Background Information Operations Research Definition, Purpose and Applicability by Lena Siegel World Vision Institute Am Houiller Platz 4 61381 Friedrichsdorf Tel.: (06172) 763–176 Fax.: (06172) 763–270 [email protected] Contents Preface ............................................................................................................................................................................3 Abstract ..........................................................................................................................................................................4 1. Why is it relevant to address Operations Research? .................................................................................5 2. Scientific Background of OR.............................................................................................................................6 3. 2.1. Purpose and Definition of OR ................................................................................................................6 2.2. History of OR .............................................................................................................................................6 2.3. Characteristics of OR ...............................................................................................................................7 Other Types of Research ............................................................................................................................... 12 3.1. 3.1.1. Exploratory Research ................................................................................................................... 13 3.1.2. Descriptive Research .................................................................................................................... 13 3.1.3. Explanatory Research ................................................................................................................... 14 3.2. 4. Applied Research .................................................................................................................................... 14 3.2.1. Evaluation Research ...................................................................................................................... 14 3.2.2. Action Research ............................................................................................................................. 15 3.3. Other aspects of research .................................................................................................................... 16 3.4. Differences between Evaluation Research and OR ........................................................................ 16 OR in the Field of Social and Health-related Research .......................................................................... 17 4.1. Characteristics ......................................................................................................................................... 18 4.2. The OR Process ...................................................................................................................................... 19 4.2.1. Problem Identification and Design of Experiment ................................................................. 19 4.2.2. Solution/Strategy Selection .......................................................................................................... 20 4.2.3. Solution/Strategy Testing and Evaluation ................................................................................. 21 4.2.4. Information Dissemination and Program Adaptation ........................................................... 22 4.3. 5. Basic Research ......................................................................................................................................... 13 An Example of Social OR ...................................................................................................................... 23 Final Remarks .................................................................................................................................................... 29 Appendix 1.................................................................................................................................................................. 31 Operations Research Checklist ........................................................................................................................ 31 Appendix II.................................................................................................................................................................. 33 Annotated Bibliography ........................................................................................................................................... 37 2 Preface At a time when the rationale for international development assistance has been seriously questioned not only by people in donor countries but also by an increasing number of representatives of recipient countries1, it seems increasingly vital to assess the impact of development programs. One of the reasons for questioning development aid is the economic stagnation and social predicament in which many of the poorest countries, notably in Africa, still find themselves despite decades of concerted efforts to reduce poverty and injustice in the developing world. The necessity of impact assessment has always been a crucial and yet ambiguous issue. On the one hand, the development business has been subjected to a multitude of evaluation appraisals to satisfy donor agencies trying to ascertain impact evidence. On the other hand, development programs rarely have an inbuilt mechanism for ongoing program adaptation for the purpose of ameliorating, let alone optimizing, their impact. Furthermore, many observers have satisfied themselves with the finding that there was a perceptible impact at all, regardless of the costs, according to the dictum: As long as there is a significant positive impact, we don’t question the costeffectiveness ratio. Since the development business is not subject to profit expectations by donors or shareholders, the issue of cost-effectiveness has often been sidelined or even completely ignored. Private non-governmental organizations such as World Vision have only fairly recently recognized the importance of demonstrating not only the effectiveness of their programs but also their cost-effectiveness, i.e. the efficiency of their approaches. This has led to the recognition that evidence-based research must be drawn upon to ascertain certain development objectives: (1) to improve or even optimize the effectiveness and efficiency of the programs; (2) to recognize that there may be more than just one solution to a particular problem and to test these alternative solutions; and (3) most importantly, to make ongoing adaption and amelioration an integral part and parcel component of the programs right from the start. Few research methods will meet these criteria. Operations Research (OR) claims to do this. Unfortunately, there is as yet a widespread lack of understanding of what OR is and what criteria it must fulfill to be OR. Encouraged by World Vision’s international Health, HIV/AIDS and WASH CoP Team, the World Vision Institute has set out to define and explicate OR characteristics and to show its applicability in the social and health field. I thank Lena Siegel for undertaking this important task. In my view, she herewith submits an excellent paper of great practical relevance and applicability. Kurt Bangert Director Research World Vision Institute for Research and Innovation 1 See, for instance: Dambisa Moyo: Why Aid is Not Working and How There is a Better Way for Africa, 2009. 3 Abstract The goal of this paper on Operations Research (OR) is to provide a basis for further research within World Vision’s programs. It should serve as a guideline for OR that is conducted to improve the effectiveness and efficiency of the programs. Therefore, it is important to understand the concept of OR in all its details. The paper starts with a short introduction describing why it is important to address Operations Research. The second part covers the scientific background of OR including the purpose, the history and the characteristics of OR. To apply this approach to social and health-related research it is crucial to understand from which scientific field OR derives and which steps are essential while conducting an OR study. The next part demonstrates other types of research. These types either fall in the category of Basic Research or Applied Research. This chapter should contribute to a better distinction between different types of research and to highlight the difference to OR. The fourth part concentrates on OR conducted in social and health-related research and displays the specific characteristics as well as the process of OR adapted to this field. This part concludes with an example of an OR process in the field of primary health care. The paper ends with final remarks about which minimum standards studies must meet to be classified as OR. 4 1. Why is it relevant to address Operations Research? Looking at the research conducted to measure the performance of programs that are implemented in the development field, we more and more see that it is classified as “OR”. But what is behind this concept, which comes from the field of mathematics, informatics and economics? What are the steps to be taken in order to conduct this kind of research? And how can this approach be adapted to the social and health-related field? What is the purpose of an OR study and how does this concept differ from other types of research? When conducting an OR study, we must answer each of these questions. With this paper I hope to contribute to a deeper understanding of the components of OR and to its adaptation to the social and health-related field. 5 2. Scientific Background of OR In this part of the paper, I will describe the scientific background and the characteristics of Operations/Operational Research. I will use the term “Operations Research”. The term “Operational Research“ is primarily used in the UK. 2.1. Purpose and Definition of OR As the name implies the basic definition of OR is “research on operations”. OR is applied to complex problems about how to conduct and coordinate the operations and activities within an organization. Organizations may seek a very wide range of operational problems in terms of efficiency, customer service, quality or cost. The objective of OR is to find practical solutions to operational or strategic problems to enhance the effectiveness of the whole organization. Thus, OR addresses the practical management and needs of the organization and therefore must provide understandable information on how to overcome the problem for the decision makers. Furthermore, rather than just improving the status quo, OR aims at finding the best solution amongst a variety of possible solutions, and that best solution is called the optimal solution. The goal is to find a best possible course of action, which means the “search of optimality”. Usually, there will be multiple solutions that need to be tested in order to find out the optimum. OR is being used in diverse areas such as manufacturing, transportation, construction, telecommunications, financial planning, healthcare, the military and public services. Thus, there exists a wide field of application for OR. The Association of European Operational Research Societies (EURO) states that OR can be described as a scientific approach to the solution of problems in the management of complex systems (cf. EURO 2011). The German Society of OR (GOR) has a more general definition as it defines OR as “the development and the application of quantitative models and methods to support decision making.”