• No results found

Inventory Management Systems for Health Facilities in Developing Countries

N/A
N/A
Protected

Academic year: 2021

Share "Inventory Management Systems for Health Facilities in Developing Countries"

Copied!
77
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Inventory Management Systems for Health

Facilities in Developing Countries

The power of vendor managed inventory to improve medicine availability

By

(2)
(3)

3 Master thesis

Inventory Management Systems for Health Facilities in Developing Countries: The power of vendor managed inventory to improve medicine availability

Student: Rik Maarten van Hees

Student number RUG: 1883291

Student number NUBS: 130589262

University of Groningen Faculty of Economics and Business and Newcastle University Business School Double Degree MSc Operations Management

Word count: 14.708

Submission date: 17-12-2014

Address: A-weg 26 E2

Zip code: 9718 CW, Groningen

E-mail address: r.m.van.hees.1@student.rug.nl or rikvhees@gmail.com Telephone number: +31 (0) 6 45 97 96 74

Supervisor RUG: Ms. Nonhlanhla Dube MSc - n.dube@rug.nl

Supervisor NUBS: Dr. Ying Yang - ying.yang2@ncl.ac.uk

University of Groningen, Faculty of Economics and Business Nettelbosje 2, 9747 AE Groningen

Tel: +31(0)50 3633741

Newcastle University Business School

5 Barrack Road, Newcastle upon Tyne, NE1 4SE Tel: +44(0)191 2081500

(4)

4

Contents

Contents ... 4

Acknowledgements ... 7

List of tables and figures ... 8

Abbreviations ... 9

Abstract ... 10

1. Introduction ... 11

Terminology ... 14

2. Literature review ... 15

2.1 Inventory management systems and functions ... 15

2.2 Medicine characteristics ... 16

2.3 Context developing countries ... 17

2.4 Theoretical framework ... 18

3. Methodology ... 21

3.1 Research design and research setting ... 21

3.2 Sampling ... 21

3.3 Data collection ... 24

3.4 Data analysis ... 24

4. Results ... 27

4.1 Inventory management systems and functions ... 27

4.1.1 Inventory management systems ... 27

4.1.2 Inventory management functions ... 30

4.2 Performance of inventory management systems ... 35

4.3 Influential aspects of inventory management systems ... 36

4.3.1 Manual system ... 37 4.3.2 Legacy system ... 37 4.3.3 Vendor system ... 38 4.3.4 Recurring aspects ... 40 5. Discussion ... 45 5.1 IMS characteristics ... 45 5.1.1 Funding ... 47 5.1.2 Outsource IM responsibility ... 47

5.1.3 Supply chain issues ... 48

(5)

5

5.1.5 Forecasting ... 49

5.1.6 Ordering & Receiving ... 49

5.1.7 Exchanging ... 50

5.2 Research and managerial implications ... 52

5.2.1 Implications for research ... 52

5.2.2 Implications for practice ... 53

5.3 Limitations and future research ... 55

6. Conclusion ... 57

Reference list ... 61

Appendix A: Interview protocol ... 68

Appendix B: Manual system ... 70

Appendix C: Legacy system ... 71

Appendix D: Vendor system ... 73

(6)
(7)

7

Acknowledgements

I would like to thank Ms. Dube from the University of Groningen and Ms. Ying Yang from the University of Newcastle for their contribution to my thesis. Their refreshing insights were very helpful in structuring and writing my thesis. Most importantly, the continuous critical perspective of Ms. Dube helped me to further improve my thesis.

I also would really like to thank Chris Weeks and Vishal Punamiya for the opportunities that they offered me to set up this thesis in collaboration with their project. The support they provided during the research really helped me to effectively collect the essential data. Especially I would like to thank them and their partners for providing me the opportunity to observe inventory management practices in real life in Gambia.

(8)

8

List of tables and figures

Table 1.1: Terminology P.14

Table 3.1: Overview of respondents, IMSs, operators,

countries, locations and size of HFs P. 22

Table 3.2: Sampled cases P. 23

Table 4.1: Components of IMSs P. 27

Table 4.2: IMSs' performance P. 35

Table 5.1: Required IMS characteristics P. 46

Figure 2.1: Theoretical framework P. 19

(9)

9

Abbreviations

Health Facility HF

Inventory Management IM

Inventory Management System IMS

Non-Governmental Organisation NGO

Stock-Keeping-Unit SKU

Supply Chain SC

Vendor Managed Inventory VMI

(10)

10

Abstract

(11)

11

1. Introduction

Medicine stock-outs in Health Facilities (HFs) in low- and lower-middle-income countries are still a major problem nowadays, highlighted by the average availability of essential medicines of 57% in public HFs and 65% in private HFs (United Nations, 2013). For example, in 2014 41% of the patients in Tanzania could not obtain the required medicines at public HFs due to stock-outs (Wales et al., 2014). Stock-outs are defined as: when a health facility temporarily

does not have supplies of medicine (Wales et al, 2014: 8). Simultaneously, medicine

availability also deteriorates when existing stock expires (Nakyanzi et al., 2010). Stock-outs and expiry are the consequence of an endless number of interrelated causes that are not easily resolved (Wales et al., 2014). However, in the past emphasis has been on exploring the causes that pertain to the Supply Chain (SC) and the role of governments (United Nations, 2013), while the role of Inventory Management (IM) at the HF level is scarcely and only indirectly addressed (Wales et al., 2014). Yao and Dresner (2008) endorse the importance of IM downstream the SC however, concerning the medical SC in developing countries this aspect seems underestimated. This research focuses on this underexposed subject and sets out to investigate the characteristics of Inventory Management Systems (IMSs) that should support the execution of IM at HFs in developing countries.

(12)

12

existing literature, while it has recently been indicated that IMSs have the potential to improve the execution of IM (Nilseng et al., 2014). In the field of humanitarian logistics, Blecken and Hellingrath (2008) assessed a number of SC management systems that contain IM modules. They discovered that these systems are developed by Non-Governmental Organisations (NGOs) in order to increase SC transparency. However, the systems they assessed did not cover HFs at the lowest level of the SC and thus these systems do not provide a solution to challenges in IM practices at the HF level. In conclusion, barely any investigation of IM and IMSs suitable for HFs in developing countries is applied (Kim, 2005), while existing studies provide evidence that improvements in IM at HFs have the potential to improve the availability of medicines (Barrington et al, 2010; Daff et al., 2014).

The above mentioned gap in the literature combined with that potential, necessitates the investigation of IM and its related IMSs through empirical research. This provides the opportunity to add to academic knowledge what the influence of IM and IMSs is on medicine availability in developing countries. In addition, this research aims to indicate what characteristics an IMS should contain to improve medicine availability in developing countries. In order to achieve the contribution for research and practice, this research sets out to investigate the characteristics of IMSs that can properly support the management of medicine stock in HFs in developing countries, resulting in the following research question with three sub research questions:

What characteristics should an Inventory Management System that manages medicine stock in Health Facilities in developing countries contain?

