Vertical or integrated health programmes? The consequences for the laboratory information system in Mozambique

Baltazar Chilundoa,b and Margunn Aanestadb

a) Faculty of Medicine, University Eduardo Mondlane, Maputo, Mozambique

b) Institute of Informatics, University of Oslo, Oslo, Norway

Corresponding author: Margunn Aanestad,

Instutute of Informatics

P.O.Box 1080, Blindern

NO-0316 Oslo, Norway

Email: margunn@ifi.uio.no

Vertical or integrated health programmes? The consequences for the laboratory information system in Mozambique

ABSTRACT:

Many developing countries are heavily dependent on foreign financial support for providing vital social services to the population. However, donor policies often limit the scope for national decision making and planning by focusing the support on vertical and stand-alone projects that exist alongside but are not integrated with the national structures. In this paper we describe some of the effects of donor-funded vertical programmes on health service provision in Mozambique, with an emphasis on clinical laboratories at the peripheral level. We also discuss the current attempts by the Ministry of Health to integrate these vertical programmes. Our claim is that such integration attempts are going to place new demands on the sector’s information systems. We discuss the challenges associated with this and whether useful lessons can be learnt from the IS literature from technology-intensive settings regarding similar attempts at integration of sub-standard and fragmented systems. However, there is little recognition in IS research to date of the special constraints that developing countries face due to a different material and financial context. Our case study also demonstrates the necessity of deploying a global perspective, as international organisations’ policies have profound influence in shaping local disparities in work practices and conditions. Our main message, based on the IS literature as well as the particularities from the case, is to emphasise the need for a profound knowledge of local work practices and work environments, and simultaneously the interdependency between them and across levels when attempting to integrate the information systems.

Keywords: health information systems, clinical laboratories, integration, data quality, developing countries, work practices, global organisations.

1. Introduction Many developing countries still depend on donors for the financial support of their health care sector. For the health care sector of Mozambique in 1997 only 22% of the funds came directly from the Government; of the rest 52% came from donors, and 26% from multilateral credits and households (family expenditure) (Chao et al. 2002). In order to allow donors’ direct control over their resources, the support has largely been organised as stand-alone, vertical programmes that have been targeted at specific goals or areas to address specific health problems. Some of the relevant health programmes in Mozambique are the ELAT/ELAL1 (Strategy for Fight Against Tuberculosis/Leprosy), Malaria, Sexually Transmitted Diseases (STDs)/AIDS and the Mother and Child Health (MCH) programme.

In this paper we describe the effects of such separate, vertical programmes on the information systems of the clinical laboratories at the peripheral level in Mozambique. The laboratories are service providers for medical personnel and they function as “hubs” in the health care institutions, as they relate to several different programmes simultaneously. This is in particular the case at the peripheral level. Laboratory services are crucial both for clinical decision making related to specific patient cases and for secondary use, e.g. epidemiological reasoning, so that shortcomings in the laboratories’ services will affect the overall quality of health care service delivery (Mallapaty et al. 2000).

In our study we have addressed the following research questions:

• What is the current situation concerning information flows and reporting?

• What is the relation between different health programmes and the laboratory’s information system?

• What are the generic operational barriers and facilitating factors of these health programmes in the laboratory system?

The paper is organised as follows: In the next section we describe the empirical study of work arrangements and information systems in use in the peripheral labs. We then go on to describe the findings, with a special emphasis on the effects of the vertical programmes in contrast to the integrated services. In section four we discuss the relevance of existing IS research and the implications of our case study.

2. Empirical study of laboratory information systems

2.1. Laboratories at the peripheral level

The National Health System in Mozambique encompasses 206 laboratories. These laboratories are all located within health care institutions. There are three laboratories at central hospital (level IV), seven in provincial hospitals (level III), seven in general hospitals (level II), 27 in rural hospital labs (level II) and 162 laboratories within health centres (level I). The degree of the complexity of the lab depends on the level of the health facility, with the most complex services requested of the labs in the health facilities at higher levels (e.g. the National Reference Laboratory). The organizational structure is illustrated in figure 1, below.

