COVID-19 as a trigger for dynamic capability development and supply chain resilience improvement

ABSTRACT

A firm’s ability to manage risk and resilience in supply chains has turned out to be an invaluable capability during the COVID-19 pandemic. Fast responsiveness, quick decision-making, and the ability to reconfigure the resource base have helped firms during the pandemic, which caused rapid disruptive effects for which they were unprepared. This research applies the dynamic capability view as a theoretical framework to study the impact of the COVID-19 pandemic on firms’ dynamic capabilities and, the influence of dynamic capabilities on supply chain resilience. These capabilities are highly needed to survive during the pandemic. Using the survey data, we found that the impacts that COVID-19 had on a firm’s upstream supply chain influence firms’ capabilities to seize opportunities or neutralise threats. Furthermore, we found that reconfiguring ability has a strong influence on supply chain resilience. Thus, the impacts of COVID-19 on the downstream supply chain pushed firms to realign resources to respond better to demand. Upstream disruptions pushed companies to react to threats and opportunities in the supply market, while downstream disruptions leveraged reconfiguring capabilities.

KEYWORDS:

• Supply chain resilience
• supply chain management
• disruption management
• questionnaire study
• dynamic capabilities

1. Introduction

Supply chain volatility and uncertainty are inherent elements of supply chain management and can occur in any company (Chiang and Feng 2007). Generally, supply chain volatility and uncertainty are generated by disruptions that are the effects of catastrophic events, such as natural disasters (hurricanes, flooding, tornadoes, earthquakes, tsunamis), man-made catastrophes (nuclear power plant disruptions, accidental toxic spills, poisonings, wars, terrorist attacks), and legal disputes or strikes (Ivanov and Dolgui 2019). The degradation of the world’s ecological system, coupled with socioeconomic instability, is increasing the intensity and frequency of catastrophic events and exposing supply chains to more disruptions – to the extent that one single event can have remarkable and unpredictable influences and lead to huge financial and non-financial damage to supply chains, companies, and society.

At the supply chain level, these disruptions can have swift and negative effects on manufacturing plants, suppliers, distribution facilities, and transport connections, all of which may become unavailable for some time. In many cases, companies realise the vulnerability of their supply chain only when the disruption has already occurred (Bier, Lange, and Glock 2020). In addition, as noted by several scholars, material and product shortages and delivery delays spread very quickly, leading to additional undesirable effects on supply chain performance and resilience deterioration (Ivanov, Sokolov, and Dolgui 2014; Jabbarzadeh, Fahimnia, and Sabouhi 2018; Pavlov et al. 2019; Li and Zobel 2020). For example, the 2011 tsunami in Japan had extensive effects on global supply chains, and Ericsson’s sub-supplier fire breakout in 2000 in Albuquerque (US) resulted in a shortage in components and raw materials in Ericsson’s supply chain (Ivanov, Sokolov, and Dolgui 2014).

Pandemic outbreaks represent a special type of supply chain disruption. They are extremely dangerous because they spread very quickly in different geographical areas. COVID-19 is a new type of extremely contagious coronavirus that is devastating supply chains worldwide (Choi 2020; Ivanov 2020a). The first country hit by the COVID-19 pandemic was China. The immediate consequence was the gradual paralysis of supply chains in several continental areas (Asia, Europe, North America) and sectors (automotive, chemicals, electronics, garments and textiles, machinery, metal, medical devices). The pandemic affected the entire world and created a worldwide health crisis, disrupting many global supply chains and shifting supply lines.

The consensus is that the COVID-19 pandemic is imposing a profound change on the way supply chains are managed. Before the pandemic, most companies managed supply chains by focusing first on cost minimisation and just-in-time delivery, resulting in low levels of inventory (Simchi-Levi 2020; Dohale et al. 2021). The COVID-19 outbreak quickly exposed the drawbacks of this approach in the manufacturing sector, forcing change. According to Choi, Rogers, and Vakil (2020), COVID-19 created severe problems and disruptions for those companies that were not well-prepared – meaning most companies – because of the complex structure of their supply chains. This comes from the massive outsourcing of production to low-cost countries, which increases the complexity and number of actors within supply networks. Simultaneously with the expected cost and efficiency benefits, the risks of failure and disruptions in global supply chains have steadily increased. Companies were, therefore, totally unprepared for the rapid disruptive effects of COVID-19 on their international supply chains.

One key element in recovering from supply chain disruptions and in building resilience is the development of firms’ capabilities. To reduce the impact and to be able to bounce back after the disruption requires a response and recovery ability (Chowdhury and Quaddus 2016). If companies fail in developing readiness, response and recovery abilities, their supply chain will be even more vulnerable (Chowdhury and Quaddus 2016). Mitigating vulnerabilities requires resilience capabilities to remain in the long run, requiring firms to develop dynamic capabilities (Chowdhury and Quaddus 2017). For example, the World Economic Forum highlighted that companies must adjust supply chains for future challenges by dynamically leveraging capabilities and strategies in digital readiness and data sharing (WEF 2020). Fosso Wamba et al. (2020) found that big data analytics-enabled dynamic capabilities are more likely to build agility and adaptability to supply chains. Moreover, Lee (2021) renewed the original triple-A concepts of agility, adaptability and alignment (Lee 2004) to the post-pandemic world and relabelled these as super-agility, architectural adaptability and ecosystem alignment. For example, for creating super-agility, companies need to be super-sensitive in sensing and super responsive in responding (Lee 2021). Ordinary capabilities with basic functional activities are not enough in building these super abilities and better resilience, but dynamic capabilities with high-level routines are needed. The widespread effects of the pandemic mean that different strategies and actions are required in different situations. One solution is to build robust supply chain resilience (Chen, Das, and Ivanov 2019), which redirects flows of materials within the supply chain and creates adapted structures for ensuring continuity in operations (Zhao, Zuo, and Blackhurst 2019).

Despite the large body of current research on supply chain resilience, the impact of a pandemic on the development of capabilities and the improvement of supply chain resilience are still unexplored. Ivanov (2021a) states that even though COVID-19 enormously affected supply chains, the existing research on adaptation strategies, viability and resilience and recovery is still scarce. In the same vein, Queiroz et al. (2020) noted that some empirical theories, such as dynamic capabilities, resource-based view, and contingency theory, combined with resilience and jointly employed with operations research/operations management approaches, may allow for the more effective examination of the effects of epidemics and pandemics on supply chains. These theories could help apply empirically-built analytics to better study the impact of different pandemic outbreaks on supply chains (Bode et al. 2011; Dubey et al., “Empirical Investigation,” 2019). Furthermore, Chowdhury and Quaddus (2017) state that existing studies that apply resource-based view or dynamic capabilities view fail to identify processes, resources and paths that increase competencies during supply chain uncertainties. Even though previous research has studied resilience as a key part of a firm’s ability to manage disruptions, more research on how firms can handle dynamism and disruptions in supply chains by developing resilience, especially during catastrophic events like COVID-19 (Yu et al. 2019), is needed. Schilke (2014) found that in moderately dynamic environments, the dynamic capabilities have the highest effect on competitive advantage, whereas, in low or high levels of environmental dynamism, the efficacy decreases. As COVID-19 enormously challenged supply chains and forced companies to react rapidly to drastic changes in the business environment, the dynamism was high. This kind of highly dynamic environment can provide lots of opportunities for resource reconfiguration but can be extremely challenging for creating long-term competitive advantage (Schilke 2014). The effects of a pandemic are enormous, creating a worldwide health crisis and serious damage to the global economy; they are different from the standard disruptions. Thus, more research on the topic is highly needed, now more than ever before.