A. What IMSs are used and how do they support IM?

B. How are these systems performing in terms of medicine stock-outs and expiry? C. What aspects of the IMSs influence that performance?

(13)

13

(14)

14

Terminology

Term Definition

Inventory Management (IM)

IM is defined as the execution of a set of IM functions, which are: monitoring inventory levels and consumption data, forecasting future demand, ordering, receiving supplies, storing and dispensing

medicines.

Inventory Management System (IMS)

A tool that supports the execution of IM functions.

Health Facility (HF) HF is defined as a facility where medicines are provided to patients, ranging from large city hospitals to a very small health posts in a rural environment. HFs are the last link in the medical SC and is thus located at the bottom of the SC.

Developing countries The IMF identifies 153 developing countries according to their criteria: the composition of export earnings and other income from abroad; a distinction between net creditor and net debtor economies; and, for the net debtors, financial criteria based on external

financing sources and experience with external debt servicing (International Monetary Fund, 2014: 158). HFs in 19 of these 153 countries are involved in this research.

Performance of Inventory Management Systems

IMS performance is measured in terms of stock-outs and expiry. Stock-outs mean that the HF temporarily does not have (certain) medicines, while expiry means that (some) existing medicines need to be disposed because of their expiry dates. A high rate of stock-outs and/or expiry thus indicates a poor performance.

(15)

15

2. Literature review

In this section, the theoretical framework in which this research takes place is developed in three steps. Firstly, the basics of IM are discussed including an overview of the different IM functions. Secondly, the specific influence of medicine characteristics is presented. Thirdly, the consequences posed by the context of developing countries are presented. By combining these steps, the theoretical framework is presented in Figure 2.1 and explained in paragraph 2.4.

2.1 Inventory management systems and functions

In the last 20 to 30 years the growth of literature in IM has been tremendous and especially focused on the development of specialised models (Prasad, 1994). Since this research aims to focus on the very elementary application of IM in HFs in developing countries especially, it is first defined what IM encompasses in that environment. One of the two most fundamental functions of inventory control is continuously or periodically monitoring inventory levels and consumption data. The second function involves forecasting future demand and evaluating forecasting errors (Axsäter, 2007). In addition, IM includes functions referred to as warehouse functions, which comprehends: ordering, receiving, storing and dispensing (Gu, Goetschalckx and McGinnis, 2007). Ordering involves communication with suppliers in order to procure new medicines. Receiving represents the link upstream when medicines are delivered to a HF. Storing encompasses the organisation of incoming supplies. Finally, dispensing means the provision of medicines to patients (Axsäter, 2007).

The objective of IM on the one hand is to minimise stock levels to free up money that otherwise would be tied up in inventory. On the other hand the objective is to remain ample inventory to serve customers as good as possible (Axsäter, 2007). Since the primary objective of this research is to increase medicine availability, it is required to remain ample stock. Simultaneously, stock levels should not be too high since overstocking could lead to the expiry of medicines (Prasad, 1994).

(16)

16

IMSs can also support multiple actors (e.g. several hospitals and one supplier), through a Vendor-Managed-Inventory (VMI) approach. When applying VMI, the execution of IM functions is transferred from the HF to the supplier (Management Science for Health, 2012). Also, medicines are now pushed to HFs since the supplier has the responsibility of ordering and resupplying (Watson, Serumaga and McCord, 2012). In order to enable VMI, Vendor Managed Inventory Systems (VMISs) are required to provide the necessary information transparency between actors (Yao and Dresner, 2008). Several researchers, often in other industries, have indicated the compelling advantages that could be gained by introducing VMIS, such as a reduction of inventory costs and improved customer service (Achabal et al., 2000;Waller, Johnson and Davis, 1999). After their case study on IM in Malaysian hospitals, Mustaffa and Potter (2009) suggested the implementation of VMI since it could achieve large efficiency gains.

The existing literature lacks to provide a detailed description of the components that an IMS should contain. Smith and Offodile (2002) explain the usefulness of using barcode scanners, barcode printers, (mobile) computers and software in developed countries. Berger et al. (2007) agree that such technology could have a contribution in the context of developing countries but also state that components such as stock-cards should be maintained.

In conclusion, IM in HFs in developing countries comprehends monitoring, forecasting, ordering, receiving, storing and dispensing. Although it is unknown from the existing literature what an IMS specifically entails, they support either single facilities or multiple facilities (VMI). Finally, IMSs could contain components such as hardware, software and stock-cards (Smith and Offodile, 2002; Berger et al., 2007).

2.2 Medicine characteristics

According to Rushton (2010) medicines have unique characteristics that have implications for ordering, receiving, storing and dispensing. For example, some SKUs are very bulky and require much storing space, while others require storing at certain humidity levels (Hayford, Privor-Dumn and Levine, 2011) or temperatures (Guire, 2011). These aspects need to be taken into account because:

(17)

17

Despite the knowledge about unique medicine characteristics and its implications, the development of systems such as wireless temperature and humidity monitoring (Yamazoe and Shimizu, 1986) falls short in HFs in developing countries. Hayford et al. (2011) argue that proper monitoring is essential for successful preservation and that only nurses, doctors or pharmacists possess the right education to do this. Finally, perishability of medicines has implications for forecasting (Nakyanzi et al., 2010).

In conclusion, characteristics of medicines seem to have implications for the execution of all six IM functions and thus are taken into account in the theoretical framework.

2.3 Context developing countries

A considerable amount of research focused on identifying the complex context of humanitarian relief operations (Thomas and Kopczak, 2005; Van Wassenhove, 2006) and most of these operations take place in developing countries (Alcántara-Ayala, 2002). Many contextual factors reviewed here are derived from that specific field (humanitarian logistics). Environmental factors are widely examined by Van Wassenhove (2006) and although some factors are particularly applicable to disaster relief environments, most of them also apply to the context of developing countries. Firstly, a lack of adequate infrastructure and hindered transportation options could lead to a delay in receiving supplies. Secondly, robust equipment is often missing (e.g. adequate systems that could work off- and online since internet it not always available). Thirdly, there is often a lack of qualified personnel due to high staff turnover. Guire (2011) adds safety and security issues that arise due to political instability and shortages of medicines that encourage corruption and theft. Kunz and Reiner (2012) also mention government situational factors such as corruption and national regulations. Secondly, they mention socio-economic factors such as the presence of local suppliers and the level of education. Van Wassenhove (2006) and Kunz and Reiner (2012) both mention that technology is often inadequately used due to the absence of logistic experts, while it also occurs that adequate systems are missing even when logistics experts are present (Thomas and Kopczak, 2005).

