Estimating the economic benefits of positive shifts in fibromyalgia severity: an exploratory analysis based on modeling of clinical trial data of pregabalin

Abstract

Objective:

To estimate the annualized differences in direct medical and indirect costs associated with improvement in fibromyalgia (FM) severity among pregabalin-treated patients.

Methods:

Data from three clinical trials of pregabalin in patients with FM were modeled; efficacy results were extrapolated. Mean annual costs (direct and indirect) were assigned based on FM severity levels (mild: $10,219; moderate: $26,217; severe: $42,456). FM severity levels were defined using established cut-points on the Fibromyalgia Impact Questionnaire (FIQ). Mean annualized costs at end-point were estimated for all patients within each cohort and the mean differences in costs were compared using a regression model.

Results:

Relative to placebo, there was a significantly higher proportion of subjects with mild FM at end-point with pregabalin 450 mg and a significantly lower proportion of subjects with severe FM. Mean total costs were lower with pregabalin (300 mg, $25,721; 450 mg, $24,103) than placebo ($26,162). The difference in mean annual costs was $2059 lower for pregabalin 450 mg (p = 0.003) and $441 lower for pregabalin 300 mg (p = 0.52). Mean direct medical costs were higher with pregabalin (300 mg, $4962; 450 mg, $4820) than placebo ($4364). The difference in mean annual direct medical costs was significantly higher for pregabalin 450 mg by $456 (p < 0.0001) and by $599 for pregabalin 300 mg (p < 0.0001). Mean indirect costs for pregabalin (300 mg, $20,783; 450 mg, $19,306) were lower than placebo ($21,735). The difference in mean annual indirect costs for pregabalin 450 mg was lower by $2429 (p < 0.0001); pregabalin 300 mg was lower by $951 (p = 0.12).

Limitations:

Use of 13-week end-point data from clinical trials to extrapolate to annual costs and an assumption of continuous therapy over the course of 1 year.

Conclusions:

Improvements in FM severity were associated with overall reductions in cost. Reductions in indirect costs may offset the costs of treatment with pregabalin.

Keywords::

• Fibromyalgia
• Severity
• Costs
• Pregabalin

Background

The chronic condition known as fibromyalgia (FM) has conventionally been defined by widespread and persistent musculoskeletal pain and the presence of tenderness at a minimum of 11 of 18 specified points1. However, FM is characterized by heterogeneity in its clinical presentation, with patients complaining of a constellation of symptoms that may include fatigue, non-restorative sleep, functional impairment, mood disturbances, and cognitive dysfunction2–4. The multidimensional nature of FM from the patient and clinical perspectives has recently been recognized by incorporation of the non-pain symptoms into revised diagnostic criteria5 and the EULAR (European League Against Rheumatism) core set of symptoms for evaluation in clinical trials6.

The complexity of symptoms not only leads to the high healthcare resource utilization that has been reported among patients with FM7–11, but also results in substantial indirect costs mainly resulting from lost work productivity9,11–14. In a study that used patient-level data to thoroughly explore direct and indirect costs, lost productivity due to absenteeism, presenteeism, unemployment, and disability among subjects with FM was determined to be the primary cost-driver, accounting for∼ 73% of indirect costs and 56% of total costs14.

Recent studies have shown that direct medical costs and indirect costs are moderately and significantly associated with FM severity11,14, as assessed using the Fibromyalgia Impact Questionnaire (FIQ), a validated patient-reported outcome that evaluates the impact of FM on daily life over a 1-week recall period using a 0–100 scale, with higher scores indicating greater impact15. These studies reported that, as FM severity increased, both direct and indirect costs also increased. Defining FM severity by using validated cut-points on the FIQ enhances our ability to characterize FM and its treatment by enabling greater assessment of treatment differences and serving as an anchor to define changes in other outcomes, including costs.

While costs were characterized for each level of FIQ-defined severity (mild, moderate, and severe)14, there is a lack of data on how changes in FM severity resulting from specific therapy may impact overall costs. Therefore, the purpose of this study was to estimate the annualized differences in healthcare costs that may be expected to result from shifts in FM severity among patients treated with pregabalin modeled from results of clinical trials. The study represents a novel approach in that it used data from three randomized, placebo-controlled, clinical trials of pregabalin for the efficacy data, and an observational, health economic study of patients with FM in the US to provide cost information.

