Abstract
Static and dynamic analyses were used for investigating the acute relationship among acute respiratory disease, meteorologic condition and air pollution. Daily incidence of the disease was obtained from the records of the National Health Insurance regarding those who lived in two small areas in Tokyo where air monitoring stations were located. Reliability of the incidence estimated was proved by a preliminary study.
After eliminating the variations due to day of week, stepwise regression analysis revealed consistent relation of temperature to the incidence and a spectral analysis showed the probable chain connecting them to be a second-order differential system analogous to protection taken against temperature spell. Dermal disease as a control showed no relation to air pollutants.
Among pollutants, only NO and its associates showed a relatively constant relation to the incidence. A spectral analysis indicated that the incidence associated with the change of NO but not with NO itself, which was not contradictory to regression analysis because NO, NO2, NOx or Oxidant were entered when the lag varied.
A further study will be needed to modify the probable error due to deviation from ‘true’ date of disease occurrence or unsuccessive air measurement, to specify the most probable agent among nitrogen oxides and to determine which is more effective, average or peak concentration.
