Prioritizing performance objectives of vaccine supply chain for future pandemics

Immunization programs are essential for preventing diseases, lowering health-care costs and reducing early deaths and disabilities by leveraging vaccines to enhance immunity (Sarigol et al., 2023). Their success depends on efficient vaccine supply chains (VSCs) that ensure vaccines are properly distributed, stored and administered on time to maintain efficacy and achieve widespread coverage (WHO, 2020). Pandemics, however, present unique challenges for VSCs that differ from those of traditional immunization programs (Dey et al., 2024). Unlike routine programs that primarily target infants, pandemic immunization efforts must address a significantly larger and more diverse population (Chakraborty and Mali, 2021). Furthermore, the urgency of pandemic vaccination requires rapid administration to curb the spread of infection. A high vaccination rate is critical to breaking the chain of transmission and reducing mortality (Rele, 2021).

Meeting these demands places immense strain on VSCs, significantly increasing the risk of failure. VSC disruptions can lead to vaccine shortages, further complicating the situation by causing higher hospitalization rates and putting additional stress on health-care systems (Kargar et al., 2024). To address these challenges, it is essential to identify and prioritize key performance objectives (POs) tailored to the specific demands of pandemic VSCs. For instance, during pandemics, the willingness to get vaccinated among beneficiaries is negatively influenced by their scepticism regarding the efficacy and potential side effects of newly developed vaccines (Yadav and Kumar, 2023). Alam et al. (2021) further suggest that this scepticism largely stems from inadequate positive vaccine marketing and poor coordination with local administrations. Also, rumors about these vaccines spread rapidly, exacerbating existing confusion (Attwell et al., 2021). To mitigate these challenges, a PO focused on “effective communication” for proper community engagement is essential.

Similarly, a comprehensive list of POs must be defined to address and classify all possible issues of pandemic VSC. While the literature on challenges faced by VSC during a pandemic has grown significantly since the COVID-19 outbreak (see Table 1), efforts to systematically define POs for addressing these issues remain limited. To bridge this research gap, this study proposes the first research question as:

To answer the proposed question, various POs of a VSC are defined by mapping them with various VSC issues in the pandemic context through a comprehensive literature review and interviews of experts who carried out mass vaccination during COVID-19. During this process, the POs were simultaneously categorized into four dimensions of the balanced scorecard (BSC) framework for evaluation. The BSC is perceived as a general tool for organizational performance measurement and various organizations can develop their own suitable performance measures as indicated by their strategic development needs (Tawse and Tabesh, 2023). This study illustrates how defining POs of pandemic VSC in terms of BSC will be helpful in the sustainable scaling up of pandemic immunization programs.

Different POs linked to pandemic VSCs tend to influence each other. For instance, “optimizing overall vaccination cost” and “ensuring safety of health-care workers (HCWs)” mutually influence each other. Several frontline HCWs operate under the risk of getting infected during a pandemic, and arranging a large number of personal protective equipment (PPE) to ensure their safety requires significant additional funds (Morishita et al., 2021). Optimizing vaccination cost is necessary to free up financial resources for procuring these PPE kits. Also, when HCWs feel safe, their efficiency of vaccine administration and crowd management at vaccination centers improves significantly, thereby reducing vaccine wastage (Andiç-Mortan and Gonul Kochan, 2023; Mukherjee et al., 2023). This, in turn, supports the goal of cost-effective vaccine administration. Hence, there exists a mutual causal relationship between the two POs.

However, one-way causal relationships between POs are also common. For example, Rai et al. (2022) suggest that the PO of digitizing vaccine management practices for more efficient compilation and monitoring of stock records can improve demand management across different layers of the VSC. However, this relationship is not reciprocal, as better demand management does not necessarily drive improvements in digital management systems.

Hence, understanding the nature of the relationship (whether a one-way or mutual causation) between POs can be instrumental while framing appropriate policies and adequate strategies for managing pandemic vaccination drives. Eventually, we can prioritize a PO based on the extent of the overall influence it exercises on other POs. This way, VSC performance can be improved by focusing on issues associated to high-priority POs as they significantly influence the outcome of other POs. An in-depth analysis that investigates these interrelationships is currently lacking and, therefore, less understood. This leads to the second and third research questions of this study:

The causal relationships among the finalized POs and their prioritization, based on their overall influence, are established using a fuzzy decision-making trial and evaluation laboratory (DEMATEL)-based analytic network process (DANP) approach. Techniques such as interpretive structural modeling (ISM) and DEMATEL are widely used for analyzing complex interrelationships among factors (Alam et al., 2021; Yadav et al., 2023). DEMATEL is preferred over ISM as it quantifies the degree of influence among heterogeneous factors (Farooque et al., 2020). Since the factors being analyzed in this study (POs of pandemic VSC) are further classified into broad dimensions of BSC, accounting for the contextual interactions between these dimensions is also essential. The fuzzy DEMATEL method does not incorporate the influence of broader groups while prioritizing the underlying factors. Hence, DANP is used as it considers the weightage of the corresponding broader group when assigning final ranks to the factors.

