financial/other relationships

NK, BJ, ZH, and JC-H are all employees and stockholders of AbbVie, Inc. Study concept and design was primarily conducted by NK, with contributions from BJ, JC-H, and ZH. BJ collected the data, which was interpreted by NK, JC-H, and ZH, with assistance from BJ. The manuscript was written by the authors with help from Maureen Lage of HealthMetrics, and revisions were completed by all the authors. JME peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Previous presentations

86th Annual Meeting of the American Thyroid Association. September 21–25, 2016. “Economic Impact of Switching Branded Levothyroxine (Synthroid^®) in Patients with Hypothyroidism” (Control ID 2546556).

Acknowledgments

Patricia Platt, Michael Treglia, and Maureen Lage of HealthMetrics Outcomes Research, LLC provided editorial assistance during manuscript preparation, for which AbbVie Inc. provided compensation.