Segmentation of the Pharmaceutical Market in Ecological Marketing

D.V. Babaskin¹*, A.N. Voronin¹, E.E. Ilycheva¹, T.M. Litvinova¹, L.I. Babaskina¹,

O.V. Savinova¹, T.I. Okonenko², G.Ya. Ibragimova³, G.K. Akhmadullina²

¹Sechenov First Moscow State Medical University, 8-2 Trubetskaya str., Moscow, 119991, Russian Federation.

²Yaroslav-the-Wise Novgorod State University,

41 Sankt-Peterburgskaya str., Velikiy Novgorod, 173003, Russian Federation.

³Bashkir State Medical University, 3 Lenina str., Ufa, 450008, Republic of Bashkortostan, Russian Federation.

*Corresponding Author E-mail: babaskin.d.v@mail.ru

ABSTRACT:

For the effective implementation of environmental measures related to the disposal of pharmaceuticals at the stage of their use, high-quality market segmentation and the selection of target segments are required. This study aims to segment the pharmaceutical market using artificial intelligence (AI) models and traditional methods to enhance environmental protection against the impact of unused and expired pharmaceuticals. Materials and Methods. Four AI models were used: ChatGPT-4o (W1), DeepSeek-V3 with Deep Think (R1) mode (W2), Qwen 2.5 Max Large Language Model (W3), and Perplexity PRO (W4). Traditional segmentation was performed using a complex faceted multifactor method. The obtained results were analyzed using the expert evaluation method based on the Likert scale, according to the following indicators: obligatoriness (A) and constructiveness (B) of the characteristic; necessity and sufficiency of the number of variables for the characteristic (C) and their constructiveness (D); segment attractiveness (E) and its alignment with the strengths of the implemented environmental measures (F). Results. When segmenting the market using the four AI models, 6 to 9 characteristics and their variables were selected (the total number of segments in each model ranged from 648 to 26,244). The highest frequency of positive expert evaluations for indicators A–D was observed in models W3 and W4. When selecting the key target segment (indicators E and F), segments with average values were prioritized, which, according to experts, was not always justified. Traditional market segmentation was conducted based on three characteristics (the most significant ones according to the AI models) and their variables, resulting in a total of 36 segments. The results of the expert evaluation demonstrated the comparability of data on similar key target segments between the traditional method and model W4. Conclusion. The use of AI models for market segmentation allows for an expanded range of segmentation characteristics and criteria for selecting the key target segment. However, special attention should be paid to the accurate and precise formulation of the task (prompt) and the careful adjustment of the number of resulting target segments.

KEYWORDS: Pharmaceutical market, Market segmentation, Target segments, Ecological marketing, Pharmaceutical ecology.

INTRODUCTION:

In recent years, the role of ecological marketing in pharmacy has significantly increased as a type of activity aimed at meeting the needs and demands of the target market segment for pharmaceutical products and services, while ensuring environmental safety^1-4. This marketing is an integral part of ecological marketing in healthcare and is closely linked to medical products and services⁵. The increasing requirements of environmental protection legislation in the Russian Federation have necessitated the further development, implementation, and evaluation of measures aimed at protecting the environment from the impact of pharmaceuticals not only during their production, storage, wholesale, and retail trade, but also during their use. Not all purchased drugs are fully used by consumers. This may be due to side effects, non-compliance with the dosage regimen, poor adherence to treatment, changes in the method of use or dosage, improvement in condition after illness, expiration of the shelf life, or the influence of manufacturers on advertising^6-8. Such pharmaceutical products need to be disposed of or destroyed. However, consumers are not always fully informed about how to dispose of them^9,10. As a result of improper disposal, pharmaceuticals enter the environment, potentially contaminating soil, water bodies, and aquifers^6-8,11-14. This ultimately has a detrimental impact on the health of humans and animals^9,10,14-18. In this regard, an urgent task is the development of effective environmental measures aimed at solving this problem in relation to pharmaceutical consumers. The success of such initiatives largely depends on their alignment with the current capabilities and demands of consumers. This, in turn, dictates the need for high-quality market segmentation and the selection of target segments. The accuracy with which the boundaries of the target market are defined will determine the detail of its components, such as accessibility, capacity, demand and supply levels, market penetration, scale of operations, resource distribution, and the potential for effective functioning both now and in the future¹⁹.