Experts with experience in managing pandemics like COVID-19 can provide insights into how specific POs influence others, although their subjective judgments often involve elements of vagueness (Yadav and Kumar, 2023). Scholars recommend using fuzzy set approaches to address such uncertainties (Alam et al., 2021). To capture the imprecise information from subjective expert responses, the spherical fuzzy sets (SFS) concept is used. Unlike other fuzzy set extensions, SFS generates a spherical surface of membership functions, offering a broader preference domain for decision-makers and capturing uncertainty more effectively (Kutlu Gündogdu and Kahraman, 2019). This approach is particularly relevant as the lessons learned from managing COVID-19 provide critical insights for developing VSC strategies for future pandemics. Recent outbreaks, such as monkeypox, declared a public health emergency by the WHO (UN, 2024), underscore the importance of robust VSC preparedness. Based on these considerations, the fourth research question (RQ4) is as presented:

This study is the first to comprehensively evaluate the priorities and interrelationships of key VSC issues in a pandemic context by defining broad POs around them. It provides a strategic roadmap for VSC practitioners to sustainably scale up vaccination programs by developing a strategy map (Figure 5) that illustrates causal relationships among POs and links intangible assets, such as organizational and human capital, to value-creating processes in internal and customer management dimensions. The study categorizes POs into cause-and-effect groups and identifies those requiring higher priority (Table 12), aiding policymakers in preparing contingency plans and improving VSC performance during future global disease outbreaks.

The remaining research paper is organized as follows. Section 2 presents the research objectives based on an extensive literature review. Section 3 outlines the research methodology used in this paper. Section 4 illustrates the results and analysis of the collected data. Section 5 delineates the discussion and research implications. Finally, Section 6 portrays the conclusion and opportunities for future work.

This section is divided into two subsections. The first subsection summarizes the research on VSC performance issues with a focus on pandemic scenarios to highlight relevant POs. A PO contributes to the overall success of an organization. Developing any decision support system for the organization involves defining relevant POs that cater to the existing issues (Chandra and Kumar, 2019). The second subsection highlights the operational issues of VSC during COVID-19 and how they are related to the various POs. We derive and explain our research gap at the end of the second subsection.

The four dimensions of BSC used to classify the identified POs are financial, customer, internal process and learning and growth. The BSC is a performance measurement tool widely adopted across business, government and nonprofit sectors (Kaplan and Norton, 2001). It aligns organizational activities with mission and strategy, enhances communication and tracks progress toward strategic goals. Implementing the BSC in nonprofit sectors, such as public immunization programs, provides similar benefits to its use in private organizations (Chandra and Kumar, 2021). In the context of pandemic VSCs, classifying issues under different POs using the BSC framework enables immunization managers to better understand the causal relationships between the various dimensions and, hence, the underlying POs. While financial POs are central in for-profit settings, customer-related issues, such as vaccine accessibility and coverage, take precedence in VSC management during pandemics. The BSC offers a structured framework to strengthen the resilience and sustainability of VSCs during crises.

It is often claimed that developing nations have weak financial provisioning for immunization programs (Ozawa and Stack, 2016). The large-scale distribution of new vaccines during a pandemic with the existing infrastructure is always challenging (Yadav and Kumar, 2023). To control a pandemic situation, installation of additional vaccination centers in remote locations for rapid vaccination of the population (Duffy et al., 2021), ensuring appropriate training and remunerations for all temporary as well as permanent HCWs (Robinson et al., 2021) and accurate estimation of overall immunization cost (Haidari et al., 2017) are required. It would increase operational costs associated with vaccine procurement, storage and distribution, workforce hiring and training, vaccine wastage (both planned and unplanned), etc.

The unavailability of funds to arrange extra infrastructure (Chandra and Kumar, 2019) and improper utilization of funds due to insufficient planning in uncertain scenarios and malpractices (Sung et al., 2021) are detrimental to pandemic VSC management. For instance, Alam et al. (2021) reveal vaccination cost and lack of funds for vaccine purchase as one of the most crucial challenges to COVID-19 VSC. To address such problems, Sung et al. (2021) emphasize redirecting the corporate social responsibility funds in addition to government funding toward public immunization programs to improve its cash flow. The dual problem of “increased operational cost” and “constrained funding” of immunization programs motivates us to explore the POs of “Adequate health financing (HF)” and “operating cost optimization (OC)” in the context of pandemic VSC from a financial perspective.

