# Table 5 Influencing factors of TB health workers per 10 000 population—OLS model

Variables Dependent variable: TB health workers per 10 000 population
Old system 0.494*** (0.141) 0.432*** (0.114)
Rural − 0.115 (0.149) 0.180 (0.125)
Missing TB health workers 0.177 (0.222) − 0.113 (0.151)
Log (inpatient visit) − 0.133 (0.105) 0.041 5 (0.041 9)
Log (outpatient visit) 0.246** (0.093 6)
Log (square) − 0.013 2 (0.091 6) − 0.006 04 (0.074 1)
Log (population) − 0.463*** (0.132) − 0.239*** (0.082 8)
Log (GDP per capita in 2015) 0.055 8 (0.121) − 0.013 5 (0.099 3)
Constant 5.417*** (1.885) 3.490** (1.364)
Observations 101 138
R-squared 0.299 0.202
1. Old system: = 1 if the county still had the traditional CDC/TB dispensary model and = 0 if the service delivery model in the county was transformed
2. Rural: = 1 if the county is a rural area and = 0 if the county is an urban area
3. Missing TB health workers: = 1 if more than one hospital in the county did not report the number of their TB health workers
4. Log (outpatient visit)/log (inpatient visit)/log (square)/log (population)/log (GDP per capita in 2015): natural logarithm of the total number of the outpatient visit/inpatient visit/total square land area/permanent population/GDP per capita in 2015 within the county
5. Standard errors in parentheses. There are 138 counties in three provinces. However, the data of hospitalization rate of TB patients is available only in 101 counties
6. ***p < 0.01, **p < 0.05