Abstract
Aims
Pharmaceutical utilization within global healthcare has significant environmental impact throughout its lifecycle. Understanding how the environmental impact of medicines influences patients' preferences relative to ease of use and cost supports interventions promoting appropriate medicine consumption, reduced waste and environmental sustainability. The aim of this study was to investigate patient preferences and attitudes regarding the environmental impact of medicines in relation to ease of use and cost.
Methods
A cross‐sectional survey (December 2022 to January 2023) was conducted using the Dutch AMP pharmacy panel (>25 000 chronic medicine users). It included a discrete choice experiment (DCE) on medicines' environmental impact, ease of use and cost, and questions on environmental sustainability in medicine use and daily life. Latent class analysis identified groups with shared preferences. An environmental sustainability attitude score was calculated from daily life behaviour.
Results
Of 25 787 invited panel members 6390 responded and 4122 were included in the analysis (16.0%, mean age 68 years, 42.6% female). Four preference groups were identified and labelled based on DCE‐derived attribute preferences: eco‐focused (55.7%), cost‐focused (20.1%), indifferent (16.4%) and eco‐sceptical (7.8%). Environmental sustainability attitude scores were highest in the eco‐focused and eco‐sceptical (median 75.0%) groups, followed by the indifferent and cost‐focused (median 66.7%) groups. Important factors influencing this were environmental concerns (eco‐focused), environmental and personal health concerns (eco‐sceptical), costs (cost‐focused) and environmental, cost and personal health concerns (indifferent). Only 8.8–15.5% reported good or very good self‐assessed knowledge of medicines' environmental impact.
Conclusions
Environmental sustainability significantly affects medicine preferences, although heterogeneity exists. Tailored initiatives are required to promote environmentally sustainable pharmaceutical care.
Keywords: discrete choice experiment, environmental impact, medicine waste, patient preferences, pharmaceutical care, sustainable medicine use, sustainable pharmacy
What is already known about this subject
Pharmaceutical utilization within the global healthcare sector has significant adverse effects on the environment throughout its lifecycle.
Studies show population preference to or willingness to switch to medicines with a lower environmental impact, support eco‐labelling of over‐the‐counter medicines and pay for CO2 compensation.
Different population profiles influence environmentally sustainable attitudes.
What this study adds
Over half of our study population value the environmental sustainability of medicines, often exceeding the importance of ease of use and cost.
An important self‐reported knowledge gap is observed about medicines' environmental impact.
A substantial heterogeneity is observed in our study population's preferences and attitudes for different medicines' attributes, implying that tailored strategies are needed to advance environmentally sustainable care.
1. INTRODUCTION
Pharmaceutical utilization within the global healthcare sector has significant adverse effects on the environment throughout its lifecycle 1 , 2 , 3 Pharmaceutical residues increasingly appear in water systems, disrupting ecosystems. 4 , 5 , 6 , 7 Improper disposal of unused or expired medicines further contaminates the environment. 8 , 9 A transition toward a circular pharmaceutical supply chain—one that minimizes waste and maximizes resource reuse—depends partly on patient choices regarding medicine use, disposal and over‐the‐counter purchases. 10 , 11 , 12 These choices directly affect the environmental footprint of pharmaceutical care and understanding the factors that drive these decisions is crucial for advancing environmentally sustainable medicine use. 13 , 14 Multiple studies suggested that many patients prefer more environmentally sustainable healthcare options in general, provided clinical effectiveness is not affected. 15 , 16 While a recent study indicates that a substantial proportion of individuals support environmentally sustainable use of medicines, 17 understanding of patient preferences and attituded regarding the environmental impact of medicines remains limited.
Different medicine attributes can be important for a patient's decision‐making regarding medicine use, including the cost of medicine and ease of use. 18 , 19 , 20 The extent to which considerations about the environmental impact of medicines may influence patient's decision‐making, alongside other attributes, remains unclear. Furthermore, it is unclear to which extent patients may differ in their preferences and attitudes for environmental sustainability of medicines.
Such insight can support stakeholders—healthcare providers, pharmaceutical companies, regulatory agencies and governments—in developing and promoting medicines and policies that lessen the environmental footprint of medicines specifically and healthcare in general. Examples of such policies are reducing waste, improving appropriate consumption and supporting environmentally sustainable medicine use. Therefore, this study aimed to investigate patient preferences and attitudes regarding the environmental impact of medicines—relative to cost and ease of use—by means of a discrete choice experiment (DCE).
2. METHODS
2.1. Study design
A cross‐sectional survey was developed containing multiple‐choice questions, open‐ended questions and a DCE. 21 , 22 , 23 , 24 The DCE methodology allows researchers to quantify how individuals value different attributes of a product or service by asking them to make trade‐offs between different options.
2.2. Study setting, population and recruitment
Members of the Dutch nationwide pharmacy patient panel managed by AMP Onderzoek & Advies in de Zorg (AMP Research & Consultation in Healthcare) 25 , 26 were invited to respond to the survey by email. This email included an explanation of the study and instructions for participation. AMP panel members are recruited via community pharmacies in the Netherlands. There are no specific inclusion or exclusion criteria for membership. The panel consisted of 25 787 members at the time of the survey, of which 48.0% reported female gender and 97.0% indicated chronic use (≥3 months) of at least one medicine.