The scientific literature fully addresses the issues of market segmentation for pharmaceutical and medical products and services^20-22. However, there are very few studies on defining the target market for conducting environmental measures related to unused and expired pharmaceuticals. These studies are usually fragmented and not always scientifically or empirically substantiated²³. Moreover, traditional segmentation methods are time-consuming and labor-intensive, which makes the use of artificial intelligence (AI) technologies appealing^24-26. The application of well-known AI models in pharmaceutical ecology will allow for the optimization of ongoing initiatives and will contribute to the formation of environmentally responsible behavior among consumers regarding unused and expired pharmaceuticals. Therefore, it is evident that addressing the issues of pharmaceutical market segmentation and selecting the key target segment for conducting environmental measures is highly relevant.

The goal of this study is to conduct pharmaceutical market segmentation using artificial intelligence models and traditional methods in the context of environmental measures aimed at protecting the environment from the impact of unused and expired pharmaceuticals.

MATERIALS AND METHODS:

Market segmentation was carried out with respect to the end consumers of pharmaceutical products. In marketing, end consumers are defined as actual and potential consumers who purchase products (goods, services) or would like to purchase them exclusively for personal use and to meet specific needs^27,28.

Market segmentation of pharmaceutical products was performed using four AI models: ChatGPT-4o, DeepSeek-V3 with Deep Think (R1) mode, Qwen 2.5 Max Large Language Model, and Perplexity PRO. The main task for the AI (prompt) included: selecting the most necessary characteristics and their variables (values) for market segmentation (consumers) of pharmaceutical products with the goal of motivating and engaging them in environmental measures related to unused and expired pharmaceuticals; determining the key target segment of the market. The models used for segmentation involved: a tokenization system – a system that allows text to be transformed into logically meaningful units (tokens), such as words, phrases, and categories^29,30; the Bayesian updating principle of behavioral probability changes; the principle of triads (division into three levels); combinatorics of segments; checking for the independence of characteristics and their variables; working with discrete variables. In selecting the key target segment, logical inference based on a probabilistic model was applied, along with criteria such as relevance, motivation, capacity, accessibility, economic efficiency, and others.

Traditional market segmentation was carried out using a complex faceted multifactor method.

The analysis of the obtained results was conducted by 19 experts in the fields of pharmacy (47.4%), medicine (26.3%), ecology (10.5%), marketing (10.5%), and other specialties (5.3%). The competence level (Kk) of each expert candidate was initially determined^31,32. It was conditionally accepted that the competence of the experts should not be lower than the average level (Kk ≥ 0.4). The experts' work experience in the specialty should be at least 3 years. The majority of experts had work experience of 10 years or more (57.9%). All experts had experience in conducting activities related to ecology (100%).

The expert survey was conducted remotely via email, using the expert evaluation map developed by us. The map contained 3 questions regarding the general characteristics of the expert (specialty, work experience in the specialty, experience in conducting environmental measures related to unused and expired pharmaceuticals), and 2 tasks: assessing the characteristics of segmentation, their variables, and the key target segment when using AI models; evaluating target segments for the purpose of selecting the key one in traditional segmentation methods. Experts expressed their opinions based on the following indicators: obligatoriness of the characteristic (A); constructiveness of the characteristic (B), which determines the accuracy, correctness, and clarity of its formulation; necessity and sufficiency of the number of variables for the characteristic (C); constructiveness of the variables for the characteristic (D); attractiveness of the key target segment (E) and its alignment with the strengths of the conducted environmental measures (F). The evaluation was conducted using the Likert scale. The "positive evaluation" group included "completely corresponds" and "rather corresponds"; the "neutral evaluation" group included "equally corresponds and does not correspond"; the "negative evaluation" group included "rather does not correspond" and "categorically does not correspond." Frequency analysis of the expert evaluation results was carried out for each characteristic and each indicator.

Statistical data processing was performed using the IBM SPSS Statistics 29.0.1 software. The average frequency of positive evaluations (N_avg) is presented as M±σ (M – arithmetic mean, σ – standard deviation). The reliability of differences between N_avg was assessed using the Student’s t-test. The critical significance level for testing statistical hypotheses in the study was set at 0.05.

RESULTS AND DISCUSSION:

Market segmentation using artificial intelligence models

Table 1 presents the results of pharmaceutical market segmentation using 4 AI models (W1-W4) in relation to environmental measures for protecting the environment from the impact of unused and expired pharmaceuticals.

