Trends in geographic distribution of nursing staff in Japan from 2000 to 2010: a multilevel analysis

The aim of this study was to examine trends in the geographic distribution of nursing staff in Japan from 2000 to 2010. We examined time trends in the rates of nursing staff per 100,000 population across 349 secondary health service areas. Using the Gini coefficient as a measure of inequality, we separately analyzed the data of 4 nursing staff types:public health nurses (PHN), midwives (MW), nurses (NS), and associate nurses (AN). Then, using multilevel Poisson regression models, we calculated the rate ratios (RRs) and their 95% confidence intervals (CIs) for each type of nursing staff per 1-year change. Overall, the distribution of PHN, MW, and NS improved slightly in terms of the Gini coefficient. After adjusting for prefectural capital and population density, PHN, MW, and NS significantly increased;the RRs per 1-year increment were 1.022 (95% CI:1.020-1.023), 1.021 (95% CI:1.019-1.022), and 1.037 (95% CI:1.037-1.038), respectively. In contrast, AN significantly decreased;the RR per 1-year increment was 0.993 (95% CI:0.993-0.994). Despite the considerable increase in the absolute number of nursing staff in Japan (excluding AN), this increase did not lead to a sufficient improvement in distribution over the last decade.