[Consultation patterns among Japanese cedar pollinosis subjects at an urban clinic report 2: a 20-year experience (1990-2009) in Chiyoda Ward, Tokyo].

To determine the consultation patterns of Japanese cedar pollinosis subjects during the pollen dispersal season, we surveyed those treated at a private clinic in Chiyoda Ward, Tokyo, from 1990 to 2009. We also studied the relationship between subject numbers and pollen count. The above relationship correlated highly with linear regression equation y = 0.0897x + 627.47 with R² = 0.7851 (p < 0.001). The relationship from 1994 to 2009, when the study began in early January, correlated highly with logarithmic regression equation y = 257.43Ln(x) - 1014.8 with R² = 0.9542 (p < 0.001). Based on these results, we concluded that estimated pollen count helps predict the year's Japanese cedar pollinosis subject numbers very highly accurately. Numbers of subjects returning did not correlate well with pollen count, and continued to decrease each year. Return visits averaging 2.6 in 1990 gradually decreased to 0.73 in 2006--a reduction we attributed to long-term medication and the increased self-payment proportion in medical expenditures.