Table 1. Criteria and results of pharmaceutical market segmentation using AI models in relation to environmental measures for unused and expired pharmaceuticals

AI Model, W	Segmentation Characteristic, S	Number of Variables, Р_i	Number of Segments, C*	Key Target Segment
W1 ChatGPT-4o	S1.1. Environmental awareness	3	26244	Environmental awareness: average. Behavioral habits: throwing away in the trash. Level of trust: average. Age: middle-aged (31-50 years). Marital status: families with children. Availability of infrastructure: average. Willingness to change habits: average. Frequency of use: average. Sources of information: internet and social media
	S1.2. Behavioral habits of pharmaceutical disposal	3
	S1.3. Level of trust in environmental initiatives	3
	S1.4. Age	3
	S1.5. Marital status	3
	S1.6. Availability of infrastructure for disposal	3
	S1.7. Willingness to change habits for the sake of the environment	3
	S1.8. Frequency of pharmaceutical use	3
	S1.9. Sources of information on pharmaceutical disposal	4
W2 DeepSeek-V3 with Deep Think (R1) mode	S2.1. Method of pharmaceutical disposal	4	648	Method of disposal: throwing away in the trash. Awareness of the problem: average. Attitude towards the environment: moderately interested. Availability of collection points: none. Household waste sorting: no. Frequency of accumulation: regular.
	S2.2. Awareness of the problem	3
	S2.3. Attitude towards the environment	3
	S2.4. Availability of collection points for unused and expired pharmaceuticals in the vicinity	3
	S2.5. Household waste sorting	2
	S2.6. Frequency of accumulation of pharmaceuticals	3
W3 Qwen 2.5 Max Large Language Model	S3.1. Age	3	864	Age: elderly (50+ years). Gender: female. Income level: average. Frequency of purchasing: frequent. Awareness of disposal: average. Disposal habits: responsible. Place of residence: city. Expenses on pharmaceuticals: average
	S3.2. Gender	2
	S3.3. Income level	3
	S3.4. Frequency of purchasing pharmaceuticals	2
	S3.5. Awareness of pharmaceutical disposal	2
	S3.6. Pharmaceutical disposal habits	2
	S3.7. Place of residence	2
	S3.8. Expenses on pharmaceuticals	3
W4 Perplexity PRO	S4.1. Awareness of the issue	3	2187	Awareness: high. Behavior: situational. Interest: high. Age: middle-aged (35-54). Socioeconomic status: average. Frequency of use: frequent. Willingness to participate: ready to participate actively (volunteering, information dissemination)
	S4.2. Behavior in disposing of unused and expired medications	3
	S4.3. Interest in environmental initiatives	3
	S4.4. Age	3
	S4.5. Socioeconomic status	3
	S4.6. Frequency of medication use	3
	S4.7. Willingness to participate in environmental initiatives	3
* C = P₁ • P₂ • P₃ • … • P_n

Despite the fact that the main task specified the selection of only the most essential segmentation characteristics, the total number of characteristics across all models amounted to 30, with the number of characteristics in each model varying from 6 to 9. As expected, many characteristics were very similar and often repeated. For example, all AI models included the characteristic "Environmental awareness" (S1.1) and its derivatives: "Awareness of the issue" (S2.2 and S4.1) and "Awareness of medication disposal" (S3.5). The characteristic "Behavioral habits of medication disposal" (S1.2) also appeared with its derivatives: "Medication disposal habits" (S3.6), "Behavior in disposing of unused and expired medications" (S4.2), and "Method of medication disposal" (S2.1). Additionally, three models repeated the characteristic "Age" (S1.4, S3.1, S4.4), "Frequency of medication use" (S1.8, S4.6) and its derivative "Frequency of purchasing medications" (S3.4). The characteristic "Socioeconomic status" (S4.5) was broad enough to encompass several other characteristics, such as "Family status" (S1.5), "Place of residence" (S3.7), and "Gender" (S3.2).

The number of variables for each characteristic across all models ranged from 2 to 4, which helped reduce the total number of segments somewhat. However, despite this, the W1 model still produced more than 26,000 segments. It should be noted that some variables had the same names but differed in content. For example, the characteristic "Age" included three variables: "Young" (18-30 years in the W1 model, 18-25 years in the W3 model, 18-34 years in the W4 model), "Middle-aged" (31-50 years in the W1 model, 26-50 years in the W3 model, 35-54 years in the W4 model), and "Elderly" (50+ in the W1 and W3 models, 55+ in the W4 model).

Based on the AI's selection of the key target market segment, consumer profiles were created for each model regarding the implementation of environmental activities related to unused and expired pharmaceutical products. For example, according to the W3 model, this includes middle-aged women living in urban areas, with an average income level, who frequently purchase pharmaceutical products with average spending on them, possess moderate awareness of the disposal of pharmaceuticals, and have a responsible attitude towards it.

To verify and evaluate the results of market segmentation of pharmaceutical products using four AI models, a survey of expert specialists was conducted.