2.3. Survey development and content
The survey comprised five sections and 18 questions (Supporting Information The Survey). The first section (questions 1‐5) addressed respondents' attitudes regarding environmentally sustainable practices in daily life. The second section (questions 6‐9) focused on respondents' self‐reported knowledge and attitudes regarding environmentally sustainable practices with medicines in daily life.
The third section concerned the DCE (question 10). The DCE is a quantitative method that elicits preferences by asking respondents to choose between hypothetical alternatives, each defined by several attributes at different levels. Our DCE examined three attributes of medicines: (1) environmental impact, whether a medicine is harmful to the environment or not; (2) ease of use: easy to use packaging vs not easy to use packaging; and (3) cost, €5 vs €10. These three attributes, each with two distinct levels, produced eight theoretical medicine profiles (23). In each of the eight choice scenarios, respondents were presented with two hypothetical medicines containing different combinations of the attribute levels. Respondents were informed that all other aspects (eg, quality, efficacy and safety) were equivalent across the choices. The selection of scenarios was randomized through an algorithm within the Sawtooth Discover 27 application. Respondents were required to choose one option for each pair, with no “opt‐out” alternative provided.
The fourth section (questions 11‐13) gathered information on respondents' preferences and attitudes toward environmentally sustainable practices of their community pharmacy, including medicine disposal and redispensing of returned unused medicines. Finally, the fifth section (questions 14‐18) captured demographic and other characteristics of the respondent (eg, gender, age, amount of medicines used chronically). In this section, respondents were asked whether they lived in a city, village or other types of area. Responses described as “other” were subsequently reviewed and recoded as either “city” or “village” derived from their descriptions. Additionally, respondents were asked to indicate their highest completed level of education. These responses were then classified into three educational levels: “low” (no education, primary school or preparatory/lower vocational education), “middle” (secondary vocational or higher general secondary education) and “high” (higher professional or university education). Responses described as “other” were subsequently reviewed and recoded as “low”, “middle” or “high” educational level derived from their descriptions. The survey was expected to take 10‐15 min to complete.
Prior to deployment, three chronic medicine users representative of typical panel members pretested the survey to ensure its clarity and comprehensiveness. Although no fixed minimum sample size exists for DCE studies, a threshold of 300 respondents is commonly recommended for robust results. 22 , 23 Given the large size of the AMP patient panel and a prior study indicating high response rates by the panel, 28 the number of respondents for this study was expected to be higher than 300.
2.4. Survey distribution and data collection
An invitation to participate was sent to panel members on 14 December 2022, using the Survalyzer software used for all AMP panel surveys. Individuals who consented to take part were then directed to the Sawtooth Discover application to complete the survey. The survey link remained open through 31 January 2023, with no reminders issued. All responses were securely recorded and processed on the Discover platform.
2.5. Data analysis
Respondents who did not complete the survey and respondents who completed the survey within 3 min were excluded from the analysis. Furthermore, respondents who completed the survey within 3‐5 min and showed an inconsistent choice pattern in the DCE, indicated by a root likelihood <0.6, were also excluded. The root likelihood is computed as the geometric mean of the probabilities of respondents' choices, with higher values indicating more consistent decision‐making patterns across the choice tasks. 29 A single implausible response to a question (eg, “9999” medicines used) was treated as missing data. More than one implausible response led to exclusion.
For the included responses, first, attribute utilities and the relative attribute importance in the DCE were assessed. Utilities quantify the degree of preference for each attribute level (positive = preferred, negative = not preferred, larger magnitudes = stronger preference). 30 , 31 Individual‐level utilities were estimated using Gibbs sampling simulation, a Monte Carlo Markov Chain technique. 32 , 33
Latent class analysis was then used to identify groups of respondents with similar attribute preferences in the DCE. 31 , 34 Model selection was guided by the Akaike information criterion (AIC), the Bayesian information criterion (BIC) and the “elbow” assessment of model fit. 35 , 36 Each group was assigned a label reflecting its distinct attribute preferences in the DCE.
Next, each identified group was characterized by demographic features and its members' responses to the questions about their self‐reported knowledge about the effects of medicines on the environment (question 6), attitudes regarding environmentally sustainable practices in daily life (questions 1 and 2) and factors influencing these attitudes (question 4). To summarize the attitudes regarding environmentally sustainable practices in daily life (questions 1 and 2), an “environmental sustainability attitude in daily life score” was calculated for each individual, awarding one scoring point for every “often” or “always” to the eight sub‐questions in question 1, and for every “yes” to the four sub‐questions of question 2. This score ranged from 0 to 12 and reflects the extent of an individual's environmentally sustainable attitude in daily life, with higher scores indicating stronger environmental sustainability‐oriented attitudes. Group‐level scores were calculated as percentages (ie, with a score of 12 equalling 100%), and the median and interquartile range (IQR) of these percentages were visualized using boxplots.
To further investigate potential differences between groups in attitudes regarding environmentally sustainable practices for leftover medicines in daily life (question 7) and attitudes regarding disposal and redispensing of returned unused medicines by the community pharmacy, bar charts were made to visualize the proportion of answers to each question that indicate more or less environmentally sustainable attitudes.
Latent class analysis was performed using Sawtooth Software's latent class module (version 4.8.2303.1610). All other data analyses were conducted in RStudio (version 4.3.0).
