Monograph on

Public Health Surveillance in the Pacific

Secretariat of the Pacific Community, Noumea, NEW CALEDONIA

Health Information Trainee, Secretariat of the Pacific Community, Noumea, NEW CALEDONIA

Twenty-two Pacific Island countries and territories (PICTs) receive various levels of technical assistance, training and financial support from international, regional and national agencies. To support the different aspects of these activities, the agencies currently request health data from the PICTs on a systematic basis in two major fields: health programme monitoring and disease surveillance. There is currently little consultation or integration between agencies. Communication exists mostly in terms of the exchange of various types of processed information such as reports, circulars, and other publications.

The Inter-agency Meeting on Health Information Requirements in the South Pacific took place in December 1995 in Noumea, New Caledonia, to discuss the potential for more integration and cooperation in order to ease the pressure on the data providers (the countries) ant to improve the relevance, quality, and timeliness of regional health information in the Pacific. As part of the effort to deal with the problems of both the pressure on data providers and the low quality and availabity of good health information, methodological tools for evaluating both health indicators and diseases subject to surveillance have been developed in order to ascertain those most suitable for public health surveillance. Finally, looking forward to the establishment of a Pacific Public Health Surveillance Network, important perspectives are outlined.

Public health surveillance typically differs from other sources of health data such as epidemiological studies, health surveys or administrative records, in terms of data collection method, mode, volume and frequency; reporting procedures and channels; type of data analysis and information dissemination; and costs and human resources involved.(Thacker, 1989 and Stroup, 1992). 'It is the ongoing collection, analysis and interpretation of health data, closely linked with the timely dissemination of these data both to those providing the data and to those who can apply the data to control and prevention programmes' (Thacker, 1989). In that regard, the complementary links existing between the potential sources of health information (e.g., epidemiological surveys cross-checking the validity of surveillance data) is of great importance. These links must be developed to enhance the decision-making process in public health spheres. However, this article focuses on the selection of a set of core indicators, that could ideally be used at both regional and in-country levels, for the ongoing measurement of the health status of Pacific Island populations, and the relevant decision-making based on such information.

In the Pacific, several international, regional, or national agencies (with a regional interest, e.g., the US Centers for Disease Control and Prevention, or CDC) request health data from the Pacific Islands countries and territories on a systematic basis, in two major fields: health programme monitoring and disease surveillance. These above organisations are bound to the monitoring of public health programmes in various aspects of technical assistance, training and financial support. Three agencies are also conducting disease surveillance activities: CDC, the World Health Organization (WHO), and the South Pacific Commission. These activities produce an ongoing quest for health information.

However, there is currently little consultation and integration in the planning and implementation of the regional public health surveillance activities. Communication does exist, taking place mostly in terms of exchange of processed information, for example reports and feedback materials between the PICTs and the various individual agencies, and information bulletins, circulars or similar publications, mostly between agencies. Nevertheless there is nothing like a planned regional public health surveillance system, or network.

This lack of coordination has harmful subsequent effects. The successful practice of public health surveillance is based on sound decisions, and these are possible only when decision-makers have accurate, timely and relevant information. Concerns about the current quantity and quality of health data requirements at regional level have been raised on many occasions by agencies such as UNICEF, WHO and SPC, as well as by health professionals from the Pacific Islands countries and territories themselves. PICTs representatives complained unanimously about the volume of the demand for health information, partly due to duplication in the requests, and asked the agencies to increase the level of integration and coordination of their data needs and requests.

On the other hand, the agencies stressed that it was rather disappointing not being able to properly monitor the development of key public health programmes at both national and regional levels (e.g., in the fields of immunisation, environmental health, etc.) nor progress toward regional public health goals, such as the reduction of measles mortality or rate of low birth weight, because regional public health surveillance was not sound enough.

Therefore, our primary two suggestions were to decrease the pressure on data providers by making concerted efforts towards integration of regional health data requirements, as no public health surveillance system can be useful if not fueled with relevant (complete, acurate and timely) health information. The SPC Community Health Programme took up the challenge of finding ways to reconcile these approaches. It was a concern and a challenge for anybody involved in health related development in the Pacific.