Expert evaluation of the market segmentation results using artificial intelligence models:

Table 2 presents the results of the expert evaluation of the proposed AI segmentation features, their variables, and key target segments related to environmental activities aimed at protecting the environment from the impact of unused and expired pharmaceutical products.

Table 2. Expert evaluation of the criteria and results of pharmaceutical market segmentation using AI models in relation to the implementation of environmental activities concerning unused and expired pharmaceutical products

Segmenta-tion Characteris-tic	Frequency of Positive Ratings (N, %)						*Average Frequency of Positive Ratings in the AI Model (N_avg, %)*
	Rating Indicator
	А	В	С	D	E	F
S1.1	73.7	42.1	84.2	94.7	52.6	42.1	W1 67.6±24.1
S1.2	94.7	84.2	21.1	47.4	31.6	26.3
S1.3	63.2	79.0	89.5	89.5	47.4	42.1
S1.4	94.7	100	84.2	31.6	84.2	57.9
S1.5	84.2	94.7	68.4	84.2	89.5	68.4
S1.6	89.5	73.7	94.7	94.7	52.6	52.6
S1.7	84.2	68.4	89.5	89.5	47.4	42.1
S1.8	94.7	94.7	89.5	79.0	52.6	42.1
S1.9	73.7	21.1	15.8	26.3	79.0	57.9
S1.1-S1.9	83.6± 11.3	73.1±26.2	70.8±30.6	70.8±27.7	59.7±19.7	47.9± 12.5	Average frequency of positive ratings for indicators A, B, C, D, E, F in the W1 model (N_avg, %)
S2.1	89.5	68.4	21.1	47.4	31.6	26.3	W2 59.7±27.4
S2.2	79.0	79.0	94.7	94.7	52.6	42.1
S2.3	52.6	21.1	89.5	52.6	47.4	42.1
S2.4	89.5	89.5	15.8	63.2	26.3	10.5
S2.5	89.5	68.4	94.7	94.7	31.6	10.5
S2.6	94.7	73.7	89.5	36.8	79.0	57.9
S2.1-S2.6	82.5±15.5	66.7±23.7	67.5±38.1	64.9±24.6	44.8±19.6	31.6± 19.1	Average frequency of positive ratings for indicators A, B, C, D, E, F in the W2 model (N_avg, %)
S3.1	94.7	100	84.2	42.1	84.2	57.9	W3 77.2±19.8
S3.2	73.7	100	100	100	52.6	52.6
S3.3	57.9	89.5	94.7	94.7	52.6	52.6
S3.4	94.7	84.2	52.6	89.5	94.7	94.7
S3.5	89.5	94.7	15.8	26.3	52.6	42.1
S3.6	94.7	84.2	52.6	73.7	89.5	94.7
S3.7	57.9	84.2	100	100	73.7	63.2
S3.8	57.9	84.2	94.7	94.7	52.6	52.6
S3.1-S3.8	77.6± 17.7	90.1± 7.1	77.0± 25.5	85.5± 19.4	69.1± 18.6	63.8± 20.0	Average frequency of positive ratings for indicators A, B, C, D, E, F in the W3 model (N_avg, %)
S4.1	79.0	79.0	94.7	94.7	84.2	94.7	W4 81.7±17.7
S4.2	94.7	94.7	52.6	84.2	89.5	94.7
S4.3	79.0	68.4	89.5	89.5	89.5	94.7
S4.4	94.7	100	84.2	31.6	84.2	57.9
S4.5	73.7	26.3	68.4	68.4	52.6	52.6
S4.6	94.7	94.7	89.5	89.5	94.7	94.7
S4.7	94.7	89.5	89.5	63.2	94.7	100
S4.1-S4.7	87.2± 9.5	78.9± 25.6	81.2± 15.2	74.4± 22.2	84.2± 14.6	84.2± 19.9	Average frequency of positive ratings for indicators A, B, C, D, E, F in the W4 model (N_avg, %)