1 In the rest of the paper we will refer to ELAT/ELAL as the Tuberculosis/Leprosy Programme

Our study mainly concerns the laboratories at the rural hospitals and the health centres, which we denote the peripheral level. At this level, most of the analyses done rely on the microscope. There are usually no facilities for biochemical tests, or other tests that require technical equipment beyond the microscope. The laboratories at the peripheral level face severe resource constraints on most areas, from reagents, manpower, training, equipment, lack of electric energy and so forth. Consequently, the information system of the laboratories at this level is exclusively paper-based.

2.2. Research method

The empirical data has been collected during the years 2001 (June and July) and 2002 (from May to August). The main methods of data collection were observations, interviews and document analyses. The case study presented here is based on analysis of the perception of the functioning of the laboratories at the district level by the laboratory users, laboratory workers, decision-makers and laboratory managers2. The research approach is interpretive research, which implies that we attempt to understand phenomena through the meanings that people (either lab workers or users and managers) assign to them.

One semi-structured interview has been conducted with two managers working in the National Section for Laboratories (NSL). Observation of work practices were conducted in laboratories at the peripheral level (six health centres and four rural hospital laboratories) in six districts in two provinces (Chóckwe, Xai-Xai, Manjacaze districts in Gaza Province and Zavala, Inharrime and Maxixe districts in Inhambane Province).

Semi-structured and unstructured interviews were conducted with seven lab managers (working in Gaza and Inhambane Provinces), nineteen lab workers (assistants, agents and technicians) and with twenty-four lab users (clinicians, nurses, medicine and surgical 2 In general, almost all laboratory managers do work as laboratory workers.

162 Health Centre Labs

Provincial Directorate of Health

Maputo Central Hospital Lab 1 National Reference

District Directorate of Health

7 Provincial Hospital Labs 2 Central (regional) Hospital Labs

27 Rural Hospital Labs 7 General Hospital Labs

National Directorate of Health

National Section for Laboratories

Level I

Level II

Level III

Level IV

Figure 1: Organizational structure of the laboratory system in Mozambique

Labs targeted in the study (shadow)

technicians and agents) and one short interview with a consultant of Planning and Cooperation Directorate in the Ministry of Health headquarters in Maputo.

3. Findings from the study

3.1. Work practices in the laboratory

We first describe shortly how information flows within a peripheral health facility, between laboratory users (mainly clinicians) and laboratory workers. To order a certain test, the lab users send a request written on paper to the laboratory. Usually (if the patient is fit to walk) each patient goes to the laboratory and gives the specimen by himself/herself. In these cases, the patient brings the request form and gives it to a lab worker who collects and registers its content in a registration book. At this point the request form is given a number (serial number) that will be the identification for further work in the lab. Then the specimen is collected, and the specimen containers are labelled with the same serial number.

Before going to the analysis room, the specimens can be kept in the refrigerator or incubator depending on the required storing conditions. The request forms are kept on a shelf according to examination type, date and priority. In the analysis room, the specimens are prepared and analysed, and the result is then written on the paper request form, which then becomes a lab report. The result is also registered in the registry book together with the information that was entered when the request was received. The report may be kept on the shelf for some time, until it is brought back to the doctor requesting it. Usually patients or servants3 act as couriers who bring the lab reports back to the doctors. The information contained in the report is usually based on abbreviations so that the couriers can not interpret the information inscribed in the report (e.g. NSE for malaria test meaning that no malaria parasites were encountered).

3.2. General situation

However these labs function with a range of constraints imposed by the lack of reagents and working equipment. Our impression is that lab users don’t rely heavily on the test results, because they know the constraints the lab works are under as well as the nature of the test themselves (whether their specificity and sensitivity is adequate). Doctors usually place more emphasis on clinical observations, and may not bother to even request lab tests in many cases. We also have the impression that the lab workers’ sense of obligation and loyalty is directed towards these primary users of the lab services. Despite the local problems (work overload, lack of equipment and spare parts) the lab workers see the provision of analyses as their main work task. The reporting of these data to health care institutions at higher levels are given secondary importance by the lab workers.

3.3. The information system in the laboratories

Within the National Health System, the main health information system (SIS) collects data from almost all health facilities in the country. The IS for laboratories is officially run as a branch (subsystem) of SIS. Each month every laboratory shall collect and report data (on standardised paper forms) regarding the number of tests that have been performed, and the amount of reagents that has been used. This information is aggregated at the district level with other kinds of data coming from infirmaries, mother and child health, immunisation and other programmes and sent to provincial and national levels.

