--- /dev/null
+<html>\r
+\r
+<head>\r
+<meta http-equiv="Content-Type"\r
+content="text/html; charset=iso-8859-1">\r
+<title>Computing global mortality using IMaCh</title>\r
+</head>\r
+\r
+<body bgcolor="#FFFFFF">\r
+\r
+<hr size="3" color="#EC5E5E">\r
+\r
+<p align="center"><font color="#004080" size="5">Estimation of\r
+the force of mortality independently of the initial health status\r
+using IMaCh</font></p>\r
+\r
+<p align="center"><font color="#004080" size="5">May 2004</font></p>\r
+\r
+<hr size="3" color="#EC5E5E">\r
+\r
+<ul>\r
+ <li><a href="#intro"><font color="#004080">Introduction</font></a></li>\r
+ <li><a href="#math"><font color="#004080">Mathematical\r
+ modelisation of the age-specific mortality</font></a></li>\r
+ <li><a href="#param"><font color="#004080">The parameter file</font></a></li>\r
+ <li><a href="#graph"><font color="#004080">Output file and\r
+ graph</font></a></li>\r
+</ul>\r
+\r
+<hr>\r
+\r
+<p><font color="#004080" size="4">Introduction</font><a\r
+name="intro"></a></p>\r
+\r
+<p><font\r
+color="#000000">Estimating mortality from the American LSOA\r
+surveys or from the French HID survey or from many recent\r
+cross-longitudinal surveys in various countries is neither easy\r
+nor accurate with classical demographical tools. Even if dates of\r
+death are checked with data from vital statistics and are of good\r
+quality the samples are often too small to be divided into\r
+subgroups. Also it is mandatory to estimate the mortality of\r
+subgroups or even of the whole sample if it suspected of biases\r
+in comparison with national mortality estimates.Using IMaCh\r
+(0.96d) we are able to estimate status based age specific forces\r
+of mortality and to derive global mortality by weighting them\r
+according to age-specific cross-sectional prevalences but here we\r
+are interested in estimating mortality directly i.e. without\r
+specifying any health status. This is obviously a simple problem\r
+but which is not so easy to solve because human mortality varies\r
+exponentially with age and age must be controlled accurately.We\r
+adapted a special program previously used to estimate mortality\r
+from centenarians surveys to implement a survival model which\r
+takes into account the exact duration between the first interview\r
+and the death if the person died before the last interview or the\r
+exact duration between the first and the last interview if the\r
+person is still alive.We included this program as a new feature\r
+of the IMaCh program version 0.97. </font></p>\r
+\r
+<hr>\r
+\r
+<p><font color="#004080" size="4">Mathematical modelisation of the\r
+age-specific mortality</font><a name="math"></a></p>\r
+\r
+<p><font color="#000000" size="3">The force of mortality is parametrized\r
+as a Gompertz fonction mu(x)=a exp(b*x) where x is age and a and\r
+b are the parameters. The model implemented in IMaCh is detailed\r
+in the pdf file </font><a href="docmortweb.pdf"><font\r
+color="#000000" size="3">docmortweb.pdf</font></a></p>\r
+\r
+<hr>\r
+\r
+<p><font color="#004080" size="4">The parameter file</font><a\r
+name="param"></a></p>\r
+\r
+<p><font color="#000000" size="3">The parameter file should be\r
+the same as the maximisation. The estimation of the global\r
+mortality is obtain when mle= -3. You can also choose the waves\r
+(firstpass and lastpass), number of observation (lastobs) and to\r
+add weights (weights).</font></p>\r
+\r
+<p><font color="#FF0000" size="3">Example of parameter file : </font><a\r
+href="mortparam.imach"><font color="#FF0000" size="3">mortparam.imach</font></a></p>\r
+\r
+<hr>\r
+\r
+<p><font color="#004080" size="4">Output file and graph</font><a\r
+name="graph"></a></p>\r
+\r
+<p><font color="#000000">When the run is finished, the user\r
+should enter the caracter 'e' to get the results in a htm file.\r
+This file contains the two parameters with confidence interval\r
+and a graph of the age-specific mortality obtained with the\r
+estimated parameters.</font></p>\r
+\r