According to the experts, the selected AI segmentation characteristics (indicator A) were deemed necessary, absolutely required, and in demand across all models (N_avgA = 82.7%). The most significant and frequently occurring characteristics and their derivatives were: “Frequency of pharmaceutical purchases” (S1.8, S2.6, S3.4, S4.6; N_avgA = 94.7%); “Behavior in the disposal of unused and expired pharmaceutical products” (S1.2, S2.1, S3.6, S4.2; N_avgA = 93.4%); “Age” (S1.4, S3.1, S4.4; N_avgA = 94.7%). The constructiveness of the examined segmentation characteristics (indicator B) varied significantly (from 21.1% to 100%). The highest expert ratings were for the W3 model characteristics (N_avgB = 90.1%), while the lowest were for W2 (N_avgB = 66.7%). It should be noted that the W2 model had the lowest results in terms of both the number of variables per feature (indicator C, N_avgC = 67.5%) and their constructiveness (indicator D, N_avgD = 64.9%). Moreover, all experts rated the W1 model characteristic “Sources of information on pharmaceutical disposal” (S1.9) poorly in terms of both indicators C and D (N_avgC = 15.8% and N_avgD = 26.3%; frequency ratio of ratings by groups, in %: for indicator C – 52.6:31.6:15.8, for indicator D – 31.6:42.1:26.3) and the W3 model feature “Awareness of pharmaceutical disposal” (S3.5) (N_avgC = 15.8% and N_avgD = 26.3%; frequency ratio of ratings by groups, in %: for indicator C – 26.3:57.9:15.8, for indicator D – 10.5:63.2:26.3). It is important to emphasize that no expert rated “categorically does not match”.

The lowest results in selecting the key target segment (indicators E and F) were observed in the W1 and W2 models. Notably, there was a predominance of negative ratings for the features S2.4 and S2.5 concerning this segment as the most aligned with the strengths of the environmental initiatives being conducted (N_F = 10.5%; frequency ratio of ratings by groups, for indicator F, in %: for feature S2.4 – 52.6:36.9:10.5, for feature S2.5 – 47.4:42.1:10.5). In the W3 model, indicators E and F also had low values within the model (N_avgE = 69.1%, N_avgF = 63.8%).

The expert evaluation of the pharmaceutical market segmentation results using AI models showed that the W3 (N_avgW3 = 77.2±19.8%) and W4 (N_avgW4 = 81.7±17.7%) models were significantly more preferred for segmentation than the W1 (N_avgW1 = 67.6±24.1%) and W2 (N_avgW2 = 59.7±27.4%) models (p < 0.05). When selecting segmentation characteristics (indicators A and B) and their variables (indicators C and D), all the AI models studied can be used, as the average frequency of positive ratings prevailed over the others. The large number of segments obtained from segmentation (648-26,244) requires adjustment, as they need to be analyzed in the next stage to identify the key target segment. To regulate the number of segments obtained, it is necessary to clearly define the boundaries in the task (prompt) regarding the number of characteristics and their variables. In our view, it is better to use AI models for the preliminary selection of characteristics and their derivatives, followed by the establishment and use of the most significant ones. As for determining the key target segment (indicators E and F), the average values of variables were selected for many features, for example, in the W1 model – S1.1, S1.3, S1.4, S1.6, S1.7, S1.8, or in the W3 model – S3.3, S3.5, S3.8. According to experts, this was not always justified, so the average frequency of positive ratings for these features in indicators E and F across the W1-W3 models ranged from 30-60%. Only in the W4 model were the ratings for these indicators above 80% (N_avgE = 84.2±14.6%, N_avgF = 84.2±19.9%). To provide a more detailed justification for selecting the key target segment in relation to environmental activities regarding unused and expired pharmaceutical products, we used the traditional segmentation method.

Market segmentation using the traditional method:

When determining the segmentation characteristics and their derivatives, the results of the previous AI-based study were used. Limitations were introduced on the number of characteristics (no more than 3) and the number of their variables (no more than 4). This was due, on the one hand, to the difficulties of establishing a single key target segment considering all the characteristics and their variables, and on the other hand, to the need to select a segment that is sufficiently large and accessible to ensure a significant impact, and in which the capabilities of pharmaceutical consumers considering disposal, but lacking a convenient method for its implementation, are taken into account. As a result, 3 segmentation characteristics were selected, which were the most significant and frequently occurring according to expert evaluation: age; behavior in the disposal of unused and expired pharmaceutical products; frequency of pharmaceutical purchases (Table 3).

According to the experts, the most attractive (indicator E) and aligned with the strengths of the environmental initiatives (indicator F) was the target segment T3 (N_avg=92.1±3.7%). This segment consists of middle-aged consumers (45 to 60 years old) who frequently purchase pharmaceuticals (every month or more often) and dispose of unused and expired products depending on the circumstances (with or without compliance to disposal regulations). The results for segment T3 on indicators E and F were comparable to similar data for the closest key target segment in model M4 (N_avg=84.2±16.8%) and did not differ significantly from segment T4 (N_avg=84.3±7.4%) (p>0.05). Although target segments T3 and T4 differ only in terms of the age of consumers, the base segment of middle-aged people (Y2) is larger than the segment of elderly people (Y3)³⁴, but in terms of the frequency of pharmaceutical purchases under variable U1, the elderly significantly outnumber the middle-aged group^20,35. This, in our opinion, explains the closeness of the results for segments T3 and T4. Target segments T1 and T2 included responsible individuals in terms of disposal of unused and expired pharmaceutical products (R1), i.e., a loyal audience, and the segmentation aimed at finding those who could be motivated and encouraged to participate in environmental activities (this goal is reflected in the prompt). Consumers who dispose of products depending on circumstances (R2), meaning they do not have established habits and are not engaged in the process, but could be involved if convenient solutions are offered, are easier to change than irresponsible individuals (R3). Therefore, target segment T3 (N_avg=92.1±3.7%) significantly outperformed segments T5 and T6 (N_avg=60.6±3.8%) on indicators E and F (p<0.05).