(18)

18

2.4 Theoretical framework

Jahre et al. (2012) stress the relationship between the complex medical SC and the context of developing countries. Complexities of that SC are long lead times, low order frequencies and unknown demand patterns. The outcomes of their research suggest that those medical SCs should be better adapted to the context. Mustaffa and Potter (2009) acknowledge the need for adaptability, especially concerning IM of HFs. One of the things they discovered is that stock-outs could be prevented by applying an IM approach that increases transparency in the SC. The need to take medicine characteristics and contextual factors into account is also confirmed by Bateman (2013). He found in the Oliver Tambo region in South Africa that stock-outs were caused by corruption and theft, while medicines expired through: a chronic

cycle of over-ordering by health facilities as the result of poor stock-keeping (Bateman, 2013:

601).

(19)

19

(20)
(21)

21

3. Methodology

The methodology section first explains the research design and setting and continues with sampling and case selection. Thereafter, the data collection is explained after which data analysis is discussed.

3.1 Research design and research setting

A multiple case study is chosen because of four reasons. Firstly, this research explores IMS characteristics in real-life contemporary situations where the researcher has no control over the complex context (Yin, 2009). According to Eisenhardt (1989) case study research can be used to pursue different research purposes, which are formulated by Voss, Tsikriktsis and Frohlich (2002) as exploration, theory building, theory testing and theory extension. Since IM in HFs in developing countries has hardly been investigated, the purpose of this research is exploration (Voss et al., 2002). Secondly, a case study is very powerful in identifying and proving causal relationships (Stuart et al., 2002). In this multiple case study the potential relationships explained in the theoretical framework are studied. Thirdly, a multiple case study is recommended because it further stimulates the exploration of new variables (Meredith and Samson, 2002). Fourthly, Eisenhardt and Graebner (2007) argue that research becomes more grounded, accurate and generalizable when multiple cases are involved.

In terms of research setting, the research focusses on IMSs that are used to support the execution of IM in developing countries’ HFs. Within this setting, there is substantial diversity between HFs. Firstly, some HFs are from NGOs, while other are from private organisations, governments or dioceses. Secondly, HFs are located in 19 different developing countries. Thirdly, HFs are located in either a rural or an urban environment. Finally, the size of HFs varies from small health posts till big city hospitals. Table 3.1 provides an overview of the respondents, IMSs, operators, countries, locations and the size of each HF. Incorporation of so many different HFs increases the generalizability of findings.

3.2 Sampling

The unit of analysis is determined as IMSs. Within this research an IMS is defined as: a tool

that supports the execution of IM. Case selection is based on two steps (Miles and Huberman,

(22)

22

Respondent

number IMS Operator Country Location Size ^

1 Manual system Public Tanzania Urban 0/300

2 Manual system Public Tanzania Urban 0/300

3 Manual system NGO Democratic Republic of Congo

Rural */*

4 Manual system NGO Democratic Republic of Congo

Rural */*

5 Manual system Public Tanzania Rural 50/150

6 Manual system Public Tanzania Rural 50/150

7 Manual system Public Tanzania Rural 50/150

8 Manual system Diocese Ethiopia Rural 0/*

9 Manual system Diocese Uganda Rural 400/ *

10 Manual system NGO Benin Rural 6/*

11 Manual system Public Tanzania Rural 120/80

12 Manual system Public Tanzania Rural 180/*

13 Manual system Public Tanzania Rural 180/*

14 Manual system Public Tanzania Rural 120/50

15 Manual system Public South Sudan Rural 20/70

16 Manual system Diocese Ghana Rural 80/150

17 Manual system Public Zambia Urban 0/60

18 Manual system Public South Africa Urban & Rural 280/150

19 Manual system NGO Somalia Urban 134/150

20 Manual system NGO South Africa Urban 0/25

21 Manual system NGO South Africa Urban 0/25

22 Legacy system NGO Ivory Coast Urban 80/250

23 Legacy system NGO Pakistan Rural 100/225

24 Legacy system NGO Papua-New-Guinea Urban 80/60

25 Legacy system NGO Myanmar Urban & Rural 120/*

26 Legacy system Private Congo Brazzaville Rural 40/60

27 Vendor system NGO Senegal Urban & Rural #

28 Vendor system NGO Senegal Urban & Rural #

29 Vendor system NGO Mozambique Urban & Rural #

30 Vendor system NGO Mozambique Urban & Rural #

31 Vendor system NGO Nigeria Rural #

32 Vendor system NGO Kenya Urban & Rural #

33 Vendor system NGO Kenya Urban & Rural #

Table 3.1: Overview of respondents, IMSs, operators, countries, locations and size of HFs

^ = in X/Y, which means that X is the number of beds and Y the number of Out-Patients-Department visitors per day * = Respondent was not able to provide measure

(23)

23

In addition to convenience sampling, respondents are approached based on two criteria (Eisenhardt, 1989; Yin, 2009). Firstly, respondents from different types of operators are interviewed (NGOs, private organisations, governments and dioceses) since operators could have a large influence on the type of IMS that is used. For example, nowadays NGOs implement new IM approaches (Daff et al., 2014). Secondly, respondents who worked in rural or urban HFs are approached because the availability of resources can significantly differ between the two settings (Strasser, 2003) (see both criteria in Table 3.1). Strasser (2003) states that resources are often concentrated in urban areas, which might influence the type of IMSs used. With those two criteria in mind, 33 semi-structured interviews are conducted with experts that told about the IMS they used in practice. The interviews show that three main types of IMSs are used in practice and that based on the criteria eight cases arise (displayed in Table 3.2). The manual system consists of stock-cards combined with a stock ledger (a book). The legacy system consists of stock-cards combined with Excel files. The vendor system consists of stock-cards combined with dedicated, customised software. More details about the IMSs are provided in 4.1 Results.