The successful functioning of a pandemic VSC faces several other issues. The massive scale of vaccination coverage required during a pandemic leads to problems like poor accessibility of vaccination camps by beneficiaries (Bansal et al., 2022), beneficiaries forced to return on vaccination days due to limited capacity (Mohammed et al., 2021), poor address of public grievances associated with the vaccination drive (Chandra and Kumar, 2021) and mismanagement at crowded vaccination camps. These issues, if left unattended, create a poor perception of the immunization program and dissatisfied customers refrain from getting vaccinated. Amidst a pandemic like COVID-19, it is crucial to manage the VSC to ensure adequate vaccine availability and minimize its adverse impact (Ocampo and Yamagishi, 2020). Reducing travel distance to vaccination centers, cooperative behavior of HCWs and improved responsiveness through a reply to queries enhance beneficiary satisfaction (Sazvar et al., 2021).

Another set of customer-centric issues revolves around societal anxiety regarding the efficacy and safety of new vaccines introduced for countering the pandemic. The spread of rumors and misinformation regarding vaccines is common among relatively less educated people (Attwell et al., 2021). Alam et al. (2021) suggested inadequate positive marketing of the new vaccine as a major challenge to the COVID-19 VSC. Effective communication about vaccination benefits to the target audience includes community engagement through coordination with the local administration and frequent reminders of important information about vaccination centers, dates and timeslots to improve vaccination coverage. Hence, achieving the POs of “customer satisfaction (CS)” and “effective communication (EC)” are instrumental to issues concerning the beneficiary (customer’s) perspective of the pandemic VSC.

From a process perspective, efficient distribution of vaccines through the cold chain remains critical. Butt (2022) shows that the distribution of COVID-19 vaccines resulted in a significant decline in the regular immunization programs for infant children. At the state, zonal and district levels, cold storage facilities are congested, and it is challenging to maintain stringent temperatures due to the addition of new vaccinations. Additionally, the network’s inefficiencies put vaccines’ availability, price and safety in peril (Zaffran et al., 2013). Maintaining a controlled temperature environment is a major challenge in developing nations’ routine immunization programs since it is necessary for the vaccines’ quality and effectiveness (Matthias et al., 2007). Ivanov and Dolgui (2021) show that the inclusion of new vaccines in immunization programs without strategic planning may affect the availability of new and regular vaccines due to limited storage capacity and transportation. Chandra and Kumar (2021) highlighted supply chain agility (AG), vaccine distribution management (VDS), vaccine demand management (VD), vaccine wastage optimization (VD), innovation management (IM) and vaccine management practices (VMP) as six POs associated to VSC of the regular immunization program. We adopt these POs in our analysis framework and define them based on issues specific to pandemics.

Also, Kaplan and Norton (2001) have defined three types of intangible capital under the learning-growth perspective of the BSC, namely, human, information and organization. We adopt them as the three POs for analyzing the pandemic VSC under the learning and growth perspective. We summarize the POs and map them to associated VSC issues from a pandemic context in Table 1. As evident from the discussed literature, the research on pandemic VSC incorporated several operational problems at an individual level. Classification of these issues under broad POs in the context of pandemics, which is missing, can help in a wholesome analysis for macroscopic strategic planning of future pandemics. Hence, an attempt is made to map the identified pandemic VSC issues under each relevant PO, which is shown in Table 1. It modestly answers our RQ1, as highlighted in the earlier section.

Several researchers have contributed to the extent of VSC literature in the COVID-19 context. Guttieres et al. (2021) proposed a framework considering multiple vaccination strategies to help plan for different scenarios while evaluating the pros and cons. Rele (2021) studied the challenges and opportunities in developing a new vaccine during a pandemic and proposed implications for future pandemics. Other researchers also contributed topics like VSC risks (Sorooshian et al., 2022), challenges (Alam et al., 2021), resilience (Sazvar et al., 2021), information sharing and coordination (Pan et al., 2022), vaccine access uncertainty (Gilani and Sahebi, 2022), prioritization of lean-agile-green practices (Yadav and Kumar, 2023), strategies for achieving herd immunity (Sinha et al., 2021) and vaccine selection (Öztürk et al., 2021) in context to COVID-19.