This paper aims to analyse the impact that COVID-19 had on capabilities development and supply chain resilience improvements in the medical device industry. To achieve this objective, quantitative data were collected by a survey conducted in the medical device industry in Italy and Finland in the summer of 2020. This study contributes to the scientific debate on strategies for dealing with supply chain disruptions during catastrophic events by developing capabilities without compromising and even potentially improving supply chain resilience. By studying the relationships between different kinds of capabilities, we contribute to the research of dynamic capabilities view (DCV) by showing that the basic logic of DCV works in the case of disruptive events, during which companies try to reconfigure their operations to survive. We show that firms’ capabilities to sense will affect their capabilities to seize, which will affect their abilities to reconfigure. This means that firms must first sense the environment by detecting and gathering information, and then capture the opportunities sensed or neutralise the perceived threats. To neutralise threats, firms need to be able to realign their resources and assets. We further contribute to the research on supply chain risk and resilience by showing that the impacts that COVID-19 had on firms’ upstream supply chain influence firms’ abilities to seize opportunities or neutralise threats by creating processes and structures for firm-level decision-making. Finally, we show that reconfiguring abilities have a strong influence on supply chain resilience.

The article is structured as follows. Next, the theoretical background is formed, and the hypotheses are justified and presented. After that, the research methodology adopted to test the hypotheses is described, and the results of the survey are shown. The results are then discussed in the context of their implications for theory and practice. The final section presents key conclusions emerging from the study.

2. Theoretical background and hypotheses

2.1. Dynamic capabilities view

COVID-19 as a supply chain disruption first forced companies to react to the disruption and then required them to develop new resources, solutions, and capabilities to survive. The DCV (Teece, Pisano, and Shuen 1997) has been commonly suggested as a framework for handling changing environments by providing a means for extending, modifying, and reconfiguring existing capabilities (Pavlou and El Sawy 2011). Therefore, it offers a theoretical background and dynamic lens to examine the development of firms’ capabilities (Eisenhardt and Martin 2000). Traditionally, the DCV has been applied in rapidly changing environments in which continuous updating of firms’ capabilities is done to develop their long-term competitive advantage (Teece, Pisano, and Shuen 1997; Eisenhardt and Martin 2000). According to Winter (2003), the word dynamic connotes change, which often comes from force majeure from the business environment. COVID-19 is an extreme example of force majeure (Govindan, Mina, and Alavi 2020), forcing companies to develop their capabilities. Thus, DCV offers an excellent background for studying firms’ abilities to respond to the changes and challenges caused by COVID-19. Very recent research by Ruel and El Baz (2021) studying resilience and readiness during COVID-19 also applied DCV as a theoretical lens. It highlighted its applicability as theoretical background for exploring how to improve resilience during the pandemic.

Winter (2003, 991) defined organisational capabilities as ‘a high-level routine (or collection of routines) that, together with its implementing input flows, confers upon an organization’s management a set of decision options for producing significant outputs of a particular type.’ Dynamic capabilities, according to Teece, Pisano, and Shuen (1997, 516), are ‘the firm’s ability to integrate, build, and reconfigure internal and external competences to address rapidly changing environments.’ Moreover, the dynamic capability is ‘the capacity of an organization to purposefully create, extend, or modify its resource base’ (Helfat et al. 2007, 4). Helfat et al. (2007) added ‘purposefully’ to the definition, following Winter’s (2003) idea that organisational capabilities need to be learned routines with a clear purpose. Firms’ ordinary capabilities are different from their dynamic capabilities. Ordinary (or operational) capabilities are defined as zero-level capabilities, which are the basic functional activities of a firm that permit its existence; as Winter (2003, 992) put it, they are ‘how we earn a living now’ capabilities. Dynamic (or first-level) capabilities are high-level routines that create the ability to understand the environment and respond to it by expanding, adapting, or creating ordinary capabilities (Eisenhardt and Martin 2000; Winter 2003).

Following Teece (2007), dynamic capabilities can be categorised into the capacities of sensing, seizing, and reconfiguring (or transforming). Dynamic capabilities can be viewed as organisational and managerial processes that help firms sense opportunities and threats to seize the opportunities and reconfigure their resources to match the environment (Teece, Pisano, and Shuen 1997; Teece 2007; Helfat et al. 2007). Sensing capability means scanning, detecting, identifying, and interpreting new opportunities (Teece 2007), which includes activities that aim to figure out what is happening in the business environment at a given time (Teece 2007). Sensing capacity encompasses dynamic managerial capabilities because the initiation of change depends on managers’ abilities to sense and interpret new opportunities (Maijanen and Jantunen 2014). Sensing is also about understanding latent demand, seeing the structural evolution of markets, and sensing signals coming from suppliers and competitors (Teece 2007); it enables companies to detect the required changes of the resource base based on the changes in the environment.

A firm’s capacity to seize is its ability to capture the sensed opportunities or neutralise the threats by creating firm-level decision-making procedures and structures (Teece 2007; 2012). It is visible in activities such as creating organisational innovations, choosing and developing new business models, and investing in needed technologies (Maijanen and Jantunen 2014). Teece (2007) states that once a firm has sensed a new opportunity, it should create new offerings and invest heavily in it.

Reconfiguring capacity refers to the alignment and realignment of certain assets so that the firm can renew and ensure that its resources are in line with the detected changes and sensed opportunities (Teece 2007; 2012). Reconfiguration is needed to maintain evolutionary fitness (Teece 2007), and the development and adoption of new organisational structures and managing knowledge with certain activities are in a key role (Maijanen and Jantunen 2014). In essence, reconfiguring capacity is the ability to reconfigure and recombine assets and organisational structures in response to changing markets (Teece 2007).

The path-dependent nature of dynamic capabilities means that all three capacities, sensing, seizing, and reconfiguring, are significant in responding to and carrying out change; they form an interrelated chain of activities (Helfat and Peteraf 2009). Path dependence can be described as learning mechanisms (Eisenhardt and Martin 2000) that require firms first to sense the market and external environment by detecting new information and knowledge and then to capture the sensed opportunities and integrate the new information. Finally firms need to align or realign needed assets to create renewal and to make sure that the resource bases are in line with the detected changes and sensed opportunities. To investigate how the basic logic of DCV works in the case of disruptive events during which companies try to reconfigure their operations to survive, such as a pandemic, we follow the logic by DVC and hypothesise:

H1a: A firm’s capability to sense affects the firm’s ability to seize opportunities.

H1b: A firm’s capability to seize opportunities affects the firm’s ability to reconfigure resources.

