> 8:7{ bjbjzz 4*
0008h0($t!r??T%%%%%%% 8^u%j0%!.!%%h!x%??%! : Crude and standardized measures are just weighted averages
Many overall (crude) measures in epidemiology are weighted averages (at least approximately). If we divide a population into subpopulations and calculate an epidemiologic measure for the whole population and for each of the subpopulations, then it is often the case* that the crude measure in the whole population can be expressed as a weighted average of that measure across the subpopulations, where each subpopulation measure is weighted by the proportion it makes up of the whole population. A simpler way to state this might be:
Suppose population P (of size p, i.e., p people) is composed of three subpopulations, A (size a), B (size b), and C (size c), so that p=a+b+c.
Suppose we calculate an epidemiologic measure (e.g., mortality rate) for subgroups A, B , and C and also for the whole population, P. Then the mortality rate for P (lets call it mP) can usually be expressed as a weighted average of the mortality rates for A, B, and C (mA, mB, mC) where the weights are the subpopulation proportions:
mP = (a/p) mA + (b/p) mB + (c/p) mC
a/p is subpopulation As proportionate size in population P. So if the populations have sizes 20,000 (a), 30,000 (b), and 50,000 (c), fora total of 100,000 (p), then the crude mortality rate for P (mP) would be:
mP = (20,000/100,000) mA + (30,000/100,000) mB + (50,000/100,000) mC
Note that 20,000/100,000 + 30,000/100,000 + 50,000/100,000 = 100,000/100,000 = 1, as they must for a weighted average, so the formula could be written:
mP = 0.2mA + 0.3mB + 0.5mC
If the mortality rates for A, B, and C are mA=40 per 1,000/year, mB=30 per 1,000/year, mC=10 per 1,000/year, then the crude mortality rate will be 0.2 x 40 + 0.3 x 30 + 0.5 x 10 = 22 per 1,000/year. (Remember to keep the units and scaling.)
You can also compute the number of deaths in each subpopulation (e.g., deaths = mA x A = 40 per 1,000/year x 20,000 people in A = 800 deaths in subpopulation A). Summing these numbers across the subpopulations yields the total number of deaths in the entire population, which can then be used to compute the crude mortality rate as the total number of deaths (2,200) divided by 100,000/year.
The above numerical example illustrates that the crude mortality rate can be expressed as a weighted average of the subpopulation-specific mortality rates weighted by subpopulation sizes. The subpopulations can be age groups, gender groups (female, male, prefer not to state), etc. The larger the proportionate size of the subpopulation, the more the crude rate reflects that subpopulations rate.
Because the crude mortality rate (as well as the crude birth rate, general fertility rate, crude rates of reportable diseases, etc.) is influenced by the composition of the population, comparisons across populations will reflect differences in population composition (e.g., age distribution). If that influence would be a distraction, as it often is, then we want to use measures that are not as affected by population composition.
We can use subpopulation-specific measures (e.g., mortality rate for age 45-49 year olds, age-specific fertility rate), but that could make for many separate comparisons. For a summary comparison we want a measure that reduces the influence of population composition. Such a measure is called adjusted (e.g., age-adjusted mortality rate).
One commonly used type of adjusted measure is a standardized measure. Standardization makes use of the concept that the crude measure is a weighted average of the measures across the subpopulations. But instead of letting the weights for that average be the subpopulation proportions, which differ across the populations being compared, the same set of weights is used for the standardized measure for each population compared. The weights are a set of proportions that sum to 1.0, but they are often expressed as percentages or as a standard population whose distribution is used as the weights. The key is that by creating weighted averages using a single set of weights, we reduce the influence of population composition on the comparison.
An important footnote is to be aware that the relative sizes of the weights do typically influence the resulting standardized measures and the comparison. For example a standard population in which group A predominates will give a different result from a standard population in which group B predominates, etc. For further explanation of this point and other aspects of standardization, see the relevant chapter in Understanding the Fundamentals of Epidemiology: An Evolving Text (at www.epidemiolog.net).
Key points:
1)Overall measures are typically influenced by population composition.
2)The values of the measure for the subpopulations are the underlying realities.
3)Standardized measures are weighted averages of those underlying realities constructed with a fixed set of weights (proportions).
4)Such weighted averages (standardized measures) are therefore unaffected by differences in composition of the populations being compared, though the choice of weights (the standard population) determines which subpopulation rates are most influential in the resulting weighted average.
Victor J. Schoenbach, 6/1/2013, www.epidemiolog.net/studymat/
P.S. A familiar example of weighted averages is the typical GPA. Here is a GPA calculator I found on the web, in case youd like to experiment: http://www.unc.edu/~chendel/GPA.html There is also a spreadsheet with an age standardization example linked from the page where this was linked.
