Attempts to apply evaluation methods to the overall assessment of health policy are subject to numerous limitations, which, if ignored, can lead to a complete undermining of the sense of making such efforts. Certainly, the problem that is most difficult to ignore is the fact that for health policy interpreted as a public policy sector it is difficult to indicate the objectives for which it is implemented. Referring to the previously quoted understanding of health policy as a public policy in Dye's interpretation, it can be seen that establishing its purpose would have to be arbitrary. Therefore, the evaluator would be forced to infer about the general purpose of health policy on the basis of the general direction of the intentions declared by public authorities, expressed in individual program documents, strategies, long-term plans etc. created in this way may prevent reproduction of the overall vision of health policy. Here, however, the question of intentional or unintentional selectivity in the actions of the state arises again, which means that the image of goals created in this way may make it impossible to recreate the overall vision of health policy. You can also try to evaluate based on a reference to externally set norms or standards (e.g. WHO guidelines or findings based on theoretical models), but in this case you should be aware that the actors directly involved in the design and implementation of health policy they do not have to treat these goals as values, and thus do not have to show their will to achieve them. Nevertheless, it seems that greater cognitive value is obtained by adopting the second of the proposed approaches, with notion that it does not have to (and often will not) entail practical consequences in the form of real impact on the shape of implemented policies.

Another caveat that should be made results from the fact that, while individual health policy programs have their own time frames, health policies taken as a whole are deprived of them. Thus, evaluation in such a context is always carried out during the course of health policy and can at best inform about the progress in achieving general objectives or, if we are able to indicate them, objectives adopted for a limited time frame (e.g. parliamentary term).

Having the above reservations in mind, we will focus on indicating the potential applications of various forms of evaluation, separated on the basis of previously defined criteria. Description of identified barriers and advantages of discussed methods is presented in table (Table 2).

The temporal aspect of evaluation boils down to determining the right moment for conducting evaluation activities. As indicated in Table 1, basing on the above criterion, there are three main types of evaluation: ex-ante (prospective), mid-term and ex-post (retrospective). The application of the ex-ante evaluation method is currently practiced at the stage of designing activities. It seems that this type of evaluation can be considered quite commonly used in the area of health-related policies, although in the case of observed practice, it should be stipulated that this usually applies to health projects financed from public funds and selected in the grant application procedure [23]. This procedures usually include the stage of evaluation of applications according to predetermined criteria. In the case of programs or policies, usually such evaluation is replaced with impact assessment and expert consultations. These processes can hardly be considered as an equivalent of evaluation, mainly due to a different reference point (as in the case of impact assessment) or unstructured nature (as in the case of expert consultations, often not requiring them to apply a methodical approach). In international literature, the prospective evaluation procedure usually occurs in the context of investment decisions or research projects [24, 25]. The subject literature also lacks studies that would indicate the limitations of this type of evaluation while being used in health policy. Nevertheless, we could use arguments of a universal nature that come from studies carried out for the needs of other areas of public policy. The relatively frequent lack of access to valid evidence from scientific research should be considered the main factor limiting the usefulness of prospective methods. This phenomenon is particularly frequent and severe in relation to the assessment of the socio-economic context, but also the state of health and health behaviour of the population [26]. In addition, it should be noted that whenever the policy objectives change, be it in relation to its mechanics or in relation to the material and subjective scope, it is necessary to refer to other evidence.

Mid-term evaluation is mainly pragmatic, and could be brought down to ongoing verification of activities carried out and their value in relation to the objectives to which they were subordinated. It is worth mentioning that this is also a procedure that is relatively often confused with ongoing monitoring, therefore it should be noted that the latter is devoid of a axiological aspect, which is the essence of all evaluation activities [27]. Due to its formative nature, mid-term evaluation can lead to a number of changes in the program, project or policy, especially in the process of their implementation. The literature review did not identify any barriers that could be considered specific to this type of evaluation. However, it seems that in this case there is a potential risk of making premature decisions about the need to change the current mode of action.