2.2. COVID-19 supply chain disruptions and dynamic capabilities

Many previous studies have focused on supply chain vulnerability, risks, disruptions, and resilience (e.g. Xu et al. 2020). However, the COVID-19 outbreak created supply chain vulnerability, risks, and disruptions that the current business environment had never faced before (Araz et al. 2020; Craighead, Ketchen, and Darby 2020; van Hoek 2020). Ivanov (2020a) states that epidemics and pandemics represent an exceptional type of supply chain risk in which the scaling cannot be predicted, and the disruption is long-term, spreading simultaneously in the population and in supply chains and causing problems with demand and supply. Upstream disruptions, such as raw material shortages and supply delays, hit supply chains dramatically, especially during the first phases of the pandemic (Ivanov 2020a; Paul and Chowdhury 2020). Companies were totally unprepared for the rapid disruptive effect of COVID-19 on their supply chains (Choi, Rogers, and Vakil 2020). The transition from deep supply chain disruption to rapid recovery and reinforced competitiveness requires enhanced transparency regarding the structure of the supply chain, an unbroken chain of information, and investments in mapping the global supply networks (Choi, Rogers, and Vakil 2020).

Supply chain disruptions are unplanned incidents that disrupt the material flows and expose companies to risks (Craighead et al. 2007). In many cases, they originate from the supply network (Kim, Chen, and Linderman 2015) and have systemic effects on the entire supply chain or network (Scheibe and Blackhurst 2018). Ivanov and Dolgui (2020, 2905) suggest that an ‘intertwined supply network that encapsulates entireties of interconnected supply chains’ could offer a perspective that is extensive enough for creating better survivability and viability in similar future disruptions. Ivanov and Dolgui (2020, 2905) highlight viability which they define as ‘the system ability to meet the demands of surviving in a changing environment’ and state that in the case of extraordinary events, such as COVID-19, firms may be better able to avoid large-scale effects by considering disruptions at a larger scale of survivability. In supply chain settings, viability can be seen as a combination of resilience, adaptability and sustainability (Ivanov and Dolgui 2020) and thus requires that firms can react and adapt by developing new abilities. Disruptions push firms to develop resilient capabilities to reduce their impact (Azadegan et al. 2020). Blackhurst et al. (2005) determined that discovery, recovery, and supply chain redesign are crucial for managing supply chain disruptions. These factors are reflected in the capacities to sense, seize, and reconfigure. Managing supply chain disruptions requires high-level capacities to scan and sense the supply market, recognise the disruptions, neutralise the threats (or capture the opportunities) by forming decision-making procedures that enable recovery and redesign the supply chain by realigning resources to renew and cope. Thus, disruptions caused by COVID-19 may have developed firms’ capacities to sense, seize, and reconfigure. Thus, it is hypothesised

H2a: The upstream supply chain disruptions caused by COVID-19 influence firms’ sensing capability.

H2b: The upstream supply chain disruptions caused by COVID-19 influence firms’ seizing capability.

H2c: The upstream supply chain disruptions caused by COVID-19 influence firms’ reconfiguring capability.

This logic and the necessity of sensing, seizing, and reconfiguring the environment do not apply only to the upstream supply chain, but the input and information coming from the downstream supply chain is highly relevant. Craighead, Ketchen, and Darby (2020) highlighted that COVID-19 had caused extreme shifts in demand and supply. Raw material shortages and supply delays in the upstream supply chain are reflected downstream, causing delivery delays, decreased service levels and lower quality (Ivanov 2020a). This phenomenon is called the ripple effect which occurs when a disruption cascades downstream and impacts the performance in the supply chain (Kinra et al. 2020; Dolgui and Ivanov 2021; Ghadge et al. 2021). Both the suddenly increased demand for essential items, such as hand sanitiser, medicines, and toilet paper, and the reduced demand for products like garments due to cancelled orders and postponed production caused downstream disruptions (Paul and Chowdhury 2020; Ali, Rahman, and Frederico 2021). Ivanov (2020a) stated that examining disruptions in the downstream supply chain caused by pandemics is key in determining how such incidents can influence the ripple effect propagations. Ivanov (2020a) continued that supply chains for products that are in high demand during pandemics, such as masks and gowns, hand sanitiser, and medical alcohol, require further study. According to Teece (2007), customers can be among the first to perceive changes, requirements, and opportunities, and firms that are vigilant and able to sense this can usually turn customer-led endeavours into new products or use their perceptions to respond and reconfigure resource bases. Thus, it is hypothesised as

H3a: The downstream supply chain disruptions caused by COVID-19 influence firms’ sensing capability.

H3b: The downstream supply chain disruptions caused by COVID-19 influence firms’ seizing capability.

H3c: The downstream supply chain disruptions caused by COVID-19 influence firms’ reconfiguring capability.

2.3. Supply chain resilience

Resilience thinking is an ‘ability of a system to absorb disturbance and still retain its basic functions’ (Walker and Salt 2006, 1). Resilience in supply chains can be understood as an ability to respond to and recover from unpredicted disruptions, such as natural disasters and terrorism (Ponomarov and Holcomb 2009). Undoubtedly, the outbreak of COVID-19 caused severe disruptions to global supply chains, which could not be predicted and where the resilience of the supply chain was highly needed (Ivanov 2020b; Paul and Chowdhury 2020). Folke et al. (2010) characterise resilient supply chains as persistent, adaptable, and transformable, and Pettit, Fiksel, and Croxton (2010) present resilience as a situation where vulnerabilities and capabilities are in balance. Adobor and McMullen (2018) divide resilience in supply chains into engineering, ecological, and evolutionary resilience. Engineering resilience focuses on the efficiency, control, and optimisation of a supply chain and its visibility and velocity in terms of supply and demand volatility. Ecological resilience refers to the capability of a firm to use additional capacities and diversity in supply systems to remove redundancy and increase flexibility. Evolutionary resilience is a supply chain’s ability to evolve and reshape for the better after a disturbance. Hosseini, Ivanov, and Dolgui (2019) describe that supply chain is a system where the capacity to absorb, adapt and restore, establish its resilience capacity, and Chowdhury and Quaddus (2016) state that the ability to respond quickly is a key in supply chain resilience. In the literature, supply chain resilience is often linked to risk management and supply chain vulnerability (Kochan and Nowicki 2018; Macdonald et al. 2018).

The resilience of the entire supply chain is dependent on the company’s capability to react to disruption and reduce its negative impacts (Brandon-Jones et al. 2014). In the context of COVID-19, Yang et al. (2021) further found that a firm’s supply chain risk management capabilities affect the supply chain resilience, meaning that in resilient supply chains the companies have risk management capabilities, which help them mitigate risks and attain continuity in severe disruptions, such as COVID-19. Information processing ability and embedding it into organisational processes through modern technologies, such as blockchain, can shorten the response and recovery time of a supply chain (Dubey et al. 2020). Improvement of visibility and monitoring of supply network by mapping the supply chain and creating digital twins enable firms to react quickly and adapt to changes in the supply network (Ivanov and Dolgui 2021). Having appropriate recovery practices leads to better operational performance in an unexpected disruption (Dabhilkar, Birkie, and Kaulio 2016). Hence, supply chain resilience requires not only organisational-level routines but also sophisticated tools for supply chain monitoring and resource reconfiguration (Ivanov and Dolgui 2021; Yu et al. 2019).