(GOR 2011). 2.2. History of OR The roots of OR can be traced back many decades. Since the industrial revolution, organizations have grown in complexity and size. The division of labor and the segmentation of management responsibilities started. This development created new problems. The different areas of an organization grew into relatively autonomous areas with its own goals and values. Thus, it may happen that different areas work for their own goals and loose the overall objectives of the whole organization so that different areas end up working at cross-purposes. 6 “A related problem is that as the complexity and specialization in an organization increase, it becomes more and more difficult to allocate the available resources to the various activities in a way that is most effective for the organization as a whole.” (Hillier & Lieberman 2010, p.1.) From these kinds of problems and the need to find a better way to solve them OR emerged. Generally, World War II is referred to as the beginning of OR because there was a need to allocate scarce resources to the various military operations and to the activities within each operation in an effective way. Therefore, the British and U.S. military management called upon scientists to apply a scientific approach to dealing with this and other strategic and tactical problems. They were doing research on (military) operations. At the end of war, the success of OR resulted in using OR outside the military as well. The problems caused by the increasing complexity and specialization in organizations were again arising. It became clear that these problems were the same that had been faced by the military but in a different context. In the early 1950s, people who were engaged in OR during the war had introduced the use of OR to a variety of organizations in business, industry and government. The rapid spread of OR soon followed. Nevertheless there are also other important factors that played a key role in the rapid growth of OR. First of all, a remarkable progress in improving the techniques of OR had been made. Many of the standard tools such as linear programming, dynamic programming, queuing theor and inventory theory were developed during that time. Second, the computer revolution played a key role in the development of OR because a large amount of computation is usually required to deal with the complex problems typically considered by OR. In the 1980s OR got a further boost because of the development of more powerful personal computers and good software packages for doing OR. The progress further accelerated in the 1990s and into the 21st century. Today, millions of people have access to OR software. 2.3. Characteristics of OR OR has some specific characteristics that are important for planning and conducting it. Quantitative mathematical techniques are the main methods to conduct OR. But that does not mean that a practical OR study is primarily a mathematical exercise. The mathematical analysis forms only a small part of the total OR study. Therefore, a group of individuals with diverse backgrounds and skills is needed to execute OR. No single individual can be an expert on all the many aspects of OR work and problems. The OR team typically needs to include individuals who are highly trained in the fields of mathematics, statistics, probability theory, economics, business administration, computer science, 7 engineering, physical sciences, behavioral sciences and the special techniques of OR. However, which experts will actually be included certainly depends on the nature of the problem and the resources available to solve it. OR uses an approach that resembles the way research is carried out in established scientific fields. This approach is often referred to as the “scientific method”. A typical OR study can be divided into six major phases that may overlap each other (cf. Hillier & Lieberman 2010, p. 8): 1. Defining the problem of interest and gathering relevant data 2. Formulating a mathematical model to represent the problem 6. Implementation 3. Developing a computerbased procedure for deriving solutions to the problem from the model 5. Preparation for the ongoing application of the model as prescribed by management 4. Testing the model and refining it as needed Although the six steps may not all together be applicable to the social field, in what follows here we are nevertheless giving a brief description of what typically constitutes OR: Phase 1 The first phase in conducting an OR study consists of the careful observation and formulation of the problem including gathering all relevant data. In this first step, it is crucial to develop a welldefined statement of the problem. The statement must include relevant considerations of the problem such as appropriate objectives, limitations on what can be done, the context of the problem and possible alternative courses of action. Another important part of the first phase is the involvement of the decision makers. An OR study has normally the goal to advise the management. The OR team first performs a detailed 8 technical analysis of the problem and then presents recommendations to the management. The management will make the final decisions. Therefore, it is important for the OR team to cooperate with the management and to identify the right problem from management’s point of view. The OR team needs the support of the management for the implementation of the study. Additionally, OR is concerned with the welfare of the whole organization. But often problems only concern one component of the organization. In this case, it is the challenge for the OR team to formulate the objectives as specific as they can be while still incorporating the main goals of the decision makers and the overall objectives of the organization. Gathering all the relevant data that is necessary to conduct the study is very time-consuming. The research team needs much data to gain an accurate understanding for the development of a mathematical model in the next phase. In many cases, it will be necessary to create a Design of Experiment. Equipped with such a Design, researcher will test certain assumptions. For more information on Design of Experiment, pls. go to p. 20. Phase 2 In the second phase, the OR team has to develop a model that represents the existing problem for further analysis. It is hypothesized that the model represents the real problem for all practical purposes so that the conclusions and solutions can be adapted to the real problem and are valid for it. Typically this model has to be a mathematical model2 and requires the formulation of a mathematical function (Figure 1). The function consists of equations and mathematical expressions. Therein, the decisions to be made are represented as variables (x1, x2, …, xn). The values of the variables have to be identified. The goal might be to choose the values of the variables in a way to maximize the whole mathematical function (e.g. maximize the profit) or to minimize the function (e.g. minimize the number of rejections). Assigned appropriate values to the variables of the model is a challenging process. It is important to really understand how the solution will change if the values for the variables are changed because often the values are only estimated from the data. In this context it is important to distinguish between variables and parameters both of which may figure in the model. Variables are concepts whose numerical value is under the control of management. Parameters are concepts whose numerical values are given by the environment or outside forces. The value of parameters can be measured but not changed. The value of variables can be set. „ [Such] models are invaluable for abstracting the essence of the subject of inquiry, showing interrelationships, and facilitating analysis. Mathematical models are [also] idealized representations but they are expressed in terms of mathematical symbols and expressions.” (Hillier & Lieberman 2010, p.11) 2 9 Figure 1: This figure shows a mathematical model with three variables (retrieved from Stat-Ease). Phase 3 In the third phase of the OR process the goal is to develop a procedure for deriving solutions to the problem from the model. That is usually a computer-based procedure that works with algorithms. OR has the goal to search for an optimal solution but in practice the researchers have to keep in mind also the cost-benefit ratio. At some time, the decision makers want to see results from the study and do not want the OR to be too time-consuming and cost-intensive. Time and work load may be a constraining factor – not for the OR-modeling itself, but for data acquisition needed to provide key information to run the model. So in practice and for various reasons, it is sometimes not possible to find the optimal solution. While it is still the goal for OR to find the optimal solution, in light of existing constraints, however, one will sometimes have to settle with a good suboptimal solution. After finding a solution, it is important to do further analyses (postoptimality analysis). In that kind of analysis the question is what would happen to the solution if different assumptions are made about future conditions. It is a process of experimenting with changes in the model by testing different values of the variables. Phase 4 The fourth phase consists of testing the whole model. In order to avoid mistakes, researchers sometimes vary the values of parameters and/or the decision variables in order to check whether the output from the model behaves in a plausible manner. The researchers try to find and correct as many errors as possible. This process is also called validation or model validation because it is the testing and improving of the model to increase its validity. It is important to document the process for future use of the model. 