Table 1.
Characteristics of selected cohort studies on association between income inequality and mortality
| Details of study | Age (years) | Follow-up (years) | Outcome (No of events) | Measure of income inequality | Area level variable | Adjusted variables in primary models other than age and sex | |
|---|---|---|---|---|---|---|---|
| Eligible papers | |||||||
| Osler et al, 2002w1 | Copenhagen City Heart Study, Glostrup Population Study (CCHS/GPS) 1976-8/1964-92 (n=28 131), Copenhagen, Denmark | ≥20 | 3-28 | All cause mortality, confirmed by national population register (n=7567) | Median share* | 149 parishes | Income |
| Blomgren et al, 2004w2 | Census 1990 (n=1.08 million men),* Finland | 25-64 | 6 | Alcohol related disease mortality, confirmed by death register (n=9820) | Gini | 84 NUT4 regions | Income, education, occupational status, and mother tongue |
| Kravdal, 2008w3 | Census 1980-2002 (n=54.31 million), Norway | 30-79 | 1-22 | All cause mortality, confirmed by population database (n=513 746) | Gini | 431 municipalities | Income, education, mean area income, and data year |
| Blakely et al, 2003w4 | Census 1991 (n=1 391 118), New Zealand | 25-64 | 3 | All cause mortality, confirmed by mortality record (n=19 128) | Gini | 35 sub-regions | Income, mean area income, and rural residency |
| Henriksson et al, 2006w5 | Census 1990 (n=1 578 186), Sweden | 40-64 | 2-7 | All cause mortality, confirmed by national cause of death register (n=49 782) | Gini | 170 parities/municipalities | Occupational position |
| Gerdtham and Johannesson, 2004w6 | Survey of Living Conditions 1980-6 (n=41 006), Sweden | 20-84 | 10-16 | All cause mortality, confirmed by national cause of death register (n=6725, 16.4% of total) | Gini | 24 counties/284 municipalities | No of children, immigrant, marital status, income, education, employment status, functional limitations, self rated health, high blood pressure, data year, urbanisation, and mean area income |
| Fiscella and Peter, 1997 w7/2000w8 | NHANES I Epidemiologic Follow-up Study (NHEFS) 1971-5 (n=13 280), US | 25-74 | 2-16 | All cause mortality, confirmed by medical records and death certificates (n=1992, 15% of total) | Median share* | 105 primary sampling units | Income and family size. Morbidity, depression, and baseline self rated health are adjusted only in primary model |
| Lochner et al, 2001w9 | National Health Interview Survey (NHIS) 1987-94 (n=546 888), US | 18-74 | 1-6 | All cause mortality, confirmed by the National Death Index (n=19 379) | Gini | 48 states | Race/ethnicity, marital status, income, and poverty rate |
| Backlund et al, 2007w10 | National Longitudinal Mortality Study (NLMS) 1979-85 (n=521 248), US | ≥25 | 4.75-10.75 | All cause mortality, confirmed by the National Death Index (n=19 049) | Median share* | 50 states | Household size, marital status, race, Hispanic origin, family income, education, employment status, and urbanisation |
| Studies not addressing data clustering (for sensitivity analysis only) | |||||||
| Daly et al, 1998w11 | Panel Study of Income Dynamics 1978-88 (sample size not reported), US | ≥25 | 5 | All cause mortality, reported by the next year survey (n=716) | Median share* | 50 states | Race, family size, and median area income |
| Kahn et al, 1999w12 | Cancer Prevention Study II 1982 (n=76 628 men),‡ US | 30-74 | 14 | All cause mortality, confirmed by the death certificates (n=15 439) | 90/10 ratio | 318 standard metro areas | Education |
*Median share—that is, % of income sum below median in total area income.