Pregabalin is a high-affinity ligand of alpha2-delta (α2-δ) sub-units of voltage-gated calcium channels in the central nervous system16 that reduces the release of neurotransmitters including glutamate17,18 and substance P19. In addition to its anticonvulsant properties, it has analgesic activity and was the first pharmacologic therapy approved in the US for the treatment of FM20.

Methods

Modeling was based on data from three randomized, placebo-controlled, clinical trials evaluating pregabalin for treatment of FM21–23. The studies were conducted in accordance with the Declaration of Helsinki and were approved by the appropriate institutional review boards or independent ethics committees; patients provided written informed consent prior to participation.

All three studies utilized twice-daily dosing of pregabalin at fixed doses of 300, 450, and 600 mg/day. While the studies were generally of similar design, one study evaluated the effects of pregabalin treatment over 13 weeks21. The other two studies used a 14-week evaluation period that was preceded by a 1-week, single-blind, placebo run-in period, after which patients with a ≥30% reduction in the pain visual analog scale (VAS) were excluded22,23. Additional details of the design of the studies have previously been reported21–23.

Although data for all doses were used in the modeling, for the current analysis our interest centered on data from the groups treated with pregabalin 300 mg and 450 mg, since these doses have been approved for the treatment of FM20. The populations for each of these active treatment groups as well as the placebo group were pooled from the three studies.

FM severity levels at baseline and end-point (using last observation carried forward) were defined using established cut-points on the FIQ24. The cut-points define mild FM severity as an FIQ score < 39, moderate FM as a score ≥39 to <59, and severe FM as a score ≥ 59. The proportions of patients showing a shift from baseline to end-point in FM severity level were estimated for the treatment groups and compared between active treatment and placebo using chi-square analysis24.

Extrapolation of the efficacy from the clinical trials to determine annualized costs was based on a 1-year open-label study25. For the cost analysis, costs (direct, indirect, and total) were assigned for every subject based on their level of severity at end-point and the study arm to which the subject was randomized. Mean per-patient annualized costs were then estimated based on all patients within each cohort and the mean differences in costs between cohorts were compared using a regression model.

Severity-based costs were derived from the US Fibromyalgia Burden of Illness Study (FM-BOI)14, which estimated average total annual costs of $10,219, $26,217, and $42,456 for mild, moderate, and severe FM, respectively, which were defined using the same FIQ cut-points in the current analysis. Similarly, annual direct costs were estimated at $4854 (mild), $5662 (moderate), and $9318 (severe), and indirect costs derived from lost productivity were $5366 (mild), $20,556 (moderate), and $33,139 (severe). In the current study, direct and total costs were adjusted for the cost of the study drug by replacing all pharmacy costs from the FM-BOI study with the cost of pregabalin using the Medicaid drug utilization data maximum allowable cost (MAC) of $3.63 for pregabalin 300 mg/day and $3.62 for pregabalin 450 mg/day26. Pregabalin daily treatment costs were applied for the 365 day period of the analysis. Zero cost was assumed for placebo and was used to replace the pharmacy costs.

To evaluate the robustness of the results and account for possible bias arising from the use of a single mean value for all subjects within a given severity level, a sensitivity analysis was performed using a distribution-based approach. In this analysis, costs were randomly assigned (50,000 simulations) from the distribution of costs actually observed among the patients at each level of FM severity in the FM-BOI study14, still taking into account the price of pregabalin.

Results

The baseline demographic characteristics were similar across the three studies and across treatment groups, and were appropriate for a pooled analysis. Consistent with an FM population, patients were predominantly female (93.4%), with a mean age of 49.1 years and a mean duration of FM of 9.3 years. Based on the proportion of patients at baseline in each treatment group who reported different levels of FM severity, mean annual total costs per patient (± standard deviation) were estimated to be $34,527 ± $9843 (Table 1).

Table 1.  Baseline characteristics of fibromyalgia patients in three randomized controlled trials relative to a real-world fibromyalgia population identified in the Fibromyalgia Burden of Illness study.