3 In Mozambican health sector, servants are the health workers who are employed to help in the health facility (cleaning, couriers, patient accompaniers…) usually without any specific training.

The general picture concerning the laboratory information flow between the different health care levels is that it is not working well. When reported at all, the collected data is not entirely reliable. It appears that the information is not used (at least not to any substantial degree) by the authorities that request this information. This impression has been confirmed by lab managers at hte district, provincial and national levels, including also the national staff of the Department of Health Information in MoH. In general the information is reported because of reasons of bureaucracy where the practices take on a self-sustaining institutional structure. Feedback from the higher levels is almost non-existent. This data collection is not valued highly by the lab workers, and the reporting task is seen as secondary to the primary task of providing results to different clinical tests. There is very little local use of the data. Only in one health centre lab (Chalucuane Health Centre) did we find that comparative graphs and tables were used to detect seasonal variations and to monitor disease patterns.

Beside the official reporting routines, there are also some unofficial and parallel routines dealing with the lab data. A lab manager will use a locally improvised form to report in more detail, indicating the total number of analyses performed in relation to the type of tests requested. These (unofficial) are the data used by provincial and national lab managers to make annual reports describing the overall laboratory situation, including the key needs and problems.

However, there is generally a weak link between inventory lab data and the request for reagents. There are fixed quantities of reagents for example, dispatched (by the provincial level) regardless of the reported needs and the requested supplies. In practical terms, whether the lab workers/managers collect data or not, they get the supplies in predetermined quantities (i.e. determined by the available amounts). To request reagents and other consumables the lab manager fills a standardised requisition form to report the existing needs, usually in a trimester basis, and sends it to the provincial level where there is a storage centre for the labs. The provincial lab manager uses the requisition forms from districts to compile the provincial needs in supplies and request to the central level. The NSL in MoH prepares an annual plan of the national needs in supplies (excluding the Tuberculosis/Leprosy lab consumables) and submits the plan to the Directorate of Administration and Management (DAG) in MoH (responsible for logistics). According to the available budget the acquisition of the reagents is made. An example given by a lab manager at the central level concerned the requisition of lancets (tools used to collect small amounts of blood, usually for malaria tests) for the year 2002. According to the list, they (lab managers) have requested 100 000 boxes with 500 lancets each (required needs), but they received only 2000 such boxes from the DAG. This situation is due to limited stock of reagents and other consumables at the national level imposed by financial constraints.

While the laboratories comply more or less strictly with this official (and unofficial) reporting system within the National Health System, there also exist other parallel routines for reporting that originates as a result of the vertical health programmes. Before the beginning of the integration process that the government wants to carry out (which we will describe in due course), all labs were supported by various vertical programmes. We describe one such parallel reporting system supported by the Tuberculosis/Leprosy programme.

3.4. The impact of the vertical programmes on local work practices

One of the major programmes in the Mozambican health care sector is the Tuberculosis/Leprosy programme (ELAT/ELAL). This programme remains outside the national health care structure in the sense that it has its own budget, almost completely funded

by a donor or group of donors with common goals. We will describe the information system of this programme in contrast to the general information system.

At the district level the programme has a health worker who is responsible for all programme-related tasks. Every month this person goes to the laboratories and collects all TB or leprosy data (e.g. the number of positive smear tests, the available amount of reagents and other consumables). The information from the district is compiled with other relevant TB/leprosy data and sent to the upper levels of the hierarchy (within the TB programme, not the National Health Information System). The supply of consumables, reagents and materials for TB test are also distributed through the programme. During our visits to the laboratories, the supplies and reagents for TB tests were available in all visited labs.

The TB/Leprosy programme (as the only one) provides the laboratories and the laboratory users with a standardised paper form, which is used when a tuberculosis test is being requested (see figure 3). This form has several tick-boxes and structures the information to be entered. In order to request any other test than a TB test, physicians have to use ordinary sheets of paper. The handwritten forms are often incomplete and illegible, which according to our informants is a factor that contributes to misunderstandings, e.g. concerning which test that is requested.

