Table 2.
Effect of hope on healthcare utilization, expenditure and survival.
| Dependent variable | Number of planned visits/admissionsa | Planned visits/admissions expenditureb | Total expenditureb | Survivalc |
|---|---|---|---|---|
| Mean, (95% CI) | Mean, S$ (95% CI) | Mean, S$ (95% CI) | Hazard ratio, (95% CI) | |
| HHI | 0.07 (−0.28 to 0.43) | −59 (−1,421 to 1,302) | −537 (−2,018 to 943) | 0.992 (0.968 to 1.016) |
| HHI quartile (ref. 1st quartile) | ||||
| 2nd quartile | 0.46 (−5.34 to 6.25) | 7,471 (−22,614 to 37,556) | 3,182 (−29,420 to 35,784) | 1.096 (0.682 to 1.762) |
| 3rd quartile | −1.89 (−8.94 to 5.15) | −14,171 (−47,979 to 7,283) | −29,157* (−60,255 to 1,941) | 0.783 (0.470 to 1.306) |
| 4th quartile | 1.23 (−6.14 to 8.61) | −20,348 (−33,725 to 23,575) | −8,901 (−41,748 to 23,946) | 1.298 (0.787 to 2.142) |
| Expect to be alive in at least 2 more years | 6.60** (0.90 to 12.30) | 11,112 (−19,026 to 41,249) | 3,320 (−34,390 to 41,030) | 0.592** (0.355 to 0.987) |
| Believe primary intent of treatment is to become cancer-free | 2.76 (−1.57 to 7.10) | 3,069 (−17,383 to 23,522) | 11,147 (−12,016 to 34,310) | 1.268 (0.898 to 1.792) |
| Number of observations | 179 | 179 | 179 | 193 |
Healthcare utilization and expenditures were measured in 12-month period around the survey. All models are controlled for age, gender, education, cancer site, whether cancer was detected at early stage, years since cancer diagnosis, and health rating today. CI, Confidence interval. Statistical significance denoted by *p < 0.1, **p < 0.05.
a
Average marginal effects from negative binomial regression model.
b
Average marginal effects from generalized linear model (GLM) with log link and gamma-distributed errors.
c
Hazard ratios on all-cause mortality using Cox proportional-hazards regression model stratified by leukemia/lymphoma cancer site.