2.6. Ethics and confidentiality
The patient data were handled in strict adherence to the Dutch law on the Protection of Personal Data for Medical Research. The research protocol was approved by the Institutional Review Board (protocol identifier UPF2212) from the Department of Pharmaceutical Sciences, Faculty of Science, Utrecht University.
3. RESULTS
3.1. Study population
Between 14 December 2022 and 31 January 2023, AMP panel members were invited to participate, of whom 6390 responded (24.8%). After excluding incomplete responses (N = 2266; Supporting Information Table S1) and complete responses that did not meet the quality criteria (N = 2), 4122 responses were included for the analysis, representing 16.0% of all invited panel members (Figure 1). The included respondents had a mean age of 68 years (standard deviation 9.8) and 42.6% reported female gender. Most respondents had a high level of education (ie, higher professional or university; 48.2%), lived in a city area (55.4%) and used between one and five different medicines chronically (62.4%) (Table 1).
FIGURE 1.
Flowchart of inclusion of responses in the study. RLH, root likelihood (ie, reflecting the consistency of a respondent's choices in the discrete choice experiment).
TABLE 1.
Characteristics of the study population (N = 4122) and of the four groups identified based on similar preferences for medicine attributes in the discrete choice experiment.
| Characteristics | Study population (N = 4122) | Eco‐focused group (N = 2294) | Cost‐focused group (N = 828) | Indifferent group (N = 677) | Eco‐sceptical group (N = 323) |
|---|---|---|---|---|---|
| Female gender, N (%) | 1758 (42.6) | 994 (43.3) | 313 (37.8) | 317 (46.8) | 134 (41.5) |
| Mean age, years (standard deviation) | 68 (9.8) | 68 (9.4) | 66 (11.0) | 70 (9.3) | 70 (9.0) |
| Educational level, N (%) | |||||
| Low | 699 (17.0) | 318 (13.9) | 157 (19.0) | 151 (22.3) | 73 (22.6) |
| Middle | 1415 (34.3) | 697 (30.4) | 316 (38.2) | 266 (39.3) | 136 (42.1) |
| High | 1985 (48.2) | 1268 (55.3) | 350 (42.3) | 257 (38.0) | 110 (34.1) |
| Missing | 23 (0.6) | 11 (0.5) | 5 (0.6) | 3 (0.4) | 4 (1.2) |
| Living area, N (%) | |||||
| City | 2283 (55.4) | 1269 (55.3) | 462 (55.8) | 373 (55.1) | 179 (55.4) |
| Village | 1826 (44.3) | 1020 (44.4) | 363 (43.8) | 299 (44.2) | 144 (44.6) |
| Other | ‐ | ‐ | ‐ | ‐ | ‐ |
| Missing | 13 (0.3) | 5 (0.2) | 3 (0.4) | 5 (0.7) | ‐ |
| Number of medicines used chronically, N (%) | |||||
| 1‐5 | 2574 (62.4) | 1463 (63.8) | 509 (61.5) | 387 (57.2) | 215 (66.6) |
| 6‐10 | 1224 (30.0) | 640 (27.9) | 261 (31.5) | 236 (34.9) | 87 (26.9) |
| 11‐15 | 166 (4.0) | 96 (4.2) | 29 (3.5) | 29 (4.3) | 12 (3.7) |
| >16 | 29 (0.7) | 15 (0.7) | 3 (0.4) | 8 (1.2) | 3 (0.9) |
| Missing | 129 (3.1) | 80 (3.5) | 26 (3.1) | 17 (2.5) | 6 (1.9) |
3.2. Latent class analysis of the DCE
The latent class analysis indicated four distinct groups of respondents derived from their similar preferences for medicine attributes in the DCE (Supporting Information Figure S1). The four‐group model was selected based on the elbow point in both the AIC and BIC curves, where further improvements in fit levelled off. Additionally, the four groups were conceptually distinct and interpretable, supporting their relevance to the research question. Models with more groups offered minimal improvement in fit.
The mean maximum membership probability was estimated at 89.7%. While each group's preferences for the ease of use of a medicine were similarly low, with utilities (ie, quantified preferences) ranging from 0.36 to 0.55, each group displayed a unique pattern of preferences for the environmental impact and cost of a medicine. These latter two medicine attributes and the related utilities for each group are plotted in Figure 2. Derived from each group's orientation in this plot, the groups were labelled as follows: the eco‐focused group (N = 2294; 55.7%), indicating a preference for medicines with a lower environmental impact; the cost‐focused group (N = 828; 20.1%), indicating a preference for medicines with a lower cost; the indifferent group (N = 677; 16.4%), indicating relatively low preferences overall; and the eco‐sceptical group (N = 323; 7.8%), indicating no preference for medicines with a low environmental impact.
FIGURE 2.
Utilities (ie, quantified preferences) for the medicine attributes cost and environmental impact across the four groups identified in the discrete choice experiment: positive = preferred; negative = not preferred; larger magnitudes = stronger preference.