In December 95, the South Pacific Commission organised the Inter-Agency Meeting on Health Information Requirements (IAMHIR) in Noumea, New Caledonia, with the support of UNICEF and WHO. The participants were representatives of several regional agencies and health professionals of Pacific Island countries and territories. The IAMHIR meeting aimed to establish the basic principles of a public health surveillance network in the Pacific. The foundations of this network are described hereafter.

Reviewing the available information from most of the international agencies involved in the region, we have compiled a listing of 178 indicators 'required' for the monitoring and evaluation of the health status of the Pacific Island populations. Their distribution in the various fields covered are listed in Table 1. To limit the scope to a manageable level, the indicators concerning provision of health care (e.g., health infrastructures, personnel), health expenditures, and health policy were not considered. Nevertheless, we do not underestimate their potential usefulness, since sound public health surveillance can lead to relevant changes in public health policies. These changes can themselves be significant, and therefore deserve appropriate measurement. If not directly connected to health activities, indicators related to socio-economic and education sectors have not been considered either.

Among the indicators listed, 65 out of 178 (37%) are requested by more than one agency, and about 10% requested by three or more. Multiple demand varies according to different areas. Looking at the whole pool of indicators, it appears that not all are specific for a single health event. In 13 instances a single public health problem or field of activity (e.g., low birth weight, obstetric care, vitamin A deficiency, access to safe water) is measured in different ways. The usual purpose of these different measurements is to refine the analysis, using disaggregated data. Depending on the topic, two, three, or up to five distinct indicators are used. Sometimes definitions are vague, only slightly different, and do not really bring in new elements to analyse. Altogether, these represented 36 individual indicators (20% of the total).

At this stage, the data providers' point of view deserves to be considered. There is little doubt that when differences in measurements are not expressly meant to increase the understanding of common public health problems (such as anaemia, vitamin A deficiency, or low birth weight), they can easily cause confusion at data collection levels, and beyond. The 178 health status indicators cover a rather wide range of data and data collection vary significantly depending on data sources, purpose and mode of collection, specificity of the information wanted, frequency and type of analysis required, expertise and resources necessary. As a result, not all data 'required' are amenable to public health surveillance.

The objectives of the IAMHIR meeting were to select a core set of health indicators relevant to public health surveillance. The first step we suggested was to look more closely at the set of 65 indicators used by more than one agency at the regional level, because a strong common demand certainly reflects a similar interest for certain indicators. Since this group of indicators represents overlapping data requirements, harmonisation and integration would contribute to alleviating the current pressure on the actual data providers.

To facilitate the evaluation of the indicators, we have developed two distinct methodological tools: (1) to appraise whether or not it is appropriate for a disease to be subject to surveillance, compared to other candidates; and (2) to determine whether a given indicator is relevant and useful for public health surveillance. The ultimate inclusion of the selected indicator into a regional core set of criteria, and the definition of the set itself, would be based primarily on the PICT's national priorities, the objectives of the surveillance, and on a comparative analysis of the diseases and other indicators.

The highest number of multiple demanded indicators is in the field of communicable diseases, representing 42% of the total pool of indicators used at the regional level to monitor communicable diseases, conditions and syndromes. It essentially covers the incidence and mortality of 14 diseases, out of 21 monitored altogether at regional level. Communicable disease control and prevention, especially including Expanded Program on Immunization target diseases, have historically been attached to public health surveillance. Even though health is not merely the absence of disease, the burden of diseases, communicable or not, on the Pacific communities remains considerable, and the huge majority of the public health programmes aim at prevention and control of diseases. For these reasons, we decided that disease surveillance, although included in the core set of indicators, deserves to be approached separately.

A system based on a series of scored criteria was developed, to evaluate the relevance of a disease, condition or syndrome to be subject to the public health surveillance. It is essential for such a priority-setting system that objectives would be first clearly defined, so that criteria for selection flow consistently. Our chief concern regarding the selection of any indicators is usefulness at both national and regional levels. Therefore, we defined the following objectives:

- to provide estimates on the incidence and mortality of diseases, conditions, or syndromes under surveillance; 

- to monitor trends in the diseases, conditions, or syndromes under surveillance, and to detect changes indicating potential needs for appropriate action; 

- to properly identify outbreaks for timely investigation and control; 

- to allow for the assessment of the effect of disease control measures, providing relevant information for re-programming more appropriately-focused public health interventions;

After drafting a theoretical model, we pre-tested its pertinence, feasibility, and acceptability at the country level, with the active collaboration of the Health Services of New Caledonia. We chose the list of notifiable diseases currently used in the territory, which was scheduled for revision. The proposed objectives were found acceptable. The pre-test allowed us to refine the number, definition and grouping of the selection criteria.

