Patent holders produce cheaper and better drugs than generic brands

Gothe, H., Schall, I., Saverno, K. et al. The Impact of Generic Substitution on Health and Economic Outcomes: A Systematic Review. Appl Health Econ Health Policy 13 (Suppl 1), 21–33 (2015). Access on 6/30/24 Through KU Library, Berto – WH

*GS – Generic Substitution

Regarding economic outcomes, we extracted reported data on drug costs for the investigated drug and/or potential co-medication, as well as costs on healthcare utilization (inpatient and outpatient) and co-payments.

Of the 40 included studies, three examined economic outcomes only whereas another five studies examined both economic and clinical outcomes. These eight studies led to 22 outcome comparisons. The most frequent study design underlying the economic analyses was retrospective database analysis (five of eight economic studies). Another study projected actual economic observations from a Canadian retrospective open cohort study to a US setting using mathematical approaches, while another study determined the relapse incidence at which switching to a generic drug was cost-neutral through a simple decision analytical model. Only one prospective cohort study was found within this economic context.

These eight studies examined economic outcomes of GS in epilepsy patients treated with AEDs (37.5 %), organ transplant recipients treated with immunosuppressives (37.5 %), atypical neuroleptics (12.5 %) and anticoagulants in patients with continuous use of warfarin (12.5 %).

Seven of eight economic studies (87.5 %) found drug acquisition costs of the investigated medication to be lower for generic drugs (Table 4). Five studies identified drug costs for additional medication during generic drug use. In four of them (80 %) the supplementary medication costs exceeded the cost savings obtained from lower costs for the investigated medication due to GS. Total costs for inpatient and outpatient healthcare utilization (e.g., physician visits, hospitalization visits, etc.) were always lower with use of original products. Only costs that arose during the study periods ranging from 180 days to 5 years were included.

In the consolidated evaluation of economic outcome comparisons, 64 % indicated that staying with an original product incurs lower costs than GS. One of these studies examined economic outcomes for de novo therapy, in this case immunosuppressives after organ transplantation, and found total costs to be lower with originator therapy. In the group of maintenance therapy, 55 % of the economic comparisons opposed our second hypothesis.

Our findings suggest that 67 % of the evidence reported clinical similarity of GS as compared to original drug therapy, whereas 64 % of the comparisons of economic outcomes suggest costs to be lower when using original drugs. Accordingly, our first hypothesis was supported and our second hypothesis was rejected.

When stratifying the groups by de novo and maintenance therapy, we found a slight difference among studies on clinical outcomes: 76 % of clinical comparisons found similar effects for generics in de novo therapy and 64 % of clinical comparisons found similar effects in generics in maintenance therapy. Likewise, all economic comparisons in patients receiving de novo therapy (one study) versus 56 % in maintenance therapy (seven studies) revealed lower cost with originator therapy. None of the economic outcome comparisons revealed similar costs. However, the low number of studies in each group limits the generalizability of these results.

The majority of economic studies are related to sensitive therapeutic categories and maintenance therapy. Data on economic consequences of generic drug substitution in less sensitive therapeutic categories such as antihypercholesterolemics or osteoporosis were missing or did not meet our inclusion criteria. Thus, our review may suffer from publication bias and the economic conclusions are only relevant to these sensitive therapeutic areas. In the absence of evidence, no conclusions on economic advantages of one policy (e.g., generic substitution) over the other (e.g., originator therapy) can be drawn for less sensitive therapeutic areas. Moreover, with only one economic study analysing the economic impact of generic substitution for de novo therapy, the result must not be generalized. Hence, more evidence is needed to examine the economic impact of generic substitution in a real-life therapy situation, where chronic generic therapy may involve multiple substitution and thus multiple switching. Total healthcare costs were calculated in seven of eight studies. Half of these found cost reductions to be realized with generic drugs, whereas the other half found costs to be lower with the original drug. In contrast, our dichotomous classification supporting or opposing GS resulted in a stronger preference for originator drugs (lower healthcare cost in general).

The high heterogeneity among the cost types reported in the economic studies is also important to note, even among those which claimed to take the payer perspective. If only drug costs of the investigated drug and concomitant medication were examined, findings were likely to show cost reductions after GS. However, if dose adjustments, co-medication or healthcare utilization were considered depending on the rates of adverse events, GS no longer realized cost reductions. Instead, economic outcomes became more preferable in original drugs. It may be speculated that such results depend on additional factors such as therapeutic area, patient age or education level, number of medications or general healthcare context.