3.3. Group characteristics
Demographic characteristics were broadly similar across groups, with mean ages ranging from 66 (cost‐focused group) to 70 years (eco‐sceptical group), proportions of female gender ranging from 37.8% (cost‐focused group) to 46.8% (Indifferent group), and proportions of respondents living in a city area between 55.1% (indifferent group) and 55.8% (cost‐focused group). However, the level of education varied notably, with the highest proportion of respondents with higher education in the eco‐focused group (55.3% vs 34.1‐42.3% in the other groups). The proportion of respondents using up to five different medicines chronically ranged from 57.2% (indifferent group) to 66.6% (eco‐sceptical group) (Table 1).
When asked to report their amount of knowledge about the effects of medicines on the environment, the indifferent and eco‐sceptical groups most often indicated good to very good self‐reported knowledge (15.5%), followed by the eco‐focused group (14.1%) and the cost‐focused group (8.8%) (Figure 3).
FIGURE 3.
The self‐reported knowledge about the effects of medicines on the environment. The four groups were derived from similar preferences for medicine attributes in the discrete choice experiment. For the indifferent group, one missing value was identified.
3.4. Environmental sustainability attitude in daily life in general and concerning medicines specifically
Respondents' environmental sustainability attitude in daily life score was highest for the eco‐focused and eco‐sceptical groups (median 75.0%, IQR 58.3‐83.3%), followed by the indifferent group (median 66.7%, IQR 58.3‐83.3%) and the cost‐focused group (median 66.7%, IQR 50.0‐75.0%) (Figure 4A). The most important factors influencing these attitudes of each group are shown in Figure 4B. These factors were concerns about the environment for the eco‐focused group (84.1%), concerns about the environment and personal health for the eco‐sceptical group (75.5% and 75.2%), and costs for the cost‐focused group (78.0%). For the indifferent group, factors were relatively balanced between personal health (71.8%), costs (66.2%) and concerns about the environment (64.1%).
FIGURE 4.
Environmental sustainability attitude in daily life score (A). This score ranged from 0 to 12 and reflects the extent of an individual's environmentally sustainable attitude in daily life, with higher scores indicating stronger environmental sustainability‐oriented attitudes. Group‐level scores were calculated as percentages (ie, with a score of 12 equalling 100%). Factors influencing these attitudes for each group (B). The four groups were derived from similar preferences for medicine attributes in the discrete choice experiment. Respondents were allowed to choose multiple answers.
When zooming in on attitudes regarding environmentally sustainable practices for leftover medicines in daily life (Supporting Information Figure S2) and attitudes regarding disposal and redispensing of returned unused medicines by the community pharmacy (Supporting Information Figure S3), most appeared similar between groups. With regard to leftover medicines, the most common decision across all groups, indicated as “always”, was to return them to the pharmacy. All groups indicated that leftover medicines were nearly never flushed down the toilet or sink. However, differences were observed between groups in the proportion of respondents that retain leftover medicines for potential future use, with 51.1% of the cost‐focused group indicating that this is done sometimes, often or always, followed by 46.8% of the eco‐focused group, 40.2% of the indifferent group and 21.7% of the eco‐sceptical group (Supporting Information Figure S2). In each group, most respondents agreed that it is important to return unused medicines to the pharmacy (64.0‐77.3%), with the highest percentage for the eco‐focused group. However, a different number of respondents considered it important that returned unused medicines are redispensed to other patients, and fewer respondents indicated willingness to use redispensed returned unused medicines themselves. Redispensing of returned unused medicines to other patients was most frequently considered important by the eco‐focused group (81.0%), followed by the cost‐focused (73.6%), indifferent (73.1%) and eco‐sceptical (67.2%) groups. Similarly, willingness to use redispensed returned unused medicines themselves was highest in the eco‐focused group (77.6%), followed by the cost‐focused (70.0%), indifferent (67.2%) and eco‐sceptical (61.0%) groups (Supporting Information Figure S3).
4. DISCUSSION
This study investigated patient preferences and attitudes regarding the environmental impact of medicines relative to cost and ease of use by means of a DCE. Our findings demonstrate that most respondents place a high value on the environmental sustainability of medicines, often exceeding the importance of cost and ease of use. Over half of the respondents (eco‐focused, 55.7%) prioritized low environmental impact above the other attributes, making this the largest of the four groups identified. The remaining groups, cost‐focused (20.1%), indifferent (16.4%) and eco‐sceptical (7.8%), illustrate substantial heterogeneity in how the environmental sustainability and cost is valued as medicine attribute. Overall, respondents expressed positive pro‐environmental attitudes and commonly returned unused medicines to the pharmacy. These findings highlight that, while sustainability is important to most patients, the diversity of preference profiles calls for differential, profile‐aligned actions rather than a single, uniform approach to promoting environmentally sustainable pharmaceutical care. Notably, only a minority of respondents expressed to have good or very good self‐reported knowledge about the effects of medicines on the environment.
