Health Promotion International, Vol. 14, No. 4, 289-299,
December 1999
© Oxford University Press 1999
Practical lessons in using indicators of determinants of health across 47 European cities
1 Department of Public Health Medicine, Guy's, King's and St Thomas' Medical School, Guy's Campus, London Bridge, London SE1 9RT and Lambeth Southwark and Lewisham Health Authority, Waterloo, London SE1 7NT, UK, 2 WHO Healthy Cities Project Office, Copenhagen, Denmark and 3 South East Institute of Public Health, Tunbridge Wells, Kent TN3 0XT, UK
Address for correspondence: Yvonne Doyle 21 Thorpewood Avenue London SE26 4BU UK
 |
SUMMARY |
|---|
 TOP SUMMARY BACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
A survey was conducted of 47 European cities applying to join the third phase of the World Health Organization Healthy Cities Project. The survey tested the feasibility of recording baseline information on health-promoting processes and activities in the cities. A broad multi-sectoral focus for health in the questionnaire presented formidable challenges to respondents. Despite goodwill and local networks, the quality of certain responses was variable. Socio-economic statistics are not universally available at city level; however, descriptive responses about services and health promotion yielded uniquely useful local information over the national level. Survey responses were not comparable across cities and such interpretation would be counter-productive. Recommendations for future surveys are made.
Key words: city health; determinants of health; indicators; measures
|
BACKGROUND |
|---|
|
TOPSUMMARYBACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
This paper describes work undertaken as part of the World Health Organization's (WHO) Healthy Cities Project in Europe. The Healthy Cities Project encompasses a variety of people, institutions, cultures and activities (Duhl, 1986
). The participating cities in the European Project are shown in Figure 1. Their diversity can be gleaned by viewing Table 1, which compares some basic parameters of six participating cities.
|
|
An early technical challenge to the Project was to devise indicators which captured a range of local initiatives addressing the wider dimensions of health. These indicators were then employed across the diverse cities in the Project. The experience of using indicators at national level in the Health For All (HFA) Programme throughout the five WHO regions of the world has shown the variable utility of indicators in that programme (Miglionico and Casas, 1990
; Nossikov et al., 1990). This is because of differing methods of data collection, and difficulties in interpreting the meaning of global indicators. The World Health Organization has recognized these problems particularly in Europe, and several cross-national consultations have taken place to develop common methods and instruments (Netherlands Central Bureau of Statistics, 1988; 1992). The experts on these consultations noted the poor comparability of many broader health indicators such as those measuring social and economic status. Consequently, many HFA indicators have some way to go before harmonization can be achieved.
|
THE PROCESS OF DERIVING THE HEALTHY CITIES INDICATORS QUESTIONNAIRE |
|---|
|
TOPSUMMARYBACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
In 1993, the Healthy Cities Project was entering its second phase for 5 years. The objectives for this phase were the adoption of policies at city level, building strategic links with sectors other than the health sector, and putting structures and systems in place to promote health. Healthy City profiles and plans were expected to cover health, social and environmental issues (World Health Organization Regional Office for Europe, 1993). Indicators were seen as a tool in providing the information towards this end. A Multi-City Action Plan on city indicators was convened in 1991 with the aim of recording a baseline of activity in the Project and measures of health in 47 applicant cities. Towards this end, a survey was proposed, and a questionnaire and guidance was published in 1992 (World Health Organization Regional Office for Europe, 1992a). The indicators of the questionnaire are shown in Table 2
. They comprise health, service, environmental and socio-economic measures, as well as general census information at city level.
|
The questionnaire was devised to reflect: (i) the Health For All targets (60% of the indicators were adapted from HFA targets); (ii) the Ottawa Charter, adapted to city level; and (iii) the 11 qualities of a Healthy City (World Health Organization Regional Office for Europe, 1992b). Twenty percent of indicators were adapted by the Multi-City Action Plan team to reflect these 11 qualities, particularly indicators numbers 15–20, 24–25, 35–37, 39–40 and 45 in Table 2
. All cities applying to join the 1993–1998 phase of the Project were required to complete the questionnaire. The questionnaire was not piloted, because it was considered that the survey itself was a feasibility study that would inform a possible future repeat of the exercise, for example after the 1993–1998 phase was complete. The questionnaire was sent to the 47 city co-ordinators for the Project, chosen because of their pivotal role in the Project and their local knowledge. They in turn used their service and academic contacts in other establishments of the cities, in order to complete the information.
