Economic burden of obesity and its complications in Germany

Abstract

This study estimated the economic burden of illness of obesity and selected comorbidities in terms of health outcome and costs to society and healthcare payer in Germany. The proportions of selected diseases (comorbidities) attributable solely to obesity were estimated using odds ratios/relative risks and prevalences based on data from the literature. The top-down approach was employed to match healthcare spending with the number of patients suffering from obesity (BMI 30+) and the major comorbidities to evaluate overall direct and indirect costs. In Germany there are approximately 12.24 million obese adults, 2.06–3.76 million of whom suffer from the selected comorbidities. From a societal perspective the total costs for obesity and comorbidities are €2,701–5,682 million per year and the direct treatment costs alone account for €1,343–2,699 million, imposing a major burden to the healthcare system. In view of the magnitude of the economic burden of illness there is a need for both further research and action at the health policy level.

Appendix: calculation of exposure-dependent prevalence

Given:

• T fields N, e, f, g, h in Table 5

  Table 5. Calculation of exposure-dependent prevalence

• Odds ratio (OR)

Only one of the fields a, b, c, or d needs to be calculated to estimate all the others:

$$ {OR = odds\;(disease\;\& \;\exp osed)/odds\;(disease\;\& \;non - \exp osed)} $$

$$ OR = {{a/b} \over {c/d}} = {{a*d} \over {b*c}} $$

(1)

Equation 1 needs to be substituted to estimate a:

$$ \matrix{ {a + b = e} & \to & {b = e - a} \cr } $$

(2)

$$ \matrix{ {c + d = f} \hfill & \to \hfill & {d = f - c} \hfill & \to \hfill & {d = f - g + a} \hfill \cr } $$

(3)

$$ \matrix{ {a + c = g} & \to & {c = g - a} \cr } $$

(4)

Substitution of Eq. 1 with Eqs. 2, 3, and 4 leads to:

$$ OR = {{a(f - g + a)} \over {(e - a)*(g - a)}} $$

Rearranging:

$$ OR = {{af - ag + a^2 } \over {eg - ae - ag + a^2 }} $$

$$ {OR = {\left( {a^{2} + (f - g)a} \right)}/{\left( {a^{2} + ( - e - g)a + eg} \right)}} $$

$$ a^2 + (- e - g)a + eg = {\textstyle{1 \over {OR}}}a^2 + {\textstyle{{(f - g)} \over {OR}}}a $$

$$ {{\left( {1 - 1/OR} \right)}a^{2} + {\left[ {{\left( { - e - g} \right)} - (f - g)/OR} \right]}a + eg = 0} $$

$$ \left({1 - {\textstyle{1 \over {OR}}}} \right)a^2 + \left[ {\left({ - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \right]a = - eg $$

$$ a^2 + {{\left[ {\left( { - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \right]} \over {1 - {\textstyle{1 \over {OR}}}}}a = - {{eg} \over {1 - {\textstyle{1 \over {OR}}}}} $$

$$ a^2 + {{\left[ {\left({ - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \right]} \over {1 - {\textstyle{1 \over {OR}}}}}a = - {{eg} \over {1 - {\textstyle{1 \over {OR}}}}} $$ <![CDATA[ $$ a^2 + {{\left[ {\left({ - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \right]} \over {1 - {\textstyle{1 \over {OR}}}}}a + {{eg} \over {1 - {\textstyle{1 \over {OR}}}}} = 0 $$

→ quadratic equation: x ²+px+q=0 where:

$$ x = a $$

(5)

$$ p = {{\left[ {\left({ - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \right]} \over {1 - {\textstyle{1 \over {OR}}}}} $$

(6)

$$ q = {{eg} \over {1 - {\textstyle{1 \over {OR}}}}} $$

(7)

→ solution to quadratic equation: $x = - {\textstyle{p \over 2}} \pm \sqrt {{\textstyle{{p^2 } \over 4}} - q}$→ substitute with Eqs. 5, 6, and 7:

$$ a = - {{{{\left[ {\left({ - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \right]} \over {1 - {\textstyle{1 \over {OR}}}}}} \over 2} \pm \sqrt {\left[ {{{\left({{{\left({ - e - g} \right) - {\textstyle{{(f - g)} \over {OR}}}} \over {1 - {\textstyle{1 \over {OR}}}}}} \right)^2 } \over 4}} \right] - \left({{{eg} \over {1 - {\textstyle{1 \over {OR}}}}}} \right)} $$

The attributable proportion in the population (AP _pop ) is then calculated by:

$$ AP_{pop} = {{\left[ {\left({{\textstyle{g \over N}}} \right) - \left({{\textstyle{c \over f}}} \right)} \right]} \over {\left({{\textstyle{g \over N}}} \right)}} $$