Also, several articles focused explicitly on vaccine distribution and allocation technical aspects during COVID-19. Abbasi et al. (2020) suggested downstream allocation of vaccines in the VSC to propose an alternate model for its distribution and allocation. Applying blockchain and the Internet of Things has effectively improved VSC’s distribution and responsiveness during COVID-19 (Kamenivskyy et al., 2022). Mathematical models for the equitable distribution of COVID-19 vaccines in developing nations have been proposed by Tavana et al. (2021). Shahparvari et al. (2022) proposed a heuristic model prioritizing the distribution of vaccines among benefactors based on their vulnerability. Fadaki et al. (2022) proposed vaccine distribution through a multi-period allocation model developed by performing linear optimization, taking into account new constraints in the COVID-19 context. We present a tabular representation of the selected studied in Table 2.

From the COVID-19 VSC literature, it can be observed and argued that the research topics directly or indirectly affect one or more POs. For instance, topics like “vaccine counterfeiting” and “challenges with the development of new vaccines” mentioned in the previous paragraph directly impact the PO of vaccine demand, whereas topics associated with vaccine distribution and allocation planning implicitly aim at improving the POs related to internal processes of the VSC. Hence, an evaluation framework for analyzing the priority of these POs provides practical insights regarding implementing topic-specific solutions in a sequence which are currently missing in the literature.

Also, developing an evaluation index comprising various VSC POs along with the developed framework shall facilitate the investigation of the interrelationship between the POs from an issues-based perspective, and the prioritization of POs in the context of sustainable rapid scaling of the COVID-19 vaccination program will be possible. A comprehensive evaluation of broad POs of pandemic VSCs to explore their macroscopic interrelationships is missing. Therefore, a study that focuses on identifying, defining and prioritizing the POs of pandemic VSC and understanding the relationship between them is necessary. The methodological framework adopted for prioritizing the identified VSC POs and exploring their interrelationship is discussed in the next section.

We performed the study in three stages. In the first stage, we have defined an evaluation framework for the POs as shown in Figure 1. In the second stage, we developed a questionnaire enquiring about the influence exercised by a given PO on the rest of the POs included in the list. As shown in Figure 2, these questionnaires were in the form of an influence interaction matrix, which the experts were supposed to fill separately for the four dimensions of the BSC framework and the identified POs. Responses were collected using a linguistic scale, whose description is again available in Figure 2. Finally, the collected data was analyzed using a spherical fuzzy extension of the DANP method (SF-DANP) for modeling the POs in the BSC framework, determining the interrelationship between the POs and ranking them for priority in a pandemic context.

Due to complex interactions between the different POs under realistic circumstances, evaluating them for priority becomes very difficult. Especially when the number of factors to be analyzed is large (Yadav et al., 2023). The fuzzy set extension of MCDM approaches like ISM, DEMATEL, analytical hierarchy process (AHP) and VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) have been extensively used while prioritizing key factors that influence a system (Ghuge et al., 2022; Vishwakarma et al., 2019; Wang et al., 2024). AHP is preferable for straightforward, hierarchical decision scenarios with independent criteria, offering simplicity and ease of use. In contrast, ANP excels in handling complex, interconnected decision environments where criteria influence one another, providing a more nuanced and comprehensive analysis (Pan and Nguyen, 2015). Although AHP and ANP are simple techniques for assigning criteria weights while analyzing and prioritizing factors, barriers or enablers impacting a process, examining complicated interrelationships between the variables is not possible. ISM and DEMATEL techniques are frequently used for this. Given a certain context, ISM segregates complex systems into smaller subsystems and then extracts possible relationships between driving and driven factors. On similar lines, DEMATEL extracts causal relationship by establishing cause and effect groups with an added advantage of tackling heterogeneous factors by quantifying the degree of influence they exercise (Kumar and Dixit, 2018; Farooque et al., 2020).

In a situation where the interrelationship among POs of pandemic VSC are analyzed in the purview of different dimensions of the BSC framework, a combination of DEMATEL and ANP (DANP) can be more useful. These are complementary methodologies that, when used together, provide a robust framework for analyzing and improving complex systems like pandemic VSCs. DEMATEL excels in uncovering and visualizing the causal relationships and interdependencies among various factors, while ANP takes care of the hierarchical structure and effectively prioritizes these factors to guide strategic decision-making. The synergy between these methods leads to more informed, efficient and sustainable supply chain strategies, ultimately enhancing the ability to respond to future pandemics with greater resilience and effectiveness (Bafandegan Emroozi and Fakoor, 2023). To account for the ambiguity and vagueness in preferences made by experts while filling their responses, a fuzzy extension of DEMATEL is used in this study. Details on the SFS used and construction of scales is discussed in Sections 3.2 and 3.3, respectively.