Ex-post evaluation is a comprehensive summary of a program, project or policy, not only in relation to its results, but in all its aspects. The main challenges related to this type of evaluation include:

The increasing complexity of the analyzed phenomena, which may make it difficult to establish cause-and-effect relationships, and thus to make judgments about the value of actions taken under the program or policy in relation to the results obtained.

The complexity of the institutional environment, which may impede the identification of valuable data sources for the evaluation process.

Summing up the thread of the time criterion in relation to the evaluation process, it should be noted that the problem of lack of time frame is a heavy burden for the application of the forms of evaluation listed above. As it was mentioned in previous paragraphs, health policy when considered as a sectoral policy, does not have clearly defined time frames. Thus, keeping strictly to the nomenclature described above, each evaluation would be a mid-term evaluation. This does not exclude the possibility of using ex-ante and ex-post evaluations, however due to their very nature they can only relate to a selected time period. Therefore decision to make either one of those evaluations should be followed by indication of the relevant premises determining the setting of a time frame.

Many authors who write about the topic of evaluation consider the purposiveness of evaluation activities as the decisive criterion, and therefore refer to the division of evaluation into formative and summative. Historically, it is also one of the oldest divisions proposed, introduced to literature in the 1960s by Scriven [6].

Formative evaluations are carried out during the development and implementation of the program, and their usefulness is manifested primarily in obtaining guidance on how to best achieve the objectives of the program or how to improve it. Summative evaluations are carried out when the programs are well established and allow determining to what extent the program achieves its goals.

Thus, formative evaluations are considered to be particularly useful during the implementation of pilot projects or the implementation of new strategies. It seems most advantageous to use them at early stages. The evaluation results obtained in this way allow the necessary changes to be made to the programs, in order to improve their efficiency [28]. The scope of such evaluation includes two areas of critical importance for the quality and value of planned interventions, which are needs assessment and process evaluation. It should be noted that such assumptions of formative evaluation are now an immanent part of many recommendations related to the planning and development of interventions in the field of health policy. Examples include A Planning Framework for Public Health Practice (National Public Health Partnership, Australia) [29] or The Health Promotion Strategic Framework (Health Service Executive, Ireland) [30]. In both cases it was pointed out that the development of a health intervention program should be initiated by recognising a health problem that is supposed to be a focal point of the activities carried out and identifying the determinants of the problem.

Summative evaluation focuses on the results of a project, program, strategy or policy. It also refers to the observed impact of the project on change in selected indicators (e.g. change in the incidence of specific disease entities). The logic of the program and the ability to see the difference between causation and coincidence become important in this case. The distinction between performance evaluation and impact assessment is highlighted, among others by CDC in its recommendations for stakeholders involved in the implementation of the Healthy Communities program. In the discussed proposal, this division was referred to the time dimension and it leads to recognizing the impact evaluation as an assessment of long-term effects, while the outcome evaluation was referred to short- and medium-term effects only [31].

At this point, one should pay attention to the rather special situation that occurs in the case of the second of the forms discussed. If negative deviation from the expected outcome is observed it must necessarily lead to the initiation of formative activities. Otherwise, an irrational situation would arise in which the observed ineffectiveness or even harmfulness of implemented policies, programs or interventions would not entail any corrective actions or conclusions for the future [17, 32].

Regardless of any kind of digressions, summative evaluation presents a special value for decision-makers, because, by linking activities with achievements, it allows to infer about the effectiveness of implemented solutions. Thus, its formative properties will also be manifested in a broader aspect, resulting from the fact that it provides the basis (evidence of effectiveness) for continuing specific actions and programs or for their termination. Auer and Kruppe among the others pay attention to the feedback between the results of the evaluation and its impact on the shape of the agenda of future activities (in their cases related to the labour market policy) [33]. In the conditions of the functioning of democratic mechanisms and the pursuit of transparency in spending public funds, any relevant decision needs to be based on solid substantive foundations. This seems particularly important in the case of spending funds on health programs, the results of which are often not directly experienced by the public, which in turn leads to raising questions about rationality of incurring such expenses.

