Table 2

RRR* from the base and health insurance models

Under 65‡Over 65‡
RRRSEp>zRRRSEp>z
Uncontrolled hypertension
Base model§
 England11
 USA0.830.100.110.830.060.02
Insurance model¶ 
 US Private insurance11
 US Government insurance0.960.200.861.020.090.80
 US no insurance1.540.380.07NANANA
 England1.250.150.071.210.100.03
Undiagnosed hypertension† 
Base model§
 England11
 USA0.570.07<0.00010.500.05<0.0001
Insurance model¶ 
 US Private insurance11
 US Government insurance0.320.130.011.110.150.45
 US no insurance1.350.370.27NANANA
 England1.660.22<0.00012.140.24<0.0001
  • *RRR are the exponential of the coefficients in the multinomial logistic regression. They give the relative risk of switching between explanatory variable categories on being in a category of the response variable compared to the reference category. For example, in the insurance model (under the age of 65), the relative risk ratio of switching from US private insurance to US Government insurance is 0.32 for being in the undiagnosed hypertension category versus the controlled hypertension category (reference). In other words, the expected risk of staying in the undiagnosed hypertension category (as opposed to the reference of controlled hypertension) is lower for respondents with US Government insurance than with US Private insurance.

  • †The reference category is controlled hypertension.

  • ‡Models are fitted for the under 65 and the 65 and over age groups separately.

  • §The base model includes explanatory variables of age, sex, country, ethnicity (white, US Non-White, England Non-White), BMI.

  • ¶The health insurance model includes explanatory variables of age, sex, wealth, ethnicity (white, US Non-White, England Non-White), BMI and insurance status (US private insurance, US Government insurance, US no insurance and England (NHS)).

  • BMI, body mass index; NA, not applicable; NHS, National Health Service; RRR, relative risk ratios.