Methodology / method
In this chapter the outline of the methodology with regards to conducting the research for the thesis will be presented. There will be reference to the research techniques used and how it may be helpful for other re-searchers when performing research of the same nature. A discussion of the research approach, strategy,
methodological choice and data collection methods will be presented. To conclude this chapter, there is also a description and discussion of the research’s limitations which should aid further work in this area as well as a discussion regarding its generalisability and ethical considerations.
The study was originally instigated based on the pharmaceutical segment of logistics and supply chain management of pharma logistics which the authors found interesting. This, combined with some initial reading, led to the defining and creation of concrete research question and suitable research strategy in order to answer them. Furthermore it was imper-ative that the research strategy be structured and tailored to the purpose of the study as this was what would assist in addressing not only the research problem but also its associated questions.
Using Figure 3.1 it is possible to highlight the entire research process from start to finish and therefore provide a guide as to how the study has been conducted. The first column under the headline ‘industry focus’ shows the proceeding of the authors that led to the ‘in-troduction’ starting with the definition of logistics & supply chain management as the field of research and narrowing it down to the pharmaceutical industry within Sweden. The sec-ond column illustrates the ‘theoretical framework’ created subsequently, which provided insights in important subjects accompanying the topic at hand and helped with making the research questions more precise. The following column covering the ‘methodological choice’ shows the decisions on the approach, research strategy, data collection method and analysis based upon the need to address the research questions. The next column deals with the empirical data gathered in the semi-structured interviews constituting the chapter called ‘empirical findings’, whereas the last column concerns the analysis of the empirical data and theory, as well as the conclusions drawn, thereby constituting the ‘analysis’ and the ‘conclusion’ chapters. Though the chapters and subchapters are depicted in a rough chronological order of their creation in Figure 3.1, they needed to be amended several times and their contents changed partly throughout the process of generating the study
In general the two alternatives deduction and induction are regarded as the major ap-proaches, which can form the basis for the reasoning of a research. Abduction, the third alternative, is usually seen as a combination of the aforementioned options (Saunders, Lew-is, & Thornhill, 2012).
Using deduction, the first of those two approaches, the logically drawn conclusion is actual-ly deemed to be true in case all of the assumptions set up beforehand. Induction on the other hand is making use of gathering impressions, which are then assessed thoroughly and can be identified as factors favouring a defined outcome – thus the conclusion is drawn. Although it is regarded as most likely to be true, there is still some uncertainty about the actual outcome remaining (Saunders et al., 2012). A third approach, abduction, starts with the observation of an interesting outcome, which is used as the basis to elaborate on fac-tors that are considered to lead to such a result in all probability. The justification is that an occurrence all these factors leads to the natural consequence of the outcome applying as well (Ketokivi & Mantere, 2010). According to Kovacs and Spens (2005) the abductive approach avoids the major disadvantages of the deductive and inductive approach, which are a lack of empirical sensitivity in the first and the risk of leading to theoretically uninterest-ing findings in the latter case (Polsa, 2013).
For the specific purpose of this thesis an abductive approach was adopted and similar to deductive reasoning a theoretical framework based on existing literature was compiled first. Unlike in deductive and rather matching with inductive reasoning, the elaboration of the thesis commenced by conducting semi-structured interviews in order to achieve insights reaching beyond what was already written in existing literature. This is described more de-tailed in chapter 3.6, ‘Analysis of qualitative data’.
Among the existing types of research strategy the case study is one that allows the investi-gation of the research subject within its real-life context and can be regarded as its most striking feature. It can be conducted in different ways, among them the selection of either single cases, because they are typical or unique on the one hand or the selection of more cases to test, whether they generate alike results on the other (Saunders et al., 2012). Case studies in general permit the utilisation of qualitative as well quantitative methods and are regarded to be suitable for exploratory, as well as explanatory studies and a good choice to facilitate the answering of ‘what?’ and ‘how?’ questions (Saunders et al., 2012; Yin, 2009).
