BACKGROUND: The willingness of patients in the rural tropics to seek medical care at primary health care facilities is influenced by the distance they have to travel, but few studies have tried to estimate these distance effects. METHODS: Distance decay effects in attendance rates were estimated from a database of 4348 attendances at a rural health centre in Papua New Guinea, linked to demographic and house position data for the catchment population. Small-scale spatial patterns and differences between diagnoses, age groups and gender are described. RESULTS: Attendance decreased markedly with distance both overall (50% decrease at 3.5 km) and for patients with malaria or acute respiratory infections. This decrease was non-linear (on log scale) with distance. Although constant over time, there were big differences in this distance effect among age and gender groups: Female patients showed less distance decay in adolescents and adults, but higher in the infant group. Spatial patterns accounted for 32% of the variation in age- and gender-specific attendance rates. Of the spatial effects more than 50% were due to distance effects. CONCLUSIONS: Distance effects were similar in magnitude to those reported elsewhere, suggesting that distance effects may be generalizable to many parts of the rural tropics. The non-linearity of distance decay implies that a bell-shaped demand function should be used in health planning.