Variable	Value
	RCTs (n = 2228)21–23	FM-BOI study (n = 203)31
Female, n (%)	2082 (93.4)	192 (94.6)
Age, years, mean ± SD	49.1 ± 11.2	47.9 ± 10.9
Duration of FM, years, mean ± SD	9.2 ± 7.9	10.7 ± 8.1
Average pain intensity, mean ± SD	6.8 ± 1.3	6.3 ± 2.1
Number of tender points, mean ± SD	17.0 ± 1.7	14.7 ± 3.4
FIQ score, mean ± SD	61.7 ± 14.7	63.2 ± 19.0
FM severity distribution, %
Mild (FIQ score < 39)	6.0	10.3
Moderate (FIQ score ≥ 39 to < 59)	37.0	24.1
Severe (FIQ score ≥ 59)	57.0	65.5
Annual FM costs per patient (based on FM severity), mean ± SD	$34,527 ± $9843	Not reported

FIQ, Fibromyalgia Impact Questionnaire; FM, fibromyalgia; FM-BOI, Fibromyalgia Burden of Illness; RCTs, randomized controlled trials; SD, standard deviation.

All three treatment groups demonstrated a general shift from baseline such that there were greater proportions of patients at lower FM severity levels at end-point (Figure 1). As shown in Figure 1, this shift led to an overall lower FM severity among subjects in both pregabalin groups relative to placebo, with the greatest proportion of patients reporting mild FM (34%) in the pregabalin 450 mg group. As shown in Figure 2, this proportion was significantly higher than the placebo group by 7.4% (95% confidence interval [CI] = 2.2–12.9%), and there was a corresponding decrease in the proportion of subjects with severe FM that was significantly lower than placebo by 10.0% (95% CI = 4.5–15.6%). While the proportion of patients with moderate FM was also higher in the pregabalin 450 mg group relative to placebo (35% and 32%, respectively), the difference of 3% was not significant. A similar trend was observed at the 300 mg dose for shifts in FM severity; differences from placebo were not statistically significant.

Figure 1.  Shifts in fibromyalgia severity categories among patients in three clinical trials of pregabalin. Distribution of patients at each severity level at BL and EP. Adapted with permission from Bennett et al.24. BL, baseline; EP, end-point.

Figure 2.  Differences from placebo at end-point for proportion of patients at each severity level treated with pregabalin. Significance is indicated if 95% CI does not cross zero. Data from Bennett et al.24.

For all subjects who were characterized at baseline as having moderate severity, 34–40% across treatment groups transitioned to mild severity by the end of the study. Similarly, for all subjects characterized at baseline as severe, 27–29% shifted to moderate, and 19–27% shifted two severity categories to mild by the end of the study.

The observed shifts in FM severity across all three treatment groups resulted in annual FM health-economic costs that were substantially lower, by 25–30%, than the annual costs estimated based on their baseline FM severity levels. As shown in Table 2, mean total costs for the three treatment groups according to reported FM severity at end-point were $26,162 for placebo, $25,721 for pregabalin 300 mg, and $24,103 for pregabalin 450 mg. When total costs were broken down into direct and indirect cost components (Table 2), both pregabalin groups were characterized by higher direct and lower indirect costs than placebo.

Table 2.  Estimated mean annual costs for patients with fibromyalgia in three clinical trials.*.

Cost category	Annual costs, $
	Placebo (n = 558)	Pregabalin 300 mg (n = 551)	Pregabalin 450 mg (n = 552)
Total costs	26,162	25,721	24,103
Direct costs	4363	4962	4820
Indirect costs	21,735	20,783	19,306

*Based on fibromyalgia severity at end-point.

Estimates of the cost differences showed that the mean difference in FM health-economic costs was significantly lower with pregabalin 450 mg by $2059 (p < 0.005) (Figure 3). Although the 300 mg group was also characterized by total costs that were lower by $441, this difference was not significant. For the direct cost component (Figure 3), the mean differences (placebo minus pregabalin) of −$599 and −$456 for the 300 mg and 450 mg groups, respectively (both p < 0.0001), show significantly higher costs with pregabalin. In contrast, the mean annual indirect cost differences of $951 for pregabalin 300 mg and $2429 for pregabalin 450 mg suggest lower indirect costs than placebo, although the difference was significant only for the pregabalin 450 mg group (p < 0.0001).