10 It is worth noting that none of the tested options may turn out to be the optimal solution. Instead, OR may find the optimal solution without actually having tested it. (For instance, if one has tested for a variable equal to 1, 3 and 5, OR may determine the optimal solution for that variable to be 2.4.) Phase 5 After testing and improving the model, the next step is to install a system for applying the model. This system usually includes the model, solution procedures and operating procedures. This is usually a computer-based system that consists of many different computer programs (e.g. databases that provide up-to-date data for the model or programs that automatically interpret the output of the model). The system for application can be called upon even when personnel may change. Once the system is well in place, variables can easily be adjusted whenever parameters may have changed. This way, an optimal solution (optimal efficiency/effectiveness) can be achieved even under altered environmental circumstances (i.e. changed parameters). Phase 6 In the final phase of the OR process the implementation of the system is carried out. It is important that the OR team participates in this phase to make sure that the whole system works. This phase depends greatly on the support of the management. Therefore, it is crucial to involve the management in the whole process of OR. The system may be used for years. Thus, it is crucial to continue to obtain feedback on how well the system is working. If necessary, modifications in the system can be made. In the whole OR process it is very important for the research team to document each step clearly for any future work. Concluding the description of the different phases of an OR study, it is important to state that it is impossible to write down any standard procedure that should always be followed by OR teams because OR requires innovation. Finally, I will conclude with my own definition of OR trying to incorporate all the aspects of OR that I mentioned above and trying to define OR in a short but precise way. It is an attempt to make it concrete and useable for the practical work: OR is the process of identifying and analyzing a problem within the operations (activities) of an organization or a program to be set up or already in place, and of finding an optimal solution for this problem by testing alternative ways of action (models) which will serve as a basis for decision making regarding successive modifications of key variables to achieve the overall goal of enhancing or even optimizing the efficiency and effectiveness of the program or organization. 11 3. Other Types of Research It is not easy to make distinctions between different forms of research and research methods. The literature shows that there are many different terms used for the same types of research and that there does not really exist a consistent terminology. Another problem in the distinction of different kinds of research derives from the mix-up of research designs and research methods. This causes the sometimes incorrect use of these terms and complicates a distinction. In the first part of this section, I will try to differentiate between the different types of research regarding their purpose. But we have to keep in mind that any research project can have more than one purpose. First we can make a distinction between Basic Research and Applied Research. Basic Research is designed to add to our fundamental understanding and knowledge of the social world regardless of practical or immediate implications, whereas Applied Research is intended to be useful in the immediate future and to suggest action, or to increase effectiveness in one area. Applied research is designed to provide information that is immediately useful to those participating in institutions or programs. This kind of research can be done for organizations and communities or with them. So also OR can be classified as Applied Research. Emily Stier Adler and Roger Clark stated that “a great deal of research can be seen as falling somewhere between the poles of basic and applied research” (Stier Adler & Clark 2011, p. 382). That statement shows the difficulty of distinguishing different forms of research from each other. Nevertheless, I assigned the following types of research to one of these forms regarding their purpose and intention. Applied Research Basic Research Descriptive Research Explanatory Research Exploratory Research Operations Research Action Research Evaluation Research 12 3.1. Basic Research 3.1.1. Exploratory Research Exploratory Research can be described as research on a relatively unstudied topic or in a new area. This kind of research uses an inductive approach. The researcher starts with observations about the subject of study and tries to develop tentative generalizations about it. Therefore, data is collected from only a few cases through observations or in-depth interviews. That qualitative data is analyzed and interpreted. During the research process, the researcher also tries to spot themes that emerged from the data for further research. Example: Managing motherhood from prison. Definition: Exploratory Research is research mostly inductive on an unstudied topic or in a new area through gathering and analyzing qualitative data that is gained from a few cases, with the overall goal to become familiar with this area, to develop some general ideas about it, to generate theoretical perspectives and statements on the topic and to spot themes for further research. 3.1.2. Descriptive Research Descriptive Research intends to describe groups, activities, situations or events with a focus on structure, attitudes or behavior. Typically, the researcher already has a general understanding of the topic before he/she starts to collect the data. The overall goal of this kind of research is to gain an accurate and precise picture of the topic under study. It describes how things are or how things have changed over time. Typically, the data is collected through surveys or content analysis and involves a large number of cases, which requires the use of quantitative data analysis. Example: Census. Definition: Descriptive Research is designed to describe a topic with a focus on structures, attitudes and behavior, about which the researcher typically already knows something before the data collection, which is done by gathering and analyzing data about a large number of cases using quantitative data analysis with the overall goal to gain an accurate and precise picture of the topic. 13 3.1.3. Explanatory Research Explanatory Research is conducted to explain why things are the way they are, and therefore looks for causes and reasons. This kind of research uses a deductive approach by moving from statements that are more general to less general. Thus, in explanatory research, pre-existing theories are often used to develop first ideas of a possible explanation for a question and then further data is collected to verify or modify the explanation. This kind of research typically requires a large number of cases so the data is collected using quantitative methods like surveys, and is then further analyzed. Example: Why are some people more likely than others to do work for pay after they retire from their regular lifetime work? Definition: Explanatory research is intended to explain variations in subjects by identifying causes and reasons for these variations using a deductive approach which implies consulting preexisting theories to identify the variables that influence the variations and to develop preliminary explanations which are then further proved by collecting data from a relatively large number of cases and using quantitative data analysis to finalize the explanations. 3.2. Applied Research 3.2.1. Evaluation Research Evaluation Research is a form of applied research because it is conducted to provide organizations with information that can be used in the immediate future. Evaluation research has the purpose of assessing the impacts of programs, policies and legal changes. Often it focuses on whether a program or policy has reached the intended outcomes. The results and implications of an evaluation research have to be communicated to the clients or decision-makers for further planning and continuance of the program. Within evaluation research, we can further find different types of evaluation research regarding their specific purpose and focus: Formative and Summative Evaluation, Outcome and Process Evaluation, Needs Assessment and Cost-Benefit Evaluation. Formative evaluation focuses on the design or early implementation stages of a program or policy. Outcome evaluation, which is also referred to as summative evaluation or impact evaluation, is designed to assess the effects or impacts of a program or policy and to determine whether the intended goals are accomplished. Needs assessment can be described as an analysis that determines the needs of a geographic area or a specific population. Cost-benefit evaluation is research that compares a program’s cost to its benefits. 14 Definition: Evaluation Research is a complex form of research that can have in itself many different purposes but its overall goal is to assess the impacts of programs, policies or legal changes and to determine whether they have achieved its intentions with the practical focus to communicate the results and their implications to clients and decision-makers for the future development of the program or policy. Formative Evaluation Process Evaluation Outcome Evaluation Evaluation Research CostBenefit Evaluation Needs Assessment 3.2.2. Action Research Action Research (synonyms: participatory research or participatory action research) has the focus to influence or change some aspect of the subject of research. Improvement as well as involvement is an essential part of action research. Central to this kind of research is the collaboration between researchers and those who are the focus of the research (e.g. community) and their participation in the process. The people in the organization or community under study participate actively with the professional researcher. They are involved in each step of the research process from the initial design to the final presentation of the results and the discussion of action implications. The action that comes from the research is intended to help the community or organization rather than the research community. Action Research has a social purpose component namely the empowerment of low status people in the organization or community to make decisions and take actions. Those who are normally the subject under study are actively engaged in the research process. The focus of action research is often an evaluation, so there are links between these two types of research. Therefore, some particular variants of evaluation have the same goals as action research (e.g. participatory evaluation, empowerment evaluation). 15 Definition: Action Research is intended to promote a change in an aspect of the subject of research and is characterized by the participation of the people who are directly affected by the research in the whole research process to empower them in decision-making and action taking. 3.3. Other aspects of research Based on which kind of purpose the research will cover, the researcher will decide which research design and research methods are appropriate. In short, research designs can range from non-experimental to true-experimental and can be cross-sectional, longitudinal or case study designs. Regarding the methods of research, one can in general differentiate between qualitative and quantitative methods. Quantitative research methods are typically methods that involve large numbers of cases such as surveys, questionnaires, standardized tests and attitude scales that need to be analyzed using statistical methods. In qualitative research, the researcher is more interested to really understand individual cases instead of questioning high numbers of people. Methods that are typically used in qualitative research involve individual or group interviews, focus group discussions and observations. Nevertheless, it is possible and reasonable to combine quantitative and qualitative research methods. 