Traditional pharmaceutical market segmentation showed that even using a limited number of segmentation features and their variables does not guarantee the reliability of selecting only one key target segment. The application of AI models in market segmentation allows for expanding the boundaries in terms of the number of segmentation features and their variables, as well as the requirements for selecting the key target segment. However, it is necessary to correctly and precisely formulate the purpose and tasks of segmentation in the prompt, identify key words, clarify as many details as possible, use simple words and phrases, provide references, and double-check the model's responses.

CONCLUSIONS:

1. Market segmentation using 4 AI models for environmental initiatives aimed at protecting the environment from the impact of unused and expired pharmaceuticals was conducted based on 6-9 characteristics and 2-4 of their variables. The models W3 (N_avg=82.6±18.3%) and W4 (N_avg=80.5±18.7%) had the highest frequency of positive expert ratings according to indicators A-D. The numerous segments resulting from segmentation (648-26,244) significantly complicated the selection of the key target segment for further environmental activities. Many of the characteristics had their average values selected, which, according to experts, was not always justified (in models W1-W3, the average frequency of positive ratings from experts for indicators E and F ranged from 30-60%). To provide more detailed justification for the selection of the key target segment, market segmentation was conducted using the traditional method.

2. When segmenting the market using the traditional method, restrictions were introduced on the number of features and variables, which significantly reduced the total number of segments to 36. The results of the expert evaluation of the most significant segments (T1-T6) regarding environmental initiatives showed that the most attractive (indicator E) and aligned with the strengths of environmental activities (indicator F) was the target segment T3 (N_avg=92.1±3.7%). The results obtained for this segment were comparable with the similar data for the most closely related segment from model W4 (N_avg=84.2±16.8%) and did not significantly differ from segment T4 (N_avg=84.3±7.4%) (p>0.05). It was shown that even the use of a limited number of features and variables does not guarantee the reliability of selecting only one key target segment.

The use of AI models in market segmentation allows for expanding the scope regarding the number of segmentation characteristics and the requirements for selecting the key target segment. However, it is essential to formulate the task (prompt) correctly and accurately.

REFERENCES:

1. Bagnyuk DV, Lebedev AV. Conceptual foundations of environmental marketing and environmental product. Economics and Management. 2020; 26(6): 665-71. (In Russ.) doi: 10.35854/1998-1627-2020-6-665-671

2. Malychev AA. The development of environmental marketing in the enterprise. Russian Journal of Resources, Conservation and Recycling. 2019; 6(1): 1-12. (In Russ.) doi: 10.15862/05ECOR119

3. Mishacheva E. Ecological marketing. Bulletin of Science and Practice. 2023; 9(3): 295-9. (In Russ.) doi: 10.33619/2414-2948/88/35

4. Shaukat F, Ming J. Green marketing orientation impact on business performance: Case of pharmaceutical industry of Pakistan. Frontiers in Psychology. 2022; 13: 940278. doi: 10.3389/fpsyg.2022.940278

5. Sinyavets TD. Prospects of ecological marketing in Russia. Zhurnal Issledovaniy po Upravleniyu. 2020; 6(3): 23-32. (In Russ.)

6. Barnett-Itzhaki Z, Berman T, Grotto I, Schwartzberg E. Household medical waste disposal policy in Israel. Israel Journal of Health Policy Research. 2016 Oct 10; 5:48. doi: 10.1186/s13584-016-0108-1

7. Gidey MT, Birhanu AH, Tsadik AG, Welie AG, Assefa BT. Knowledge, attitude, and practice of unused and expired medication disposal among patients visiting Ayder Comprehensive Specialized Hospital. BioMed Research International. 2020 Aug 24; 2020:9538127. doi: 10.1155/2020/9538127