The Tuberculosis programme also provides a standardised registry book. This is the book in which the basic information on an order is registered, typically including the date, the patient’s name, the test required, results, as well as the serial number. Both the Tuberculosis and the Sexually Transmitted Diseases/HIV programme have their own specific standardised registry books; however we found the TB programmes books in all labs. For the rest of the tests requested there are no books provided by the MoH. Often the lab workers will purchase a locally available exercise book and separate the pages into sections themselves (see figure 4). Thus the structure and information content of these general registry books may differ from lab to lab. One example is that all laboratories will register the number of malaria tests

(B)

(A)

Figure 2: The two paper-based request forms most used in peripheral labs (A) general order form; (B) TB order form

performed, but not all will provide the additional information on whether the test was positive or negative. This implies potential problems for secondary users of this information, e.g. for epidemiological studies of malaria prevalence.

The supervision visits and the external quality control for TB are frequent. Each trimester the local lab manager selects some slides and sends these to the reference laboratory and the feedback report is often sent back. We have seen the TB feedback reports in many peripheral labs. In addition, the TB programme offers regular training sessions. This is one of the few chances a laboratory worker gets to receive training, but the training typically concerns tuberculosis tests only. Participants in the training seminars usually receive per diems, financial support for travel and lodging, which contribute for more interest in participating as the workers often complain about their salary being a “misery”. This results in the other programmes getting less importance. Many lab workers interviewed were not trained regularly for malaria tests, despite tests for malaria being the most commonly requested outnumbering TB tests by far. The figure 2 shows this picture from Chicuque Rural Hospital Laboratory (note that all of these tests are usually performed by the same laboratory worker).

Figure 3: Part of registry books used in the peripheral labs (A) general registry book, (B) TB registry book

(A)

(B)

Malaria

Syp

hili

sU

rine

Fae

ces

TB oth

ers

Malaria

Syp

hilis U

rine

Fa

ece

s

TB oth

ers

Malaria

Syp

hilis

Urin

eF

aece

s

TB oth

ers

Malaria

Syp

hili

sU

rine

Fae

ces

TB oth

ers

Malaria

Syp

hili

sU

rine

Fa

eces

TB oth

ers

Malaria

Syp

hilis

Urin

eF

aece

s

TB othe

rs

0

500

1000

1500

2000

2500

3000

N.o of tests

January February March April May June

Comparison chart: Types of tests performed in the 1st semester 2002 Chicuque RH

Malaria

Syphilis

Urine

Faeces

TB

others

Thus there are significant differences between tests according to the specific programme that supports the lab. Some tests (e.g. biochemistry, haemoglobin) are performed with a range of constraints and without the logistic and practical support that the vertical programme’s tests receive. The Ministry of Health is aware of the implications of the activities of the vertical programmes, and has attempted to overcome the problem by aiming to integrate the different health care programmes.

3.4 The efforts to integrate health programmes The issues of vertical and specialised donor programmes are well known from many sectors, not only health care. Such arrangements do not facilitate transparency, or comprehensive and long-term planning and may be lead to duplication of functions and lack of coordination, and wastage of resources (Hutton 2002). Currently there is a more global trend, backed by major actors such as the World Bank and the WHO, towards the integration of such stand-alone programmes into larger sector or sub-sector programmes. “Integration” could mean the interaction of vertical programmes with the general health structure, or the integration of two or more such programmes with each other. In addition to different objects of integration there are also different levels of integration:

• integration of services at the point of delivery • integration of resource management (making local management systems efficient) • integration of organisation and policy (making the service structure consistent with

policy) (Briggs et al. 2002) In an integrated programme, functions that would have been performed by different staff assigned to separate health initiatives are now combined, reducing costs to the client (time, training, transportation, etc.) (Miller et al 1996) and requiring far fewer people on government payrolls, a fundamental requirement of Structural Adjustment Programmes (WHO-CHD, 2002). The aim is to utilise the existing infrastructure, resources and personnel resources more fully and to avoid the duplication of efforts in both administration and service delivery (Cossa and Gerritsen, 2000). The integration attempts are followed by hopes that the local processes of prioritisation shall be facilitated and supported, leading to increased local ownership, engagement and motivation. The increased financial stability shall allow planning with longer time horizons (Chao and Kostermans, 2002). Like many other developing countries, Mozambique is implementing such integration programmes. The Ministry of Health (MoH) has decided to implement a sector-wide approach in the health care sector. (..SWAp4..) So far the implementation has focused on establishing high-level policy discussion with MoH’s funding partners, as is reflected in the “Kaya Kwanga commitment” from May 2000, which was signed by all major donor organisations (MoH Mozambique, 2000). One important issue of this commitment is the transition from several vertical health programmes into one common structure. The aim is to pool the donor grants into a common account and to allocate resources not tied to specific programmes. This integration is planned to be carried out at the provincial level, so that authorities at the provincial level resume responsibility for allocation of resources (Provincial Department of Planning or to the Provincial Health Departments). The reorganisation of the whole sector