+<p><font color="#FF0000" size="3">Example of output file :</font><font\r
+color="#FF0000"> </font><a href="mortparam-mort.htm"><font\r
+color="#FF0000">mortparam-mort.htm.</font></a></p>\r
+\r
+<ul>\r
+ <li><font color="#000000">Results : In this example, the two\r
+ parameters of the Gompertz fit (with confidence interval\r
+ in brackets) are<br>\r
+ <br>\r
+ p[1] = 0.026565 [0.022327 ; 0.030802]<br>\r
+ p[2] = 0.087517 [0.075484 ; 0.099550]<br>\r
+ <br>\r
+ So that the Gompertz equation modelling the force of\r
+ mortality expressed in years is :</font></li>\r
+</ul>\r
+\r
+<pre><font color="#000000"> </font><font color="#000000"\r
+size="4">mu(age) =0.026565*exp(0.087517*(age-71))</font><font\r
+color="#000000">\r
+</font></pre>\r
+\r
+<ul>\r
+ <li><font color="#000000">Graph : The figure </font><a\r
+ href="graphmort.png"><font color="#000000">graphmort.png</font></a><font\r
+ color="#000000"> represents the age-specific mortality\r
+ mu(age) according to the previous equation.</font></li>\r
+</ul>\r
+\r
+<p><font color="#004080"><img src="graphmort.gif" width="415"\r
+height="311"></font></p>\r
+\r
+<hr>\r
+\r
+<p> </p>\r
+</body>\r
+</html>\r
--- /dev/null
+<body>\r
+<title>IMaCh rmortparam.txt</title>\r
+ <font size="2">Imach version 0.97b, May 2004, INED-EUROREVES <br> $Revision$ $Date$</font> <hr size="2" color="#EC5E5E"> \r
+Title=Mortality <br>Datafile=data1.txt Firstpass=2 Lastpass=4 Stepm=1 Weight=1 Model=.<br>\r
+\r
+<hr size="2" color="#EC5E5E"> <ul><li><h4>Parameter files</h4>\r
+ - Copy of the parameter file: <a href="ormortparam.txt">ormortparam.txt</a><br>\r
+ - Log file of the run: <a href="mortparam.log">mortparam.log</a><br>\r
+ - Gnuplot file name: <a href="mortparam-mort.gp">mortparam-mort.gp</a><br>\r
+ - Date and time at start: Tue Jun 29 12:04:57 2004\r
+</ul>\r
+\r
+<br>Total number of observations=8000 <br>\r
+Youngest age at first (selected) pass 71.75, oldest age 95.08<br>\r
+Interval (in months) between two waves: Min=1 Max=38 Mean=23.24<br>\r
+<ul><li><h4>Result files </h4>\r
+ Force of mortality. Parameters of the Gompertz fit (with confidence interval in brackets):<br> mu(age) =0.026565*exp(0.087517*(age-71)) per year<br><br> p[1] = 0.026565 [0.022327 ; 0.030802]<br>\r
+ p[2] = 0.087517 [0.075484 ; 0.099550]<br>\r
+<br><br><img src="graphmort.png"></ul><br>Local time at start Tue Jun 29 12:04:57 2004\r
+<br>Local time at end Tue Jun 29 12:05:13 2004\r
+<br>
\ No newline at end of file
--- /dev/null
+#Imach version 0.97, May 2004, INED-EUROREVES \r
+title=Mortality datafile=data1.txt lastobs=8000 firstpass=2 lastpass=4\r
+ftol=1.000000e-009 stepm=1 ncovcol=2 nlstate=2 ndeath=1 maxwav=4 mle=-3 weight=1\r
+model=.\r
+# Parameters nlstate*nlstate*ncov a12*1 + b12 * age + ...\r
+12 0. 0.\r
+13 0. 0.\r
+21 0. 0.\r
+23 0. 0.\r
+# Scales (for hessian or gradient estimation)\r
+12 0. 0.\r
+13 0. 0.\r
+21 0. 0.\r
+23 0. 0.\r
+# Covariance matrix \r
+# 121 Var(a12)\r
+# 122 Cov(b12,a12) Var(b12)\r
+# ...\r
+# 232 Cov(b23,a12) Cov(b23,b12) ... Var (b23)\r
+#121 Var(a12)\r
+#122 Cov(b12,a12) Var(b12)\r
+#131 Cov(a13,a12) Cov(a13,b12) Var(a13)\r
+#132 Cov(b13,a12) Cov(b13,b12) Cov(b13,a13) Var(b13)\r
+#211 Cov(a21,a12) Cov(a21,b12) Cov(a21,a13) Cov(a21,b13) Var(a21)\r
+#212 Cov(b21,a12) Cov(b21,b12) Cov(b21,a13) Cov(b21,b13) Cov(b21,a21) Var(b21)\r
+#231 Cov(a23,a12) Cov(a23,b12) Cov(a23,a13) Cov(a23,b13) Cov(a23,a21) Cov(a23,b21) Var(a23)\r
+#232 Cov(b23,a12) Cov(b23,b12) Cov(b23,a13) Cov(b23,b13) Cov(b23,a21) Cov(b23,b21) Cov(b23,a23) Var(b23)\r
+121 0.\r
+122 0. 0.\r
+131 0. 0. 0. \r
+132 0. 0. 0. 0. \r
+211 0. 0. 0. 0. 0. \r
+212 0. 0. 0. 0. 0. 0. \r
+231 0. 0. 0. 0. 0. 0. 0. \r
+232 0. 0. 0. 0. 0. 0. 0. 0.\r
+# agemin agemax for life expectancy, bage fage (if mle==0 ie no data nor Max likelihood).\r
+agemin=70 agemax=100 bage=70 fage=100 estepm=12\r
+begin-prev-date=1/1/1986 end-prev-date=1/1/2000 mov_average=0\r
+pop_based=1\r
+prevforecast=1 starting-proj-date=1/6/1988 final-proj-date=1/1/2004 mobil_average=0\r
git://henry.ined.fr / .git / commitdiff