We included 30 VSC experts and practitioners who worked at various levels of public organizations (including senior doctors, medical professionals and bureaucrats from the Indian Administrative Services) and had direct involvement in critical decision-making of VSC management during COVID-19. The list of experts and their details is provided in Table 3. The experts also included the following midlevel practitioners who have greater on ground experience: ten district immunization officers, ten district-level WHO representatives and six vaccine cold chain managers. Including members from multiple administrative layers in an expert survey ensures a comprehensive analysis of organizational processes by capturing diverse perspectives and expertise (Markus et al., 2002). The experts were given the prepared questionnaires (refer Figure 2), which had the description of each PO along with its underlying issues. The overall analysis framework had 13 POs under the four dimensions of BSC. The experts were asked to perform a pair-wise comparison that measures the direct impact between VSC POs and BSC dimensions separately. They were asked to use the linguistic terms mentioned in Table 4. It took six months to collect experts’ responses, from December 2023 to the end of May 2024. The Delphi method was adopted to establish a consensus in the opinions of a panel of experts through several rounds of interviewing and sharing the answers of others. The process map of research methodology is shown in Figure 3.

The SFS combines Neutrosophic (Smarandache, 2002) and Pythagorean fuzzy sets (Yager, 2013), using a three-dimensional membership function that integrates membership, nonmembership and hesitancy (or indeterminacy) as parameters. SFS has been developed by Kutlu Gündogdu and Kahraman (2019) that generate a spherical surface of a membership function to enable a large preference domain for decision-makers. SFS gathers hesitancy information as an extra parameter and includes the advantages of Neutrosophic and Pythagorean fuzzy sets to develop an efficient solution. SFS with MCDM methods can express uncertainty and vagueness on subjective responses with enhanced accuracy. The following properties of an SFS are useful for scale construction and analyzing data using the MCDM methods.

Property 1: a SFS $Q˜s$ is defined in an infinite collection of dialogues, I as follows:

where

$μQ˜S:I→[0,1],ϑQ˜S:I→[0,1],πQ˜S:I→[0,1]$ and

Property 2: the SFSs are added as per the following definition:

We developed a constant-valued linguistic scale using a nonlinear optimization method for assigning the fuzzy values to the corresponding linguistic terms for our proposed methodology. A constant-valued linguistic scale is widely used in the literature (Karasan et al., 2019). We decided to have five linguistic terms to manage the complexity of our study based on the frequently used words by experts during the interviews, as shown in Table 4. Each linguistic term is represented by a score index (SI) defined by equation (4). The corresponding parameters of the spherical fuzzy number (SFN), namely, membership (μ), nonmembership (v) and hesitancy (π), are selected in such a way that they must also satisfy equation (2):

The following calculation shows that the developed SFNs follow the sequence of SI defined for each linguistic term:

The calculations provided in the response show that the derived SFNs align with the SI values for each linguistic term, as demonstrated through the numeric results for “Very Low,” “Low,” “Moderate,” “High” and “Very High” influences. This process guarantees that the chosen scale not only fits the theoretical structure but also remains practically meaningful for the decision-making context of this study.

The DEMATEL is a widely adopted MCDM method to examine the dependencies of the evaluated criteria among themselves (Erdoğan et al., 2021). DEMATEL-based ANP or “DANP” establishes the causal relationship between the criteria and finds criteria weights or ranks using an interrelationship structure (Pan and Nguyen, 2015). Few papers have explored fuzzy DEMATEL, fuzzy ANP or both methods to increase their ability to reflect the same (Alam et al., 2021; Erdoğan et al., 2021). We applied SF-DANP to determine the dependencies of the VSC POs in a pandemic situation, considering the uncertainty and vagueness of experts’ responses, and then prioritize them based on the evaluated weights considering the dependency structure. The steps of the suggested SF-DANP method are inspired by the work of Gül (2020), and are given below:

• Step 1: The direct relation matrix (D)_n × n is constructed to incorporate the expert opinion using linguistic terms.

• Step 2: Convert linguistic terms to their corresponding SFNs based on Table 4.

• Step 3: Constructing the initial direct influence matrix (X).

Corresponding to the three parameters of an SFS (membership, nonmembership and hesitancy degrees), the (D)_n×n matrix is divided into three submatrices. Then, the three matrices are normalized separately as per equation (5). Equation (6) represents the outcome of this step in matric form:

• Step 4: Determining the total influence matrix (T).