According to Durach and Machuca (2018), socially embedded supplier management and relationships impact the resilience of a firm. Supply chain resilience can be improved significantly by reducing the uncertainty between the supply chain actors through increased cooperation and trust (Dubey et al., “Antecedents of Resilient,” 2019; Dubey et al. 2020). Therefore, investments in the relational skills of supply chain professionals can improve their ability to sense changes in the supply environment. Integrating information flows and processes into a supply chain and utilising technological innovations foster supply chain resilience (Kwak, Seo, and Mason 2018) help firms seize the changes, opportunities, and threats adequately. Furthermore, in the context of humanitarian supply chain and relief operations, it has been found that agility and resilience are complementary dynamic capabilities that add proactivity and are related to sensing capability and performance in supply chains (Altay et al. 2018). Thus, supply chain resilience can be considered a reasonable organisational response to the dynamism of a supply chain (Yu et al. 2019).

Ivanov (2021b) developed an AURA (Active Usage of Resilience Assets) framework to bring more dynamics to the supply chain resilience discussion because, according to Ivanov (2021b), resilience needs to be seen as ‘an inherent, active and value-creating component of the operations management decisions rather than as a passive shield to protect against rare, severe events.’ The optimisation, control, and determination of the best supply chain management practices may no longer be sufficient because of the complex socioecological system of the current world and its ever-changing conditions (Walker and Salt 2006). Firms now need additional capabilities to sense opportunities and challenges, capture or neutralise them and reconfigure their supply chains and resource bases for renewal to have resilient supply chains (Teece 2007, 2012; Maijanen and Jantunen 2014). If a firm cannot develop internal capabilities such as the capability to sense, it may also fail to manage supply chain resilience (Mohammed, Jabbour, and Diabat 2021). Thus, a strong inherent resilience in a supply chain, an anticipative resilience to face crises such as COVID-19, and an adaptive resilience in decision-making (Azadegan and Jayaram 2018) are intertwined with the sensing, seizing, and reconfiguring capabilities of a firm. Based on this, it is argued as

H4a: A firm’s capability to sense influences supply chain resilience.

H4b: A firm’s capability to seize influences supply chain resilience.

H4c: A firm’s capability to reconfigure influences supply chain resilience.

The conceptual model of the study and its hypotheses are shown in Figure 1. Next, the hypotheses are tested using a survey as a research method.

Figure 1. The conceptual model.

The conceptual model showing the hypotheses where COVID-19 upstream and downstream impacts lead to sensing, seizing, and responding capabilities, which further lead to supply chain resilience.

3. Methodology

3.1. Data collection

A survey was conducted in the medical devices industry in Italy and Finland during the summer of 2020. The medical device industry was selected because COVID-19 triggered an extremely rapid increase in demand for certain medical devices, such as masks, gloves, gowns, and ventilator machines, causing significant disruptions to global supply chain operations. These products can be called essential items, and the significance of those drastically increased during COVID-19 (Singh et al. 2021). These supply chain disruptions caused considerable issues for the population and substantial stress on the healthcare systems of many countries during the early part of the pandemic (March–May 2020). The pandemic also created exceptional circumstances for the medical device industry because the medical device supply network became intertwined with the automotive supply network to meet the sudden demands for ventilators, thereby creating the intertwined supply network to improve resilience (Feizabadi, Gligor, and Choi 2021). In the medical devices sector, the appropriate management of supply chains is critical for ensuring that products arrive safely and securely to customers/users. The pandemic exposed major weaknesses in the global medical devices supply chain, especially for companies that rely heavily on Chinese manufacturing for specific products. The medical devices supply chains are different from those in many other sectors because they are quite complex and involve medical material with high value; most importantly, they deal directly with human lives.

Managing the medical devices supply chain requires dynamic processes in which procurement, manufacturing, and the delivery of goods and services to customers are integrated. It must involve a continuous flow of products, orders and information between different phases. The COVID-19 pandemic affected all parts of the supply chain, especially at the clinical frontline, causing a shortage of personal protective equipment and various medical devices (Iyengar et al. 2020). When the flow of these products from manufacturers to patients and clinicians was disrupted, there were severe consequences (Haleem, Javaid, and Vaishya 2020). In Italy, for instance, there was a significant concern regarding the shortage of critical one-time-use PPE, particularly in the northern part of the country, during the early phases of the pandemic. Just before the COVID-19 outbreak, Aldrighetti et al. (2019) found that for healthcare supply chains of Northern Italy, the most suitable mitigation strategy for supply disruptions has a backup supplier. It seems that in normal scale disruption, this could have helped, but in the case of extremely wide disruption of COVID-19, mitigation strategies designed for normal disruptions did not work.

The survey was sent to 466 companies, of which 384 operated in Italy and of which 82 operated in Finland; the numbers reflect the difference in the size of the medical device industry in these countries. We received 110 valid responses (92 from Italy and 18 from Finland), the total response rate was 24%. Of the respondents, 59% represented micro-, small-, and medium-sized companies. 35.4% were medical device manufacturing companies and biomedical technology and diagnostics providers, and 6.4% were services and software companies. Approximately 30% of the respondents did not specify their industry. Most of the respondents were in top or middle management positions (52.7%) (see Table 1).

Table 1. Summary of the characteristics of the respondents.

Features of the respondents	n	%
Company size
Micro	17	15.5
Small	24	21.8
Medium	24	21.8
Large	13	11.8
N/A	32	29.1
Total	110	100
Industry
Medical devices	22	20
Biomedical technology	4	3.6
Laboratory testing and in vitro diagnostics	13	11.8
Services and software	7	6.4
Other	32	29.1
N/A	32	29.1
Total	110	100
Respondents’ position
Top management	33	30
Middle management	25	22.7
Operative task	17	15.5
Expert	1	0.9
Other	1	0.9
N/A	33	30
Total	110	100

3.2. Research instrument

The survey instrument for the study consisted of six components: Resilience, Sensing Capability, Seizing Capability, Reconfiguring Capability, COVID-19 Upstream Impact, and COVID-19 Downstream Impact. Respondents were asked to rate individual questionnaires related to the sub-areas of the survey on a 5-point Likert scale.

The concept of Supply Chain Resilience was measured by metrics that were asked to assess how well each company was able to adapt to the supply chain disruptions caused by COVID-19 and is based on the research of Ambulkar, Blackhurst, and Grawe (2015) (see Appendix 1). Dynamic capabilities were measured by examining three domains: Sensing Capability, Seizing Capability, and Reconfiguring Capability; the metrics were based on Lee and Rha’s (2016) study of abilities. Sensing Capability measures how a company maintains awareness of the business environment during COVID-19, and Seizing Capability measures how well a company performs in strategic decision-making during COVID-19. Reconfiguring Capability measures how well a company aligns its operations with a changing business environment during COVID-19.

This survey also created a metric for how much COVID-19 affects a company and its supply chain, directly and indirectly, divided into two constructs to examine COVID-19’s effects on the upstream and downstream of the supply chain. These sections were named in the survey instrument as COVID-19 Upstream Impact and COVID-19 Downstream Impact (see Appendix 1).