8. Habte HY, Bahta M, Russom N, Kibreab F, Andemariam A, Tewelde T, Russom M. Household pharmaceutical disposal practices, community understanding, and readiness for medicines take-back in Asmara, Eritrea: A cross-sectional analysis. Risk Management and Healthcare Policy. 2024 Nov 11; 17:2777-88. doi: 10.2147/RMHP.S478511

9. Michael I, Ogbonna B, Sunday N, Anetoh M, Matthew O. Assessment of disposal practices of expired and unused medications among community pharmacies in Anambra State southeast Nigeria: A mixed study design. Journal of Pharmaceutical Policy and Practice. 2019; 12: 12. doi: 10.1186/s40545-019-0174-1

10. Nairat LL, Abahri NA, Hamdan YA, Abdel-Khaliq RT, Odeh SM, Abutaha S, Al-Jabi SW, Koni AA, Abushanab AS, Zyoud SH. Assessment of practices and awareness regarding the disposal of unwanted pharmaceutical products among community pharmacies: A cross-sectional study in Palestine. BMC Health Services Research. 2023; Sep 27; 23(1): 1035. doi: 10.1186/s12913-023-09888-5

11. Tegegne AA, Genet G, Limenh LW, Yohannes L, Mohammed Seid A, Alemayehu TT, Ayenew W, Simegn W. Public awareness, knowledge, and attitude regarding proper disposal of unused medicines and associated factors in Gondar city, northwest Ethiopia. Frontires in Public Health. 2024 Jun 12; 12:1372739. doi: 10.3389/fpubh.2024.1372739

12. Yimer A, Moges G, Kahissay MH. Household storage and disposal of unused and expired medicines in Dessie, Ethiopia: A cross-sectional study. Frontiers in Public Health. 2024 Oct 24; 12:1422304. doi: 10.3389/fpubh.2024.1422304

13. Giunchi V, Fusaroli M, Linder E, Villén J, Wettermark B, Nekoro M, Raschi E, Lunghi C, Poluzzi E. The environmental impact of pharmaceuticals in Italy: Integrating healthcare and eco-toxicological data to assess and potentially mitigate their diffusion to water supplies. British Journal of Clinical Pharmacology. 2023; Jul; 89(7): 2020-7. doi: 10.1111/bcp.15761

14. Nepal S, Giri A, Bhandari R, Chand S, Nepal S, Aryal S, Khanal P, Moktan JB, Shastry CS. Poor and unsatisfactory disposal of expired and unused pharmaceuticals: A global issue. Current Drug Safety. 2020; 15(3): 167-72. doi: 10.2174/1574886315666200626164001

15. Kaya A, Genc F. Rational drug use and the disposal cost of unused medications within house; a descriptive-cross-sectional study. BMC Public Health. 2025; Feb 14; 25(1): 622. doi: 10.1186/s12889-025-21501-4

16. Kоksoy S. Unused, expired pharmaceuticals and their disposal practices among the general public in Burdur-Turkiye: A cross-sectional study. BMC Public Health. 2024; May 13; 24(1): 1303. doi: 10.1186/s12889-024-18788-0

17. Rogowska J, Zimmermann A. Household pharmaceutical waste disposal as a global problem-a review. International Journal of Environmental Research and Public Health. 2022; Nov 27; 19(23): 15798. doi: 10.3390/ijerph192315798

18. Singh A, Pratap SG, Raj A. Occurrence and dissemination of antibiotics and antibiotic resistance in aquatic environment and its ecological implications: A review. Environmental Science and Pollution Research. 2024; Jul; 31(35): 47505-29. doi: 10.1007/s11356-024-34355-x

19. Kumar V, Kotler P. Transformative marketing: Combining new age technologies and human insights. Palgrave Macmillan. 2024; 412 p.

20. Analysis of the pharmaceutical market in Russia - 2025. Indicators and forecasts. TEBIZ GROUP. 2025; 280 p. (In Russ.) Available at: https://tebiz.ru/assets/pdf/mi/rynok-lekarstvennykh-preparatov-v-rossii.pdf

21. Babaskina LI, Litvinova TM, Babaskin DV, Veber VR, Okonenko TI. Identifying target markets in the development of rehabilitation methods for the patients based on phytocomplex phonophoresis method. Journal of Pharmaceutical Sciences and Research. 2018; 10(7): 1814-7.

22. Petrov AL, Andrianova GN, Mashkovtseva NA, Ponomarev GA. Methodical approaches to the search for promising segments of the pharmaceutical market for import substitution in the aggregate assortment portfolio of pharmaceutical plants in the Urals Federal District. Ural Medical Journal. 2018; 4: 136-40. (In Russ.)