Location Rural Urban Operator NGO/ Private Manual system (3) Legacy system (3) Vendor system (7) Manual system (3) Legacy system (3) Vendor system (6) Public/

Dioceses Manual system (12) Manual system (4)

Table 3.2: Sampled cases( (x) represents the number of respondents)

(24)

24

3.3 Data collection

Data is collected through the application of three different techniques. Firstly, 33 semi-structured interviews are conducted according to a predetermined interview protocol which includes questions concerning the variables that were determined in the theoretical framework (see appendix A). Semi-structured interviews are used because it enables the discovery of new variables and relationships between variables (Meredith and Samson, 2002). In addition, some freedom during the interviews was required because of the different professional backgrounds of respondents (Barriball and While, 1994). The respondents are logisticians working for different NGOs, HF pharmacists, healthcare providers (medical directors, doctors, clinical officers, nurses, medical- and pharmaceutical students) and finally project managers working for NGOs that were implementing new IM practices. Respondents have been (or are still) working at HFs in developing countries. Using informants with diverse experience and knowledge of IM strengthens the construct validity by not being biased by one specific respondent group (Patton, 2005). Interview audio recordings are transferred into transcripts, which are stored in an interview database. Secondly, triangulation of evidence is achieved through observations in Gambia in two public urban hospitals, one public rural health clinic and one NGO urban health clinic. During visits the IMSs and the execution of IM functions were inspected to validate findings from the semi-structured interviews. Thirdly, in order to compare the findings, data is collected from archival sources such as NGO reports and literature in the relevant field. Through the use of data from multiple sources, triangulation is achieved. This strengthens the reliability of the findings (Voss et al., 2002) and increases the construct validity of this research (Yin, 2009). The use of retrospective as well as real-time cases through interviews, observations in Gambia and archival sources increases the internal validity through mitigation of bias that could arise when using only one of the two (Leonard-Barton, 1990).

3.4 Data analysis

(25)

25

theoretical replication are investigated by using pattern matching (Yin, 2009), which helps to find similarities and differences. Based on the cross case analysis, unique characteristics – as well as recurrent characteristics of IMSs can be found. Finally, it is analysed which characteristics influence (both positively and negatively) the performance in terms of stock-outs and expiry.

(26)
(27)

27

4. Results

Paragraph 4.1.1 presents the components of the three different IMS after which paragraph 4.1.2 shows how the IMSs are used to execute IM functions. Paragraph 4.2 presents the performance per IMS in terms of stock-outs and expiry. Finally, paragraph 4.3 addresses the causes for performance.

4.1 Inventory management systems and functions

Note that sampling in this research resulted in eight cases in which three different IMSs are represented. Some IMSs are missing in certain environments (legacy and vendor system in public facilities), though others are present in all circumstances (manual system). Interestingly, the components and usage of each of the three systems did not significantly differ between a rural and urban environment, neither did it differ between NGO or public operators. For example, a manual system used in an urban NGO facility is not really different from a manual system used in a rural public facility. On the other hand, very interesting differences were found between the three IMSs. Hence, the results section emphasises these differences and systematically covers the manual, legacy and vendor system.

4.1.1 Inventory management systems

In order to answer sub research question A, first the components of each IMS are presented in Table 4.1 and explained below.

What is within the IMS? Manual system Legacy system Vendor system

Stock-cards X X X

Stock ledger X

Excel files X

Customised software X

Table 4.1: Components of IMSs

(28)

28

stock. Finally, paper based sheets or phone calls are used to order new medicines from suppliers.

Again, the components of the manual system discussed are representative for all four cases (see Table 3.1). In other words, manual systems used in either a rural or an urban environment, with either a public or a NGO operator, are very similar.

A legacy system is a more developed IMS in which the stock ledger is replaced by standardised Excel files (see appendix C). Stock-cards are retained in order to generate the overviews in Excel. Inventory status and consumption data in Excel files are easily shared between HFs and logistic coordinators through internet or by retrieving it physically. Logistic coordinators are trained logisticians that steer IM of multiple HFs from a higher level in the SC (a central office). For example, logistic coordinators aggregate orders from multiple HFs.

Excel sheets are similar to stock ledgers but contain more data points. For example: classification of medicines, expiry dates of every batch of the same SKU and manual stock counts. Moreover, discrepancies and remarks can be listed (see appendix C). In addition, Excel enables transforming monthly overviews into yearly overviews very easily. This might be helpful in derive consumption trends, for example.

Some respondents mentioned a traffic light system implemented in Excel that supports HF personnel and/or logistic coordinators in determining reorder points. As one respondent stated it also helps to evaluate forecasts:

“To reduce the amount to which it is messed up we have an early warning system that monitors closely and helps to understand the reason behind low consumption or high consumption. (…) This tool helps medics by telling them what was forecasted compared to the actual consumption. It gives a red sign if they are consuming more compared to their forecasted consumption. They have to smoothen that because (even while they are not running out now) otherwise they will run out soon.” (R#25, legacy system)

Most respondents mentioned that orders are placed by using the same Excel files. If HFs did not possess computers, paper sheets were used and later the information was put in Excel by a logistic coordinator:

(29)

29

that manual information from paper into a computer and then send it further by email with an Excel file. This is more or less where technology ends at the HF level.” (R#23, legacy system)

Again, the components of the legacy system discussed here hold for both cases that were represented earlier (see Table 3.1). In other words, the same legacy system is used both in rural and urban environments.

A vendor system is unique compared to the former two systems since most IM responsibilities are outsourced to the supplier. This implies that the IMS is also kept by the supplier. VMISs contain customised software installed on tablets, laptops or desktop computers. The customised software could have different names but the options are very similar. Therefore all these customised software packages are referred to as a vendor system or VMIS. Some respondents explained a vendor system in which they tried to remove the stock-cards:

“With the vendor system we are trying to remove the paper trail by a tablet or a laptop that could go with the field coordinator to the healthcare facilities.” (R#30, vendor system)

All other respondents explained vendor systems in which stock-cards were retained at HFs:

“The data collection is one of the big advantages of the VMIS. (…) When a field coordinator comes back to the provincial capital after a distribution loop, stock-card information is imported into a computer.” (R#29, vendor system)

VMISs that run on a tablet or laptop go along with the vendor during a distribution loop in order to simplify data collection. A distribution loop is predetermined route for the supplier in which a number of HFs is supplied with medicines in a number of days. According to the respondents, on average 20 to 40 HFs are supplied in a distribution loop that takes about two weeks.

“With the vendor system and the transport loops it essentially follows a logical loop for the transport. It takes the medicines from the province level (…) just logically from A to B to C to D, directly to the health centres in a logical fashion, where the roads go essentially.”(R#29, vendor system)

(30)

30

“You can drill down into the data by clicking on a particular line to see what is behind it. For example, you can see that regions near the capital region perform way better than regions further away. Within the software you can drill down per province, per delivery zone, per district and even towards each specific healthcare facility. Whatever you choose to drill down, the system will always show how many healthcare facilities were visited, what the delivery intervals were, what the reasons for not visiting were and what the percentages of stock-outs per SKU were.” (R#30, vendor system).

Information that is collected with the VMIS is automatically synchronised with a server when the vendor temporarily has internet connection during a distribution loop. In that way, aggregated consumption data from multiple HFs is easily shared to parties such as the government. Also, aggregated data is used to analyse consumption data and generate forecasts based on historical consumption data.

One of the respondents provided a very comprehensive tour through the VMIS. Appendix D provides some screenshots of that particular VMIS.