Regular review of the survey progress was provided by a standing group on indicators comprising technical and professional advisers from academic establishments throughout the European Region and the WHO itself. Coordinators drawn from the Multi-City Action Plan team and from cities where progress had been made in profiling the health status of the local residents were also involved (World Health Organization Regional Office for Europe, 1994). A researcher at the WHO Regional Office for Europe sought whatever missing information could be retrieved subsequent to the return of the questionnaires. The results of the survey were analysed independently by a multi-disciplinary group of analysts chosen because their disciplines reflected the sections of the questionnaire.
|
LESSONS LEARNT FROM THE SURVEY |
|---|
|
TOPSUMMARYBACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
There are five questions which warrant reflection in the light of the initial aims of the survey in 1991. These questions are about: the need for a survey; the definition of ‘health’ measures at city level; the interpretation of the indicators across different cultures and health systems; the sources and adequacy of data; and the utility of the results in comparing and contrasting cities.
Was there a need for a survey?
There was some debate about the utility of a survey that was a considerable undertaking. The counter-arguments were that: (i) very little data of the type collected in this survey existed in the public domain at city level across Europe; (ii) it was important to know the feasibility of obtaining a baseline if measuring outcomes from the Project was desired in future; and (iii) the indicators could aid the local production of city profiles and plans which were part of the work of the 1993–1998 phase. Although there were real difficulties in collection and interpretation of data, the survey was broadly welcomed by co-ordinators as an initiative which forged useful local links towards the governance of health-promoting processes. It was not undertaken as a research exercise, but to give local impetus to the project and to demonstrate the degree of local activities and the quality of health information available.
Defining ‘health’ measures at the city level
The proposal that broad health-promoting processes can contribute to the health of city residents is not an unreasonable tenet; and these urban processes are multi-sectoral (Duhl, 1986). The four sections of the questionnaire sought to capture the wider dimensions that are relevant to city health. The survey sought available data from primary sources, rather than any attempt to measure health through local population interviews. Nevertheless, 53 indicators emerged consequent on the broad scope; and this was too many. Some indicators (e.g. ‘percentage of the population covered by health insurance’, no. 10, Table 2) had poor discriminatory power, while many in the socio-economic section were relevant to health (Davey-Smith and Eggar, 1993) but yielded poor quality data at the city level. More attention was required in validating indicators during the development of the questionnaire.
Similar indicators had been discussed at the national level in a first consultation on methods in 1988 when it was noted that collection might be more successful at city level using the known good networks in the Healthy Cities Project (Netherlands Central Bureau of Statistics, 1988). However, by the time of a third consultation (after the commencement of the Healthy Cities survey), it was remarked that even when a small core set was applied trans-nationally, only three–four indicators were fully available per country (Netherlands Central Bureau of Statistics, 1992). The dilemma of keeping the ‘health’ focus broad but measurable was not fully surmounted in the current survey.
Interpretation of indicators by respondents
It has been noted at a national level in several WHO regions that there is a lack of conceptual uniformity in the interpretation of indicators (Miglionico and Casas, 1990; Nossikov et al., 1990; Wysocki et al., 1990). That is, the meaning of a global term, e.g. ‘within easy reach’ or ‘poison control system’ is different in different countries. This problem was also found at city level in the survey. The HFA indicators 26–31 deal with health care. These describe processes as well as outcomes, and processes are particularly subject to variability in content. Several were adapted to city level and yielded more descriptive analysis than data. Unsurprisingly, there was also confusion in the interpretation of global statistical indicators as applied to local health services. Table 3 shows comments about the coverage of city populations by the nursing profession (indicator no. 9, Table 2).
|
It is probably impossible to eliminate different interpretations of such indicators, but on the positive side, the qualitative information returned on some of the health-service measures gave insight into the local infrastructure to support health promotion in the various cities. Table 3
also shows examples of interesting information returned by respondents in their attempts to answer the question on health information communication (no. 13, Table 2).