3.4. Differences between Evaluation Research and OR In considering each definition of the different types of research, I think it is easier to see the difference between OR and the first three types of research (exploratory, descriptive, explanatory research). Because evaluation research is often used in social and health-related fields I want to highlight this distinction again. Evaluation aims to measure the impact of a program or strategy. Generally, there is one solution/strategy implemented to overcome a problem and to reach a desired outcome. With an evaluation one wants to assess whether that solution/strategy was successful for achieving the desired outcome or not. In OR however, several solutions/strategies are tested and modified to find the optimal solution to overcome the problem. This optimal solution is then implemented to enhance the effectiveness and efficiency of the program. Another important distinction between Evaluation Research and OR is that typically an evaluation does not entail a control or comparison group, whereas OR ideally makes a comparison between a target group and a control group. In Evaluation Research a change of the program may or may not be one of the consequences of the research; in OR, however, the change or adaption of the program is part and parcel of the research. Evaluation Research aims to take a retrospective look at a program to assess its impact3; whereas OR is characterized by a prospective orientation and focus on action-oriented decision making and management (International Healthcare Consultants 2011). 3 However, when an evaluation contains both a prospective baseline survey and a retrospective follow-up survey, then such evaluation may, at least in part, be a prospective exercise. 16 4. OR in the Field of Social and Health-related Research In the field of social and health-related research, OR becomes more and more important. Programs and projects that are implemented to improve the health and nutrition situation, the educational situation and the social situation need to be effective and efficient because NGOs and other social services providers need to maintain and strengthen their legitimacy to operate. These programs always face complex tasks. Therefore, detailed planning at all levels, close coordination of program implementation efforts, careful training and supervision of personnel, and continuous evaluation of program development and impact is needed. OR is a way to support and inform these essential planning, coordinating, training, and evaluation functions. The Global Fund to Fight AIDS, Tuberculosis and Malaria, the Special Program for Research and Training in Tropical Diseases (TDR) and an inter-agency technical working group developed a Framework for Operations and Implementation Research in Health and Disease Control Programs. The framework was finalized within a meeting in April 2008 attended by over fifty participants representing the Global Fund, TDR, the World Health Organization, the Joint United Nations Program on HIV/AIDS (UNAIDS), United States Agency for International Development (USAID), the World Bank, field-based programs, policy-makers and research communities from all over the world (cf. Global Fund 2008, p.3). The participants in this meeting agreed on a rather broad definition of OR that is used in this framework: “Any research producing practically-usable knowledge (evidence, findings, information, etc) which can improve program implementation (e.g., effectiveness, efficiency, quality, access, scaleup, sustainability) regardless of the type of research (design, methodology, approach) falls within the boundaries of operations research.” (Global Fund 2008, p.12) The scope of this paper is to describe the characteristics of OR, its process and especially to distinguish it from other forms of research. Therefore, this definition is a bit problematic because it is a more universal definition that may apply to all forms of applied research as described in the previous chapter and which makes a distinction difficult. For this reason, I think other definitions of OR (e.g. from WHO or Fisher et al.) are more useable in the practical work. Also with my own definition I tried to develop a more specific definition: OR is the process of identifying and analyzing a problem within the operations (activities) of an organization or a program to be set up or already in place, and of finding an optimal solution for this problem by testing alternative ways of action (models) which will serve as a basis for decision making regarding successive modifications of key variables to achieve the overall goal of enhancing 17 or even optimizing the efficiency and effectiveness of the program or organization. In this part of the paper I will try to describe the characteristics and methods of OR in the social and health-related field. From my point of view, OR in these fields needs to be adjusted but still has to meet minimum standards to be classified as OR. 4.1. Characteristics The World Health Organization defines OR as “the use of systematic research techniques for program decision-making to achieve a specific outcome. It provides policy-makers and managers with evidence that they can use to improve program operations (WHO 2003, retrieved from Global Fund 2008, p. 49).” OR in this field focuses on specific problems within the operations or activities of the programs that are controlled by managers and administrators. Operations or activities in this field could be for training of personnel, logistics, public information or counseling.. Through OR, program problems can be identified and solutions can be developed through comparing different approaches against another to increase the efficiency, effectiveness, quality of services delivered by providers and the availability, accessibility, and acceptability of services desired by users (cf. Fisher & Foreit 2002, p.2). The practical objective of OR is to provide managers, administrators and policymakers with the information they need to apply the results of the OR to the programs. It helps them to design and implement new programs and to review, redirect and restructure programs that have already been in place (cf. Population Council 2002, retrieved from Global Fund 2008, p. 49). The goal should be to identify as early as possible any shortcomings of a program that may affect the expected outcomes (The International Food Policy Research Institute 2005, retrieved from Global Fund 2008, p.49). OR utilizes systematic data collection to accumulate evidence for supporting decision-making. But there is no single set of methods or designs that are unique for OR. Especially in the social and health-related field, the applied methods range from the qualitative to the quantitative, from the nonexperimental to the true experimental. It is not the application of a particular set of methods or the preference of one design over another that distinguishes OR from other forms of research; to a greater extent, it is the focus or objective of the research that distinguishes OR from other forms of research, namely to obtain a better understanding of the operations of a program so that needed improvements can be made to strengthen the program and maximize its effectiveness (cf. The International Food Policy Research Institute 2005, Global Fund 2008, p. 49). An OR study is only successful if the results are used to make program decisions. To conduct an OR study, collaboration between managers and researchers is necessary during the whole research process to identify the research problem, to develop the study design, to 18 implement the study and to analyze and interpret the results. Furthermore, an interdisciplinary research team is required that ideally should include professionals from diverse fields such as mathematics, statistics, economics, sociology, anthropology, management, information technology, health sciences and social sciences. 4.2. The OR Process In the following I shall describe the process of OR in health and development fields as follows:4 1. Problem Identification and Design of Experiment 2. Solution/Strategy Selection 3. Solution/Strategy Testing and Evaluation 4. Information Dissemination and Program Adaptation 4.2.1. Problem Identification and Design of Experiment The most important task for starting the OR process is to identify and clearly formulate the problem. That step is important because it leads to the statement of research objectives, to the hypotheses, to a definition of key variables and to a methodology for measuring the variables. It is crucial to analyze the problem in detail to define its scope, magnitude, seriousness, characteristics and probable causes. Therefore, the researchers first need to gather all information and data about the problem, which is already available. It can be useful to review relevant literature, examine current statistics, seek educated opinions from persons concerned about the problem and obtain probable reasons for the problem from theory. A good diagnosis of the problem will help to establish its parameters and variables. In this first step of the OR process, it is important to involve program managers but also other people such as village chiefs, teachers and health personnel. This ensures that the researchers really understand the problem as a whole. Another benefit of early involvement 4 Fisher & Foreit (2002) distinguish the last phase in information dissemination and information utilization. 19 of all key stakeholders in the OR process is that it increases their interest later in reviewing and using the results. Once the problem has been identified, the researchers and decision makers need to determine the desired objectives, which should be achieved through implementing one of the alternative solutions. Only after accomplishing the objectives one can search for various possible alternative solutions. Following the definition of the problem and the determination of the desired objectives, it is useful, if not mandatory, to perform a Design of Experiment. In fact, determining the desired objectives may also be considered part of such a design. A Design of Experiment is needed for testing hypothetical assumptions. With their experiments, researchers will try to establish a causal relationship between variables. A well-designed experiment will be robust under questioning and will yield good statistical data. A good Design of Experiment is needed because experiments can be costly and may not be repeated just because the experiment did not yield good statistical data or because a control group was not really useful for comparison. Building a design means carefully choosing a small number of experiments that are to be performed under controlled conditions. A Design of Experiment will usually involve three interrelated steps Define the objective of the experiment Describe the desired outcome variables (response variables) Define the variables to be controlled during the experiment (design variables) The experiments designed in such a way may quite often have to be carried out first before a mathematical model is developed. But to actually do the right experiments and the right amount of experiments, a Design of Experiment is advised to be done at the beginning. 