Using equation (7), the submatrices of X are converted to submatrices of T by executing equation (7). In this step, we performed Euclidian normalization for each matrix to fit them into the logic of SFS. Equation (8) shows the merged T matrix:

• Step 5: Calculating row and column sums with SFNs.

The row (r_i) and column (c_j) sums are obtained as per equations (9) and (10), respectively. The addition operation for SFNs is performed using equation (3):

The (r_i) and (c_j) values depict the general strength and general weakness for each criterion in terms of its influence on other criteria. Subsequently, we defuzzified r_i and c_j ∀i, j for finding prominence and relation values using equation (11):

• Step 6: Producing a causal diagram.

The r and c values obtained in Step 5 are added for a given criterion, which indicates its total influence interaction with other criteria and is known as “Prominence,” whereas (r – c) values known as “Relation” separate criteria into “cause” and “effect” groups based on their positive or negative values, respectively. The causal diagram is obtained by plotting the “Prominence” and “Relation” values on the horizontal and vertical axis, respectively. The causal diagram depicts the causal dependencies or influences among attributes. A threshold (α) determined using equation (12) finds significant influence relations among the attributes used to draw the causal diagram. This filtering of the T matrix is performed after the SF numbers in it are defuzzified using equation (11):

Subsequently, ANP is applied to get the weights of VSC POs using the results obtained from SF-DEMATEL. These weights are used for the global ranking of the POs. ANP provides more accurate weights when the criteria are interrelated within and among the clusters. The following steps describe the procedures of weights computation using ANP:

• Step 7: Establish an unweighted supermatrix.

The ANP process is initiated by adopting the defuzzified total influence matrix T obtained as per equation (8) as the initial supermatrix, which is further used to obtain the unweighted supermatrix W_u. The normalization of W_u is then performed by making the sum of its columns as unity:

• Step 8: Obtain the weighted supermatrix.

The clusters have different influences on each other, which is taken into account by obtaining the weighted supermatrix. For this, the product of the unweighted supermatrix (W_u) and the transpose of the normalized total influence matrix is done. The resulting matrix is then defuzzied and filtered with respect to α for BSC dimensions, T_s.

where $tijs=tijα/di$ indicates the total influence matrix that is filtered with α after normalization, whereas W_ij indicates the unweighted supermatrix for the given criteria:

• Step 9: Obtain a long-term stable supermatrix.

Now, the weighted supermatrix is raised to power k, which tends to infinity, as shown in equation (14). This stabilizes the weights by converging them:

First, the SFS concept was used to find the total influence matrix from the linguistic direct relation matrix. Next, the DEMATEL method determined prominence and relation values and defuzzified the total influence matrix that maps the causal relationship diagrams. Subsequently, we developed a strategy map for scaling up the public vaccination program during a pandemic. Finally, we applied the ANP method to determine the POs weights to prioritize them considering the interrelationship structure.

The linguistic direct relationship matrix capturing experts’ consensus about the influence relationship between the BSC dimensions is developed using terms mentioned in Table 4. The same exercise is performed for VSC POs. Next, linguistic terms are replaced with their equivalent SFNs based on Table 4. Using equations (5), (6) and (7), we computed the total influence matrix for BSC dimensions as well as VSC POs, as mentioned in Steps 3 and 4 in the methodology section. Due to space constraints, the total influence matrix for the BSC dimensions is only shown in Table 5. Now, we computed r_i and c_j using equations (9) and (10), respectively, and defuzzified them using equation (11) in Step 5. In Step 6, the prominence and relation values are calculated as (r_i + c_i) and (r_i – c_i), respectively, by taking i as equal to j. The corresponding values in Steps 5 and 6 for BSC dimensions and VSC POs are shown in Tables 6 and 7. The influence groups for BSC dimensions and VSC POs are determined considering the sign of relation values and are given in the last column of Tables 6 and 7, respectively.

A defuzzified T matrix (shown in Table 8) is estimated by using equation (11) to know the important influences among BSC dimensions. A similar computation is performed for VSC POs. The defuzzified T matrix is then filtered using a threshold value of α that is obtained using equation (12), which is eventually used to draw the causal diagrams. Figure 4 shows the cause and effect diagram among the four BSC dimensions. After that, a causal relationship between the various POs is depicted through a VSC strategy map in Figure 5. Practitioners could follow the illustrated connections to develop a strategic plan and procedures to implement them for sustainable scaling up of pandemic vaccination programs at the desired rate. Sub-Sections 4.1.1 and 4.1.2 answers the RQ2 of this study.

