Table 3.
Multivariable associations between patient characteristics and employment-related changes in health insurance coverage in a sample of employed patients with cancer who received assistance from a national non-profit (Oct – Nov 2019).
| VARIABLES | Change in Health Insurance1 | |||
|---|---|---|---|---|
| Clinical Characteristics Only2 | Clinical & Sociodemographic Characteristics3 | |||
| Average Marginal Effect4 | 95% Confidence Interval | Average Marginal Effect4 | 95% Confidence Interval | |
| Observations | 619 | 619 | ||
| Race/Ethnicity (ref = NH White) | ||||
| NH Black | 0.093 | (-0.007 – 0.193) | 0.070 | (-0.025 – 0.166) |
| Hispanic/Latinx | 0.100 | (-0.030 – 0.230) | 0.050 | (-0.069 – 0.169) |
| Other | 0.108 | (-0.062 – 0.278) | 0.131 | (-0.035 – 0.297) |
| Not reported | -0.009 | (-0.170 – 0.153) | 0.010 | (-0.153 – 0.173) |
| Clinical Characteristics | ||||
| Cancer Stage at Diagnosis (ref = Stage 1 or 2) | ||||
| Stage 3 or 4 | 0.12 | (0.040 – 0.200) | 0.103 | (0.026 – 0.180) |
| Unknown stage | 0.056 | (-0.044 – 0.155) | 0.044 | (-0.053 – 0.142) |
| Cancer Site (ref = Solid tumor) | ||||
| Blood | 0.195 | (0.084 – 0.306) | 0.179 | (0.072 – 0.287) |
| Not reported | 0.094 | (-0.052 – 0.240) | 0.059 | (-0.077 – 0.196) |
| Age at Diagnosis (ref = 19-40 years old) | ||||
| 41 – 60 years | -0.008 | (-0.108 – 0.092) | -0.051 | (-0.149 – 0.046) |
| 61 years or older | -0.153 | (-0.265 – -0.041) | -0.132 | (-0.247 – -0.016) |
| Time Since Diagnosis (ref = < 1 year ago) | ||||
| 1 to 4 years ago | 0.147 | (0.052 – 0.242) | 0.146 | (0.051 – 0.241) |
| 5 or more years ago | 0.139 | (0.038 – 0.240) | 0.116 | (0.017 – 0.215) |
| Sociodemographic Characteristics | ||||
| Gender (ref = Female) | ||||
| Male | 0.101 | (-0.005 – 0.206) | ||
| Marital Status (ref = Married, Living with Partner) | ||||
| Single | 0.088 | (0.005 – 0.171) | ||
| Divorced, Widowed, or separated | 0.123 | (0.036 – 0.211) | ||
| Employment Status at Diagnosis (ref = Part-time) | ||||
| Full-time | 0.141 | (0.039 – 0.243) | ||
| Educational Attainment (ref = 2 year degree or less) | ||||
| College degree (BA/BS) or more | -0.068 | (-0.140 – 0.003) | ||
| Other or not reported | -0.23 | (-0.357 – -0.102) | ||
| Insurance Status at Diagnosis (ref = Private Insurance) | ||||
| Public (Medicare, Medicaid, Military) | -0.233 | (-0.320 – -0.145) | ||
| Uninsured | -0.176 | (-0.288 – -0.064) | ||
| Other or not reported | -0.031 | (-0.208 – 0.146) | ||
1Did the change to your employment status impact your insurance coverage? Yes vs. No/Not sure (referent).
2The first column includes results from a multivariable model controlling for clinical characteristics only, specifically cancer site, stage and age at diagnosis, and time since diagnosis.
3The second column includes results from a multivariable model additionally controlling for sociodemographic characteristics, specifically gender, marital status, employment status at diagnosis, educational attainment, and insurance status at diagnosis.
4Multivariable logistic regression used to estimate average marginal effects (95% confidence intervals reported in parentheses). Average marginal effects represent the average difference in the predicted probability of experiencing income disruption, or a change in health insurance, holding all other covariates constant, across all observations in the analytic sample.