1. Surveillance interest for WHO and EPI (Expanded Program on Immunisation) target diseases 

2. International Health Regulations

3. Incidence 

4. Hospital costs 

5. Mortality 

6. Case-to-death ratio⁽¹⁾ 

7. Socio-economic impact (mainly absenteeism from work and long-term disability) 

8. Risk perception (public, decision-makers, media, health personnel)

9. Communicability 

10. Epidemic potential 

11. Vaccine preventability

12. Necessity for immediate action 

13. Usefulness as a health indicator 

14. Amenable to public health measures⁽²⁾ 

15. Priority status (with decision-makers) 

16. Ease of diagnosis

Some of the definitions (see Annex 1), and the disease surveillance method drew upon a Canadian experience in setting priorities for communicable disease surveillance.(Carter, 1992 and Wetterhall, 1992). If any of the diseases met either of the first two criteria (Surveillance interest for WHO and EPI target diseases; international health regulation), they were automatically included in the set subject to surveillance. The rest of the criteria have been rated either on a scale of 0-2 (criteria nos. 3, 5-9, 11, 13, 15, 16), or 0-5 to increase their relative weight, for those criteria found to be critical to the objectives of the surveillance system (criteria nos. 4, 10, 12, 14)⁽³⁾.

All diseases proposed for surveillance (in practice, those already included in the regional and national surveillance systems, plus any additional requests) can be evaluated according to all criteria, then ranked according to their total score. More important than the absolute score is the relative ranking of the diseases. The cut-off point for recommending that a disease be included in regional/national surveillance, though crucial, remains based on a balance between information needs and resources available. In pre-testing on New Caledonia's list of notifiable diseases, we agreed on a cut-off of 15 points.

It is not our intention to claim that this scored system of selection would be in any way fully objective. Subjectivity interferes in many occasions: in the score given to diseases for every criteria; in the different weights given to the criteria; in the choice of the cut-off point; even in the choice of diseases proposed for evaluation. We nevertheless believe that it provides an original and reliable standardised method, with enough flexibility for use at both national and regional levels. Moreover, it provides plenty of opportunities for ongoing refinement and improvement, through field epidemiology studies and hands-on training.

The epidemiological transition from a prevailing pattern of communicable to non-communicable diseases, mostly in terms of mortality, has been extensively described and documented in the Pacific since the early 1980s (Taylor, 1989). However, currently existing surveillance systems favour communicable diseases over lifestyle-related pathologies. To our knowledge, with the exception of cancer registries, none of the existing regional disease surveillance systems include non-communicable diseases (or NCDs), such as diabetes mellitus, cardiovascular diseases or asthma. Can NCDs be assessed in the same manner as communicable diseases, using the same tools? Nothing in the disease surveillance system's objectives precludes such an approach.

The method proposed for the determination of health indicators is based on the same scored framework used to select diseases subject to surveillance. The criteria are chosen according to the objectives of public health surveillance, bearing in mind basic principles such as acceptability and affordability at national and regional levels (see Annex 2).

We believe a standardised method is essential to identify a consistent set of indicators wich will ensure the following: (1) national and regional standards for public health surveillance through reference tools and guidelines; (2) flexibility to allow different communities to decide on target-levels tailored to their specific situation and priorities; (3) public health surveillance to be planned and operated at operational level, in harmony with regional data requirements; (4) better planning of resources for complementary health information activities that are not appropriate for public health surveillance, or more specific of the needs of certain programmes or agencies (e.g., management programme information systems, focused epidemiological studies, programme-specific cross sectional surveys). 