It was observed from the prominence relation map (Figure 4) that customer (C), learning and growth (L) and internal process (I) perspectives of BSC are influenced by the financial perspective (F). The right-side causal relationship diagram showed that there exists a one-sided causal relationship from F to L and from L to I, as depicted by a single-head arrow. In contrast, C is mutually interrelated with the other three BSC dimensions. It indicates the POs that define C are not only affected by but also influence the POs under F, L and I. The threshold value (α) of 0.249 is used to find the influence relations among the attributes for drawing the causal diagram. The highest score value of r, as observed in Table 6, is 1.298, whereas the highest score value of c is 1.130. It ascertained that the financial perspective (F) has the strongest influence on others, and the customer perspective (C) is the most affected perspective among all. It can be justified that sustainable health financing would positively impact the POs under learning and growth perspectives, which would improve performance indicators associated with internal processes. All of these would eventually lead to efficient and timely vaccination of customers. On the other hand, empowered customer satisfaction indicators motivate people to take vaccines. It would enable effective communication of important information via different communication channels and help the beneficiary reach the vaccination centers, which, in turn, help improve the financial and internal process aspects associated with the vaccination program.

The strategy map, shown in Figure 5, includes four dimensions of BSC and corresponding POs of each dimension. A single head or double head arrow represents, respectively, the one way or mutually causal interrelation between the POs within a dimension or between the different BSC dimensions. The POs are segregated into the cause or effect group based on their (r – c) values, as observed in Table 7. Sustainable health financing (HF), innovation management (IM), optimizing overall vaccination cost (OC) and ensuring effective communication (EC) have the highest positive (r – c) values in this sequence. These POs exercise maximum influence on the sustainable scaling up of pandemic vaccination and hence are most vital. Other POs in the cause group are information capital (IC), human capital (HC) and VSC agility (AG). In contrast, the most highly influenced POs include vaccine wastage (VW), organizational capital (OGC) and vaccine demand (VD), as they possess the highest negative value of (r – c). Achieving satisfactory results for these POs heavily depends on the functioning of POs belonging to the cause group.

The results from SF-DEMATEL are used in determining the global relative influence weights of the POs following the ANP procedure. The interactions between POs and BSC dimensions are incorporated to determine the global weights. The methods section describes Steps 7, 8 and 9 are used to compute the weighted supermatrix and the long-term stable supermatrix, which are then displayed in Tables 9 and 10 correspondingly. The long-term stable supermatrix provides the global weight for each POs. Table 11 summarizes the research findings that include prominence, relation, global weight and corresponding ranks of the POs.

SF-DANP, a combination of SFS and DANP, helps an organization decide on the high-priority POs that significantly influence the other POs. The ranks of the POs highlighted in Table 11 suggest that the issues related to the two POs defined under the customer dimension, that is, “effective communication” and “ensuring customer satisfaction,” must be prioritized during pandemic vaccination programs. Next, performance associated with sustainable health financing and operating cost optimization under the financial perspective must be focused upon to sustainably scale up the pandemic vaccination drive and achieve the customer-centric goal of achieving high coverage in the stipulated time while optimizing the operating cost. The global ranks obtained for the different POs helps us answer the RQ3 of this study.

This article explores the relationships among various VSC POs and prioritizes them during a pandemic using the BSC framework. The analysis involves aligning strategic POs of pandemic VSC with the four BSC perspectives and assessing their interdependencies and performance metrics. In profit organizations, financial outcomes act as lagging indicators driven by leading factors such as customer satisfaction, process optimization and learning and growth. For nonprofits, mission success is the lagging indicator, while financial health and process efficiency serve as critical enablers rather than primary objectives (Kaplan and Norton, 2001).

Our findings for VSCs in nonprofit public immunization programs during pandemics reflect similar dynamics. The strategy map in Figure 5 aligns with Kaplan and Norton’s (2001) model for nonprofits, demonstrating the interconnectedness of the four dimensions. POs related to “adequate and consistent financing” and “cost optimization” within the finance dimension form the base of the VSC strategy. These elements drive advancements in “Learning and Growth,” leading to improved internal processes and ultimately enhancing outcomes for beneficiaries. Mission success, defined as vaccinating the maximum number of people in the shortest time, is directly linked to the customer dimension that is observed as a lagging or most influenced indicator.