In order to deliver reliable results case studies require triangulation based on multiple sources of evidence and/or methods of data collection (Williamson, 2002).
It can be stated that this thesis constitutes an analytical study that, due to the scarcity of existing research, is exploratory in nature as it attempts to analyse “what is happening; to seek new insights; to ask questions and to assess phenomena in a new light” (Robson, 2002, p. 59). Due to the study’s requirements matching the above mentioned characteristics of case studies, the adoption as the research strategy was reasonable. The study has been conducted using a dual case study, selecting companies among the Swedish bricks-and-mortar pharmacy chains, thus ensuring triangulation, as company specific and concurrent answers could be identified. Employing this research method allowed the depiction of the circumstances under which the study was conducted, which is important, due to the possi-bility of influencing the results.
Company selection process
As the emphasis of the study lies on the potential introduction of the home delivery of PO medicine to patients in Sweden and the necessary realigning of a pharmacy’s supply chain, the focal companies needed to be pharmacy chains, operating bricks-and-mortar outlets and not yet offering such a service. Invitational letters to participate in the study were sent to the six largest pharmacy chains by store number in Sweden via mail. In appropriate time these companies were called and asked about their willingness to participate, however, sev-eral pharmacy chains declined due to the fear of an indirect revealing of their further busi-ness plans even in case of anonymisation as the identifiability is high owing to the limited number of competitors. As soon as one senior manager of Pharmacy 1 and Pharmacy 2 agreed to give an interview a pharmacy manager of Pharmacy 1 granted an interview as well.
Aiming to get an overview of logistics related issues relevant in pharmacy supply chains, invitational letters were sent to ten LSPs potentially offering solutions to the pharmacy sec-tor in Sweden via mail. The companies were contacted via follow-up phone calls later on with the majority of LSPs needing to be excluded since they did not offer pharmacy specif-ic solutions in Sweden at all or focused on business to business (B2B) deliveries only. Nev-ertheless an interview with a senior manager at an LSP offering both B2B and B2C services for the pharmacy and life science sector in Sweden was arranged.
Furthermore an interview with a manager at a pharmaceutical company, which already of-fered a delivery service, could be established as well. This was deemed to be beneficial for deepening the understanding of the requirements of such services, as well as the processing and the supply chain aspects involved. The company is referred to as Anonymous Pharma-cy throughout the thesis, as it requested its data to be removed later on.
Quantitative methods are considered to be especially suitable when a deductive approach is employed by mainly harnessing numerical data and statistics and are collected via the em-ployment of research strategies such as experiments, surveys or standardised interviews. Qualitative methods rely on inductive processes (Saunders et al., 2012) and in contrast make use of data other than numerical, which is gathered by methods such as semi-structured and in-depth interviews utilised in research strategies such as case studies, grounded theory, narrative inquiries or action research.
These can either be utilised in mono method designs, using only one data collection meth-od or designs using multiple. Those designs can further be distinguished into two groups; the first one, which employs two collection methods generating the same kind of data, so either qualitative or quantitative, are called multi method designs and the second one, which employs two collection methods generating different kind of data, so both qualita-tive and quantitative are called mixed method designs.
In the case at hand a multi method qualitative design was adopted to gather primary quali-tative data through several semi-structured interviews supported by secondary qualitative data obtained from documentation, stemming mainly from public agencies, such as the Swedish Medical Products Agency4, the Swedish government’s information platform and the UK’s NHS.
Interviews are considered one of the most important and common methods of collecting data when it comes to case studies and are most often associated with qualitative research studies (DiCicco-Bloom & Crabtree, 2006). Many modern texts differentiate between three types of interviews, namely structured, semi-structured and unstructured. Furthermore the data collection in interviews can follow a standardised or non-standardised approach (Saunders et al., 2012). The use of interviews in research work requires knowledge and con-tact to key players in the prevailing area and direct access to them, in order to be able to fulfil the purpose (Denscombe, 2010).