Our results suggest a broad preference for environmentally sustainable pharmaceutical care (ie, “greener” pharmacy). This is in line with multiple studies that show an increasing patient preference for environmentally sustainable, patient‐centred care in general 17 , 18 , 37 , 38 , 39 , 40 as well as pharmaceutical care specifically. A Swedish population‐based study showed general preference to medicines with a lower environmental impact and broad support for related policies. There was strong support for eco‐labelling of over‐the‐counter medicines, followed by acceptance of higher prices for environmentally harmful medicines and green prescribing requirements. 17 In addition, a study in the United Kingdom on inhaler users showed that 70% of patients were willing to switch to an inhaler with a lower carbon footprint. 41 However, our analysis of the DCE results identified four groups of patients with distinct preference patterns. Among them, the eco‐focused group stood out as both the largest (ie, comprising more than half of the study population) and most clearly expressing a preference for medicines with a lower environmental impact as well as a broader environmental sustainability attitude in daily life and regarding leftover medicines. Notably, this group showed highest amount of willingness to use redispensed returned unused medicines. While redispensing returned unused medicines may offer environmental and cost benefits, it involves complex practical, regulatory and safety challenges, and cannot be equated with established practices like secure waste management. To support informed decision‐making, the survey clarified that redispensed returned unused medicines would have the same efficacy and safety as newly dispensed medicines. In addition, the eco‐focused group possessed the highest educational level compared to the other groups. This finding aligns with prior research indicating that educational level is the single strongest predictor of climate change awareness. 42 , 43 Furthermore, the eco‐focused group demonstrated a stronger preference for the environmental impact of medicine compared to its cost. These findings align with findings from the Swedish study discussed before, 17 as well as a German survey, in which 86% of respondents favoured climate‐friendly healthcare options and 36% expressed willingness to pay additional costs for CO₂ compensation. 16 In contrast, the cost‐focused group, comprising approximately one‐fifth of the study population, demonstrated the strongest preference for lower cost of medicines relative to other attributes. They also reported the lowest scores on environmental sustainability attitude in daily life, with cost considerations appearing to be the primary factor underlying this attitude. Interestingly, the cost‐focused group comprises the lowest proportion of females and reported the least self‐reported knowledge about the effects of medicines on the environment. These findings are partly in line with previous research. A Dutch study found that females generally hold more environmentally sustainable attitudes than males, 43 while the earlier‐mentioned Swedish study reported a greater affinity for “greener prescribing” among females. 17 Although patients have limited influence over the selection of prescription medicines, increasing interest in the environmental impact of medicines may be reflected through shared decision‐making, offering a pathway to incorporate environmental considerations into pharmacotherapy. 44 , 45
Both the eco‐focused and cost‐focused groups expressed relatively strong preferences, while the indifferent and eco‐sceptical groups—together comprising less than 25% of the study population—expressed weaker preferences, with the eco‐sceptical group exhibiting no preference for medicines with a lower environmental impact. Both groups were more balanced in the factors that influenced their environmental sustainability attitude in daily life, also indicating personal health in addition to concerns about the environment and costs. For the indifferent group, these findings concerning attributes of medicines and environmental sustainability attitude in daily life are relatively in line with each other. However, for the eco‐sceptical group, these findings indicate that one's specific preferences regarding medicine attributes may not necessarily be in line with one's attitude regarding environmental sustainability in daily life. This underlines the complexity of preferences regarding environmentally sustainable use of medicines, which may be guided by more than straightforward “green” motives. For example, the high importance of personal health as a factor for these groups' environmental sustainability attitude in daily life may also influence preferences for medicine use, as shown by the Swedish study in which patients predominantly preferred environmentally friendly medicines for less severe conditions. 17 However, based on our data, the reason for the discrepancy between non‐preference for medicines with a lower environmental impact and a high environmental sustainability attitude in daily life in the eco‐sceptical group remains unclear. The label “eco‐sceptical” was derived from the DCE results and retained for consistency with the labelling of other groups, although alternative labels might better reflect this group's broader environmental perspectives.
Our results also highlight an important self‐reported knowledge gap: only 8.8‐15.5% of respondents reported good to very good self‐reported knowledge of how medicines can affect the environment, with the least self‐reported knowledge indicated by the cost‐focused group. Focused educational initiatives may further shape patient preferences and attitudes regarding environmental sustainability of medicine use. Moreover, it may also inform patients about potential actions to reduce the environmental impact of medicines, including safe disposal practices and redispensing of unused medicines that are returned to the pharmacy. 13 , 14 , 46 However, as shown by the indifferent and eco‐sceptical groups, not every patient may be sensitive to arguments about environmental impact and may thus need other information to be empowered to show environmentally sustainable behaviour regarding medicine use. Notably, while a large majority across all groups in our study returns leftover medicines to the pharmacy for proper disposal, research indicates that this habit varies significantly across the world. 47 Interestingly, respondents in our study found it more important that redispensed returned unused medicines are used by other patients than using them themselves.
A key strength of this research is the inclusion of both a DCE—focused on preferences regarding environmental impact, cost and ease of use of medicines—and survey questions addressing patients' attitudes toward environmental sustainability in daily life, as well as their self‐reported knowledge and attitudes regarding environmentally sustainable practices related to medicines. The considerable alignment between the DCE attributes and patient‐identified attitudes in the broader survey enhances confidence in the robustness of the findings. However, several limitations should be noted. Our sample skewed older and included disproportionally more men, potentially reducing generalizability. To assess the representativeness of our population, we compared key demographic characteristics with data from the Dutch Central Agency for Statistics, which showed that our population was older, included more men and more chronic medicine users, and had a higher educational level than the general Dutch population. 48 In addition, recruiting through the AMP could also introduce selection bias, favouring digitally literate and health engaged individuals. Although the number of responses exceeded our expectation, a broader national sampling approach with a higher response rate would likely yield more representative results. Furthermore, focusing on Dutch respondents may restrict the applicability of our findings to an international context. By restricting DCE scenarios to equally efficacious and safe medicines, we did not capture how patients would trade off environmental impact if it meant compromising other aspects, such as treatment effectiveness. Relatedly, some respondents might have provided socially desirable answers, inflating the apparent level of green preferences, which also do not necessarily reflect actual behaviour as reported before for the Dutch population. 43 Additionally, due to the older age of respondents, this study could not assess a potential association between age and environmentally sustainable medicine preferences. However, the earlier‐mentioned Swedish study reported that older age was linked to a greater affinity for “greener prescribing”. 17 Furthermore, respondents reported their self‐estimated knowledge of the environmental impact of medicines, which is a subjective measure and should be interpreted with caution.