However, it is interesting to note that the customer dimension exercises a mutually causal relationship with all other dimensions, which means despite being the most influenced dimension, it also exercises significant influence on other dimensions. It is interesting to note that the two POs classified under this dimension are of opposite nature. The PO of “effective communication” comes under the cause group, whereas “achieving customer satisfaction” is in the effect group. This explains the mutual causation. Hence, this study extends the theory linked to the causal relationship of BSC dimensions for nonprofits and proposes unique findings from a pandemic VSC perspective.

Sub-Section 4.1.2 mentions the POs belonging to the cause-and-effect group and Section 4.2 mentions the final weights or rank assumed by the POs. Table 12 classifies these POs into four groups using causal relationships and computed weights to discuss the results further. Group 1 includes high-weight “cause” POs and Group 2 has low-weight “cause.” Groups 3 and 4 are high and low-weight “effect” POs, respectively.

“Effective communication (EC)” has been identified as the cause group PO that should have the highest priority. Lack of coordination with the local administrative bodies and inadequate positive vaccine marketing are critical issues of pandemic VSC (Alam et al., 2021). These eventually lead to the spread of rumors and misinformation amongst relatively less educated people (Attwell et al., 2021). Prioritizing EC eliminates these issues (Vasudevan et al., 2020) and keeps the beneficiaries updated about the vaccination centers and schedule. Finally, it improves customer experience by minimizing their effort and helps in optimizing operating costs by reducing vaccine wastage and efficient use of vaccination centers and workforce. Satisfied customers improve the vaccination process through positive word of mouth via social media and oral communication (Kareklas et al., 2015). Prioritizing these two POs shall minimize unwillingness to take the vaccine and create a positive public perception of the vaccination program. It also eases the process of mass vaccination, thus assisting in the achievement of the United Nation’s third sustainable development goal focuses on the general health and well-being of the population.

The next two important POs in VSC are “Adequate health financing (HF)” and “Operating cost optimization (OC),” and they are mutually interrelated within the financial perspective BSC dimension. The transition from conventional to large-scale vaccination during a pandemic requires additional infrastructure, technology adoption and training for existing as well as newly inducted HCWs (Gupta et al., 2022). Further, operating cost is a function of vaccine storage, distribution, purchase, wastage and workforce cost. It can be optimized by managing the internal processes by redesigning cold chain storage facilities, health zones and transportation routes (Lee et al., 2015).

In the current era of the internet and technology, innovation can improve the efficiency of VSC. Indigenous technology innovations like eVIN and COWIN in India greatly assisted in keeping track of vaccine stocks at various levels, vaccination coverage of different age groups as well as priority groups, the temperature of storage and transportation devices and the allocation of the most convenient vaccination center along with a time slot for people (Gurnani et al., 2020; Saha et al., 2022). The efforts put into the POs of innovation and information capital management can help forecast accurate vaccine demand and monitor vaccinated populations during a pandemic.

Our findings offer several directions for policymakers and practitioners. First, underdeveloped or emerging countries should emphasize implementable health financing solutions. We argue that practitioners, health-care statutory bodies and/or the government should consider sustainable health financing methods such as advance market commitments and cofinancing policies (Ozawa and Stack, 2016). The transition to large-scale vaccination can also be enforced by redirecting corporate social responsibility funds toward public immunization programs, government funding and their corruption-free utilization (Sung et al., 2021). Second, practitioners are also suggested to hire more professionals and provide better training programs that regularly facilitate the health-care professionals to cope with a pandemic’s dynamic nature, leading to optimized operating costs (Khurshid et al., 2020). To establish effective large-scale vaccination programs during the pandemic, the strategy map proposed in this study and the prioritized POs should assist VSC practitioners, policymakers and the government in developing a strategic plan with specific implementation procedures. Third, the impact of digital technology can significantly reduce the adverse impacts of similar pandemics in the future, provided they are integrated and redesigned to manage and cater to different stakeholders, including hospitals, health-care professionals and the government, among others (Ting et al., 2020; Rai and Bera, 2024).

The limitations in our paper may serve as a foundation for future research. First, the causal relationship established by this study can be empirically validated through a statistical model, such as structural equation modeling, by establishing a suitable hypothesis. Although it will be difficult to simultaneously assess all the different POs, but still phase-wise studies can be conducted. Second, our respondent pool could have been more diversified from a supply chain perspective. For instance, we limited our interaction to experts, while through interaction with customers or beneficiaries, future research could address their perspective on different identified POs. Third, as an extension of our findings, individual POs could be analyzed separately to understand the nature of deployment from an effectiveness perspective. For instance, specific studies on health financing options in developing countries, defining and designing effective communication in diverse beneficiary groups across regions, and mathematical models for efficient vaccine distribution, storage and delivery could serve as future research studies.