4.2.2. Solution/Strategy Selection Developing a Model In the first step of this phase a model of the problem needs to be constructed that represents the problem and the expected outcome. The model should describe the relationships between different variables that are considered relevant to the problem. It is expressed in terms of objectives and hypotheses about the problem. If it is possible, the model should be a mathematical model consisting of equations and mathematical operations. That can make it easier to test different solutions to the problem. It is important to define the key variables of the model because that is a prerequisite for the measurement of the variables and gives meaning to them. It is also necessary because otherwise the variables might be interpreted in different ways. Additionally, it is necessary to clearly define the terms of the hypothesis (e.g. meaning of terms like more than, significant, increased) that describe the relationships between the variables. Therefore, a clear terminology is crucial. 20 Alternative Solutions Another step in this phase is the development of solutions to overcome the problem. It is crucial to first identify the problem and develop a model of it to select the solutions on this basis. It can be useful to review existing theories or strategies that other people have already used to solve a similar problem. It can be also useful to consult with people who are affected by the problem because they often have good ideas for strategies to solve it. At this point, it is also important to look for strategies that are possible and manageable for the organization to implement in terms of facilities, training, personnel, time, finance, logistics and management, and that can be sustained over time. The strategy needs also be acceptable for the key stakeholders and the community, and needs to be in line with the organization’s goals and objectives. When an approach is identified that may useful to overcome the problem and to reach the desired goals, the research team has to think about alternative solutions, which are in line with this overall approach. So the alternative solutions differ from each other in terms of changes in the variables such as the number of visits to pregnant women by Community Health Workers or the interval of training for Community Health Workers or both. 4.2.3. Solution/Strategy Testing and Evaluation In this phase of the OR study the alternative solutions need to be tested to find the optimal solution to overcome the problem and achieve the desired outcome. Before testing the solutions the researchers have to think about how to measure the outcome of each alternative solution to later make a decision which solution is the optimal one. If the problem can be represented by a mathematical model, mathematical methods could be used to find the optimal solution as described in chapter 2 to measure the outcome. However when OR is conducted in the social and health-related field, it might not always be possible to develop a mathematical model to test alternative solutions. Therefore, it may be necessary to immediately apply the selected alternative solutions and test them directly in the field. In this case, it might be more difficult to measure the outcome such as the effectiveness of a primary health care program. The researchers have to determine criteria that show the effectiveness of the program. These criteria can be changes in attitude, behavior and knowledge. After defining criteria, measures need to be identified to determine the effectiveness. For example, community health workers are appointed to pass on knowledge about HIV/AIDS transmission to pregnant women. That can be measured through a questionnaire or an interview during which questions regarding HIV/AIDS transmission are posed. The number of right answers will in this case demonstrate the knowledge acquired (8 or more = High knowledge; 4-7 = Medium knowledge; 1-3 = Low knowledge; No correct answers = No knowledge). In the case that there are so many alternative solutions that it is not possible to test them all directly in a field situation the researchers may need to go another way. They can estimate the 21 outcomes of each alternative using previous experience from the implementation of similar strategies. On the basis of this knowledge they can rank the alternatives related to the desired outcome(s). After this ranking, they can conduct a field test for example of the three best alternatives to test them in real situations. At the end of the testing, the alternative solutions need to be compared to identify the most optimal solution in terms of achieving the desired outcome(s). However, one cannot expect that one of the tested solutions will turn out to be the optimal solution. Rather, it is the model which, given the results of the tests, will calculate the optimal solution. The last step in this phase should be the sensitivity analysis to assess the reliability of the findings. As the researchers analyzed the problem and developed alternative solutions to overcome it, they made assumptions about uncertain quantities. For example they have to estimate certain costs (e.g. costs for personnel or items) and outcomes (e.g. births, deaths or dropouts). Before making recommendations to the decision makers, they should test whether variations in those assumptions would change the solution they would recommend to the decision makers. If this analysis indicates that some variables are sensitive to changes in the assumptions, the researchers should discuss that topic with the decision makers in the next phase. During the third phase, it is also important to consider ethical issues. This is especially true when it is not possible to test alternative solutions through mathematical methods. In that case, it is necessary to test the solutions in the field so that the most effective in terms of overcoming the problem and reaching the objective can be identified. We have to keep in mind that when testing different strategies, some target groups may not benefit from optimal services. On the other hand, only through testing is it possible to identify the most effective strategy, which can then be implemented on a grander scale. 4.2.4. Information Dissemination and Program Adaptation Because OR is primarily conducted to inform program or policy decision makers about the changes needed to optimize impact, it is crucial to present the findings to them. It can even be very helpful to have frequent small meetings with key managers and stakeholders throughout the research process to keep them informed. The researchers also have to think about an appropriate way to present the findings to the decision makers. Finally, once the tests have been carried out and the model has calculated the best solution under the design of the experiments, the decision makers will need to take the necessary steps to implement that solution. The solution to be implemented may not be the best solution under any circumstances but only the best solution given certain constraints. Such a constraint could be a budgetary constraint which would set upper limits to certain experiments and, hence, the solution to be found. But these constraints should be considered before the experiments are designed. Once the 22 experiments have been carried out, a number of variables have been tested, and an optimal solution has been calculated, decision makers should implement that solution. It may be noted that the process by which the optimal solution is put into operation may be considered change management; in exact parlance, it is no longer part of OR per se; however, since OR is to be carried out not for academic reasons but for the purpose of actually achieving an optimization of impact, that process certainly must be a necessary consequence of OR and has therefore been considered part of OR. 4.3. An Example of Social OR An example of social OR is the “Kenya Adolescent Reproductive Health Project”5 that was conducted from 1999 until 2002 by the “Frontiers in Reproductive Helath Population Council” in collaboration with the “Program for Appropriate Technology in Health” (PATH) and the Government of Kenya, with funding support from USAID. This project is a multi-sectoral approach to providing reproductive health information and services to young people in Western Kenya. Background Informing adolescents about appropriate and acceptable behaviors, and ways to protect themselves against unwanted and unprotected sex has proved problematic in Kenya. Education programs for in- and out-of-school adolescents are lacking, and there is controversy around providing services to sexually active adolescents and a pervasive concern that sexuality education and contraceptive services may lead to promiscuity. In response to this situation the Population Council and the three Kenyan ministries planned and designed this OR project. 1. Problem Identification The researchers started with a diagnostic phase to understand adolescents’ reproductive health issues. Therefore, they first reviewed existing statistics that showed that many Kenyan youth experience reproductive health problems. After that, the researchers tried to understand adolescents’ reproductive health issues in the socio-cultural context of the two districts where the study was to be conducted through: Securing clearance/approval from community gatekeepers Securing gatekeepers’ support and commitment Creating an inventory of existing health facilities and providers, schools, adult, and youth networks, and agencies providing health and youth services in the intervention sites 5 A Multi-Sectoral Approach to Providing Reproductive Health Information and Services to Young People in Western Kenya: Kenya Adolescent Reproductive Health Project. 23 Conducting interviews and discussions with community target groups, including religious leaders, teachers, health providers, parents, and community leaders to assess their perceptions of and attitude towards adolescents reproductive health issues and to get their advice on what should be emphasized in the intervention Conducting interviews with service providers, especially private service providers, to assess their ability and willingness to participate in the study and any constraints that may hinder their successful participation. Interviewing youth leaders, teachers and health providers to identify information, education and counseling material needs and reproductive health training needs for various categories of health and educational service providers. Assessing existing adolescents’ (10-19 years) utilization of reproductive health services through a review of clinical and community-based providers’ service data in both control and intervention sites. Summing it up, the results from the diagnostic phase were that many Kenyan youth face reproductive health problems such as: unwanted pregnancy unsafe abortion pregnancy-related school drop out high age-specific fertility rates sexually-transmitted diseases, including HIV/AIDS Additionally, the researchers identified a number of factors that caused these problems: early onset of sexual activity low use of contraceptives and STD/HIV/AIDS preventive measures lack of correct and reliable sources of reproductive health information which leads to poor reproductive health knowledge poor access to convenient, confidential and low cost reproductive health services lack of effective prevention programs for youth in and out-of-school the wider environment such as religious and community leaders is not supportive of youth acquiring reproductive health knowledge and services. 