Figure 3.  Differences from placebo in cost savings resulting from shifts in severity from baseline to end-point. *p < 0.0001 and †p < 0.005.

In the sensitivity analysis, the distribution-based approach (Table 3) resulted in cost estimates that were comparable to the primary analysis. Likewise, the differences in costs between placebo and the pregabalin groups were similar in magnitude and direction to those estimated in the primary analysis; indirect and total costs were lower with pregabalin, but direct costs were higher. The only significant difference in costs was observed for direct costs between placebo and the pregabalin 300 mg group, with pregabalin costs higher by $522 (95% CI = $59–$988); none of the other differences were significant, since the 95% CIs for the difference in costs included zero.

Table 3.  Sensitivity analysis using a distribution-based method (50,000 iterations) for estimating mean annual costs per patient.*.

Cost component	Placebo	Pregabalin 300 mg	Difference from placebo (95% CI)	Pregabalin 450 mg	Difference from placebo (95% CI)
Total costs	$26,152	$25,734	−$418 (−3329, 2487)	$24,119	−$2032 (−4988, 922)
Direct costs	$4417	$4939	$522 (59, 988)	$4801	$384 (−82, 853)
Indirect costs	$21,737	$20,785	−$952 (−3,786, 1,855)	$19,313	−$2424 (−5283, 411)

*Differences expressed as pregabalin minus placebo.

Discussion

This analysis builds on two previous studies that have suggested new approaches to evaluating treatment effects of FM therapies including economic outcomes. By defining FM severity using the FIQ, one of the previous studies characterized the shifts in severity levels that can be obtained with treatment in clinical trials of pregabalin24. The other study quantified the health-economic burden in dollars per year per patient associated with each level of FM severity, and showed that indirect costs account for the majority of this burden across severity levels14. By combining these techniques in the current analysis, we demonstrated that the shifts to lower FM severity resulting from treatment with pregabalin provide economic benefits in the form of decreased indirect costs that may offset the increase in pharmacy costs relative to placebo treatment.

The overall shift to lower FM severity at both approved doses of pregabalin resulted in lower mean annual costs per patient than placebo. However, direct medical costs for a patient treated with pregabalin were significantly higher than placebo. In particular, total costs for the pregabalin 450 mg dose were lower by $2059 per year than placebo. These results can be interpreted as, on average, it would cost $2059 per year less to treat a subject with pregabalin 450 mg compared with placebo. Direct medical costs for the same patient would be $456 higher than placebo treatment with a gain in productivity of $2429. The difference in total costs between pregabalin and placebo subjects was driven by indirect costs.

Similar trends were observed in subjects treated with pregabalin 300 mg, but the only significant difference from placebo was in the higher direct costs associated with active treatment. Nevertheless, incremental benefits were obtained at this dose relative to placebo. The differences observed between the pregabalin doses are consistent with the suggestions from meta-analyses that pregabalin 450 mg most likely represents the optimal dose for the treatment of FM27,28.

Our results regarding total cost savings are to be expected for two reasons. First, improvement in FM may potentially increase productivity, thus reducing the indirect costs. Second, although the higher direct costs may be due in part to the costs of pregabalin therapy applied daily over a 1-year period, patients with FM are characterized by a high prevalence of comorbidities2,8,9,11,29, which may contribute to substantial medical resource utilization (i.e., direct medical costs) in spite of a shift to lower FM severity levels. Indeed, while medication use has been reported to be substantial in patients with FM, evidence supports non-pharmacy resources as the primary component of direct medical costs8,9,12,13,29,30.

In the sensitivity analysis, when costs were reiteratively taken from the distribution of values to account for inter-subject variability, no apparent bias was observed for our estimations using simple means for each treatment group. Although no bias was found, it should be noted that most of the 95% CIs included a zero value, indicating that this model did not provide evidence of statistical significance for most differences. Nevertheless, the consistency of the values obtained using the two different methods demonstrates the robustness of the results.