In current practice, the tendency to introduce summative methods can be clearly seen at the level of national or international health strategies. As a primary example one can indicate, among others, evaluation activities conducted by WHO as part of monitoring and evaluation of the Health for All strategy or similar activities related to the implementation of the Sustainable Development Goals. In both cases, the evaluation activities are based on reference to the goals and measures adopted at the stage of building the strategy, describing current state of their implementation. At the same time, there is an overlap between two dividing lines, because the evaluation is carried out at both global and national levels. The choice of evaluation approach was influenced by the high complexity of the analyzed matter, as well as the diversity of situations in individual countries and regions [34, 35]. In this case, summative evaluation, as focused mainly on results, allows to omit the context of the conducted activities.

Regarding the barriers of formative and summative evaluation, it should be noted that in this case the strong relationship between both forms is emphasized. The value of formative evaluation is largely dependent on the previously carried out summative evaluations for programs of a similar nature, scope and area of activity. At the same time, it should be emphasized that, unlike the formative evaluation, the summative approach does not provide grounds for drawing conclusions about what has failed in the implemented activities or what has proved helpful, due to the fact that it focuses on assessing the outcomes of the program in relation to previously defined goals.

The use of the formative or summative nature of evaluation for a holistic assessment of health policy raises several questions. In a first place we need to ask what will be the benchmark for formative evaluation? If we assume that formative evaluation should be supported by evidence from summative evaluations, then the question should be asked which summative evaluations will meet the conditions allowing to consider them as valid evidence? If we take a look at the level of health programs, then there is observable common practice of utilizing reference points in the form of experiences of other countries (or regions, or international programs) introducing similar solutions in environments with similar specificity. While such operations, even on a project or program scale, are subject to significant risk, there is doubt whether they can be carried out at all if we aim at adopting holistic approach to health policy. This is due both to the incomparability of the conditions in which the policy is implemented and to the complexity of the policy itself. In this sense, the scope constituting the subject of evaluation would have to be limited at best to selected components of health policy. A similarly troublesome issue may be the issue of reference to the goal, which, as indicated earlier, in the case of overall health policy is not clearly defined, and if attempts to determine it are made, the level of generality of the proposed interpretations would hamper the implementation of evaluation studies.

Within the scope of the evaluation criterion, meta-evaluation and cluster evaluation deserve to be singled out. The first of these forms is used to determine the value of evaluation in relation to ethical, methodological and praxeological standards. Any evaluation can be subject to this type of assessment, moreover, a basic set of evaluation criteria is easy to determine based on generally available standards of national or international evaluation societies (e.g. PTE standards) [36]. The use of meta-evaluation brings with it the advantages of particular utility, primarily in the context of designing activities in the field of health policy, allowing for explicit rejection as potential premises of these evaluation studies that do not meet the abovementioned standards.

The issue of cluster evaluation is slightly different and is assumed to be a solution to the problem of the complexity of the evaluated issues. The method created by the Kellog Foundation is mainly used to evaluate projects that are subject to one problem area and are implemented in parallel. The cluster then includes a set of projects, each of which is subject to evaluation, however, the very assumptions of the evaluation are formulated at the cluster level, not individual projects level. Due to the unification of assumptions and the possibility of adapting them to the evaluation needs of specific projects, a scheme is created to capture a broader picture of the issue [18]. This would not be the case if the evaluation was transferred into a series of unrelated operations dedicated to individual projects. Undoubtedly, the limitation of the aforementioned form of evaluation is the fact that it was created as a solution for project evaluation, i.e. easily identifiable projects that remain limited in time and have clearly defined objectives. However, it seems that the philosophy of cluster evaluation itself presents the potential that can be used in the overall assessment of health policy. Identifying the components of the health policy "cluster" would remain a major challenge.