2. Solution/Strategy selection Resulting from the problem identification and description of causes the researchers defined the overall objective of the OR project as to determine the feasibility, effectiveness and cost of a systematic intervention to foster a supportive environment that will address reproductive health problems of both in-school and out-of-school adolescents (10-19 years) by making existing services more accessible to adolescents and youth, providing them with reproductive health information and skills. The initiators of this study wanted to obtain sufficient information to guide the future policies of the Ministries regarding provision of youth reproductive health education and services. 24 Resulting from the reproductive health problems Kenyan youth face and their causes, the researchers identified three specific goals that needed to be addressed by possible interventions: Improve knowledge about reproductive health and encourage a responsible and healthy attitude towards sexuality among adolescents Delay the onset of sexual activity among younger adolescents Decrease risky behaviors among sexually active adolescents To reach these goals a supportive environment at all levels was identified as being crucial. The researchers hypothesized that there are three fields that are important in which to carry out the interventions: the environment, the health services and the school. In detail, the hypotheses are as follows: Environment a) Providing adolescents with reproductive health knowledge and skills at the community will increase reproductive health knowledge, preventive behaviors and health seeking behavior among both married and unmarried male and female youth (10-19 years). Health Services a) Training health care providers on provision of youth friendly services will improve providers’ attitudes towards provision of reproductive health services to sexually active adolescents and the quality of reproductive health services they provide to the youth. b) Strengthening the ability of health care providers to provide youth friendly reproductive health services will increase married and unmarried male and female adolescents (10-19 years) utilization of reproductive health services. School Providing adolescents with reproductive health knowledge and skills at school will increase reproductive health knowledge, preventive behaviors and health seeking behavior among both male and female youth (10-19 years). The researchers expected that advocacy in support of adolescent access to reproductive health information, counseling and services through sensitization workshops and public meetings targeting opinion leaders and parents will increase community support for provision of adolescents’ reproductive health services in the intervention sites. Therefore, the following interventions (solutions) to the problem were designed with regard to all of the three fields: Environment: Sensitization, advocacy and community mobilization targeting gatekeepers and opinion leaders On-going community campaigns on reproductive health issues Training of peer educators and adult helpers 25 Health Services: Training health providers Improving the youth-friendliness of facilities School: Training teachers, supervisors/inspectors, parents and students Forming reproductive health clubs/ integrating reproductive health into existing clubs Sensitizing parents and teachers Establishing links between schools and health facilities 3. Solution Testing and Evaluation The field study was conducted in two districts in Kenya. The researchers used a quasiexperimental design with two experimental strategies, pre-intervention and post-intervention measurements to test the effectiveness of the interventions. To control for cultural and other sociodemographic diversity, the experimental and control sites would be expected to be similar with respect to: Socio-cultural characteristics Rural/urban characteristics Accessibility to health care services Population density In each district there were three study sites (A, B, C) to which the community was randomly assigned. Each of the three sites got different interventions to determine the effectiveness of each intervention. Design Site A Site B Site C (Intervention site) (Intervention site) (Control site) Environment Intervention Yes Yes No Youth Friendly Health Services based Intervention Yes Yes No School-based Intervention No Yes No To measure the changes in the key variables and thus the effectiveness of each intervention, a baseline and end-line survey was conducted in all three sites of each district. To measure the key variables indicators were determined: 26 Variables Measurement Indicators Feasibility: The ability of community opinion leaders, parents, school and health care officials to provide support to the project and participate in its implementation. Focus group discussions (FGDs), key informant interviews and informal discussions/meetings with community leaders before and after the onset of the interventions. Reports of outcome of scheduled sensitization and advocacy workshops/ meetings and periodic monitoring reports. The cost analysis will determine the incremental cost of the components of the interventions to the MOH and MOE and will include startup costs for training and materials development and recurrent costs to carry out day-to-day operations. This will be measured through a comparison of a pre-intervention and post-intervention survey of married and unmarried adolescents (age 10-19 years), assessing RH knowledge, attitudes, behavior and utilization of services in the two intervention sites and one control site. This will be measured through in-depth interviews and FGDs with community leaders, parents, health providers and youth. Cost: What it costs to build the capacity and run the new RH programs in schools, with health care providers and with community-networks to provide adolescents RH information and services. Effectiveness: The effect of the interventions on RH of adolescents. Environment: Refers to attitudinal aspects of the community that either facilitate or hinder the successful provision of RH services to young people. This includes attitudes of community leaders, parents, health providers and youth towards providing RH information and making RH services to married and unmarried adolescents aged 10-19 years more accessible. Service Utilization: The proportion of sexually active adolescents who report use of RH services. Youth-friendliness: Youth friendliness will be defined as the ability of service providers to treat youth seeking RH information and services with respect, ensure privacy and confidentiality of the consultation and provide the required service in a competent, nonthreatening and non-judgmental manner. RH knowledge: This includes being able to correctly describe the fertile period, body changes at puberty, various STIs and symptoms and signs of STDs and HIV/AIDS, ways of pregnancy and STD/HIV infection prevention, correct condom use, sources of various RH services in the community, and consequences of early (adolescent) pregnancy and child birth as well as unsafe abortion. . RH attitudes: These include opinions about ideal family size, use of modern contraceptive methods and condoms among youth and adult population, sex education for youth, youth engagement in sexual intercourse, and provision of RH services (maternal and child health, family planning, post-abortion care, STD diagnosis and treatment and emergency contraception) to married and unmarried adolescents. Service statistics on number of married and unmarried adolescents provided with RH services will be used to measure utilization of services. This information will be gathered before and at the end of the intervention through a review of service statistics and interviews with adolescents. Youth friendliness will be assessed through indepth interviews and FGDs with adolescents, use of mystery clients to collect data from providers, and pre-intervention and post-intervention surveys with unmarried and married adolescents 10-19 years. Knowledge will be measured through preintervention and post-intervention surveys focusing on unmarried and married adolescents aged 10-19 years These attitudes will be assessed through FGDs and in-depth interviews with parents, community leaders, adolescents and health care providers. In addition, parents and youth’s attitudes will be assessed through the pre-intervention and postintervention survey of married and unmarried adolescents aged 10-19 years. 27 RH behavior: The kinds of behavior we will be focusing on include reported age at initiation of sexual activity; sexual activity within the past six months; number of sexual partners; consistent use of condoms and use of modern contraceptive methods for the sexually active; obtaining diagnosis and treatment from recognized health practitioners for those who report having contacted STDs; accessing services for prenatal and postnatal care for those who have been/are pregnant and/or with child; and accessing emergency contraception and post-abortion care for the girls requiring such services. This information will be assessed through the preintervention and post-intervention survey of married and unmarried adolescents aged 10-19 years. Gender attitudes: These include opinions about equal education and work opportunities, family size and family planning decision-making, physical and sexual abuse and performance of household tasks, regarding males and females. These attitudes will be assessed through FGDs and interviews with youth as well as through the preintervention and post-intervention survey. 4. Information Dissemination/Stakeholder Involvement Research findings and lessons learnt were disseminated at various levels, nationally and locally and at different times during project implementation. Key audiences for district and provincial level dissemination were identified during the diagnostic phase. At the national and international level, an update of the project’s achievements, constraints, lessons learned and research findings were disseminated in project update briefs. These were mailed periodically to policy-makers in the involved Ministries and to program managers in other organizations that are implementing youth programs, donor agencies and other interest groups. Midway into the project, stakeholders and other key audiences were contacted to provide a progress report and information on any preliminary findings and to get feedback on the research interventions. On completion of the research, the final report was published and distributed to both technical and non-technical audiences. Also a summary was prepared for policy makers in the Ministries and donor agencies. National dissemination activities included a national seminar, use of the media, publications in professional journals and presentations to professional and interest groups such as educators, health providers and youth advocates. Planning for sustainability and scale up had already begun while the country project was being designed. The Population Council had begun seeking partners who were able to support scale up efforts. Scale up was also enhanced by having an advisory committee of influential people and involved throughout the project. The members of the advisory committee played an advocacy role to influence government and other youth serving organizations to scale up or replicate the project if it is successful. 