3.3 Reliability of the research instrument

The reliability assessment involves the analysis of the factor structure in terms of the significance and weight of factor loadings, validity and reliability, and cross-loadings of the latent variables (see Appendix 2). The operationalised constructs were assessed using (1) measurement construct reliability (CR), (2) factor structure validity by average variance extracted (AVE), and (3) the measurement model’s discriminant validity (Fornell and Larcker 1981; Gefen and Straub 2005; Henseler, Ringle, and Sinkovics 2009). The reliability assessments for all the constructs are found in Table 2. The critical limit for reliability by CR is .05, which a valid measurement model should exceed (Kline 2011; Little et al. 2002). The latent constructs’ CRs varied from .790 to .888, indicating good reliability. The structural analysis of the measurement model shows that all factors loadings of the measurement model are significant at p < 0.01, and the validity of the latent constructs by AVE is greater than .50, ranging from .563. to .726 (Fornell and Larcker 1981). The assessment of the discriminant validity includes three dimensions (1) the cross-loadings of the observed items, (2) the square root of AVE, and (3) HTMT criterion (Henseler, Ringle, and Sinkovics 2009; Hair, Sarstedt, and Ringle 2019). According to the tests, the observed items formed unequivocal latent factors where the cross-loadings varied from −.506 to .796. Furthermore, discriminant validity demonstrates an acceptable level in which the AVE square roots are higher than the correlations between any of the latent factors. Lastly, the HTMT ratios of the latent constructs do not exceed the critical value of < .90 varying from 0.179 to 0.657.

Table 2. Measurement reliabilities.

	Loading	t-value	p-value	Mean	SD	CR	AVE
CovUI (COVID-19 Upstream Impact)						0.790	0.563
SUI1	0.900	9.410	****	3.02	1.498
SUI2	0.683	3.268	***	1.862	2.121
SUI3	0.641	4.360	****	2.532	1.628
CovDI (COVID-19 Downstream Impact)						0.844	0.645
SDI1	0.795	4.616	****	0.473	1.128
SDI2	0.706	2.954	***	2.042	2.056
SDI3	0.898	5.050	****	1.432	1.946
SEN (Sensing)						0.805	0.587
SEN1	0.600	2.683	***	3.49	1.551
SEN2	0.924	4.539	****	3.44	1.227
SEN3	0.741	3.036	***	3.29	1.806
SEIZ (Seizing)						0.848	0.653
SEIZ1	0.742	5.449	****	3.61	1.582
SEIZ2	0.915	23.313	****	3.66	1.589
SEIZ3	0.755	10.755	****	3.18	2.151
RECO (Reconfiguring)						0.834	0.636
RECO1	0.541	3.285	***	2.63	1.939
RECO2	0.888	13.961	****	3.13	1.867
RECO3	0.910	19.811	****	3.47	2.159
RESL (Resilience)						0.888	0.726
RESL1	0.799	7.213	****	3.88	2.054
RESL2	0.881	27.990	****	3.52	1.636
RESL3	0.873	16.369	****	3.22	2.241

Note: n: not significant; *statistically significant at p < 0.1; **statistically significant at p < 0.05; ***statistically significant at p < 0.01; ****statistically significant at p < 0.001.

3.4. Empirical results from PLS path modelling

The key effects in the default model were analysed according to the hypotheses (Table 2). In the analysis, the size of the bootstrap sample was n = 110, based on the original sample. We repeated the resampling of the data 5000 times (basic bootstrapping) in the analysis, which is considered sufficient for parameter estimating in the model (Henseler, Ringle, and Sinkovics 2009). The quality of the structural model was tested and validated using the steps of (1) collinearity and overall fit, (2) explanatory power of the model, and (3) path significances. Lastly, an assessment of potential endogeneity issues of the empirical model is discussed.

The collinearity and goodness were evaluated to validate the structural model. The variance inflation factor of the latent constructs (inner-VIF) did not show any serious collinearity issues, as the highest-value, inner-VIF = 1.394, remained below the critical value of VIF = 5 (Hair, Sarstedt, and Ringle 2019). The explanatory power and goodness of the model can be evaluated by the proportion of the variance, explained by an endogenous variable (R²), the predictive relevance of the model for an endogenous variable (Q²), and by the sizes and significances of the path coefficients (Astrachan, Patel, and Wanzenried 2014). In practice, the R² is an indicator of the proportion of the variance captured in the endogenous constructs and the Q² is an indicator of whether the endogenous construct can be accurately predicted by the structural model (Sarstedt et al. 2014; Hair et al. 2019). The Q² for the endogenous constructs must be positive to signal any predictive relevance, whereas other critical are at .25 and .50, depicting medium and high accuracy of the structural model (Hair et al. 2019; Inigo, Ritala, and Albareda 2020). A remarkably high level of R² can also signal existing collinearity issues in the model, which should be examined with VIF before interpreting the results. In the path model, the R² values for the latent variables were SEN = .029, SEIZ = .299, RECO = .300, and RESL = .438. The Q² values for the endogenous constructs were SEN = .008, SEIZ = .147, RECO = .115, and RESL = .245. Overall, the explanatory power of the model was high and predictive accuracy was medium regardless of the phenomenon in focus, which is complex in that it includes multiple influences outside of the tested model (Abelson 1985; Prentice and Miller 1992).

Endogeneity issues in empirical research grounds on omitted variables, simultaneity, measurement error, and selection bias which potentially cause risks of faulty conclusions (Busenbark et al. 2021; Hill et al. 2021). In this study, the empirical model is built around the established theoretical framework of the DCV, from which none of the concepts has been dropped. The expected relationships between the concepts in the model have also received some support from the earlier literature by which bias from unexplained variance caused by omitted variables can be considered low. Simultaneity and measurement errors can be caused by the relatedness of the measured dimension, method bias caused by the applied survey method and shared context of environmental shift in the industry. To address those issues, full collinearity tests and the Gaussian Copula procedure have been applied to reveal potential endogeneity. Common method bias measures the risk of contamination of the model by variation emerging from the survey instrument or contextual factors of respondents in terms of dependency inflation (Baumgartner, Weijters, and Pieters 2021). The full collinearity test procedure was applied to assess common method bias because of the selected PLS-SEM approach and reliable results were achieved (Kock 2017; Baumgartner, Weijters, and Pieters 2021). The test procedure targets to assess to construct VIFs that should be equal to or lower than the critical value of 3.3 to judge a model not contaminated by common method bias (Kock 2017). The full collinearity test procedure for common method bias shows that VIF varies between VIF_min  > 1.042 and VIF_max < 1.921, indicating no common method bias in the model. The Gaussian Copula procedure addresses to reveal potential endogeneity and identification issues of the empirical model, which is needed because the goal of the study is to test less explained relationships between concepts and testing hypothesis (Hult et al. 2018; Hair et al. 2019). By the results (see Appendix 3), any copulas did not have significant effects at p < .05, indicating low risks of pathological endogeneity in the empirical model.