:;<KPG H l
09FkBCMNXYK
L
a
b
w
x
"#$%,-./(r{}7RYZʻh^CJOJQJaJh[CJOJQJaJh&TCJOJQJaJh$h$5CJOJQJaJh$h$CJOJQJaJ h$h[CJOJPJQJaJ#h[h$5CJOJPJQJaJ#h[h[5CJOJPJQJaJ2;<T
V
68\^4
6
{
}
68)+CENP$a$gd[gd$BEFe
hOlx~;Lg})9u8:GIO奕zh!CJOJQJaJh$h$CJOJQJaJhk]hk]6CJOJQJaJhk]CJOJQJaJh$h$5CJOJQJaJhvCJOJQJaJhCJOJQJaJhlCJOJQJaJh$h$CJOJQJaJh CJOJQJaJ0P;=9;Gx
gd gd(6F$a$gdc^`gd$gd$zKLX_p h&jh&Uhch(6FCJh(6Fh(6FCJ
hM]CJ
h(6FCJ
hj^8CJ
hcCJhchcCJh(6FCJOJQJaJh$h$CJOJQJaJh!CJOJQJaJh$h$CJOJQJaJ21h:pq/ =!"#$%j666666666vvvvvvvvv666666>6666666666666666666666666666666666666666666666666hH6666666666666666666666666666666666666666666666666666666666666666662 0@P`p2( 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p 0@P`p8XV~ OJPJQJ_HmH nH sH tH @`@NormalCJ_HaJmH sH tH DA D
Default Paragraph FontRiR
0Table Normal4
l4a(k (
0No ListH@H 0Balloon TextCJOJQJ^JaJNoN0Balloon Text CharCJOJQJ^JaJBB$apple-converted-spaceV@"V$ List Paragraphdd[$\$CJOJQJaJ6U`16 $0 Hyperlink>*B*ph4B4 0Header
H$.Q. 0Header Char4 b4 0Footer
H$.q. 0Footer CharPK![Content_Types].xmlN0EH-J@%ǎǢ|ș$زULTB l,3;rØJB+$G]7O٭V$!)O^rC$y@/yH*)UDb`}"qۋJחX^)I`nEp)liV[]1M<OP6r=zgbIguSebORD۫qu gZo~ٺlAplxpT0+[}`jzAV2Fi@qv֬5\|ʜ̭NleXdsjcs7f
W+Ն7`gȘJj|h(KD-
dXiJ؇(x$(:;˹!I_TS1?E??ZBΪmU/?~xY'y5g&/ɋ>GMGeD3Vq%'#q$8K)fw9:ĵ
x}rxwr:\TZaG*y8IjbRc|XŻǿI
u3KGnD1NIBs
RuK>V.EL+M2#'fi~Vvl{u8zH
*:(W☕
~JTe\O*tHGHY}KNP*ݾ˦TѼ9/#A7qZ$*c?qUnwN%Oi4=3N)cbJ
uV4(Tn
7_?m-ٛ{UBwznʜ"ZxJZp;{/<P;,)''KQk5qpN8KGbe
Sd̛\17 pa>SR!
3K4'+rzQ
TTIIvt]Kc⫲K#v5+|D~O@%\w_nN[L9KqgVhn
R!y+Un;*&/HrT >>\
t=.Tġ
S; Z~!P9giCڧ!# B,;X=ۻ,I2UWV9$lk=Aj;{AP79|s*Y;̠[MCۿhf]o{oY=1kyVV5E8Vk+֜\80X4D)!!?*|fv
u"xA@T_q64)kڬuV7t'%;i9s9x,ڎ-45xd8?ǘd/Y|t&LILJ`& -Gt/PK!
ѐ'theme/theme/_rels/themeManager.xml.relsM
0wooӺ&݈Э5
6?$Q
,.aic21h:qm@RN;d`o7gK(M&$R(.1r'JЊT8V"AȻHu}|$b{P8g/]QAsم(#L[PK-![Content_Types].xmlPK-!֧60_rels/.relsPK-!kytheme/theme/themeManager.xmlPK-!0C)theme/theme/theme1.xmlPK-!
ѐ' theme/theme/_rels/themeManager.xml.relsPK]
* PL#@0(
B
S ?8:NPY[&(68bdxzy{[_
=@8:^`68
33333GH;yO3;|Oq/1$xU`.5C%;|A_eitK
R+4=tK
>!yO3#'.&D;[RY`.5#-*s :&fLi5!xU$;[ddC%#--~%7,T0)kA.)"{v@*ip)37,VQPE.#-M0;[/1 Cip)3tK
C8qb9@qb9*?Y(=xU+4=Iem>!kA+CknD CxU'.&D/1knDkAVQ_PV&fN+^xU_Y~YagagaRWHd+4={5e;|A_e+C-fxU&f}x7^m;[IemPVgBoxU#pYl=5rxUv-Mv>"{v+C}x7,6w}#p8Q`.59&T!{T!&j^8ZB(6FM]k]^xhqB~l[ $lcEqv@____@Unknown G*Ax Times New Roman5Symbol3.*Cx Arial3*Ax Times7.@Calibri7@CambriaG=
jMS Mincho-3 fgC
PLucida GrandeA$BCambria Math"qhH'H'a20KQKP $P$2!xx,kapuaola gellertVictor J. SchoenbachOh+'0t
0<
HT\dlkapuaola gellertNormal.dotmVictor J. Schoenbach4Microsoft Office Word@Ik@^@-^H՜.+,0hp|
unc'Title
!"#$%&()*+,-.01234569Root Entry F 8^;1Table!WordDocument4*SummaryInformation('DocumentSummaryInformation8/CompObjr
F Microsoft Word 97-2003 Document
MSWordDocWord.Document.89q