28 5. Final Remarks OR is the process of identifying and analyzing a problem within the operations (activities) of a program or an organization, and of finding an optimal solution for this problem by testing alternative ways of action (models) which will serve as a basis for decision making regarding successive modifications of key variables to achieve the overall goal of enhancing or even optimizing the effectiveness and efficiency of the program or organization. Starting from this definition, a research study that is classified as OR needs to meet all conditions that are mentioned in this definition. So first a problem within the operations of a program or of the whole organization must be identified. Then, the problem needs to be analyzed. For this purpose, all data relevant to this problem need to be gathered and analyzed to identify the key parameters and variables influencing the problem. This is necessary to gain a deeper understanding of the whole problem and to develop solutions for overcoming it. Next, different alternative solutions are tested in terms of the desired outcome. The results need to be presented to decision makers who will then decide which solution should be implemented or which modifications to the program should be made to overcome the problem and to enhance the effectiveness and efficiency of the program as the overall goal. Finally, I will mention the main characteristics that distinguish OR from other forms of research according to the definition of OR from the WHO that is included in the Framework for Operations and Implementation Research in Health and Disease Control Programs (Global Fund 2008, p.49): It addresses specific problems within specific programs, not general health issues. It addresses those problems that are under control of managers, such as program systems, training, pricing and provision of information. It utilizes systematic data collection procedures, both qualitative and quantitative to accumulate evidence supporting decision-making. It requires collaboration between managers and researchers in identification of the research problem, development of the study design, implementation of the study and analysis and interpretation of results. It succeeds only if the study results are used to make program decisions; publication alone is not a valid indicator of successful OR. 29 By restating these criteria at the end I want to highlight the main characteristics of OR again to make the distinction to other forms of research clear. If a study does not consist of all the components that are necessary for conducting an OR, it has to be clearly stated that the study only contains some aspects of an OR study. It should be the goal within World Vision’s research to conduct OR to identify program problems and to find a solution to overcome them because that will enhance the effectiveness and efficiency of the programs. For the funding component of such OR within World Vision, this means that the OR project should set aside funds to (i) conduct a problem analysis and alternative solutions/develop a model (ii) involve researchers, program managers and policy makers in the design of the OR project (iii) support the implementation of the identified solution(s) or model(s) (iv) support the implementation of the research component (v) support participation of researchers, program managers and policy makers in interpreting and disseminating the results and planning for utilization of the results. In the context of Child Well-Being Outcome (CWBO) measurements, Area Development Programme (ADP) and non-ADP baseline and follow-up surveys, categorized as program evaluation research, will present evidence on contribution that WV is making in improving CWBOs. The gap is in identifying which of the varying strategies/ models that can be applied to meet specified CWBOs are more effective/ cost-effective. OR projects relying on prospective control or comparison group design offer the best evidence to attribute change to particular intervention models. Research that does not contribute to improving the impact of World Vision’s programs may be fine for the sake of research; but the goal of research within World Vision’s programs should be to provide information for enhancing the effectiveness and efficiency. So ideally OR should be the method of choice. 30 Appendix 1 Operations Research Checklist 1. Problem Identification and Diagnosis Yes No Is the problem identified, analyzed and clearly formulated? Is all relevant information and data gathered? Are the program managers involved? Are the objectives determined that should be achieved through implementing a solution? 2. Solution/Strategy Selection Has a model that represents the problem been developed? Are the key variables of the model defined? Are multiple solutions developed to overcome the problem? Are the solutions manageable for the organization to implement? Has a Design of Experiment been drawn up? 31 3. Solution/Strategy Testing and Evaluation Are criteria defined to measure the outcome of each alternative? Are ethical issues regarding the testing considered? Are the alternative solutions tested? Are the alternative solutions compared regarding the desired outcome(s)? Is further analysis conducted to assess the reliability of the findings? Is the optimal solution identified? 4. Information Dissemination and Program Adaptation Are the results of the study presented to the decision makers? Have the decision makers decided on a solution/strategy to implement? Is the solution/strategy implemented or are adaptations to the program made? 32 Appendix II Note: The following example was given at the end of the text of an earlier version of this paper; it has been criticized for not truly representing OR. Hence, the example was put into an appendix where it is followed by Discussion/Critique at the end. The Ministry of Health has implemented a primary health care program. The overall goal of this program is to reduce child mortality and morbidity. The program has stalled because a large number of the community health workers that are appointed to visit the population have not been trained even though the Ministry of Health developed a systematic plan of training. Through OR the most cost-effective way of training should be found out. First, the researchers analyze the problem in collaboration with the Ministry of Health. They assess that the system does not work in terms of quality of the training and the number of trained Community Health Workers. From other experience the Ministry of Health and the researchers find out that an approach that uses a core staff of trainers and audiovisual materials has proved to be very successful. Now they want to find the most cost-effective way to adapt this method. Although they have decided on a general training approach, the actual structure of that approach has not been determined. Therefore they have to identify the variables of the strategy that include: 1. Training content 2. 3. 4. 5. Training methods Training materials Trainers Duration of Training 6. Class size 7. Training location 8. Level of supervision following training The researchers decide to standardize the variables 1, 2 and 3 and they determine that the training can be completed in one day (variable 5) based on past experience. For budgetary reasons they decide that the training will be held in district offices. So three variables are left. Next, the researchers determine the options for the remaining variables. Trainers: regional staff or district staff (variable 4) Class size: 20 or 40 promoters per class (variable 6) Supervision: monthly or semi-annual supervisory visits (variable 8) When combining the different variations of these variables there are 16 alternatives. These alternatives are tested in terms of the defined outcomes namely cost, effectiveness and cost33 effectiveness ratio. The economist on the research team identifies the relevant costs and collects the data needed to test all alternatives. This can be done through mathematical methods. But also the effectiveness of the alternatives needs to be measured. Although the ultimate goal of the program is to reduce child mortality and morbidity the research team determines that it is impossible to measure any significant change in the short term of the research. Therefore, the training of the Community Health Workers who provide the needed service to reduce child mortality and morbidity is selected as effectiveness measure. The team selects knowledge, skills and behavior of the promoters and the knowledge, skills and behavior of parents as the components for measuring the effectiveness. They also define these terms, so for example the knowledge of promoters is determined by their understanding of symptoms of malaria and their ability to administer tablets and so on. Most of these data will be collected and analyzed during the field test of the alternative. Before that field test, the research team will have to rely on two measures to determine the most-effective alternative, which will then be tested: the estimated percentages of all promoters that will have to be trained and the estimated percentage of the trainees that will pass a knowledge and skills test at the end of training. Estimates for these two measures can be obtained from experience with current training programs and from the recent experience with the new training method among census workers. Thus, the team makes estimates on the percent of all promoters that will be trained and the estimated percentage of the trainees that will pass a knowledge and skills test at the end of training. These estimates can be obtained from experience. No estimates can be made on the effects training will have on promoters’ behavior or on parents. Immediate output : Number of promoters expected to complete training Total number of promoters Immediate effects: Number of promoters expected to pass a knowledge skill test Total numbers of promoters The results will lead to a ranking of the 16 alternatives. In the next step the researchers will identify the least costly alternative and the most effective alternative. If this examination would not lead to the same alternative, the research team will calculate the cost-effectiveness ratios for each alternative in terms of promoters expected to be trained and promoters expected to pass the test. So for each alternative there are two calculations. Based on these calculations the research team recommends the most cost-effective alternative. This alternative may not be the most effective but including the costs it is the best alternative that considers both costs and effectiveness. Because the research team made assumptions about costs and effectiveness they will conduct a sensitivity analysis by varying their assumptions. They are confident about the estimated costs and do not think that changes in their assumptions would influence the ranking of the alternatives. But 34 they are not sure about their assumptions related to the effectiveness of the alternatives. They discuss this topic with the minister and decide that a new alternative should be examined. Also for this alternative they calculate the costs, expected effectiveness and the cost-effectiveness ratio and present the data to the Minister. Finally, they decide to test the new alternative and the alternative, which was previously determined as the most cost-effective alternative. After the field test the research team will calculate the actual cost, effectiveness and cost-effectiveness and the most cost-effective alternative will be implemented. Discussion/Critique: 1. I am not sure where the number 16 came from, as there should only be 8 experiments from the presentation. 