The quality of the sample in the PLS-path modelling can be assessed by requisite sample size, non-response bias, and selection bias. Model configuration sets requirements for the minimum sample size, for which ‘10-times rule’ has widely been used as an estimation method, saying a minimum count of observations equals to 10 times the maximum number of paths pointing the latent in the inner or outer model (Hair, Ringle, and Sarstedt 2011). By following the rule, the maximum count of incoming links in the estimated model is six, which leads to the requirement for a sample size of 60 as a minimum. Furthermore, the sufficiency of the sample is assessed by the effect sizes (i.e. f²) of significant paths in the inner model, which has critical values of 0.02, 0.15, and 0.35 termed as small, medium, and large effects (Sullivan and Feinn 2012; Hair et al. 2017; Haverila et al. 2021). The test statistics show that effect sizes vary from small to medium effect (f²_min > .0112, f²_min > .246), indicating meaningful relations and enough potential of the sample to provide enough statistical power. ANOVA test was applied as an estimator in assessing the non-response bias based on a comparison of the early and late responses in the questionnaire (Armstrong and Overton 1977). The tests results show that no serious non-response bias exists in the data by mean comparison of the latent factors’ scores at p < .05. Overall, the sample size is valid in technical terms. Last, slight selection bias is probable in the data because of non-probabilistic sample was drawn from registers, leading to self-selection and coverage bias (Li and Hitt 2008; Lehdonvirta et al. 2021). Regarding the selection bias, it is likely that data are somehow emphasised to represent (i) organisations that have already recognised certain changes in their business, and (ii) companies that have traditions to use internet channels because e-mail contacts were not complete considering the smallest businesses.

According to the default model (see Table 3), the relationships between the core concepts of the dynamic capability framework exist as predicted by the literature, where positive relationships were found between sensing and seizing and between seizing and reconfiguring capabilities. The results confirm hypotheses H1a and H2b. The impact of COVID-19 on dynamic capabilities was analysed using a two-dimensional model in which the upstream and downstream impacts formed separate latent factors. Later, the direct effects of the COVID-19 impact latent constructs, CovUI and CovDI, the core concepts of dynamic capabilities were tested in the path model. The analysis revealed two distinct impact patterns of COVID-19 on firms’ expectations regarding their dynamic capabilities depending on whether market disruptions come from the upstream or downstream sides of the supply chains. The results show that the only statistically significant impact of CovUI is its relation to expected seizing capability, confirming hypothesis H2b, whereas hypotheses H2a and H2c were rejected by the path model. Similarly, only one statistically significant impact of CovDI was found, namely its influence on reconfiguring capabilities, confirming hypothesis H3c, whereas hypotheses H3a and H3b did not receive support from the tested statistical model. Finally, the impact of dynamic capabilities on supply chain resilience was tested, revealing the statistically significant effects of seizing and reconfiguring capabilities on supply chain resilience, confirming hypotheses H4b and H4c, whereas the model did not indicate support for hypothesis H4a.

Table 3. Main model to test the hypotheses and results of the post hoc analysis.

Hypothesis	Path	β	T Statistics	P Values
Main model
H1a	SEN → SEIZ	0.367	3.142	**
H1b	SEIZ → RECON	0.307	2.913	**
H2a	CovUI → SEN	−0.181	1.051	n
H2b	CovUI → SEIZ	−0.368	4.074	***
H2c	CovUI → RECON	−0.205	1.535	n
H3a	CovDI → SEN	0.063	0.407	n
H3b	CovDI → SEIZ	0.104	0.911	n
H3c	CovDI → RECON	0.421	2.948	**
H4a	SEN → RESL	0.102	0.786	n
H4b	SEIZ → RESL	0.344	2.450	*
H4c	RECON → RESL	0.402	3.607	***
Post hoc tests for total indirect effects
	SEN →→ RECON	.059	1.913	n
	CovUI →→ RESL	−.304	4.421	****
	CovDI →→ RESL	.235	2.469	*

Notes: n: not significant; *statistically significant at p < 0.05; **statistically significant at p < 0.01; ***statistically significant at p < 0.001.

The post hoc analysis was conducted to verify the total effects of the COVID-19 disruption on supply chain resilience. The COVID-19 disruption has a statistically significant full impact on supply chain resilience, following the pattern recognised in the main model. Upstream COVID-19 disruption has a strong negative total impact on supply chain resilience, which is partly mediated by seizing. In contrast with previous studies, downstream COVID-19 disruptions have a statistically weak significant positive influence on supply chain resilience, which is partly mediated by reconfiguring capability.

4. Discussion

The COVID-19 pandemic is an extreme example of force majeure, creating supply chain vulnerability, disruptions, and risks that the modern business environment has never faced before (Araz et al. 2020; Craighead, Ketchen, and Darby 2020). Companies were unprepared for the rapid disruptive effect that COVID-19 had and still has on global supply chains (Choi, Rogers, and Vakil 2020). This study used the DCV as a theoretical framework to study first the impact of COVID-19 on firms’ dynamic capabilities and then the influence of those on supply chain resilience. Our study also tested the relationships between the core concepts of dynamic capabilities.

4.1. Theoretical implications

Our study provides empirical evidence and theoretical insight regarding the impacts of extensive supply chain disruptions on the firm’s dynamic capabilities and their influence on supply chain resilience. Following the basic assumptions of DCV, we investigated the relationships between the core concepts of the dynamic capability framework – sensing, seizing, and reconfiguring. As hypothesised in H1a and H1b, the path-dependency from the firm’s capabilities of sensing and seizing to the ability to reconfigure was significantly true. The results illustrate a positive relationship between sensing and seizing and between seizing and reconfiguring capabilities. This shows that the basic logic of DCV works in the case of disruptive events during which companies try to reconfigure their resource bases to survive. The results follow the logic of previous studies of DCV (Helfat and Peteraf 2009) and show that these capabilities form an interrelated chain of activities, meaning that firms need to first sense markets and the environment by detecting and gathering information from the business environment. Then the firms need to capture the opportunities sensed or neutralise the sensed threats. As highlighted by Lee (2021), super abilities are needed in today’s business environment and thus, ordinary capabilities are not enough, but dynamic capabilities are needed to respond to supply chain disruptions effectively. It could be argued that firms that could detect the severity of COVID-19 in its early stages and had systems for monitoring interconnected supply chains may have had better opportunities to mitigate disruptions and risks and ensure viability of their supply chains. Similar findings were also shown by Ivanov and Dolgui (2020). However, to neutralise the threats and risks, firms need to realign their resources and assets.

In this paper, the impacts of COVID-19 were reflected in the firms’ upstream supply chains (hypotheses H2a,b,c) and the downstream supply chains (hypotheses H3a,b,c). According to our results, the impacts of COVID-19 on upstream supply chains did not influence sensing or reconfiguring capability, as suggested in H2a and H2c. This result indicates that at least in the medical device industry, which was the empirical context of our study, firms could have been aware of the risk and were thus prepared and had recovery mechanisms for their supply chains. However, the results of testing H2b show that the impacts of COVID-19 on firms’ upstream supply chains towards the suppliers and supply market influence firms’ capabilities to seize the opportunities or neutralise the threats by creating decision-making processes and structures. Logically, firms’ abilities to neutralise threats and seize opportunities were the most heavily influenced because as Choi, Rogers, and Vakil (2020) stated, the disruptions caused by COVID-19 were sudden and rapidly evolving. In addition, the magnitude of the disruption in the upstream supply chain may have had a negative effect on the firm’s seizing ability because the first phases of the pandemic hit, especially the upstream supply chains as Ivanov (2020a) and Paul and Chowdhury (2020) have reported. Previous studies found that seizing capabilities are visible in firms’ activities, such as creating innovations and selecting new business models (Maijanen and Jantunen 2014), and the results of our analysis show that, in the context of a global pandemic, if deliveries were late or some products were impossible to obtain, firms had to seek out new solutions and innovations.

