Table 3 Logistic regression analyses of emergency hospital admissions for assaults utilising individual and residence-based characteristics for individuals resident in England admitted in 2004/05*
Adjusted odds ratio 95% CI (lower to upper)p Value
Sex
    Female (Ref)1.000
    Male5.585 (5.417 to 5.757)<0.001
Age category
    0–14 (Ref)1.000
    15–299.476 (8.999 to 9.978)<0.001
    30–445.309 (5.034 to 5.598)<0.001
    45 and over1.446 (1.365 to 1.532)<0.001
People & Places category†
    Mature Oaks (Ref)1.000
    Country Orchards1.315 (1.212 to 1.427)<0.001
    Blossoming Families1.042 (0.946 to 1.148)0.405
    Rooted Households1.578 (1.477 to 1.687)<0.001
    Qualified Metropolitans1.694 (1.553 to 1.848)<0.001
    Senior Neighbourhoods1.745 (1.611 to 1.890)<0.001
    Suburban Stability2.469 (2.319 to 2.629)<0.001
    New Starters3.135 (2.924 to 3.361)<0.001
    Multicultural Centres3.889 (3.654 to 4.138)<0.001
    Urban Producers4.238 (3.969 to 4.525)<0.001
    Weathered Communities3.876 (3.623 to 4.147)<0.001
    Disadvantaged Households6.617 (6.191 to 7.072)<0.001
    Urban Challenge7.687 (7.129 to 8.289)<0.001
    Unclassified2.896 (2.571 to 3.263)<0.001
English Regions
    North East (Ref)1.000
    North West1.118 (1.068 to 1.171)<0.001
    Yorkshire and the Humber0.857 (0.815 to 0.901)<0.001
    East Midlands0.679 (0.641 to 0.719)<0.001
    West Midlands0.654 (0.620 to 0.690)<0.001
    East of England0.618 (0.583 to 0.655)<0.001
    London0.707 (0.671 to 0.745)<0.001
    South East0.661 (0.627 to 0.697)<0.001
    South West0.810 (0.766 to 0.857)<0.001

  • Ref, Reference category; 95% CI, 95% confidence intervals.

  • *Data included all individuals admitted to hospital for assault in England in 2004/05. Data for individuals not requiring such admissions were calculated by subtraction of admitted individuals from age- and sex-specific populations in each Lower Super Output Area. Only 2004/05 data were utilised in the model to avoid pseudo-replication of population data, with 2004/05 being chosen as it was closest to the latest national census projection and consequently provides best population estimates.

  • †Due to the strong correlation between People and Places (P2) categories and Index of Multiple Deprivation (IMD; table 2), P2 categories were also utilised as proxy for deprivation and, therefore, deprivation quintiles were not entered into the model as a separate variable.