2. The authors select several variables in the beginning. As presented in the example section, the variables are trainers (x), class size (y) and supervision (z). We have two measures of result, namely cost (c) and effectiveness (e). We thus desire a mathematical model (c,e) = f(x,y,z). For experimentation, each variable takes on two different values. This would result in 8 different experiments as each combination must be tested. 3. However, we may assume that the three variables are independent of each other. Thus we do not need to engage in all 8 experiments. A design of experiment would have produced the following experiments: X = regional, y = 20, z = semi-annual. X = district, y = 20, z = semi-annual. X = regional, y = 40, z = semi-annual. X = regional, y = 20, z = monthly. 4. These 4 experiments are sufficient in the presence of independence. As each dimension has only two values (both in this design of experiment as well as in the original design), we can only model the system linearly. Thus the function takes the form (with the a_i as parameters): C = a_1 x + a_2 y + a_3 z + a_4 E = a_5 x + a_6 y + a_7 z + a_8 35 5. As we have 4 experiments and two outcome measurements per experiment, we have 8 experimental facts. We also have 8 unknowns in our model and so everything is welldetermined. 6. The example now claims that the experimenters chose the best of their 16 experiments. This is not generally correct. We have used the experiments to get the parameters of the model above. Now we wish to compute the optimal point. As the outcome measures C and E interact and are mostly probably constrained to each other (effectiveness and cost are related to each other), the optimal solution is a negotiation between these two dimensions. The constraints must be formulated. 7. One way out of the conflict between two goals is to determine which is more important and to introduce weights for each. This is a management decision and should have been made in the beginning of the project. First, we must measure both C and E in comparable units and then introduce a relative weight. Let us assume that we have common units already. Second, we then combine them into a single goal function Goal = w_1 c + w_2 e However, we have to consider the constraint that C and E are not independent but rather constrained to each other and perhaps each constrained in their own right. There may, for instance, be a maximal budget. The goal must now be maximized by finding the right values of x, y and z such that goal is maximal subject to the constraints. This is a problem for optimization theory. 8. It is very likely indeed that the final position of the optimum is not one of the experimental support points tested in the experiment. For instance it may be that the optimal answer has X = regional, y = 30, z = quarterly. This depends upon the outcome of the experiments. 9. Thus, we must consider the elements of design-of-experiment, constraints and multiobjective optimization theory in our efforts to locate the true optimal point. Before engaging in an experiment, we should formulate the constraints as part of the problem definition and then perform a design-of-experiment. Only then do we have the tools to perform the (and only the) experiments needed and to be able to do the computations leading to the optimum. (Author of Discussion/Critique: Patrick D. Bangert, PhD) 36 Annotated Bibliography Non World Vision publications Hillier, F. S. & Lieberman, G. J. (2010). Introduction to OR. New York: MC Graw Hill Education. For over four decades, this is the standard work on Operations Research. It includes a detailed description of the fundamentals and the process of Operations Research and of the mathematical models that are used in Operations Research. Population Councils Frontiers in Reproductive Health Program, PATH & Government of Kenya (2004). A Multi-Sectoral Approach to Providing Reproductive Health Information and Services to Young People in Western Kenya: Kenya Adolescent Reproductive Health Project. This is a project proposal and summary of an example for social OR that was conducted by the Population Councils’ Frontiers in Health Program in collaboration with PATH and the Government of Kenya. Robson, Colin (2002). Real World Research.2nd Edition. United Kingdom: Blackwell Publishing. A book on research approaches outside the laboratory in “real world” situations relevant for behavioral and social science disciplines. Stier Adler, E. & Clark, R. (2011). An Invitation to Social Research. Belmont, CA: Wadsworth, Cengage Learning. This is a book about the methods of social research. Describing the research process in all its details this book provides guidance through the whole process of conducting social research. USAID, Bureau for Global Health Office of Health, Infectious Disease, and Nutrition (2010) Guidance for Detailed Implementation Plans (DIPs) Innovation & Tuberculosis Category Child Survival and Health Grants These are guidelines for project implementation and management with a detailed guidance for a OR concept paper. 37 Online Sources Fisher, A. A. & Foreit, J. R. (2002). Designing HIV/AIDS Intervention Studies: An OR Handbook. New York: The Population Council. Retrieved April 20, 2011 from http://www.popcouncil.org/pdfs/horizons/orhivaidshndbk.pdf This is a Handbook on Operations Research especially for conducting it in health programs related to HIV/AIDS. It provides guidance through the whole research process through a detailed description of each step. Gesellschaft für OR (GOR) e.V. GOR: OR. Retrieved April 20, 2011 from https://gor.unipaderborn.de/ This is the website of the German Society for Operations Research. It contains some basic information on Operations Research. International Healthcare Consultants. inhcc. Retrieved May 10, 2011 from http://www.inhcc.com/Operations_Research/operations_research.htm International Healthcare Consultants is a Healthcare Information System Software Developer and International Healthcare Management Consulting Company that offers cost-effective and practical integrated solutions to a wide range of healthcare professionals and organizations. Especially the information on a Primary Health Care Operations Research project is useful. Reynolds, J. & Gaspari, K.C. (1985). Operations Research Methods: Cost-Effectiveness Analysis. PRICOR – Primary Health Care Operations Research. Chevy Chase, Maryland: Center for Human Services. Retrieved May 23, 2011 from http://pdf.usaid.gov/pdf_docs/PNAAX968.pdf This document contains a description of the process of conducting a cost-effectiveness Operations Research that is supplemented by an example of a cost-effectiveness Operations Research in the field of primary health care. Stat-Ease – statistics made easy. DOE for Six Sigma/Lean. Retrieved June 20, 2011 from http://www.statease.com/sixsigma.html (Figure 1) 38 The Association of European Operational Research Societies. (2011). Operational Research: What is Operational Research? Retrieved April 20, 2011 from http://www.euro-online.org/web/pages/197/whatis-or This website of the Association of European Operational Research Societies provides general information on Operations Research and the European Societies that are engaged in Operations Research. The European Working group on Operational Research Applied to Health Services. ORAHS: What is Operational Research. Retrieved April 20, 2011 from http://orahs.di.unito.it/ This working group is specialized on Operations Research applied to health services and is intended for practitioners in this field to share their ideas and experiences. The website contains basis information on Operations Research conducted to problems in this field. The Global Fund. Framework for Operations and Implementation Research in Health and Disease Control Programs. Retrieved May 10, 2011 from http://www.theglobalfund.org/documents/me/FrameworkForOperationsResearch.pdf This is a framework for Operations Research in Health and Disease Control Programs and was developed by the Global Fund to Fight AIDS, Tuberculosis and Malaria, the Special Program for Research and Training in Tropical Diseases (TDR) and an inter-agency technical working group. World Vision Sources Chege, J. & Ghee, A. (2010). Draft Outline. A spectrum of inquiry: The role of operations research in World Vision’s work. Retrieved April 15, 2011 from https://www.wvcentral.org/cop/hnh/researchdme/Documents/Forms/AllItems.aspx?RootFolder=%2fcop%2fhnh %2fresearchdme%2fDocuments%2fOperations%20Research%2fPreparation%20for%20July%204%2d6%20e vent&FolderCTID=0x012000D8B5EEF256DDB84B9B3CF120A0B1A12D00B2F174B14066DD47B1E63A 961E004E39&View=%7b56EEBC5B%2dD5E0%2d4356%2dA79C%2dC500FE3E08C7%7d This is a first draft on Operations Research that describes the characteristics of Operations Research and the steps that need to be taken while conducting it. 39 Initiatives Inc. for The Healthcare Improvement Project (2010). Operations Research: Influence of CHW Program Functionality Strengthening on CHW Engagement and Performance – Research Protocol. Retrieved May 12, 2011from https://www.wvcentral.org/cop/hnh/researchdme/Documents/Forms/AllItems.aspx?RootFolder=%2fc op%2fhnh%2fresearchdme%2fDocuments%2fOperations%20Research%2fPreparation%20for%20July% 204%2d6%20event&FolderCTID=0x012000D8B5EEF256DDB84B9B3CF120A0B1A12D00B2F174B14 066DD47B1E63A961E004E39&View=%7b56EEBC5B%2dD5E0%2d4356%2dA79C%2dC500FE3E08C7 %7d This is a document from Initiatives Inc. for the Healthcare Improvement Project that offers a framework for the assessment of Community Health Workers (CHW) programs and for the development of appropriate interventions to program problems. Carboni, I. (2011). Guiding Principles for National Offices – Measuring and Reporting the Child Well-being Targets. Retrieved May 12, 2011 from https://www.wvcentral.org/cwb/Documents/Child%20Wellbeing%20Partnership%20Targets/Principles%20for%20Measuring%20and%20Reporting%20CWB%20T argets%20FINAL%20May%202011.pdf This document outlines guiding principles for national offices to consider before measuring and reporting on the Child Well-being targets and includes information about other forms of research such as impact evaluations. Monbiot, E. (2006). Agreed DRAFT standards for research. This is document arose from a meeting held in 2006 and defines the purposes of research within World Vision and describing the steps that need to be taken while conducting research. 40
© Copyright 2026 Paperzz