In developing this part of our proposed method, we referred to a CDC experience identifying a set of 18 health status indicators for public health surveillance in the United States (Stroup, 1992). To allow for a comprehensive measure of community health, the set of relevant selected indicators should include general measurements of community health such as overall morbidity, mortality, and disabilities, and specific measurements of community health related to identified priority public health problems. This set, or one derived subset of measurements, has to be consistent in-country (PICs) and at regional levels. The series of 10 selection criteria we retained therefore emphasise consistent characteristics. They have been grouped into three categories:

1. Appropriate for surveillance (ongoing process) 

2. Data readily available at country level 

3. Data source

4. Estimated accuracy 

5. Validity in small population

6. Purpose for being requested 

7. Usefulness in decision-making 

8. Possible use at country level 

9. Specific interventions implied (if changes are detected) 

10. Outcome oriented

Each of the categories would be summarised by a score 0-5, and therefore each of the indicators would be rated on a scale 0-15. At this point we must bear in mind that a core set of health indicators should include some of the disease surveillance indicators identified, together with other measures of the health status of the Pacific Island communities.

In looking forward to the establishment of a regional surveillance framework and network for the Pacific Island countries and territories, there are few basic principles that should be discussed.

By definition, before becoming a regional health data, all data are national health data. The only reason a country or territory would be able to sustain the necessary efforts to collect, analyse, interpret and disseminate health data is because national health professionals and decision-makers are convinced of the usefulness of these data for their own use.

The second important point is that a regional initiative must be regional in approach, and as comprehensive as possible. A regional surveillance framework must be planned in concert with all interested bodies; managed and implemented in collaboration; and monitored and evaluated with a common concern for ongoing improvement of regional public health surveillance. Finally, isolated, sectoral or opportunistic activities have little chances of having a significant impact on the health of Pacific Island populations, whether in public health programmes or public health surveillance systems.

Through the IAMHIR meeting, and by proposing methods and tools for a common approach to public health surveillance at national and regional levels, the aim of the SPC Community Health Programme was to contribute to a concerted approach to regional surveillance. Within a framework, a network may flourish and activities be integrated. Within a network, activities may be expanded and regional surveillance may become real. With both a framework and a network, public health surveillance in the Pacific becomes sustainable.

Towards the establishment of a regional public health network , three type of actions have to be promoted.

It is crucial to keep the momentum going. Pacific Islands countries and territories have to be more thoroughly consulted. A Pacific surveillance framework could be developed by building on the (adjusted) standardised methods and tools discussed at the IAMHIR meeting. In order to coordinate subsequent public health surveillance activities, a working group on public health surveillance in the Pacific was established at the IAMHIR meeting, and named Pacific Public Health Surveillance Working Group (PacPHSWG). It is composed of 10 people, with a mixed representation from PICTs and agencies who meet regularly (twice a year) to continue the work started in Noumea.

There are many possibilities for refining and improving the proposed methods. The baseline information necessary for using accurately some of the selection criteria applied to diseases and indicators is currently rather poor. It is important to fill this gap. A number of small-scale field epidemiology projects could be planned for and implemented in Pacific Islands countries and territories. These PICTs would also participate in refining both national and regional data (e.g. studies on the sensitivity and specificity of an indicator or disease (case) definition; incidence and mortality of tuberculosis; incidence and hospital costs incurred by rheumatic fever; frequency, mortality and hospital costs incurred by complications of sexually transmitted diseases).

In summary, we should endeavour to keep the communication and collaboration between parties and the ongoing work needed to improve regional surveillance (planning, implementation, coordination and evaluation).

Based on a defined framework, and taking advantage of a concerted approach, public health surveillance activities may expand in a more consistent and useful manner: : the experience acquired in surveillance efforts in some parts of the Pacific may be usefully applied in others; findings from a field study carried out in one of the PICTs could complement the information gathered in others; and the skills acquired by some Pacific Islanders may be used in other Pacific countries or territories. Exchanging information and skills would create a useful network.

Telecommunications technologies have evolved rapidly during the past ten years. Some of the most modern means are readily available in the Pacific. We strongly believe computer networking would enhance the information exchange in regional public health surveillance. Provided their use would be properly planned, resources such as the Internet, and other telecommunication networks currently in use throughout the Pacific (e.g., PEACESAT, PACTOK), provide an invaluable opportunity for supporting the networking of information(4).

Against the background of a sound framework and regional network, with reference methods and tools available in-country and regularly improved, the development of a field epidemiology training programme could reasonably be explored, in association with national and regional health training institutions (e.g., the Fiji School of Medicine, the University of Auckland Department of Community Medicine and the National Center for Population Health of the Australian National University). The development of a 'Pacific Field Epidemiology Training Programme' would greatly contribute to make national and regional surveillance sustainable.