These considerations also apply to our analysis on clinical outcomes. If we focus on patient-relevant outcomes such as additional medication, adherence and adverse events only, disregarding surrogate outcomes or others, 67 % (12 out of 18 clinical outcome parameters for these three dimensions) of the categorized data on clinical outcomes remain in favour of our hypothesis of similar clinical effects. Evidence on all clinical outcomes did support the first hypothesis more often in de novo therapy (76 % of the relevant 33 clinical outcome comparisons) than in maintenance therapy (64 % of the relevant 59 clinical outcome comparisons).

This review has multiple limitations which should be considered when interpreting the results. We used a dichotomous classification of evidence, whether studies reported a statistically significant difference between generic and original drugs or not. We used this simplifying approach as the broad topic of this review resulted in a very heterogeneous pool of studies which impaired comparability and information had to be condensed in order to be able to provide a meaningful summary on the topic. Moreover, studies were not critically appraised regarding methodological aspects in a formal manner. For example, a comparison of therapy with one generic drug versus therapy with the originator drug in an 8-week randomized clinical study may answer a different question to a 2-year observational study in a real-life setting including multiple switching under a generic substitution policy. Among the included studies, three compared original drugs with their generic counterpart as well as drugs of different active ingredients [13–15]. Of these only the data comparing generics with the original reference drug were included. Additionally, our analysis did not include substitution between biological originator drugs and biosimilars. Finally, we took a semi-quantitative approach to summarize the evidence. Rather than pooling results in a formal meta-analysis, we simply counted outcome comparisons in favour of or against the compared strategies, and therefore did not weigh studies by sample size or precision. We also ignored dependency between multiple comparisons within one study and did not perform formal statistical tests or uncertainty analysis. Therefore, our analyses are merely exploratory, and should be interpreted with appropriate caution.

Of the included studies, 22 displayed a direct financial interest in their investigated drugs. In 55 % of these studies, the potential interest was linked to the original product, in 31 % of the studies this was linked to a generic and three studies (14 %) could not be specified, for instance due to a variety of author honoraria from several industry partners. Among those with a link to an original product, ten of twelve studies (83 %) found original drugs to generate preferable outcomes (clinical or economic), and of the studies with a link to a generic product, six of seven studies (86 %) found generics to be the dominant drug. Therefore, detection bias should be further explored in subsequent analyses.

We found economic outcomes to be more favourable in original drugs than in generics, which is in accordance with Duh et al. [16] who detected higher than expected healthcare costs in generic AEDs and revealed that GS may even increase healthcare costs. The included studies of this review mainly suggested further underpinnings of their specific findings, or extensions of the covered methodology such as reporting of additional outcomes, broader or different populations or different medications. Considering the variety of original research that we identified with our systematic literature search, few systematic reviews or meta-analyses have been published focusing only on specific aspects of GS, mainly within sensitive therapeutic classes [16–19]. Kesselheim et al. [18], for instance, investigated the impact of generic AEDs on clinical outcomes, namely seizure control, and hypothesized superiority of original products. Although the meta-analysis showed “no difference in the odds of uncontrolled seizure for patients on generic medications compared with patients on original-name medications”, observational studies found health services utilization to be slightly increased after GS. The authors attributed this insight to a detection bias caused by “concern from patients or physicians about the effectiveness of generic AEDs”. Similarly, economic outcomes of AEDs were assessed by Duh et al. [16] who found generic AEDs to cause higher healthcare costs than their original counterparts in both stable and unstable epilepsy patients. Multiple-generic substitution increased this effect even more.[ The findings of these two reviews apply to our results in regard to both clinical and economic outcomes. Desmarais et al. [17] investigated the clinical equivalence of original and generic psychotropic medications such as anticonvulsants and mood stabilizers. Besides raising compliance issues, generics caused clinical deterioration, adverse effects and changes in pharmacokinetics while leading to lower than expected cost savings. The authors therefore suggested generic switching of psychotropic medication to be advised only on an individual basis while simultaneously monitoring the switch [17].

Another systematic review by Kesselheim et al. [18] focused on clinical evidence in cardiovascular disease while also examining related opinions of editorialists. The analyses did not reveal superiority of either drug type, but found a considerable amount of editorials opposing generic drugs. While these reviews addressed mostly sensitive therapeutic categories, most of them found generics to result in similar clinical effects. In accordance with our findings, these reviews found economic consequences to be higher in periods of generic use compared to periods of original use. Altogether, these reviews revealed reservations and concerns in treatment routines of generics.