Another possible source of “temporal” bias is the coincidental data collection after the 27th Conference of the Parties to the United Nations Framework Convention on Climate Change. 49 Heightened media attention may have influenced respondents' affinity toward environmentally sustainable behaviour.
Utilities score for ease of use were relatively homogeneous across all groups, probably because the DCE example focused specifically on packaging. This may have led respondents to interpret the attribute narrowly, rather than considering broader aspects of medicine use such as administration. Given this homogeneity, we focused on comparing cost and environmental impact. Furthermore, respondents were not provided with contextual details in the DCE tasks, such as whether the medicine was prescription or over‐the‐counter, or the amount of financial insurance coverage. This could influence how respondents evaluated the cost and environmental attributes. Finally, our primary analytical approach was latent class analysis, which focuses on interpreting group‐level preference profiles rather than estimating statistical uncertainty around utility scores. The absence of confidence intervals means that utility scores should be interpreted with caution.
In conclusion, this study highlights the relevance of environmental aspects in patients' preferences regarding medicines. Taken together, our DCE shows that environmental sustainability already weighs heavily in medicine users' decisions, but preferences are far from uniform. The four distinct groups—eco‐focused, cost‐focused, indifferent and eco‐sceptical—differ markedly in how they trade off environmental impact against cost. This heterogeneity means that differential initiatives (eg, policies, education, interventions), rather than a one‐size‐fits‐all approach, is needed to advance environmentally sustainable pharmaceutical care. Aligning future initiatives with the distinct characteristics and self‐reported knowledge levels of each profile—and closing the evident self‐reported knowledge gaps—will provide a sound, evidence‐based foundation for supporting patients in making environmentally sustainable medicine choices.
AUTHOR CONTRIBUTIONS
Milad Sadreghaemy: Conceptualization; methodology; data curation; formal analysis; writing—original draft; writing—review and editing. Daphne Philbert: Conceptualization; methodology; writing—review and editing. Eibert R. Heerdink: Conceptualization; methodology; writing—review and editing. Marcel L. Bouvy: Conceptualization; methodology; writing—review and editing. Toine C.G. Egberts: Conceptualization; methodology; formal analysis; validation; supervision; writing—review and editing. Lourens T. Bloem: Conceptualization; methodology; data curation; formal analysis; validation; supervision; writing—review and editing.
CONFLICT OF INTEREST STATEMENT
The authors report no conflicts of interest.
DECLARATION OF GENERATIVE AI AND AI‐ASSISTED TECHNOLOGIES IN THE WRITING PROCESS
During the preparation of this manuscript, the first author used large language models (ChatGPT‐4, Claude 3.5 Sonnet) to improve readability. After using these tools, the authors reviewed and edited the content as needed. The authors take full responsibility for the content of the publication.
Supporting information
The survey.
Table S1. Respondent dropout rate for each section of the survey.
Figure S1. Plots used for the elbow assessments of model fit.
Figure S2. Attitudes regarding environmentally sustainable practices for leftover medicines.
Figure S3. Attitudes regarding disposal and redispensing of returned unused medicines by the community pharmacy.
ACKNOWLEDGMENTS
The authors express their gratitude for the support provided by Sawtooth's experts in conducting this research, specifically for their assistance with using the Sawtooth software. The authors also express their gratitude for the support provided by dr. Ellen Koster (University Medical Center Utrecht, Utrecht, the Netherlands).
Sadreghaemy M, Philbert D, Heerdink ER, Bouvy ML, Egberts TCG, Bloem LT. Patient preferences and attitudes regarding the environmental impact of medicines: A discrete choice experiment. Br J Clin Pharmacol. 2025;91(10):2771‐2781. doi: 10.1002/bcp.70244
The authors confirm that the Principle investigator for this paper is Lourens T. Bloem, PhD.
Funding information The researchers received an unconditional grant from Sawtooth Software for using the Sawtooth Software Discover application and CBC Latent Class application to perform the discrete choice experiment.
DATA AVAILABILITY STATEMENT
The research data cannot be shared due to limitations in the permission granted by AMP Onderzoek & Advies in de Zorg.