Although meant ultimately to generate savings by avoiding certain aspects of health expenditure and improving efficiency of health programmes, public health surveillance has a cost. This might be viewed as a serious obstacle, in particular when starting up surveillance activities at the country level.

The Pacific public health surveillance group could play an essential role in preparing and submitting appropriate project proposals to institutional donors. We believe funding proposals issued by a consortium of international agencies and Pacific Island countries would be very strong ones, as aid donors should be appreciative of the coordinated efforts put into the planning, implementation and evaluation of regional projects. These proposals could be seen as even stronger if they associate national and/or regional training institutions such as the Fiji School of Medicine, or the University of Auckland Department of Community Medicine.

The IAMHIR Meeting was the first step in a long process ultimately aiming at the establishment of a Pacific Public Health Surveillance Network (PPHSN). Five operational strategies developed within the SPC have now been adopted in order to guide the development of the network:

- harmonisation of health data needs and development of adequate surveillance systems, including operational research, 

- development of relevant computer applications, 

- providing field epidemiology and public health surveillance training, 

- promoting the use of E-mail, opening the Network to new clients, new services and other networks, and 

- publication of health information bulletins, technical studies, applied research findings, information on resources available in the network, etc.

Data harmonisation efforts we initiated in December 1995 have now been on trial in the field in four Pacific countries and territories. The validation of the proposed method - now called PacSel - by the PacPHSWG, led to the identification of a sub-set of 88 indicators relevant to public health surveillance, out of the 175 originally reviewed. Furthermore, these 88 remaining health indicators could be broken down into three groups of varying priority: high (33); medium (32) and low (23). These findings validated the relevance of PacSel as a method able to establish priorities in public health surveillance, and ultimately releasing pressure from the data providers.

In refining the available surveillance information, operational research activities have started in 1997, focussing on dengue fever field diagnostic capabilities. More thoughts have been put into the concept of an innovative regional training programme in public health surveillance and field epidemiology. A feasibility study has recently been completed in this field. An E-mail (and fax) listserver called PACNET was launched in April 1997 to become the 'voice' of the Pacific Public Health Surveillance Network. PACNET's present focus is on 'early warning', allowing the networking of individuals (PICTs health professionals and other surveillance-related specialists), information and resources, in order to facilitate the early diagnosis, and better prevention and control, of outbreaks of communicable diseases in the Pacific.

The progress towards the development of the Pacific Public Health Surveillance Network has come a long way since December 1995, due the collaborative efforts of the members of PacPHSWG. This group is now called the Coordinating Body of the Pacific Public Health Surveillance Network, and the South Pacific Commission is its official focal point.

CARTER AO. (1992). Setting priorities: the Canadian experience in communicable disease surveillance. MMWR 41 Suppl.:79-84.

TAYLOR R., N. DAVIS LEWISand S. LEVY. (1989). Societies in transition: mortality patterns in Pacific Island Populations. Int J Epidemiol. 18:634-646.

THACKER S.B. and R.L BERKELMAN. (1988). Public Health Surveillance in the United States. Epidemiol. Rev; 10: 164-90.

THACKER S.B, R.L. BERKELMAN and D.F. STROUP. (1989). Science of public health surveillance. J. Public Health Policy; 10: 187-203.

STROUP D. (1992). Surveillance data for policy: a national and state approach. MMWR 41; Suppl.: 135-141

WETTERHALL S.F., R.E. CHURCHILL, E.A BRANN., M. PAPPAIOANOU, D.F. STROUP and N.E. STROUP. (1992). Proceedings of the 1992 International Symposium on Public Health Surveillance. MMWR 41 Suppl 1992.

Footnotes

(1) Subsequently withdrawn, as possibly derived from incidence and mortality. (Back to the text)

(2) Eventually determined as the first criteria to be considered. (Back to the text)

(3) The scoring system has since been simplified and all criteria are scored on the same scale. On the light of several field applications of the method, the relative weight of criteria appeared to have no effect on the relative ranking of the disease. (Back to the text)

(4) To initiate such a move, all members of the PacPHSWG committed themselves at the IAMHIR meeting to endeavour access, then use, the electronic mail technology to support distance collaboration within the group. (Back to the text)