The impacts that COVID-19 had on the firms’ downstream supply chains, towards customers, statistically influence only their reconfiguring capabilities as proposed in H3c significantly. The ripple effect in the supply chain means that disruptions in the upstream supply chain are reflected towards the downstream supply chain. Thus, reconfiguration capability in a firm is vital to keep the existing customers and ensure the deliveries in the downstream supply chain, as also suggested by Dolgui and Ivanov (2021). Even though disruptions caused by COVID-19 created huge risks, many firms also noticed opportunities that the pandemic created because of the disruptions and severe shortages of medical and hygiene products. For example, Finnish distilleries turned their alcohol production lines towards hand sanitisers and package manufacturers began making plastic face shields. A recent survey conducted by Randstad Professionals (2020) indicated that companies in the Italian fashion sector temporarily converted , donating masks and gowns to hospitals and civil protection. Companies in the medical device, automotive, beverage, cosmetics, and manufacturing sectors have implemented partial and temporary conversions, while the textile, plastics, chemicals, and printing sectors have implemented long-term reconversions.

As the capability to reconfigure means the ability to realign and recombine organisational structures and firm assets as a response to changing markets (Teece 2007), it was shown that the effects of COVID-19 on downstream supply chains pushed companies to realign their resources to better respond to customer demand. It seems that the upstream disruptions pushed companies to react to threats and opportunities in the supply market, while downstream disruptions leveraged reconfiguring capabilities. In the studies of the humanitarian supply chain, it has been found that having resilience in supply chains improves performance before and after the disruption event because of better preparedness which is natural in the humanitarian setting (Altay et al. 2018). Obviously, for private companies preparing for unpredictable disruptions beforehand would have helped companies to adapt to the customer demand during the COVID-19 pandemic better.

Seizing and reconfiguring capabilities influenced supply chain resilience as proposed in hypotheses H4b and H4c. However, sensing capability did not have a significant relationship with supply chain resilience (hypothesis H4a). This might be because sense is not easily transformed into direct actions by supply chain management, particularly in the context of a rapidly spreading pandemic. Even though some previous studies showed that relational skills improve resilience (Durach and Machuca 2018), indicating that the ability to sense changes in the supply market has the potential to improve, our results did not support that. Altay et al. (2018) have shown that dynamic sensing ability relates to proactivity and agility in the humanitarian supply chain, but that was not the case in our medical devices industry data. The changes due to the pandemic might have been too severe and too rapid, meaning that companies in the medical devices industry could not sense the market and turn that into supply chain resilience. The seizing capability, however, is evident in our empirical context. The use of new technological innovations fosters supply chain resilience, as Kwak, Seo, and Mason (2018) stated, and helps firms seize possible changes in supply markets. According to our results, reconfiguring capability has the strongest influence on supply chain resilience. This is in accordance with studies that connect risk management with resilience in supply chains (e.g. Brandon-Jones et al. 2014; Dabhilkar, Birkie, and Kaulio 2016). The ability to reconfigure supply chains and implement recovery practices and routines influence supply chain resilience, which improves the viability of the supply chain beyond the COVID-19 pandemic (Ivanov 2020b). We can conclude that supply chain resilience during a sudden global market disruption seems to depend on firms’ capabilities to re-orchestrate their supply chains reactively, whereas proactive sensing capability did not have a direct effect.

4.2. Managerial implications

From a managerial point of view, the most interesting results are those that will help companies learn from the pandemic and develop their operations for the future. Regarding the upstream part of supply chains, the results indicate that medical device companies, given the effects of COVID-19, leveraged their seizing capabilities. This indicates that companies need to make serious decisions involving their supply base and capture recognised opportunities in their supply chain. Seizing and reconfiguring supply chain operations improves resilience and therefore helps companies recover faster, reduce their vulnerability, and also generate benefits such as faster lead times and innovative new sources.

Another implication concerns the visibility of the lower tiers of supply chains. To manage and reduce the impact of the pandemic, companies need to build up strong relationships with key suppliers and invest in supply chain planning systems that can provide visibility across the entire supply network, not just to the first tier. Achieving a higher visibility, regarding production schedules, inventories at the supplier location, and shipment statuses, is essential for predicting supplier shortages and responding accordingly. Companies could also learn and adopt practices and tools from operations of humanitarian supply chain organisations to improve their sensing capabilities and preparedness for disruptions (e.g. Iakovou et al. 2014; Bhusiri et al. 2021). Finally, COVID-19 can be a trigger for reshoring decisions. It seems to have accelerated companies’ decisions to relocate production (or part of it) to Europe due to disruptions in Chinese production.

Regarding the downstream supply chain, the pandemic pushed companies to leverage their reconfiguring capabilities. Decisions concerning the realignment of resources to better respond to customer demand were heavily involved. Reconfiguration decisions may involve several areas, such as logistics, where it may be crucial to evaluate alternative outbound logistics arrangements. Firms need to ensure that their logistics service providers have enough capacity to ensure continuous production activity. Another critical area is inventory management; the pandemic could result in a potential inventory shortage, but communication with key customers is necessary to find alternative supply arrangements, avoid conflicts with customers, and minimise losses.

5. Conclusions

The COVID-19 pandemic caused large-scale, rapid effects on all operations globally, making it important to study its impact on supply chain disruptions to learn how to best adapt to such disruptions in the future. This study specifically investigated the impact of the COVID-19 pandemic on the resiliency of the medical device supply chain in Italy and Finland in 2020. The topic was approached from the perspective of DCV. The conceptual model of the study included three layers: the scale of the perceived COVID-19 impact, the expected dynamic capabilities of firms, and expected supply chain resilience. The COVID-19 impact was examined from the upstream and downstream supply chains, enabling the analysis of disruptions to demand and supply. Dynamic capabilities were studied from two perspectives according to the study's goals, which were to understand the links between sudden global disruptions, dynamic decision-making performances of firms, and supply chain resilience. To deploy the theoretical model, the hypothesised relationships and the theoretical concepts were operationalised into a path model, through which we studied how COVID-19 impacts firms’ dynamic capabilities and the influence of those capabilities on supply chain resilience.

Our findings increased the understanding of the effects of dynamic capabilities on supply chain resilience, particularly supply chain management during global pandemic disruptions. This paper contributes to DCV by showing that global disruptions have different effects on dynamic capabilities depending on whether the cause emerges from downstream or upstream. Based on the findings, the link between firms’ dynamic capabilities and supply chain resilience is biased towards operative perspectives in sudden disruptions. This emphasises problem-solving and scaling performance during decision-making, whereas awareness of the supply environment has a minor role. Nevertheless, the COVID-19 upstream and downstream impacts have direct effects on perceived resilience. The findings also show the existing mediating role of dynamic capabilities in the given structure. Finally, the critical decisions during emergent disruptions seem to be related to maintaining supply chain adaptability, which provides necessary flexibility against threats by considering assets, resources, and demand.