REFERENCES
- 1. Booth A. Carbon footprint modelling of national health systems: opportunities, challenges and recommendations. Int J Health Plann Manage. 2022. Jul;37(4):1885‐1893. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Or Z, Seppänen AV. The role of the health sector in tackling climate change: a narrative review. Health Policy (New York). 2024;143:105053. [DOI] [PubMed] [Google Scholar]
- 3. Pichler PP, Jaccard IS, Weisz U, Weisz H. International comparison of health care carbon footprints. Environ Res Lett. 2019;14(6):064004. [Google Scholar]
- 4. Nemcova M, Pikula J, Zukal J, Seidlova V. Diclofenac‐induced cytotoxicity in cultured carp leukocytes. Physiol Res. 2020;69(Suppl 4):S607‐S618. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Boxall ABA. The environmental side effects of medication. EMBO Rep. 2004;5(12):1110‐1116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Chow L, Waldron L, Gillings MR. Potential impacts of aquatic pollutants: sub‐clinical antibiotic concentrations induce genome changes and promote antibiotic resistance. Front Microbiol. 2015;6:803. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Polianciuc SI, Gurzău AE, Kiss B, Ştefan MG, Loghin F. Antibiotics in the environment: causes and consequences. Med Pharm Rep. 2020;93(3):231‐240. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ortúzar M, Esterhuizen M, Olicón‐Hernández DR, González‐López J, Aranda E. Pharmaceutical pollution in aquatic environments: a concise review of environmental impacts and bioremediation systems. Front Microbiol. 2022;13:869332. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Gunnarsson L, Snape JR, Verbruggen B, et al. Pharmacology beyond the patient – the environmental risks of human drugs. Environ Int. 2019. Aug;129:320‐332. [DOI] [PubMed] [Google Scholar]
- 10. Law AV, Sakharkar P, Zargarzadeh A, et al. Taking stock of medication wastage: unused medications in US households. Research in Social and Administrative Pharmacy. 2015;11(4):571‐578. [DOI] [PubMed] [Google Scholar]
- 11. Weaver E, O'Hagan C, Lamprou DA. The sustainability of emerging technologies for use in pharmaceutical manufacturing. Expert Opin Drug Deliv. 2022. Jul;19(7):861‐872. [DOI] [PubMed] [Google Scholar]
- 12. Alshemari A, Breen L, Quinn G, Sivarajah U. Can we create a circular pharmaceutical supply chain (CPSC) to reduce medicines waste? Pharmacy (Basel). 2020;8(4):221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Bekker C, van den Bemt B, Egberts TC, Bouvy M, Gardarsdottir H. Willingness of patients to use unused medication returned to the pharmacy by another patient: a cross‐sectional survey. BMJ Open. 2019;9(5):e024767. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Alhamad H, Patel N, Donyai P. How do people conceptualize the reuse of medicines? An interview study. Int J Pharm Pract. 2018. Jun;26(3):232‐241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Cohen ES, Kringos DS, Kouwenberg LHJA, et al. Patient perspectives on climate friendly healthcare: an exploratory study in obstetrics and gynaecology. Patient Educ Couns. 2025;130:108427. [DOI] [PubMed] [Google Scholar]
- 16. Scholz F, Börner N, Schust SA, et al. Focus on patient perspectives in climate action policies for healthcare. A German survey analysis on what patients are willing to do. Front. Public Health. 2024;12:1477313. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Håkonsen H, Dohle S, Rhedin H, Hedenrud T. Preferences for medicines with different environmental impact – a Swedish population‐based study. Environmental Advances. 2023;12:100358. [Google Scholar]
- 18. Drumond N, van Riet‐Nales DA, Karapinar‐Çarkit F, Stegemann S. Patients' appropriateness, acceptability, usability and preferences for pharmaceutical preparations: results from a literature review on clinical evidence. Int J Pharm. 2017;521(1‐2):294‐305. [DOI] [PubMed] [Google Scholar]
- 19. Tseng CW, Waitzfelder BE, Tierney EF, et al. Patients' willingness to discuss trade‐offs to lower their out‐of‐pocket drug costs. Arch Intern Med. 2010;170(16):1502‐1504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Anderson P. Patient preference for and satisfaction with inhaler devices. Eur Respir Rev. 2005;14(96):109‐116. [Google Scholar]
- 21. Hauber AB, González JM, Groothuis‐Oudshoorn CGM, et al. Statistical methods for the analysis of discrete choice experiments: a report of the ISPOR conjoint analysis good research practices task force. Value Health. 2016;19(4):300‐315. [DOI] [PubMed] [Google Scholar]
- 22. Johnson FR, Lancsar E, Marshall D, et al. Constructing experimental designs for discrete‐choice experiments: report of the ISPOR conjoint analysis experimental design good research practices task force. Value Health. 2013;16(1):3‐13. [DOI] [PubMed] [Google Scholar]
- 23. Bridges JFP, Hauber AB, Marshall D, et al. Conjoint analysis applications in health—a checklist: A report of the ISPOR good research practices for conjoint analysis task force. Value Health. 2011;14(4):403‐413. [DOI] [PubMed] [Google Scholar]
- 24. Christian Homburg MKAV. Handbook of market research. Springer International Publishing; 2022. [Google Scholar]