4 A SWAp project means: Develop a common vision, set common priorities and allocate resources, rationalise and improve the coordination of external assistance (i.e. donor support) (Cassels et al, 1998). The aim is to support national governments in reaching their development goals within a coherent sector, defined by an appropriate institutional structure and national financing programme (Schreuder, KIT 2001)

(and thus these potential benefits) is as yet still underway and is not yet fully implemented. In Mozambique the STDs/AIDS, malaria, MCH, Family Planning (FP) and Nutrition programmes have entered into an integration process. The Tuberculosis/Leprosy is now the only stand-alone vertical programme left in the health care sector. The reasons for this are not known to us at this time.

4. Integration challenges

4.1 The need for integration of the information systems Comprehensive, relevant and reliable information on resource use will be required if the Mozambican Ministry of Health (MoH) shall succeed in enrolling the donor organisations into pooling their grants. The MoH will also have to document that their planning and budgeting practices follow the norms of what is perceived as good management, i.e. that planning and resource allocation are being done in a “rational” manner, based on a comprehensive (and accurate) picture of needs (This for reasons of efficiency as well as issues of transparency etc.). These demands will imply significant changes in the information systems in health care. As we have seen, reporting information for management purposes is not highly valued, neither by the reporting institutions nor by the receivers. Consequently, the reported information is both incomplete and unreliable, and there does not exist a strong culture for using it in the decision-making processes. We have also seen that there exist considerable differences between vertical (e.g. TB) and integrated programmes, e.g. concerning the degree of successful data collection. The fragmented information systems need to be “upgraded” if the aim of integrating the different vertical programmes shall be reached. Such an “upgrading” is not a trivial undertaking, as it demands resources and skills which are not easily available in this particular context.

4.2. The challenges of top-down integration However, the attempt to integrate diverse information systems is extremely challenging also for other reasons. During the last decades, the IS community have seen numerous attempts at deploying ICT to achieve integration within corporations, whether they are called Business Process Reengineering, Enterprise Resource Planning, Customer Relationship Management, warehouses etc. The aim is to integrate technology (equipment, software, applications, systems), processes, and people. The promised benefits of these integration processes are not easily achieved, and such change processes have been notoriously difficult. The notion of “drift” has been suggested to describe the aspect of unexpected and unmanageable trajectory of such projects (Ciborra, 2000). There seems to be limited scope for governance, mainly because new interdependencies are being introduced through tighter integration of systems. (MORE) (The result is a certain ambiguity; control technologies that resist control, that simultaneously enhance and curb our governance capabilities). One can learn that attempts to support management with “better” and more comprehensive information will be a non-trivial issue. Integration projects often encounter drift and unexpected side effects. Very ambitious attempts seem to be worst off, and rather than “starting with a clean slate”-approaches, gradual changes seem to be the best way forward. Integration programmes are in general initiated, planned and implemented at the central level. It is a top-down approach, centred on policy issues, e.g. defining and implementing policy and

structural changes. This is also the case with the integration of the health care sector in Mozambique, where the process in led from the Ministry of Health. We argue that this approach will be insufficient. In order to succeed, a policy-oriented approach needs to be complemented with detailed, empirically based understanding of the information systems in the health care sector at the micro level.