A strength of this study is our use of FIQ for categorization of FM severity. Although such categorization was based on a pain anchor24, the FIQ itself captures impairment of daily activity due to FM as a condition15, thus taking into account a more global patient perspective that is not necessarily limited to pain as the primary symptom. Consequently, the shift in FM severity levels observed among the treatment groups is more representative of patients’ functional abilities; transition to lower FM severity at end-point was observed for patients in the moderate and severe FM categories at baseline.

Positive shifts in FM severity and greater ability to carry out activities of daily living are also associated with improvements in quality-of-life and higher health utility scores. A large patient-level US cross-sectional study used the EuroQol 5D (EQ-5D) to assess health-related quality-of-life (HRQoL) across five dimensions (mobility, self-care, performance of usual activities, pain, and anxiety or depression) among subjects with FM31. The study reported very low EQ-5D health valuation scores (range of −0.594 [worst possible health] to 1.0 [best possible health] for subjects with severe FM (0.20) relative to those with moderate (0.57) and mild FM (0.76). There is thus a clear association between lower FM severity and better HRQoL.

Limitations of this study should be noted, including our use of 13-week end-point data from clinical trials to extrapolate to annual costs and an assumption of continuous therapy over the course of 1 year. However, efficacy was maintained among patients in the 1-year open-label study who remained on treatment25, suggesting the clinical relevancy of our assumptions. It should also be noted that the characteristics of the patients in the clinical trials were generally similar to what may be expected for an FM population such as that identified in the FM-BOI study31.

This research is based on pivotal Phase 3 studies in which placebo served as the sole control group and an active comparator was not included. Despite this, we sought to capitalize on the evolving and incremental nature of economic evaluations. In this context, it should be noted that several retrospective, observational studies using claims databases have demonstrated no significant differences in either overall or FM-related costs associated with the initiation of pregabalin relative to standard of care or duloxetine32–35. While one study did suggest lower costs with duloxetine relative to pregabalin, that study was characterized by exclusion criteria that created an imbalance in relevant comorbid conditions (depression, post-herpetic neuralgia, and painful diabetic neuropathy) between the cohorts36. Thus, the content of this manuscript remains important as a stepping stone and a foundation or basis upon which to build with more research, including with an active comparator, although currently no head-to-head trials have been performed comparing currently approved FM therapies.

The combination of two different datasets, from an RCT and an observational study, also conveys some potential limitations. FM health-economic costs were derived from a patient population generally representative of the US, but not specifically treated with pregabalin. Thus, the actual direct medical costs and indirect costs of pregabalin-treated patients may differ from those reported in this analysis.

Another potential limitation is the generalizability of the results to everyday clinical practice. It may be considered that the primary barrier to generalizability is the inter-patient variability in costs even among patients within the same FM severity level. This variability not only extends to resource utilization but also to differences in indirect costs including those related to productivity. However, our use of a distribution-based sensitivity analysis to account for this variability for direct, indirect, and total costs, and the resulting demonstration of robustness, suggest that similar economic benefits are likely to be obtained in the clinical setting.

Conclusions

This study demonstrates that appropriate treatment of FM results in a shift toward reduced FM severity in a proportion of patients. While it has previously been shown that costs for patients with FM increase as FM severity increases, our data demonstrate that a shift toward lower FM severity through treatment can result in an overall economic benefit. This benefit arises despite the higher costs that may be associated with treatment, since indirect costs are the primary cost-driver in patients with FM. These data may be important for both employer and payer stakeholders, each of which assume different components of the economic burden of the consequences of FM and its treatment.

Transparency

Declaration of funding

This study was funded by Pfizer Inc. All authors contributed to the study design, statistical analysis plan, interpretation of results, and review of the draft manuscript; the final manuscript was read and approved by all authors.

Declaration of financial/other relationships

JCC, AGB, GZ, and AC are employees and stockholders of Pfizer Inc, the sponsor of this study.

Acknowledgments

Medical writing support was provided by E. Jay Bienen, which was funded by Pfizer, Inc. Editorial support to prepare this manuscript for submission was provided by Charlotte Kenreigh, PharmD of UBC Scientific Solutions and funded by Pfizer Inc.