Again, the components of the vendor system discussed here hold for both cases that are represented earlier (see Table 3.1). There are no differences between vendor systems used in rural or urban environments because during a distribution loop HFs in both rural and urban environments are supplied.

4.1.2 Inventory management functions

Now that the components of the three IMSs are known, this paragraph presents how the IMSs support the execution of IM functions. In the theoretical framework the IM functions

monitoring, forecasting, ordering, receiving, storing and dispensing were identified from the

existing literature. During the semi-structured interviews and the observations in Gambia, a new function was identified: the exchange of inventory between HFs, referred to as

exchanging in this research.

(31)

31

department). Stock-cards can be updated directly after dispensing, although during the observations it became clear that this is not applied in practise. As respondent 19 explains, it is more realistic to update stock-cards once per day:

“At the pharmacy they record once per day what is dispensed to the outpatients. Then we have the inpatients at the different departments, they order two-weekly at the pharmacy and are then supplied by the pharmacy.” (R#19, manual system).

An inventory overview in a stock ledger or Excel sheet is generated monthly or weekly, while the VMIS automatically creates overviews after each distribution loop. Overviews are used to prepare forecasts, determine orders and exchange inventory. Observations in Gambia, as well as interview respondents, showed that analysis of inventory data also belongs to monitoring of the data. Excel provides a traffic light system to know when reorder levels are reached and also has options to analyse consumption data of multiple years. VMISs contain additional analysis tools that could provide more details, for example: searching for causes of stock-outs for a particular region (put in the VMIS by the vendor) and displaying service levels per region (or per HF).

Solely in the legacy system monitoring implies giving feedback. For example, logistic coordinators could provide feedback about consumption of certain medicines to HF personnel. Monitoring also comprehends regular (every 1-3 months) stock counts to determine discrepancies between the information in the IMS and the actual stock levels. Stock counts are often made just before new orders are placed, regardless of the system used. Finally, stock-level updates are generated when manual systems are in place. These updates can inform HF personnel about stock levels in the warehouse and/or pharmacy.

Forecasting is done in two ways, depending on the IMS. The first method is based on historical consumption data (subtracted from stock-ledgers, Excel files or the VMIS), provided that the HF already has consumed medicines for a few months at least:

(32)

32

Secondly, forecasts can be made based on a formula that is included in VMISs. Of course the figures presented here differ per location, however the method is the same. Respondents called this the morbidity method:

(catchment area ∗ target group ∗ doses per patient ∗ wastage factor) ∗ required buffer percentage

1. The catchment area = the size of the population that the HF serves 2. The target group = % of the population that requires a specific medicine 3. Wastage factor = % items for that SKU that are likely to expire

4. Required buffer level = indication of the importance of that SKU

“Each health facility knows his target population, for example that more or less 60 children should get the BCG vaccination per month. That is multiplied with the buffer level and wastage rate. This is automatically calculated by the VMIS for each health facility after which it gives a total for the vendor when he starts his distribution loop” (R#29, vendor system)

Accurate forecasting requires incorporation of consumption trends. In order to take trends into account it is required that previous consumption data is carefully retained. Legacy and vendor systems facilitate online data storage, while manual systems require the stock-cards and stock ledgers to be physically retained.

Ordering is not supported by the manual system. According to nearly all respondents and all observations in Gambia, using the manual system implied that ordering is done by physically visiting the supplier. If not, paper form orders were sent by post or retrieved by the supplier. With the legacy system orders are put in Excel and sent by e-mail. If the internet connection is not available at the HF, Excel files are physically retrieved by the logistic coordinator. Subsequently, he aggregates orders from multiple HFs before sending it to the supplier. Vendor systems do not involve ordering since the responsibility to decide if new supplies are required is outsourced to the supplier.

(33)

33

simultaneously during a distribution loop. During a visit the vendor fills in various data points in the VMIS, for example:

“The software will always show how many healthcare facilities are visited and what the delivery intervals were.” (R#30, vendor system).

Storing includes that some medicines are stored between two and eight- and some below zero degrees Celsius and that a certain humidity level is assured. During the observations it became clear that these are not tracked, partly because a system for doing this is missing. In contrast, a vendor system enables a vendor to list current storing conditions when he visits a HF.

In order to simplify storing, a stock-card could be connected to the shelve where the particular medicine is stored, resulting in less searching for stock-cards when new supplies come in or when medicines are dispensed. However, this means that an important part of the IMS is scattered through the pharmacy or warehouse.

Dispensing is also not represented in the manual and legacy system, neither in the vendor system. Outgoing medicines are of course listed in all three systems, but that is actually the execution of monitoring. According to one of the respondents, knowing what is dispensed goes as follows:

“Well, in terms of keeping track of what has been dispensed, right now we do that based on the prescriptions.” (R#19, manual system)

Hence, medicine prescriptions are required to update stock-cards or VMISs that do not use stock-cards. Secondly, prescriptions could be used to check if any discrepancies exist between the IMS information and actual stock levels.

When using VMISs without stock-cards, checking for the amount of medicines dispensed could also be applied as follows:

“The operator is managing the data collection (basically by saying I delivered 100 to you last month, today you only have 50 so I know your consumption was 50), (…) regardless if there was anything stolen.” (R#28, vendor system)

(34)

34

“With technology it is easier to make the linkages between locations. For example, if we retrieve X number of goods at location A we can give location B X number of goods extra. The exchange of data is not possible unless you compare two Excel sheets. (…) With advanced technology it is easier to see what is where available.” (R#22, legacy system)

IMSs enable the exchange of overstock from location A to the understocked location B, for example, by providing the inventory statuses of multiple facilities in one overview. The manual system does not provide this information. The legacy system does provide logistic coordinators with the possibility to compare Excel sheets of multiple HFs. Vendor systems simplify the exchange of stock because of two reasons. Firstly, the VMIS provides an overview of the inventory status of all medicines from multiple HFs within a few clicks. Secondly, vendors can decide to exchange stock between HFs because they already visit the facilities during a distribution loop when they supply them with new medicines:

(35)

35

4.2 Performance of inventory management systems

Now it is established what IMSs are used and how they support IM functions, this paragraph presents how the different systems perform. In order to explain this, sub research question B is answered. All the interviewees were asked how the IMS was performing in terms of stock-outs and expiry, but as mentioned in the methodology, this was based on a scale: ‘None – Barely – Sometimes – Often’. To compare the performance of the three systems, the scales were given numbers from 1 to 4 (None = 1, Often = 4), after which averages per system were derived. The results per system are presented in Table 4.2. The differences between the averages provide an indication of how a system is performing compared to the others. Thus it is not about the absolute value of the numbers presented in Table 4.2 but about the performance relative to each other.