4.3 Complementing the top-down approach The work practices related to the information systems, as well as the connectedness and interdependencies between the various elements of the systems, should be explicitly addressed in this integration process. The empirical material from our case study describes the work practices related to the information systems at the peripheral level. These work practices are situated, occurs in a constrained setting and are often informal and ad hoc. Such local work practices are not easily standardised, streamlined and integrated. However, they ar ealso related to factors beyond the local context (logistics, training, availability of resources). An extensive case study of an integration project in a global company illustrates how challenging such an standardisation and integration endeavour was (Rolland and Monteiro, 2002). The authors describe how local workarounds and adaptations were necessary in order to “root” the standard. However, this also partially undermined the central standardisation attempts, and these local changes did not just stay local, but propagated throughout the network and had “global” consequences. Despite such problems a certain level of standardisation will be required for the different parts in the system. Before defining these standards and requirements it is of crucial importance to develop an awareness of the information flows and the potential pitfalls and problems in an integration process. Nothing less than a holistic approach can address the source of the problem, which relates to compromised data quality. One of our main points in this paper is that the low data quality is not just a result of inadequate local work practices. Equally important is the cross-level dymanics, e.g. the weak links within the national health information system and the disparities in contrast to the vertical programmes. It is thus important to not just analyse one level at a time, but to focus on the interdependencies and relations between different levels in the health care structure simultaneously. For example, as long as the reported information on the amount of stocks used is not used in the supply process (which is a currently organised as a dispatch of predetermined quantities of reagents), there will be no incentive to improve the quality of information on resource use at the local level. On the other hand, if the authorities want to examine the malaria prevalence, not just based on clinical judgment (as is today), but look at malaria cases that have been verifies with laboratory test, this kind of information is not readily available from the labs, because the registry books are not standardised. Improving the data quality and achieving an integration of these systems requires a profound understanding of both inter- and intra-level dynamics. There is a heterogeneous multitude of different actors, systems, logics, and it is crucial to understand their interdependencies and the ramifications throughout the network.

4.4 Stepwise integration at the practical level We have seen that non-technical or non-policy issues play an important role in the use of the information systems, e.g. the local work practices, social relationships and institutional arrangements in a given setting. The role of local improvisations related to the use of information technologies in developing countries is also explicitly emphasised by other researchers, see e.g. (Silva, 2002; Heeks, 2002; Bada, 2002).

For the case of the Mozambican health care, this accrues to a setting where centrally mandated directives will not have the desired effect of “magically” producing high quality data. We suggest that a stepwise strategy aimed at changes may be best suited in order to improve the information systems. Rather than trying to reform everything through a centrally mandated project, it might be wiser to gradually improve bits and pieces of the existing information system. This should be done within a sensitivity of the local work practices and the complex interdependencies. One example of a stepwise approach would be to start with a suggestion to cooperate with the TB programme on training rather than demand full integration of the whole programme. The national authorities could ask for some lectures on other lab tests (e.g. malaria) to be included in the training courses. A further step would be to create similarly standardised order forms and registry books, and to distribute these through the same channels as the TB programme now does (the district representative). Providing such standardised forms and books might also have the effect of enhancing the value assigned to data collection, registration, and documentation by the lab workers. However, in achieving this, it will be of immense importance that the reporting task is perceived as relevant, necessary and useful. For this to happen, it is necessary to implement stronger links between the reporting activities and the resource supply activities, as these are interdependent.

5 Final remarks We have emphasised the need for an approach that at the same time is sensitive to the local work practices and constraints, and at the same time is aware of the complex interdependencies that exist between these local practices and other factors beyond the local context. Beyond the practical implications, we also believe that this study can be of value to the broader IS community. The paper describes the challenges related to information systems issues in a particular context which has not been extensively researched. Information systems development in this context of financially constrained developing country will face different challenges from ordinary systems development. These challenges are related to inadequate financing, lack of well functioning infrastructures, managerial problems, excessive bureaucracy, and not the least: the country’s external dependency. Our research findings have demonstrated that the international donors’ policies and strategies drastically influence the micro-level issues of day-to-day work in peripheral clinical laboratories. The different parts of the health care services display different degree of “connectedness” to global actors, and these differences have very tangible effects. It is through these linkages with the extra-national agencies that vertical programmes can continue to exist and pursue their own agenda with minimal government coordination and control. This “trans-national” or global dimension of “the context” can not be left out from an analysis. This is very different from the context where most IS literature have emerged. We think this paper offers an example that may help us understand how the macro issues connected to processes of globalisation play out at micro–level, as well as a demonstration of how interweaved non-technical issues are, the local work practices, social relationships and institutional arrangements. We think this perspective should be a topic of interest to the IS research community.

References:

1. Bada, A.O. (2002): Local adaptations to global trends: A Study of an IT-based organisational change program in an Nigerian bank. Information Society, vol. 18, no. 2, March 2002, pp. 77-86.