Sources and adequacy of data
Two main parameters are examined here: the degree of missing data; and the reliability of the information. It is unclear whether a non-response to the question meant that the information was not available, the respondent could not easily access it, or the respondent did not attempt to access it. Table 4 presents information demonstrating the variable value of the information supplied on six indicators from the four sections of the questionnaire. The classification in Table 4 is based on the experience of using Health for All indicators in the European region (Nossikov et al., 1990).
|
It is known from work on national responses to WHO indicators that completeness of response can range from 0 to 91% overall, and even for individual indicators, the range can vary more than 20-fold (Hammoud and Ouakrimi, 1990
). It is clear why responses to the indicator on disabled people in the workforce were low in our survey. This indicator was considered important in previous literature (van Oers and Reelick, 1992), but equally response rates are known to be low. However, it is not clear why apparently little use was made of cancer registry data in the ‘cause of death’ returns. The best cancer responses were submitted for breast and respiratory cancers; however, even here there were the 10 non-respondent cities each for lung and breast cancer, and cancer registries that do not appear to have been accessed serve almost all of these cities. Although we express doubt about the scale of differences in incidence of cervical cancer in Table 1, it is known from recent research that there are considerable differences in the incidence of this cancer between countries, reflecting differences in risk factors, particularly the prevalence of human papilloma virus (De Sanjose et al., 1997).
The Healthy Cities survey has sought to widen the sources of data on ‘health’, which has been recommended in literature (van Oers and Reelick, 1992). This led to gathering data from such widely dispersed sources as local city political administrations, environment offices, the police, emergency services, local surveys, academic institutions and the voluntary sector. This led to exhaustion in the respondents and few of the questionnaires were completed for every section. Moreover, it is likely that certain city respondents made great efforts to collect data while others did not use all the sources available to them.
As regards validity, a range of problems identified at the regional level in WHO literature are somewhat in evidence at the city level also. These include: under-registration of conditions, e.g. HIV/AIDS; multiplicity of systems for data collection; wide dispersal of sources; and difficulties for us as recipients of responses in interpreting the data consequent on the poor face value of the indicator (e.g. the non-specific denominator definition in ‘low birth weight’, no. 3, Table 2). The latter is amenable to improvement by more precise specification at the outset.
All-cause mortality data should be readily available in the European region, and it is disappointing that five cities did not respond at all to this section. The use of cause of death statistics for health assessment has been studied before (Ruzicka and Lopez, 1990), and has been found wanting. Some of this is because the utility of mortality as a proxy for ‘health status' is questionable, although the Second European Atlas of Avoidable Mortality documents a welcome decline in constituent conditions in western Europe around 1991, and this must contribute to the health of the young and middle-aged populations in Europe (Commission of the European Communities, 1991).
It is not surprising that so little of the mortality data presented were completely satisfactory. There are well-documented problems with certification of cause of death (Kelson and Farebrother, 1987; Mackenback et al., 1987; Anonymous, 1994). It has been noted that there appear to be large differences between countries, e.g. in diabetes mortality (Balkau et al., 1993) and suicide (Anonymous, 1994; Schmidtke, 1997). We noted large ranges of crude death rate for suicide in our own survey. For example, the rates are high in the Hungarian cities noted before in the literature (Schmidtke, 1997), but remarkably low in the Polish cities which responded. This may have more to do with known religious taboos on suicide or reporting it than what is happening in reality. In relation to AIDS/HIV, apart from St Petersburg, the countries of eastern Europe and the newly independent states reported no mortality, less than 1/100 000, or did not respond to this section. This was unlikely to be the reality in the early 1990s.
Due to incomplete population data submitted, standardized all-cause mortality ratios could only be performed on 23 cities, with a good mix of geography. Figure 2 shows the contrast in pattern between crude rates and standard mortality ratios for these cities. There is an east–west division at city level which is well noted in the literature at national level as reduced life expectancy in eastern Europe and the former Soviet Union (Boback and Marmot, 1996; Walberg et al., 1998).
|
Can results from this type of survey be used to compare and contrast cities?
