Correlation and Regression

Why would an increase in doctors reduce death rates in the community (deaths/ 1,000 people)?

2) What reasons can you think of that as number of doctors/ 10,000 people increases deaths might increase (note the upswing with >/= 3 doctors per 10,000 people)?

3) What other information could be collected to explain the increase of deaths with an increase of doctors?