5.1 Limitations and suggestions for future research

Our research has limitations that may affect its results. First, the survey was measured over a single period, meaning that all the potential changes caused by the COVID-19 pandemic may not have been observed. In addition, it would be necessary to investigate a larger sample of firms to validate the results further. The medical device industry has been central in the pandemic, but the generalisability of the results to other industries requires further research. It can be expected that the COVID-19 pandemic has affected different sectors in different ways although similarities are likely to be found. Responding to the pandemic has presumably been quite complex for many companies, and a quantitative survey can only reveal some features of the operation. A qualitative study with an empirical case study analysis could provide further insight into how medical device companies reacted to COVID-19 and why some specific decisions were made.

The extensive effects of COVID-19 on supply chains show that still much can be done to better adapt to such major disruptions in supply chain risk management. Future research should explore practices that firms could employ to adapt and respond to such serious supply chain disruptions. It would also be interesting to develop a comparative analysis to determine how the effects of a pandemic differ across sectors and countries. As the COVID-19 pandemic is on-going and has not lasted long, its relevant research is still in early phases. In the future, focusing on lessons that can be learned would be important for better preparing companies for similar catastrophic situations and disruptive events.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

Data are not available due to industry-specific restrictions. Due to the nature of this research, participants of this study did not agree for their data to be shared publicly, so supporting data are not available.

Additional information

Notes on contributors

Anni-Kaisa Kähkönen

Anni-Kaisa Kähkönen is a Professor of Supply Management at LUT University, School of Business and Management (Finland). Her current research interests include sustainable purchasing and supply management, management of multi-tier sustainable supply chains, supply strategies, and value creation in purchasing and supply management.

Pietro Evangelista

Pietro Evangelista is Research Director in Logistics and SCM at the Institute for Research on Innovation and Services for Development (IRISS) of the Italian National Research Council (CNR). His current scientific interests focus on de-carbonisation strategies for freight transport and logistics, the impact of digitalisation and ICT on supply chain management, and the role of knowledge management in logistics.

Jukka Hallikas

Jukka Hallikas is a Professor of Supply Chain Management at LUT University, School of Business and Management (Finland). His current research interests focus on purchasing and supply operations, supply chain risk management, and digitalisation of supply chains.

Mika Immonen

Mika Immonen is Associate Professor at LUT University, School of Business and Management (Finland). He has worked in various projects focusing on health care technology and consumer behaviour in health-related services. His accomplished academic works are services systems’ structures, emerging business models and customer value creation in multi-stakeholder environments.

Katrina Lintukangas

Katrina Lintukangas is Professor of Supply Management at LUT University, School of Business and Management (Finland). Her current research focuses on innovations, sustainable purchasing and supply management, risk management and supply chain resiliency, and value creation in public procurement.

Appendices

Appendix 1. Research instrument

Table

Supply Chain Resilience	Ambulkar, Blackhurst, and Grawe (2015)
Assess how well your company has adapted to the supply chain disruptions caused by COVID-19. Please rate the following items using the following scale: 1 = strongly disagree, 5 = strongly agree.
RESL1: We are able to cope with changes brought by the supply chain disruption caused by COVID-19.
RESL2: We are able to provide a quick response to the supply chain disruption caused by COVID-19.
RESL3: We are able to maintain high situational awareness at the time of COVID-19.
Sensing Capability	Lee and Rha (2016)
Assess how well your company maintains awareness of the business environment during COVID-19. Please rate the following items using the following scale: 1 = strongly disagree, 5 = strongly agree.
SEN1: We can feel the major potential opportunities and threats in our supply chain.
SEN2: We have good observation and judgment abilities in our supply chain.
SEN3: We have frequent interactions throughout the organisation to acquire new knowledge related to product development, process innovation, or logistics and distribution practices.
Seizing Capability	Lee and Rha (2016)
Assess how well strategic decision-making is performed in your company during COVID-19. Please rate the following items using the following scale: 1 = strongly disagree, 5 = strongly agree.
SEIZ1: We can quickly deal with conflicts in strategic decision-making in our supply chain.
SEIZ2: We can make timely decisions to deal with supply chain problems.
SEIZ3: We can reconfigure resources in time to address environmental change.
Reconfiguring Capability	Lee and Rha (2016)
Assess how well your company aligns operations with the changed business environment during COVID-19. Please rate the following items using the following scale: 1 = strongly disagree, 5 = strongly agree.
RECO1: We can successfully reconfigure supply chain resources to come up with new productive assets.
RECO2: We are able to align (or re-distribute) skills to meet the current needs of the whole supply chain.
RECO3: We can effectively integrate and combine existing resources into novel combinations in the supply chain.
COVID-19 Upstream Impact	Authors modified based on Pettit, Fiksel, and Croxton (2010)
How much has COVID-19 directly or indirectly affected the following areas in your company? Please rate the following items using the following scale: 1 = not at all, 5 = very much.
SUI1: Overall supply chain
SUI2: Procurement costs/prices for the purchased items
SUI3: Delivery reliability (on-time delivery, order accuracy)
COVID-19 Downstream Impact	Authors modified based on Pettit, Fiksel, and Croxton (2010)
How much has COVID-19 directly or indirectly affected the following areas in your company? Please rate the following items using the following scale 1 = not at all, 5 = very much.
SDI1: Product quality of our final product(s)
SDI2: Responsiveness to customer demands
SDI3: Customer satisfaction

Appendix 2. Cross loadings

Table

	CovUI	CovDI	SEN	SEIZ	RECO	RESL
SUI1	0.9	0.12	−0.12	−0.4	−0.28	−0.51
SUI2	0.68	0.58	−0.24	−0.25	0.08	0.03
SUI3	0.64	0.2	0.01	−0.16	−0.1	−0.15
SDN1	0.23	0.79	−0.02	0.05	0.27	0.19
SDN2	0.46	0.71	0.03	−0.14	0.07	0.06
SDN3	0.28	0.9	0.01	−0.04	0.35	0.29
SENS1	0.05	0.22	0.6	0.23	0.21	0.24
SENS2	−0.2	−0.09	0.92	0.43	0.17	0.25
SENS3	−0.16	−0.05	0.74	0.29	−0.02	0.23
SEIZ1	−0.36	−0.1	0.3	0.74	0.42	0.38
SEIZ2	−0.24	−0.05	0.34	0.92	0.26	0.39
SEIZ3	−0.33	0.09	0.38	0.76	0.23	0.51
RECO1	−0.32	−0.01	0.08	0.04	0.54	0.18
RECO2	−0.16	0.24	−0.05	0.32	0.89	0.41
RECO3	−0.11	0.42	0.27	0.41	0.91	0.58
RESL1	−0.14	0.37	0.16	0.14	0.43	0.8
RESL2	−0.26	0.28	0.42	0.6	0.51	0.88
RESL3	−0.48	0.08	0.15	0.51	0.45	0.87

Appendix 3. Results of the Gaussian copula procedure for testing endogeneity

Test procedure follows guidelines published in article and tutorial by Hult et al. (2018).

Tutorial available at https://www.pls-sem.net/pls-sem-academy/gaussian-copula-files/