The answer to this question is no. The results could not be judged at face value across so many cultures, economies and data collection systems without considerable interpretation which the official feedback to the city respondents was at pains to explain (Doyle et al., 1996
). Respondents used the feedback on their individual city to identify areas in need of further local data collection; and in some cases, to prepare city profiles and plans. We were aware from published work in one participating city, that local health information collection was considered to have potential in guiding local health policy; but even here problems arose when data were collected from several municipal and regional institutions (van Oers and Reelick, 1992). It seems that even within a single city, the potential for confounding of results is present.
Some authors have attempted to overcome comparability problems at the trans-national level by undertaking health surveys with generic survey instruments including clearly defined definitions and instructions. These have been used to some effect in surveys developed by the WHO itself (World Health Organization, 1986). Generic health surveys are being used in diverse applications and populations (McHorney, 1997). The WHO has examined problems in cross-national comparability in multi-centre studies on topics, e.g. schizophrenia (Jablensky et al., 1992). Despite these initiatives, the literature continues to note the lack of validly comparable cross-national surveys, and a continued plea is made for agreement on survey methods (Anonymous, 1994). Moreover, while field surveys may yield interesting clinical information, it is acknowledged that many still lack the precision required for health policy development (McHorney, 1997), a central reason for developing the approach of the Healthy Cities survey.
|
CONCLUSIONS |
|---|
|
TOPSUMMARYBACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
It was probably not adequately appreciated at the outset that the barriers to uniformity of reporting at a city level are as formidable as at national level. The information was not comparable across cities nor do we believe that it should be used in this way. Given the tremendous variation between communities and the inevitable range of indicator results, there is a danger of causing rivalry between locations (United Nations Association of Great Britain and Northern Ireland, 1993) and a ‘league table’ mentality. Certainly we could not validly state that one city was ‘healthier’ than another, and no city was completely confident of its performance across all dimensions of the survey. The process of completing the survey forged healthy alliances locally. It was interesting to receive examples of health promotion initiatives locally. This ‘intimate’ information brought out the strength of a city-level survey.
If a future survey of the cities is planned, we would recommend: (i) a smaller core of the indicators adapted from those identified as useful in the current survey; (ii) keeping the focus nevertheless as broad as possible; (iii) paying more attention to documenting local innovations in the Project; and (iv) training of co-ordinators who would act as chief respondents to a survey in the required methods of completing the survey. City project leaders do learn from each other's innovations and methods, therefore recording this information is a good initiative for the Project. As well as a small number of statistical measures, some questions adapted to capture the ongoing work of the Project in these diverse European locations should be considered.
|
APPENDIX |
|---|
|
TOPSUMMARYBACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
|
|
REFERENCES |
|---|
|
TOPSUMMARYBACKGROUNDTHE PROCESS OF DERIVING...LESSONS LEARNT FROM THE...CONCLUSIONSAPPENDIXREFERENCES |
|---|
Anonymous (1994) Health issues for Europe. Lancet, 343, 245–246.[ISI][Medline]
Balkau, B., Jougla, E. and Papoz, L. (1993) European study of the certification and coding of causes of death of six clinical case histories of diabetic patients. International Journal of Epidemiology, 22, 116–126.
Bobak, M. and Marmot, M. (1996) East-West mortality divide and its potential explanations: proposed research agenda. British Medical Journal, 312, 421–425.
Commission of the European Communities, Health Service Research No. 6 (1991) European Community Atlas of Avoidable Death. Second Edition, Volume 1. Oxford Medical Publications, Oxford.
Davey-Smith, G. and Eggar, M. (1993) Socio-economic differentials in health and wealth. British Medical Journal, 307, 1085–1086.
De Sanjose, S., Bosch, F. X., Munoz, N. and Shah, K. (1997) Social differences in sexual behaviour and cervical cancer. IARC Sci. Publ., 138, 309–317.
Doyle, Y. G., Brunning, D., Cryer, P. C., Hedley, S. and Russell-Hodgson, C. (1996) Healthy cities indicators. Analysis of data from across Europe. WHO Regional Office for Europe, Copenhagen.
Duhl, L. J. (1986) The healthy city: its function and its future. Health Promotion, 1, 55–60.