23. Naini SR, Mekapothula RR, Jain R, Manohar S. Redefining green consumerism: A diminutive approach to market segmentation for sustainability. Environmental Science and Pollution Research. 2024; Feb; 31(9): 12916-32. doi: 10.1007/s11356-023-31717-9

24. Chang YT, Fan NH. A novel approach to market segmentation selection using artificial intelligence techniques. The Journal of Supercomputing. 2023; 79: 1235-62. doi: 10.1007/s11227-022-04666-2

25. Kotler F, Kartajaya H, Setiawan A. Marketing 6.0: Future of immersive, merging digital and physical worlds. Trans. from English. BOMBORA, Moscow. 2024; 212 p. (In Russ.)

26. Wang G. Customer segmentation in the digital marketing using a Q-learning based differential evolution algorithm integrated with K-means clustering. PLoS ONE. 2025; 20(2): e0318519. doi: 10.1371/journal.pone.0318519

27. Hoffmann S, Akbar A. Consumer behavior: Understanding consumers – Designing marketing activities. Springer Wiesbaden, Wiesbaden. 2023; 185 p. doi: 10.1007/978-3-658-39476-9

28. End consumer. The Great Russian Encyclopedia. (In Russ.) Available at: https://bigenc.ru/c/konechnyi-potrebitel-e9aa87?ysclid=m7il5ibgho709735905

29. Andryushin SA. Tokenization of real assets: Classification, platforms, applications, opportunities and challenges of development. Russian Journal of Economics and Law. 2024; 18(1): 88-104. (In Russ.) doi: 10.21202/2782-2923.2024.1.88-104

30. Popov V, Romanenko E. Web 3.0: Tokenization, or the world of tomorrow today. Ridero, Yekaterinburg. 2021; 210 p. (In Russ.)

31. GOST R 56041-2014. Conformity assessment. Requirements for certification experts of products, services, processes. Standartinform, Moscow. 2019; 7 p. (In Russ.)

32. Vasnetsova OA, Babaskin DV. Editors. Medical and pharmaceutical merchandising. Practicum. Federal State Autonomous Educational Institution of Higher Education First Moscow State Medical University named after I.M. Sechenov Ministry of Health of the Russian Federation (Sechenov University), Moscow. 2017; 324 p. (In Russ.)

33. Medicines and drugs. How often do Russians buy medicines? What part of the family budget is spent on medicines? FOM: Fond "Obshchestvennoe mnenie". April 20, 2023. (In Russ.) Available at: https://fom.ru/Zdorove-i-sport/14863?ysclid=m7q0itet1k205384626

34. Population of the Russian Federation by gender and age. Federal State Statistics Service. (In Russ.) Available at: https://rosstat.gov.ru/compendium/document/13284

35. Mukhina SM, Baranova MI, Gomon YuM. Drug utilization research at the individual level as the main concept of pharmacoepidemiology. Real-World Data and Evidence. 2023; 3(2): 21-30. (In Russ.) doi: 10.37489/2782-3784-myrwd-33

Received on 07.04.2025 Revised on 14.08.2025

Accepted on 22.10.2025 Published on 13.01.2026

Available online from January 17, 2026

Research J. Pharmacy and Technology. 2026;19(1):396-403.

DOI: 10.52711/0974-360X.2026.00058

This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. Creative Commons License.

Segmentation Method, W	Segmentation Characteristic, S	Variable	Number of Segments, С
W5 Traditional	S5.1. Age	Y1. Young (18 to 45 years old)*¹	36
		Y2. Middle-aged (45 to 60 years old)
		Y3. Elderly (60 to 75 years old)
		Y4. Senile (75 years old and older)
	S5.2. Behavior in the disposal of unused and expired pharmaceutical products	R1. Responsible (with compliance to disposal regulations)
		R2. Circumstantial (with or without compliance to disposal regulations)
		R3. Irresponsible (without compliance to disposal regulations)
	S5.3. Frequency of pharmaceutical purchases	U1. High (every month or more often)*²
		U2. Medium (every few months)
		U3. Low (twice a year or less)

Rating Indicator	Frequency of Positive Ratings (N, %)						Average Frequency of Positive Ratings for Indicators E and F (N_avg, %)
	Target Segment
	Т1	Т2	Т3	Т4	Т5	Т6
Е	79.0	73.7	94.7	79.0	63.2	57.9	74.6±13.1
F	68.4	73.7	89.5	89.5	57.9	63.2	73.7±13.3
Е-F	73.7±7.5	73.7±0	92.1±3.7	84.3±7.4	60.6±3.8	60.6±3.8	Average frequency of positive ratings for the target segment (N_avg, %)