According to the averages in Table 4.2, the manual system performs worst in terms of stock-outs, the legacy system performs a little better and the vendor system performs best. In terms of expiry the legacy system performs worst, while the manual and vendor system have a similar performance.

Manual system Legacy system Vendor system

Stock-outs 3.1 2.6 1.4

Expiry 1.8 2.6 1.7

Table 4.2: IMSs' performance

These results indicate that the capability of the manual system to prevent for stock-outs is relatively moderate, while the capability of the legacy system to overcome stock-outs is not really exceeding the manual system’s capability. On the other hand, the results suggest that the vendor system is much better in preventing for stock-outs compared to the other two.

(36)

36

4.3 Influential aspects of inventory management systems

Now it is established what IMSs are used, how they support IM functions and what their performance is relative to each other, this paragraph presents what aspects related to each IMS influences (positively and negatively) performance. In order to answer sub research question C, paragraph 4.3.1 till 4.3.3 presents unique aspects of the systems, after which recurring aspects are presented in paragraph 4.3.4. Figure 4.1 provides a visual overview of all the aspects that will be explained below.

(37)

37

4.3.1 Manual system

Monitoring with the manual system provides inadequate inventory status overview, partly as a result of the large amount of stock-card data that need to be transferred into a stock ledger weekly or monthly. This is a time consuming task in which human mistakes are easily made:

“Nowadays many things are overwritten several times and that causes mistakes.” (R#12, manual system)

The results also indicate that a manual system implies that HF personnel are responsible for using the IMS. Many respondents mentioned that HF personnel are already overburdened with medical tasks, leading to inaccurate monitoring of dispensed medicines. Even when it was applied accurately, the observations in Gambia showed that stock ledgers lack to provide sufficient options to list multiple expiry dates. The manual system also revealed a positive aspect, namely weekly stock level updates. Those updates try to steer doctors’ prescribing behaviour in a way that stimulates them to prefer a certain substitute for a while when that medicine was about to expire:

“What helped was the weekly update that already told us in the beginning of the week what was available for the coming week.” (R#14, manual system)

Forecasting is nearly impossible because trends are hard to derive from stock-cards and stock ledgers. Complicating the situation, pharmacists and storekeepers in HFs in Gambia declared that stock-cards and stock ledgers easily disappear, resulting in lost historical consumption data.

Ordering often delays the process of being supplied with new medicines which increases the chance of medicine stock-outs. Many respondents complained about difficulties in reaching out to suppliers (by e-mail, telephone, post or courier service), forcing HF personnel to visit their suppliers to procure new medicines:

“They go to the head office, get their funds, go to the supply agency in the capital, pick the drugs they need, put it in their car and go back to the facility.” (R#10, manual system)

4.3.2 Legacy system

(38)

38

“Per item it is indicated whether it is critical or a low consumable item. (…) Low consumables need at least one or two months of stock, [while] the critical items have a higher buffer. This is all implemented within the system.” (R#24, legacy system)

These buffer levels are linked with the traffic light system and so critical medicines have a higher reorder level compared to low consumables. Secondly, a unique aspect of the legacy system is the feedback. Through sharing of Excel files, a logistic coordinator could provide feedback about consumption data. With such feedback (similar as with the stock-level updates in the manual system), doctors could be stimulated to prefer certain medicines for a while to overcome stock-outs and expiry of medicines.

Forecasting also means the evaluation of previous forecasts. Besides the feedback mentioned above, logistic coordinators also provide feedback about the comparison between the doctor’s forecast and the actual consumption. This improves the capability of HF personnel to generate forecasts in the future.

Ordering by sending an order via an e-mail (if internet connection is available) to the logistic coordinator simplifies and speeds up the ordering process, which could overcome stock-outs in case of an emergency order. Also, logistic coordinators could provide feedback on orders generated by HF personnel to remove forecasting mistakes being made before the order is send to the supplier.

Exchanging inventory with support of the legacy system goes as follows. Each HF shares his inventory overview in Excel with the logistic coordinator on a monthly basis. The logistic coordinator analyses differences between HFs and consequently decides if an exchange of stock between locations is required. If it is beneficial to exchange part of the inventory, the logistic coordinator should arrange transport to make this happen.

4.3.3 Vendor system

(39)

39

“They are supposed to track inventory on paper stock-cards, (…) but they did not receive any training to do that. You can easily find someone that does not know how to count stock or how to find the expiry date on a product. They really do not have the capability to track inventory. So one of the things the vendor system actually does is that it takes that responsibility out of their hands and shifts it to the third-party logistics provider” (R#28, vendor system)

The downside of this is that inventory data is only gathered once a month since the vendor only collects data when he visits a HF during a distribution loop.

Forecasting, according to the morbidity method, is integrated in the vendor system and enables to forecasts for multiple HFs together. In addition, with a VMIS forecasts can be made based on historical consumption data from multiple HFs. The combination of two methods minimises forecast errors:

“On an individual level it is hard to predict, however at higher levels you can predict things based on statistical data. For example, in the Democratic Republic of Congo you will need 10,000 cases to treat fever per year. That is the amount of medicine you need in the general warehouse. Hence, per country it is possible to make some forecast.” (R#3, manual system)

Based on the morbidity method a vendor can decide upon the amount to load for one distribution loop, while the forecast based on consumption data can help the vendor to decide how much to deliver to each HF. The latter can be adjusted on an ad hoc basis when the vendor visits a HF during a distribution loop.

Storing is supported in a VMIS by a temperature monitoring module that measures if storing conditions are properly maintained. If temperatures are not maintained properly over a longer period of time (a number of times the temperature was measured too high or too cold), the VMIS will display this and suggest to take action (e.g. send a technician) and so possibly avoid medicine wastage. The VMIS makes HF personnel aware of the fact that it should be monitored and motivates them to regularly and accurately track these conditions. Simultaneously, one respondent suggested further development of temperature monitoring within the VIMS:

(40)

40

Exchanging is simplified with the help of a VMIS and a distribution loop that facilitates the exchange of stock between facilities. The VMIS provides an overview similar to the traffic light system in the legacy system. Based on that overview a vendor decides to displace stock from one facility to another during a distribution route:

“If any commodities are about to expire within the coming six months, we take those commodities to facilities that have higher consumption rates so that they can be utilised. In that way we are able to reduce wastage.” (R#31, vendor system)

4.3.4 Recurring aspects

Monitoring is complicated when multiple stock-cards are used for one particular medicine. For example, pharmacists in the HFs in Gambia complained that HIV medicines could be supplied by the government and by a commercial supplier. Separate stock-cards needed to be maintained because both suppliers want to report consumption rates of the HIV medicines they supplied. Using separate stock-cards is a source for errors because historical consumption data becomes fragmented:

“You can make an entirely accurate usage calculation, but if the HF you are doing this for is receiving medicines from other sources at the same time, then the whole system is odd” (R#3, manual system)

Multiple stock-cards could also complicate expiry management. However, the legacy and vendor system provide an option to list expiry dates for each per batch of the same SKU. This minimises the chance of medicines expiring without notice. Even without multiple stock-cards for one SKU, human mistakes in monitoring activities seem to be insurmountable regardless of the system. However, the next paragraph explains the applications of some measures to increase the reliability of the data.