Hammoud, E. I. and Ouakrimi, M. (1990) Monitoring and evaluation of the Health-For-All strategy: experience in the Eastern Mediterranean Region. World Health Statistics Quarterly, 43, 32–36.
Jablensky, A., Sartorius, N., Ernberg, G., Anker, M., Korten, A., Cooper, J. E., Day, R. and Bertelsen, A. (1992) Schizophrenia: manifestations, incidence and course in different cultures. A World Health Organization ten-country study. Psychological Medicine Monographs, 20 (suppl.).
Kelson, M. and Farebrother, M. (1987) The effect of inaccuracies in death certification and coding practices in the European Economic Community on international cancer mortality statistics. International Journal of Epidemiology, 16, 411–414.
Mackenbach, J. P., vanDuyne, W. M. and Kelson, M. C. (1987) Certification and coding of two underlying causes of death in The Netherlands and other countries of the European Community. Journal of Epidemiology and Community Health, 41, 156–160.[Abstract]
McHorney, C. A. (1997) Generic health measurement: past accomplishments and a measurement paradigm for the 21st century. Annals of Internal Medicine, 127, 743–750.
Miglionico, A. and Casas, J. A. (1990) The second round of monitoring the Health-For-All strategy. The experience of the Americas. World Health Statistics Quarterly, 43, 8–13.
Netherlands Central Bureau of Statistics/WHO Regional Office for Europe (1988) Consultation to develop common methods and instruments for health interview surveys. EUR/ICP/HST 120. Report on a WHO Meeting. Netherlands Central Bureau of Statistics/Copenhagen: WHO Regional Office for Europe, pp. 9–12, 19.
Netherlands Central Bureau of Statistics/WHO Regional Office for Europe (1992) Third consultation to develop common methods and instruments for health interview surveys. Report on a WHO Meeting. EUR/ICP/HST 120. Netherlands Central Bureau of Statistics/Copenhagen: WHO Regional Office for Europe, pp. 6–8, 15–17.
Nossikov, A., Prokhorskas, R. and Shabanah, M. (1990) Using regional indicators: the experience of Europe. World Health Statistics Quarterly, 43, 25–31.
van Oers, J. A. M. and Reelick, N. F. (1992) Quantitative indicators for a healthy city—the Rotterdam local health information system. Journal of Epidemiology and Community Health, 46, 293–296.[Abstract]
Ruzicka, L. T. and Lopez, A. D. (1990) The use of cause-of-death statistics for health situation assessment: national and international experiences. World Health Statistics Quarterly, 43, 249–257.[Medline]
Schmidtke, A. (1997) Perspective: suicide in Europe. Suicide and Life Threatening Behaviour, 27, 127–136.
United National Association of Great Britain and Northern Ireland. Sustainable Development Unit (1993) Measuring Sustainability. UNA-UK, 3, Whitehall Court, London.
Walberg, P., McKee, M., Shkolnikov, V., Chenet, L. and Leon, D. A. (1998) Economic change, crime and mortality crisis in Russia: regional analysis. British Medical Journal, 317, 312–318.
World Health Organization (1986) MONICA Manual, Version 1.1, December 1986, CVD/MNC. World Health Organization, Geneva.
WHO Regional Office for Europe (1992a) Healthy Cities. Guide note for the Health Cities indicators. MCAP indicators. WHO Regional Office for Europe.
WHO Regional Office for Europe (1992b) Twenty steps for developing a Health Cities project. WHO Regional Office for Europe, 4, Copenhagen.
WHO Regional Office for Europe (1993) Setting standards for WHO project cities. The requirements and the designation process for WHO project cities. WHO Healthy Cities Project Phase II 1993–1997. WHO Regional Office for Europe, 5, Copenhagen.
WHO Regional Office for Europe (1994) Healthy Cities indicators. Report of the first technical group meeting. EUR/ICP/RUD 168. WHO Regional Office for Europe, Copenhagen.
Wysocki, M. J., Krishnamurthi, C. R. and Orzeszyna, S. (1990) Monitoring the progress of Health-For-All Strategies in South-East Asia. World Health Statistics Quarterly, 43, 16–23.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||