DiCicco-Bloom and Crabtree (2006) state that the questions in semi-structured interviews remain open-ended, with other questions developing from the dialogue between the inter-viewer and the interviewee. According to Sorrell and Redmond (1995) suggestions it can be assumed that in order for interviews to generate valuable data the interviewer should main-tain control over the interview and not contribute too much extra.
Due to the nature of the study and the lack of available information on the research topic, it could be argued that the interviews required should be semi-structured. This method of data collection was favoured, due to the importance of gaining more in-depth insights into the motivations of the pharmacy actors to enter the market of home delivery of PO medi-cine and their opinion on adaptations needed to their supply chains as a result. To render a comprehensive picture about the research subject, the utilisation of semi-structured inter-views is supported by their interactive characteristics, which on the one hand allow prepa-ration of key questions in advance, but also enable their revisions later on and the ability to ask additional probing questions during the data collection process. In essence those key questions are the ones destined to obtain the data to be able to answer the research ques-tions, whereas additional questions aim at gathering data that enables to reveal the context of the research questions.
Table 3.1 describes the order of interviews carried out and introduces the interviewee and company shortcuts used throughout the study. All of the interviews were recorded, after having been granted permission, and, as suggested (Saunders et al., 2012), transcribed sub-sequently in order not to lose information The availability of official documentation related to the topic of the thesis was scarce and access to company internal documents remained prohibited, justified by their confidentiali-ty as they included details about recent market monitoring and further business plans.
Analysis of qualitative data
Ideally, data analysis should occur concurrently as the data is being collected when adopting case studies as a research strategy. Yin (2009) stated that in order to devise a theoretical framework, one should identify the main variables, components, themes and issues in the study. This will form a basis to explain the expected outcomes later on when mixed with the author’s own experience and knowledge. Once this has been created, it is then possible to perform a direct analysis of the collected data.
As using an abductive approach is very much about moving from theory to empirical data and vice-versa constantly throughout the study (Dubois & Gadde, 2002) a theoretical framework was comprised to understand the basics of the pharmaceutical sector, its char-acteristics, the Swedish market, as well as home delivery in general. Thus, a list consisting of key questions and additional ones6, as well as possible follow-up questions, to be poten-tially asked in the first interview with PM1 were created. After the transcription, an analysis was carried out and entirely new upcoming topics or content that had not been covered yet in necessary detail in the theoretical framework at that time were identified. Subsequently the theoretical framework of the study was extended making use of secondary data and ex-isting literature. In a set of iterative processes a list of potential questions for the second interview (SM1) was created based on the amended theoretical framework. The same pro-cedure was followed for the commencing interviews with LSM, SM2, AM. Since the inter-viewed LSP (LSM) and Anonymous Pharmacy (AM) showed different company character-istics and were meant to provide insights from different angles compared to the question lists created and therefore displayed less commonalities with those for Pharmacy 1 (PM1, SM1) and Pharmacy 2 (SM2). Although question lists were created prior to the interviews, these were still semi-structured and not structured in nature as the lists served as a starting point for reorientation. Frequently questions were skipped, whereas others were added dur-ing the interviews depending on the engaging topics that arose as a result of the dialogue.
The analyses, which were carried out after each interview, were conduced according to the ‘template analysis’ procedure suggested by King (2012). The striking feature of template analysis, which was the decisive factor to adopt it for this study is its flexibility. It leaves the possibility to be adapted to the respective research’s needs and leaves freedom to amend the template throughout the procedure (King, 2012). This was especially deemed important as the emergence of topics that were thought of before throughout the course of the study were expected by the authors. Based on the theoretical framework categories were devel-oped and a code hierarchy consisting of three levels was created. Two and three code levels for more accurate classification were only applied to codes that needed to be analysed in greater depth due to their high importance to the study as they usually addressed the research questions directly, whereas codes rather providing additional background constitut-ed of one level. Accurate classification was for instance done with the level 1 code ‘home delivery’, which was narrowed down to ‘key drivers for an introduction’ on level 2, and ‘demand’ on level 3, whereas ‘pick-up service for PO medicine’ only consisted of level 1.