- 25. Apotheekonderzoek [Internet]. [cited 2025. Mar 9]. https://www.apotheekonderzoek.nl/
- 26. Coppes T, Philbert D, van Gelder T, Bouvy ML, Koster ES. Medication management during sick days: no differences between patients with and without impaired renal function. Eur J Clin Invest. 2024;54(9):e14231. [DOI] [PubMed] [Google Scholar]
- 27. Sawtooth Discover Modern Survey Software for Beginners and Professionals [Internet]. [cited 2025. May 8]. https://sawtoothsoftware.com/discover
- 28. van de Pol JM, Heringa M, Koster ES, Bouvy ML. Preferences of patients regarding community pharmacy services: a discrete choice experiment. Health Policy. 2021;125(11):1415‐1420. [DOI] [PubMed] [Google Scholar]
- 29. Software for Hierarchical Bayes Estimation for CBC Data CBC/HB v5. 1999. [cited 2025 Feb 10]. http://www.sawtoothsoftware.com
- 30. Defining Attributes and Levels [Internet]. [cited 2023. Sep 11]. https://sawtoothsoftware.com/help/lighthouse-studio/manual/cva-attributes-and-levels.html
- 31. Ng‐Mak D, Poon JL, Roberts L, Kleinman L, Revicki DA, Rajagopalan K. Patient preferences for important attributes of bipolar depression treatments: a discrete choice experiment. Patient Prefer Adherence. 2018;12:35‐44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. The CBC/HB System Technical Paper V5.6 The CBC/HB System for Hierarchical Bayes Estimation. 2021. [cited 2023 Sep 11]. http://www.sawtoothsoftware.com
- 33. Orme B. Formulating Attributes and Levels in Conjoint Analysis. 2002. [cited 2023 Sep 11]. http://www.sawtoothsoftware.com
- 34. The CBC/HB System Technical Paper V5.6 The CBC/HB System for Hierarchical Bayes Estimation. 2021. [cited 2023 Sep 11]. http://www.sawtoothsoftware.com
- 35. Lezhnina O, Kismihók G. Latent class cluster analysis: selecting the number of clusters. MethodsX. 2022;(9):101747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Sinha P, Calfee CS, Delucchi KL. Practitioner's guide to latent class analysis: methodological considerations and common pitfalls. Crit Care Med. 2021;49(1):e63‐e79. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Cohen ES, Kringos DS, Grandiek F, et al. Patients' attitudes toward integrating environmental sustainability into healthcare decision‐making: an interview study. Health Expect. 2025;28(1):e70155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Ten Buuren AAA, Poolman TB, Bongers MY, et al. Patient preferences for disposable and reusable vaginal specula and their willingness to compromise in the era of climate change: a cross‐sectional study. BJOG. 2024. Apr;131(5):684‐689. [DOI] [PubMed] [Google Scholar]
- 39. Quitmann C, Griesel S, Nayna Schwerdtle P, Danquah I, Herrmann A. Climate‐sensitive health counselling: a scoping review and conceptual framework. Lancet Planet Health. 2023;7(7):e600‐e610. [DOI] [PubMed] [Google Scholar]
- 40. Duxbury K, Alvarez Garcia D, Rutherford S. Public Polling on Climate Change and Health. 2021.
- 41. Rothwell E, McElvaney J, Fitzpatrick A, et al. Evaluating inhaler technique, patient preferences and opportunities for improvement in hospitals in the UK. Future Healthc J. 2024;11(2):100141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Lee TM, Markowitz EM, Howe PD, Ko CY, Leiserowitz AA. Predictors of public climate change awareness and risk perception around the world. Nat Clim Chang. 2015;5(11):1014‐1020. [Google Scholar]
- 43. Kloosterman R, Akkermans M, Reep C, Wingen M, Molnár‐In′t Veld H, Van Beuningen J. Klimaatverandering en energietransitie: opvattingen en gedrag van Nederlanders in 2020. 2021.
- 44. Rostoft S, van den Bos F, Pedersen R, Hamaker ME. Shared decision‐making in older patients with cancer ‐ what does the patient want? J Geriatr Oncol. 2021;12(3):339‐342. [DOI] [PubMed] [Google Scholar]
- 45. Ossin DA, Carter EC, Cartwright R, et al. Shared decision‐making in urology and female pelvic floor medicine and reconstructive surgery. Nat Rev Urol. 2022;19(3):161‐170. [DOI] [PubMed] [Google Scholar]
- 46. Smale EM, Egberts TCG, Heerdink ER, van den Bemt BJF, Bekker CL. Key factors underlying the willingness of patients with cancer to participate in medication redispensing. Res Social Adm Pharm. 2022;18(8):3329‐3337. [DOI] [PubMed] [Google Scholar]
- 47. Paut Kusturica M, Tomas A, Sabo A. Disposal of unused drugs: knowledge and behaviour among people around the world. Rev Environ Contam Toxicol. 2017;240:71‐104. [DOI] [PubMed] [Google Scholar]
- 48. Dashboard bevolking|CBS [Internet]. [cited 2025. Jul 12]. https://www.cbs.nl/nl-nl/visualisaties/dashboard-bevolking
- 49. Sharm el‐Sheikh Climate Change Conference ‐ November 2022|UNFCCC [Internet]. [cited 2025. Jul 17]. https://unfccc.int/cop27
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
The survey.
Table S1. Respondent dropout rate for each section of the survey.
Figure S1. Plots used for the elbow assessments of model fit.
Figure S2. Attitudes regarding environmentally sustainable practices for leftover medicines.
Figure S3. Attitudes regarding disposal and redispensing of returned unused medicines by the community pharmacy.
Data Availability Statement
The research data cannot be shared due to limitations in the permission granted by AMP Onderzoek & Advies in de Zorg.