2. Briggs CJ, Capdegelle P, Garner P (2002). Strategies for integrating primary health services in middle- and low-income countries: effects on performance, costs and patient outcomes. The Cochrane Library, Issue 3. Available at www.cochrane.org/cochrane/revabstr/ab003318.htm, last accessed 22/09/02

3. Brown A, Foster M, Conway T (2000). Sector Wide Approaches for Health Development – a Review of Experience. In Conference “Sector-Wide Approach in Health: Moving from Policy to Practice”, Royal Tropical Institute, Amsterdam.

4. Cassels A, Janovsky K. (1998). Better health in developing countries: are sector-wide approaches the way of the future? The Lancet 352(9142): 1777-1779

5. Ciborra, C. (2000): From Control to Drift. The Dynamics of Corporate Information Infrastructures. Oxford University Press, UK.

6. Chao S, Kostermans K (2002). Improving Health For the Poor in Mozambique: The Fight Continues. The International Bank for Reconstruction and Development/ The World Bank

7. Chilundo B, Sahay S (2002). Clinical Laboratory Information Systems in Mozambique: The great challenge. In: Proceedings of the IFIPWG9.4 working conference on Information & Communication Technologies and Development: New Opportunities, Perspectives & Challenges (Krishna S, Madon S, editors). IIMB. 477-487

8. Chrisanti A (2002). The Global, the Local, and the Disembedded. Information Systems and Global Diversity. Oxford University Press. 98-119.

9. Cossa H, Gerritsen M (2000). The Case of Mozambique:�The conductor of an orchestra, who moon-lights as a taxi driver to pay off the loan for his car. In Conference “Sector-Wide Approach in Health: Moving from Policy to Practice”, Royal Tropical Institute, Amsterdam.

10. Egwaga SM (2000). SWAP and Vertical Programme: Experience from National TB/Leprosy Programme. In Conference “Sector-Wide Approach in Health: Moving from Policy to Practice”, Royal Tropical Institute, Amsterdam.

11. Hanseth O, Ciborra C, Braa K (2001). The Control Devolution: ERP and the Side-effects of Globalization. In: The Data base for advances in information systems, special issue on ERP systems. 32 (4) 34-46

12. Hanseth O, Lundberg N (2001). Designing Work Oriented Infrastructures. Computer Supported Cooperative Work (CSCW). The Journal of Collaborative Computing. Vol. 10 Nos. 3-4, 2001, pp. 347-372

13. Heeks, R. (2002): Information Systems and developing countries: Failure, success and local improvisations. Information Society, vol. 18, no. 2, March 2002, pp. 101-112.

14. Hutton G (2002). Introductory text and bibliography for discussion group: Issues in integration of vertical health programmes into sector-wide approaches. Swiss Tropical Institute. Available at:http://www.dgroups.org/groups/sdc-health-swaps/docs/Introductory/. Last accessed 27/09/02

15. Mallapaty G, Shrestha GB (2000). Situation Analysis: Health Laboratory Services in Nepal. Department of Health Services. National Public Health Laboratory. 2- 26 (unpublished report).

16. Miller J, Wolff JA ed. (1996) Management Strategies for Improving Family Planning Services: the family planning manager compendium. Management Sciences for Health. Boston, MA. 187-207.

17. Ministry of Health Mozambique (2000). The Kaya Kwanga Commitment: A Code of Conduct to Guide the Partnership for Health Development in Mozambique

18. Monteiro, E. & Hanseth, O. (1996). Social shaping of information infrastructure: on being specific about the technology. In: Proceedings of the IFIPWG8.2 working conference on information technology and changes in organizational work, (Orlikowski, W.J., Walsham, G., Jones, M.R., Eds.). Chapman & Hall, pp. 325-343.

19. Rolland, K. and Monteiro, E. (2002). Balancing the local and the global in infrastructural information systems. Information Society, vol. 18, no. 2, March 2002, pp.87-100.

20. Schreuder B (2001) SWAps approaches for sustainable development in practice.available at: http://www.kit.nl/specials/html/sw_paper_bert_schreuder.asp last accessed 18/09/02 12:00

21. Silva, L.O (2002): Outsourcing as an Improvisation: a case from LatinAmerica. Information Society, vol. 18, no. 2, March 2002, pp. 129-138.

22. Technical Advisory Group (1998). Towards a Sector-Wide Approach to Programming in the Health Sector: Options for Dialogue and Action. MoH Mozambique.