The template needed to be revised and restructured several times after interviews had been carried out, as new topics needed to be assigned new codes, codes were merged and old codes were deleted or reclassified. To keep track of the different codes different font col-ours were assigned to them in the transcripts. The constantly changing template served as the principal element for identifying relationships and key topics, which was used to answer the research questions.
The empirical data gathered from the interview with the Anonymous Pharmacy and it sub-sequent analysis needed to be removed completely later on, due to the company’s wishes.
The aims of the literature review were to gain insights on the subject of home delivery of PO medicine in Sweden, e-commerce and ETPs with an emphasis on the logistics perspec-tive, as well as to provide the audience of this study with a requisite understanding. Com-bined with this, Knopf (2006) suggests that a literature review aides in the validation of whether or not the research questions have already been answered, and, showing the audi-ence that the study will add new knowledge to the area.
At the point of the review there had been little published research within Europe on the introduction of PO medicine home delivery via an ETP system and with an e-commerce platform in Sweden. Although there was a considerable body of literature on electronic prescribing, few of these actually move beyond how the PO medicine reaches the patient via the pharmacy or healthcare provider.
There are multiple sources and types of literature that are available, even more so with the availability of the internet. In this case, the types used included ‘primary literature’ which The University of California Los Angeles (2014) states this is ‘first-hand’ information from sources as close as possible to the origin of the information or idea under study. The study made use of process guidance documents from the NHS in England and reports from the Swedish Government7. All the primary litera ture sources were supported by ‘tertiary litera-ture’ sources which consist of tools such as online databases and indexes that allowed the authors to locate primary literature and introduce the topic in the first place.
The authors’ intention has been to collect data that would enable them to answer the re-search questions of the study. However, Saunders et al. (Saunders et al., 2012) suggest that there are data quality issues that need to be addressed, especially because the lack of stand-ardisation associated with semi-structured interviews could lead to concerns regarding the reliability of the data collected and the risk for potential bias.
Due to the technical and sensitive nature of the study and the fact that the home delivery service of PO medicines in Sweden is only offered by a very limited number of private ac-tors and the state-owned Apoteket AB, the authors agree with Long and Johnson (2000) that this study must be open to evaluation and critique. Quite simply put; failure to do so could result in findings that cause the adoption of dangerous or harmful practices.
In order to avoid interviewee error and bias the interviewees could both select the time and date for the interviews as well as a place they felt comfortable in. Furthermore pre-information about the researchers, the topic and planned treatment of gathered data was offered prior to the interviews, as well as anonymisation and providing of a copy of the interview transcript and final thesis. Nevertheless due to the sensitive topic being covered, the danger of having been provided with a partial picture cannot be neglected, as revealing information might have been unfavourable for the interviewees. To ensure researcher error and bias were reduced a short meeting to discuss the question list and focal points was car-ried out before each interview. Both phone and face-to-face interviews were carried out in pairs, where one of the researchers acted as main enquirer, but the other came up with questions as well. Attention was paid not to ask leading questions. After the interviews they were discussed among the researchers and the transcript was cross checked by both of them.
Table of Contents
1.2 Problem definition
1.4 Research questions
2.1 Pharmacy supply chains and the market
2.3 Electronically Transmitted Prescriptions
3 Methodology / method
3.1 Research process
3.2 Research approach
3.3 Research strategy
3.4 Methodological choice
3.5 Data collection
3.6 Analysis of qualitative data
3.7 Literature review
3.10 Ethical considerations
4 Empirical findings
4.1 Company and interviewee background
4.2 Pick-up service for PO medicine
4.3 Logistics costs and profitability
4.4 Home delivery of PO medicine
4.5 Supply chain
4.6 Referencing the processing in England
5.1 Home delivery of PO medicine
5.2 Supply chain
7 Discussions / concluding reflections
7.1 Future research
GET THE COMPLETE PROJECT
Introducing the home delivery of prescription medicine in Sweden An analysis of private pharmacies and their supply chains