Data sourceTop-selling SODF

On the basis of the report of the Wissenschaftliches Institut der AOK (WidO) 2021 [14], we have determined the 50 drugs with the highest number in sales to patients with statutory health insurance in 2020. Three metered dose inhalers and two multidose containers were excluded, yielding 45 medicinal products of the 50 top-selling drugs whose primary packaging was an alveolus (Supplemental Table S1). Accordingly, the blister cards of these 45 brands provide the basis for all comparisons, extrapolations, and general statements made in this analysis.

Variability of generic products

Because the 50 top-selling brands consist of a range of different active ingredients and strengths, it did not seem meaningful to only compare the blister cards of these brands. Rather, it was also of interest to analyse how big the differences in packaging were between exchangeable brands (generics) of the same strength. For this purpose, eight off-patent (generic) active ingredients (acetylsalicylic acid, ibuprofen, metformin, omeprazole, pantoprazole, sertraline, tamsulosin hydrochloride, and valproic acid) were selected based on their availability in a large German community pharmacy (Frosch-Apotheke, Lörrach, Germany) where the data collection was carried out (convenience sample). This second sample included 44 brands, one of which was one of the 45 top-selling brands in blister packs and was therefore also included in the other sample. Together with the 45 top-selling blister-packed brands, this therefore resulted in a total of 88 blister cards analysed (Supplemental Table S2).

Data acquisition

For all analyses, we used the blister cards of the largest package size available in the market, which was aimed for long-term therapy (~ 100 d) and occasionally, if no such package was available, for a treatment duration of 30 d. By examining the largest marketed package size, this study evaluated the presumed most efficient packaging with the least amount of (primary and secondary) packaging material per dosage form and the least likelihood of empty areas without blisters (Fig. 1), resulting in a conservative estimate of waste production.

Fig. 1

To assess blister dimensions and void space in blister packs (primary packaging) of a representative sample of solid oral dosage forms currently on the market in Germany, the characteristics and spatial dimensions of blister packs were assessed by recording the material of the blister cards (aluminum, plastic) and measuring the blister cards (a, b, c, i), the blister chambers (alveoli; d, e, f), and the interstitial spaces (c’, g, h), as well as recording the topography of the alveoli thereon and determining the spatial relationships. The dimensions of the blister chambers were always measured at the base of the alveolus

In the course of data collection, the same parameters were surveyed for all selected brands and collected in a Microsoft Excel file (Excel Version 16.62, Microsoft, Redmond, WA, USA). In addition to general conditions such as the number of dosage forms or the number of blister cards per pack, further measurements were performed (Fig. 1), which can be divided into 4 subanalyses:

1.

Blister cards (weight (incl. dosage forms), length (i in Fig. 1), width (2 × b), and height)

2.

Blister chamber (alveolus) (form (elongated/round), diameter (d) (if round) or length (f) and width (e) (if elongated or capsule), and largest (g) and smallest distance (h) between two blister chambers)

3.

Dosage form (form, diameter (if round) or length and width (if elongated/capsule), height, and mass), and (if available)

4.

Single blister compartments (number of large and small compartments per blister card), length (largest (a) and smallest (c)), width (b), and number and area (b × c’) of compartments without alveolus (empty blister field)

All weight measurements were taken with a digital scale (Sartorius Entris, Sartorius, Göttingen, Germany) accurate to the mg, or extracted from the drug information system AiDKlinik® (Dosing GmbH, Heidelberg, Germany), and all linear measurements were performed with a digital caliper gauge (Carbon Fiber Composites Digital Caliper, Apfelkiste, Zeiningen AG, Switzerland) that allowed for a measurement accuracy of one mm.

Quality assurance of data acquisition

Before the actual data collection was started, measurement accuracy and consistency of the investigator (OF-W) was tested in a pre-test after a learning phase and compared with the measurements of an experienced pharmacist (ADF). Based on this pre-test containing 15 SODF, which revealed complete agreement for all categorical assessments (Cohen’s kappa value 1.0) and almost complete agreement (97%) between both investigators in numerical data (222 measurements), it was decided that a single measurement per category and substance, carried out by one person was reliable.

Calculation of ideal blister card dimensions

Because the distances between individual alveoli (g and h in Fig. 1) and between the alveolus and the edge of the blister card varied considerably, we also calculated the minimum dimensions of the blister card, assuming that the SODF were grouped into 2 rows, that the current alveolus dimensions remained unchanged, and that the distances between the SODF and to the edge were uniformly either 2, 3, or 4 mm (Eq. 1 for round SODF and Eq. 2 for elongated SODF).

$$\mathrm= (2 \;d + 3 \;k) \times (k + 0.5\; n \;(k + d))$$

(1)

$$\mathrm= (2 \;f + 3 \;k) \times (k + 0.5 \;n \;(k + e))$$

(2)

where d is the diameter of round alveoli, f the length, e the width of elongated alveoli, k is the distance between the alveoli, and n is the number of SODF on a blister card (see also Fig. 1).

To calculate the savings in surface area, the blister area measured in practice was taken as 100% and then the area of the optimized blister card was compared to the original area of the marketed product. The difference between these two values then corresponds to the savings in surface area and material.

Estimation of blister card waste of solid oral dosage forms of the German pharmaceutical market

To be independent of the different package sizes prescribed, actual waste quantities of the best-selling 45 products (Table 2) were estimated based on the number of defined daily doses (DDD) of the corresponding brands dispensed in 2021 (Supplemental Table S1). The number of SODF consumed yearly was calculated by dividing the number of DDD prescribed [15] by the strength that was analysed in this study. This number was then divided by the number of dosage forms contained in the largest marketed package size as assessed in this study, giving the equivalents of drug packages. With these results and using the information on the number of SODF per blister card, the number of blister cards was calculated. These results were then multiplied by the weight of a blister card, giving the mass of waste per year in Germany for each preparation studied.

Assuming that their primary packaging characteristics are comparable to the observations in the group of the 45 most prescribed products, an extrapolation to the total market of reimbursable SODF was carried out. For this purpose, drug groups with mainly (> 50%) topically (dermatics, ophthalmics, and inhalatives) or parenterally administered medicines (e.g., immunostimulants, blood products, and allergens) were excluded. For the remaining drugs, we extracted the DDD reimbursed by the statutory health insurance funds in 2020 [15]. Using a rule of three, we extrapolated the amount of waste, assuming that the average amount of waste per DDD of the 45 best-selling products is also representative for the remaining preparations. Since, in addition, about 10.5% of the German population is privately insured [15], the result of the rule of three was multiplied by 1.105, which then gives the value for blister packaging waste for SODF of the entire German pharmaceutical market.

StatisticsDescriptive statistics

Data are expressed as means ± standard deviation and range using Microsoft Excel. To account for possible uneven distribution of values, the relationship between variables was analyzed using Spearman correlation (Prism 9, Version 9.4.1., GraphPad Software Inc., San Diego, CA, USA). A p value < 0.05 was considered significant.