(41)

41

Forecasting, predicting annually recurring trends such as the increase of demand during a malaria season, can be done based on historical consumption data that should be carefully retained. The vendor system assures data storage through automatic synchronisation with an online server. In contrast, manual and legacy system users complained about a lack of historical consumption data through inadequate storing:

“Currently, forecasting is done by taking the average of last year, if you are lucky… (...) If you have a system that can tell you about the previous years, you can get a much more concrete pattern.” (R#25, legacy system)

Ordering involves physical stock-counts that help to improve the accuracy of forecasts by showing if any discrepancies exist between actual stock and the stock level in the IMS. All three IMSs involve options to list the discrepancies on either a stock-card, in an Excel file or in the VMIS.

In the legacy system ordering is supported by visualising when a reorder point is reached by means of the traffic light system. This method prevents stock-outs. Even though the ordering responsibility is shifted to the supplier in the vendor system, it still incorporates reorder levels. The VMIS automatically indicates to the supplier if resupply is needed:

“They go to each facility on their particular route and then see at each facility what their current stock level is. They put that information in the laptop, which then automatically calculates the portion that is needed to resupply.” (R#31, vendor system)

Receiving is not supported in the manual, legacy and vendor system because they do not provide any information about the time of delivery. The obscureness about the delivery of new supplies increases lead time uncertainty and thus influences performance. For example, a pharmacist in a HF in Gambia explained the lead time uncertainty by explaining that the time between ordering and receiving could vary from three to six months. In addition, the manual and legacy systems fall short in providing any information about tracing supplies:

“Much of the medicines ordered at the supplier are not directly available. On the receipt is written: ‘to follow’. For those medicines it is completely unknown when they will be delivered.” (R#18, manual system)

(42)

42

delivered with the next distribution loop. However, even with a fixed distribution route the time of delivery can still be a little obscure.

Besides obscureness of lead times, respondents of manual and legacy systems complained about lead times that could be nine months in some situations. Even though this aspect could hardly be influenced by an IMS, it is interesting to know that these long lead times result in long order cycles. Respondent 22 explains why orders are only placed three or four times a year:

“More often ordering is inefficient. Ordering requires to make stock calculations, do the reorder process, do the receiving process, store it in the warehouse, do the procurement process. (…) Finally, you want a full container if you send something, which adds to the argument that large orders are better.” (R#22, legacy system)

Interestingly, the implementation of VMI resolves this problem by removing lead times and providing a steady distribution frequency (e.g. every month).

Storing is simplified, regardless the system, by organising commodities on shelves and racks based on the type of medicine or based on alphabet. In addition, stock-cards are often connected to these shelves or cabinets. During the observations in Gambia, some pharmacies stacked all cards on a pile, resulting in more time spent on searching for the right stock-card before a simple transaction could be made.

Dispensing requires preservation of doctor’s prescriptions. Inaccurately updating stock-cards when medicines were dispensed negatively influences performance. As mentioned before, manual systems are often solely used by HF personnel who are too busy to accurately track dispensed amounts. This also accounts for both the legacy and vendor system if stock-cards are retained and monitoring is still solely in the hands of HF personnel.

(43)

43

connection since only temporarily internet connection is sufficient to synchronise with the server:

(44)
(45)

45

5. Discussion

This research explored what characteristics an IMS should have to support IM in developing countries’ HFs so that medicine availability can be improved upon by reducing stock-outs and expiry. As aforementioned, the availability of medicines in developing countries is very low due to medicine stock-outs and expiry. Causes are often sought in SC issues and the role of governments (United Nations, 2013), while the role of IM and IMS is barely addressed (Wales et al., 2014)

In order to fill the gap in the literature, this research posed the following research question:

What characteristics should an Inventory Management System that manages medicine stock in Health Facilities in developing countries contain? A key finding of this research is that all

IMSs can be categorised into three systems of which the vendor system is best adapted to the characteristics of medicines and the context of developing countries. Moreover, findings about the manual and legacy system comprehend a few interesting insights that help to determine the characteristics of an IMS.

The findings presented earlier are now transformed into the essential IMS characteristics, after which they will be discussed and compared against existing literature. Emphasis is put on the findings that are contrary to existing literature and findings that are completely new to literature. At the end of the discussion, the implications of the findings for research and practice, limitations and directions for further research are presented.

5.1 IMS characteristics

(46)

46

System characteristic Explanation

NGO or private funds Sufficient funding is a precondition for the development and

implementation of IMSs with characteristics as presented in Table 5.1.

Outsource IMS usage to a third party

Usage of an IMS should be outsourced to a vendor that possesses logistical knowledge. This enables the combination of IM tasks with distribution of medicines.

Hardware Wireless and portable laptops or tablets simplify collection of data during a distribution loop.

Software Automatic synchronisation of the collected data with the system through internation connection. Software should be able to run in online and offline modus.

Overview inventory data Automatically generate inventory status and inventory consumption overviews based on input and output.

List multiple expiry data List expiry dates of multiple batches of the same SKU at the same time.

Stock level updates Generate stock level updates that could be communicated with HF personnel.

Classification of medicines

Classification of the whole medicine list, transferred into buffer levels.

Reorder points Automatic indication when reorder point is reached based on the buffer levels that are determined and included in the system.

Forecasting multiple methods

Display consumption trends of individual SKUs for a period of time based on historical consumption data that is stored on the server. Also, formulas of the morbidity method should be implemented in the system and figures should be configurable by the vendor.

Physical stock count To know how much resupply is needed for a HF, the vendor has to do a count stock at each HF during a distribution loop. Thus, the physical stock count is included in the application of VMI.

Delivery date information Provide indication of the delivery date of new supplies. Max-min inventory levels Stimulate vendors to balance stock levels.

Remote temperature- and humidity monitoring

Continuous and automatic monitoring of storing conditions. Measures are directly sent to IMS by internet connection. This enables vendors to directly take action when conditions are not properly maintained.

Suggest medicine exchange

Provide suggestions for medicine exchange based on information about expiry dates, potential overstock and potential stock-outs.

