Meteorological conditions and incidence of Legionnaires' disease in Glasgow, Scotland: application of statistical modelling.

This study investigated the relationships between Legionnaires' disease (LD) incidence and weather in Glasgow, UK, by using advanced statistical methods. Using daily meteorological data and 78 LD cases with known exact date of onset, we fitted a series of Poisson log-linear regression models with explanatory variables for air temperature, relative humidity, wind speed and year, and sine-cosine terms for within-year seasonal variation. Our initial model showed an association between LD incidence and 2-day lagged humidity (positive, P = 0·0236) and wind speed (negative, P = 0·033). However, after adjusting for year-by-year and seasonal variation in cases there were no significant associations with weather. We also used normal linear models to assess the importance of short-term, unseasonable weather values. The most significant association was between LD incidence and air temperature residual lagged by 1 day prior to onset (P = 0·0014). The contextual role of unseasonably high air temperatures is worthy of further investigation. Our methods and results have further advanced understanding of the role which weather plays in risk of LD infection.

Spatial modelling and the prediction of Loa loa risk: decision making under uncertainty.

Health decision-makers working in Africa often need to act for millions of people over large geographical areas on little and uncertain information. Spatial statistical modelling and Bayesian inference have now been used to quantify the uncertainty in the predictions of a regional, environmental risk map for Loa loa (a map that is currently being used as an essential decision tool by the African Programme for Onchocerciasis Control). The methodology allows the expression of the probability that, given the data, a particular location does or does not exceed a predefined high-risk threshold for which a change in strategy for the delivery of the antihelmintic ivermectin is required.

Analysing spatially referenced public health data: a comparison of three methodological approaches.

In the analysis of spatially referenced public health data, members of different disciplinary groups (geographers, epidemiologists and statisticians) tend to select different methodological approaches, usually those with which they are already familiar. This paper compares three such approaches in terms of their relative value and results. A single public health dataset, derived from a community survey, is analysed by using 'traditional' epidemiological methods, GIS and point pattern analysis. Since they adopt different 'models' for addressing the same research question, the three approaches produce some variation in the results for specific health-related variables. Taken overall, however, the results complement, rather than contradict or duplicate each other.

Predicting malaria infection in Gambian children from satellite data and bed net use surveys: the importance of spatial correlation in the interpretation of results.

In line with the renewed World Health Organization Global Malaria Control Strategy, we have advocated the use of satellite imagery by control services to provide environmental information for malaria stratification, monitoring, and early warning. To achieve this operationally, appropriate methodologies must be developed for integrating environmental and epidemiologic data into models that can be used by decision-makers for improved resource allocation. Using methodologies developed for the Famine Early Warning Systems and spatial statistics, we show a significant association between age related malaria infection in Gambian children and the amount of seasonal environmental greenness as measured using the normalized difference vegetation index derived from satellite data. The resulting model is used to predict changes in malaria prevalence rates in children resulting from different bed net control scenarios.

Do new cases of rheumatoid arthritis cluster in time or in space?

OBJECTIVES: To examine for evidence of clustering in time, in space and in space/time in the occurrence of rheumatoid arthritis (RA). SETTING: A population-based incidence register of RA in the East Anglian region of the UK: population size 413,000. SUBJECTS: In all 687 new cases of inflammatory joint disease registered between 1 January 1990 and 31 December 1994 were studied. Population data were obtained from postcode areas by age and sex. ANALYSIS: Time trend analysis was conducted over the first 36 months and observed and expected distributions compared. Spatial clustering was based on comparison of observed distribution using map grid references to random expectation based on simulation. A similar procedure was undertaken for time/space clustering. RESULTS: There was no evidence of a time trend. There was only modest evidence of spatial clustering with non-random distribution observed in one area but there was no evidence of time/space clustering. CONCLUSION: Although a viral aetiology is the strongest candidate for RA, no evidence of a localized event in time was associated with disease development in this population.

Pinpointing clusters of apparently sporadic cases of Legionnaires' disease.

OBJECTIVES: To test the hypothesis that many non-outbreak cases of legionnaires' disease are not sporadic and to attempt to pinpoint cases clustering in space and time. DESIGN: Descriptive study of a case series, 1978-86. SETTING: 15 health boards in Scotland. PATIENTS: 203 probable cases of non-outbreak, non-travel, community acquired legionnaires' disease in patients resident in Scotland. MAIN MEASURES: Date of onset of disease and postcode and health board of residence of cases. RESULTS: Space-time clustering was present and numerous groups of cases were identified, all but two being newly recognised. Nine cases occurred during three months within two postcodes in Edinburgh, and an outbreak was probably missed. In several places cases occurred in one area over a prolonged period--for example, nine cases in postcode districts G11.5 and G12.8 in Glasgow during five years (estimated mean annual incidence of community acquired, non-outbreak, non-travel legionnaires' disease of 146 per million residents v 4.8 per million for Scotland). Statistical analysis showed that the space time clustering of cases in the Glasgow and Edinburgh areas was unusual (p = 0.036, p = 0.068 respectively). CONCLUSION: Future surveillance requires greater awareness that clusters can be overlooked; case searching whenever a case is identified; collection of complete information particularly of date of onset of the disease and address or postcode; ongoing analysis for space-time clustering; and an accurate yet workable definition of sporadic cases. Other researchers should re-examine their data on apparently sporadic infection.