(47)

47

5.1.1 Funding

In the existing literature it is indicated that large differences exist between high- and low-income countries in terms of money spend on healthcare by governments (Peters et al., 2008). The relatively low health expenditures in developing countries are often supplemented by NGOs who are financing additional health services (Pfeiffer, 2003). This research stresses that the engagement of NGOs improves medicine availability, partly because of their financial support that enables the introduction of advanced IMSs. Respondent 28 (working for a NGO) explained the introduction of consignment inventory, which means that the vendor supplies commodities and remains owner (Gümüs, Jewkes and Bookbinder, 2008). Consequently, the NGO finances the inventory, while the HFs only pays for medicines upon use:

“What would happen in the old system is that the health centre would get their annual state budget [from the government], which would arrive late and would not be enough. Hence, the health centre is limited to get the products they need. (…) [In the new system the] clients enter the facility and pay a dollar for some pills. Consequently, the health centre has to reimburse the money for the dispensed medicines.” (R#28, vendor system).

In conclusion, on the one hand this research reveals that NGO commitment improves medicine availability since NGOs enable the introduction of advanced IMSs. On the other hand, it should be mentioned that NGO commitment also improves medicine availability because their funds are used to finance medicines.

5.1.2 Outsource IM responsibility

In all eight cases, logistical knowledge is insufficiently available. However, this problem was already stressed by multiple researchers in humanitarian relief operations (Perry, 2007) and in single-country research (Wales et al., 2014). This research gained insight in the fact that the lack of logistical knowledge complicates the usage of IMSs and proper execution of IM functions. The lack of knowledge can be problematic because a certain amount of knowledge is essential for using an IMS optimally. The number of stock-outs could increase as a result of inaccurate tracking of dispensed medicines or through the inadequate retention of prescriptions. In order to reduce the dependability on overburdened HF personnel (Aiken et al., 2001), respondents 29 explained the advantage of VMI:

(48)

48

Engaging vendors still requires training of inhabitants and Kovacs and Spens (2009) argue that insufficient logistic knowledge in developing countries is the result of a shortage of domestic education. The results of this research indicate otherwise. It is found that NGOs, who are implementing vendor systems, provide the required training for inhabitants. Through the application of centralised logistics, less people have to spend time and effort on IM (Kowalski, 1986), while logistical knowledge can be maintained in the country.

5.1.3 Supply chain issues

VMI also provides a solution for SC issues that were mentioned during the interviews. Firstly, the results display that many HFs face difficulties in communicating with suppliers. VMISs are a proper solution for this issue because it integrates the SC, removes communication issues and improves the transparency between actors. This was already suggested by Mustaffa and Potter (2009), however it was not investigated in detail before. Secondly, this research indicates there is high uncertainty about the date of delivery and the long lead times in the medical SC. Shah (2004) argues that lead times in developed countries can take up to six months, whereas some respondents in this research mentioned that it could take even longer before supplies arrived in their HF. A vendor system can remove concerns of HF personnel about long lead times by integrating a part of the SC. In addition, it partly removes the uncertainty about the data of delivery by applying distribution loops. Thirdly, the findings of Van Wassenhove (2006), such as the lack of transport options and poor infrastructure in rural areas of developing countries, are confirmed in this research. Consequently, this research suggests a solution to these issues: the distribution loop. This resolves SC issues, such as a lack of transportation options at the HF level.

5.1.4 Monitoring

Almost all respondents indicated that inaccurate monitoring of inventory data is a fundamental problem. IMSs can make a difference in obtaining the required data, while the IMS users significantly influence the reliability of data:

“At HFs in Mozambique health workers are just not using stock-cards. They just do not keep track, which really complicates to get consumption data.” (R#29, vendor system)

(49)

49

Berger et al. (2007) are not mentioned by one of the respondents and were not found during the observations in Gambia as an IMS characteristic. The same applies to medicine classifications. Even though the medicines were sometimes classified in terms of importance, this was not converted into buffer and reorder levels and thus also not included in the IMS. Normally, classification is used to fix service levels for groups of SKUs and to determine different inventory levels to achieve these service levels (Teunter, Babai, Syntetos, 2010). This research reveals that in the context of HFs in developing countries, classification is applied with a slightly different purpose:

“What you need is a list with essential medicines. That list already exists and is called the WHO essential medicines list. If the medicines from that list are available, you could operate in a hospital. If we are almost out of money, we should only spend money on products that are on that list and in that manner we are still able to operate. In situations without money shortages, products on this list require an extensive buffer.” (R#14, manual system)

5.1.5 Forecasting

Respondents state that forecasting is the most difficult part of IM, while they admit that it is also the most important one. Forecasting is complicated by aspects such as long lead times, uncertainty about delivery dates and obscureness about when ‘to follow’ items will be delivered. In the existing literature, Armstong (2001) argues that the combination of forecasting methods improves the accuracy of forecasts and helps to avoid large errors, especially when uncertainty is high and when it is unknown which forecasting method is the best to use. He also states that methods should differ substantially and should be derived from different information sources. Based on the combination of the argument of Armstrong (2001) and the findings of this research, it is argued that IMSs should provide a combination of methods. This could be done through forecasting at a higher level of the SC so that forecasting according to the morbidity method is possible. Besides, forecasting based on historical consumption data should be applied since depending exclusively on the morbidity method is not very reliable:

“We experience that the inventory levels (that are the outcome of the morbidity method formulas) are quite good in order to forecast at a national level. However, these formulas are not really great for forecasting at a lower level of the supply chain.” (R#30, vendor system) 5.1.6 Ordering & Receiving

Referenties

GERELATEERDE DOCUMENTEN

Download date: 21.. Deze vraag wordt dan gesteld door een marktonderzoeker, die binnen het vakgebied van de aandrijftechniek op zoek is naar een nog niet ontdekt goudveld voor

Hoewel de vraagstelling voor dit project is gericht op het mestbeleid moet duidelijk zijn dat op gebiedsniveau natuurbeleid, milieubeleid, waterbeleid en ruimtelijke

Er zijn dan ook gegevens van meer bedrijven nodig om de effecten van bijvoorbeeld productiesysteem (elke week of elke drie weken dekken), speenmethode (konijnen wel of

As we look to the future, higher education in South Africa, and certainly on the African continent, will face new challenges that include access to an even greater diversity of

This study aimed to explore the knowledge, perceptions and skills of Foundation Phase English Second Language educators as determinants for developing a support programme

Voor de beoordeling van de gunstige effecten van glycerolfenylbutyraat ten opzichte van natriumfenylbutyraat zijn beschikbaar: resultaten van de relevante korte termijn fase

Bennett and Krebs (1991:1) further argue that LED concerns the wide range of factors which underpin the growth and development of local economies, while

Nu de effectiviteit van de dolfijntherapie niet in v oldoende mate evidence based lijkt te z ijn, kan niet gesproken w orden v an een door de beroepsgroep als effectief