Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Mismacth between plotted restricted cubic spline and predicted HR values with 95% CI

    I am conducting a restrited cubic spline for the association between vigorous physical activity and specific mortality, but the predicted estimated values and the plotted spline do not match since there is a part of the spline with the upper bound lower than 1, which do not happen with the predicted value. I am using the Orsini procedure.

    These are the values I am getting with the xbrcspline command and below the plot for the spline.

    VPA_min exp(XB) LB UB
    0 1.28 1.06 1.55
    10 1.26 1.06 1.50
    11 1.25 1.06 1.49
    12 1.25 1.06 1.49
    13 1.25 1.05 1.48
    14 1.25 1.05 1.47
    15 1.24 1.05 1.47
    16 1.24 1.05 1.46
    17 1.24 1.05 1.46
    18 1.24 1.05 1.45
    19 1.23 1.05 1.45
    20 1.23 1.05 1.44
    22 1.23 1.05 1.43
    23 1.22 1.05 1.43
    24 1.22 1.05 1.42
    25 1.22 1.05 1.41
    26 1.22 1.05 1.41
    27 1.21 1.05 1.40
    28 1.21 1.05 1.40
    29 1.21 1.05 1.39
    30 1.21 1.05 1.39
    32 1.20 1.05 1.38
    33 1.20 1.04 1.37
    34 1.20 1.04 1.37
    35 1.19 1.04 1.36
    36 1.19 1.04 1.36
    37 1.19 1.04 1.35
    38 1.19 1.04 1.35
    39 1.18 1.04 1.34
    40 1.18 1.04 1.34
    42 1.18 1.04 1.33
    44 1.17 1.04 1.32
    45 1.17 1.04 1.32
    46 1.17 1.04 1.31
    48 1.16 1.04 1.30
    49 1.16 1.04 1.30
    50 1.16 1.04 1.29
    51 1.16 1.04 1.29
    52 1.15 1.04 1.28
    53 1.15 1.04 1.28
    54 1.15 1.03 1.27
    55 1.15 1.03 1.27
    56 1.14 1.03 1.27
    57 1.14 1.03 1.26
    58 1.14 1.03 1.26
    60 1.14 1.03 1.25
    64 1.13 1.03 1.23
    65 1.12 1.03 1.23
    66 1.12 1.03 1.22
    68 1.12 1.03 1.22
    69 1.12 1.03 1.21
    70 1.11 1.03 1.21
    72 1.11 1.03 1.20
    74 1.11 1.03 1.19
    75 1.10 1.03 1.19
    76 1.10 1.02 1.18
    77 1.10 1.02 1.18
    78 1.10 1.02 1.18
    80 1.09 1.02 1.17
    81 1.09 1.02 1.17
    82 1.09 1.02 1.16
    84 1.09 1.02 1.16
    85 1.08 1.02 1.15
    87 1.08 1.02 1.15
    88 1.08 1.02 1.14
    90 1.08 1.02 1.14
    91 1.07 1.02 1.13
    92 1.07 1.02 1.13
    93 1.07 1.02 1.13
    96 1.07 1.02 1.12
    98 1.06 1.02 1.11
    99 1.06 1.01 1.11
    100 1.06 1.01 1.10
    102 1.06 1.01 1.10
    104 1.05 1.01 1.09
    105 1.05 1.01 1.09
    106 1.05 1.01 1.09
    108 1.05 1.01 1.08
    110 1.04 1.01 1.08
    111 1.04 1.01 1.07
    112 1.04 1.01 1.07
    114 1.04 1.01 1.07
    117 1.03 1.01 1.06
    119 1.03 1.01 1.06
    120 1.03 1.01 1.05
    121 1.03 1.01 1.05
    124 1.02 1.01 1.04
    125 1.02 1.01 1.04
    126 1.02 1.01 1.04
    128 1.02 1.00 1.04
    129 1.02 1.00 1.03
    130 1.02 1.00 1.03
    132 1.02 1.00 1.03
    133 1.01 1.00 1.03
    135 1.01 1.00 1.02
    136 1.01 1.00 1.02
    138 1.01 1.00 1.02
    140 1.01 1.00 1.01
    141 1.01 1.00 1.01
    144 1.00 1.00 1.01
    145 1.00 1.00 1.01
    147 1.00 1.00 1.00
    148 1.00 1.00 1.00
    150 1.00 1.00 1.00
    152 1.00 1.00 1.00
    154 1.00 0.99 1.00
    156 1.00 0.99 1.00
    159 0.99 0.99 1.00
    160 0.99 0.99 1.00
    161 0.99 0.99 1.00
    162 0.99 0.99 1.00
    165 0.99 0.98 1.00
    168 0.99 0.98 1.00
    170 0.99 0.98 1.00
    171 0.99 0.98 1.00
    172 0.99 0.98 1.00
    174 0.99 0.97 1.00
    175 0.99 0.97 1.00
    176 0.99 0.97 1.00
    177 0.99 0.97 1.00
    180 0.99 0.97 1.00
    182 0.99 0.97 1.01
    184 0.99 0.97 1.01
    185 0.99 0.97 1.01
    188 0.99 0.96 1.01
    189 0.99 0.96 1.01
    190 0.99 0.96 1.01
    192 0.99 0.96 1.01
    195 0.99 0.96 1.01
    196 0.99 0.96 1.01
    198 0.99 0.96 1.02
    200 0.99 0.96 1.02
    203 0.99 0.95 1.02
    204 0.99 0.95 1.02
    207 0.99 0.95 1.02
    208 0.99 0.95 1.03
    210 0.99 0.95 1.03
    212 0.99 0.95 1.03
    216 0.99 0.95 1.03
    217 0.99 0.95 1.04
    220 0.99 0.94 1.04
    222 0.99 0.94 1.04
    224 0.99 0.94 1.04
    225 0.99 0.94 1.05
    228 0.99 0.94 1.05
    230 0.99 0.94 1.05
    231 0.99 0.94 1.05
    234 1.00 0.94 1.06
    235 1.00 0.94 1.06
    238 1.00 0.93 1.06
    240 1.00 0.93 1.07
    244 1.00 0.93 1.07
    245 1.00 0.93 1.07
    250 1.00 0.93 1.08
    252 1.00 0.93 1.09
    255 1.00 0.92 1.09
    259 1.01 0.92 1.10
    260 1.01 0.92 1.10
    264 1.01 0.92 1.11
    265 1.01 0.92 1.11
    266 1.01 0.92 1.11
    268 1.01 0.92 1.11
    270 1.01 0.92 1.12
    273 1.01 0.91 1.12
    275 1.01 0.91 1.13
    280 1.02 0.91 1.14
    285 1.02 0.91 1.14
    288 1.02 0.91 1.15
    290 1.02 0.91 1.15
    300 1.03 0.90 1.17
    308 1.03 0.90 1.19
    315 1.03 0.89 1.20
    318 1.04 0.89 1.21
    320 1.04 0.89 1.21
    322 1.04 0.89 1.22
    325 1.04 0.89 1.22
    329 1.04 0.88 1.23
    330 1.04 0.88 1.23
    336 1.05 0.88 1.24
    340 1.05 0.88 1.25
    345 1.05 0.87 1.26
    350 1.05 0.87 1.27
    356 1.06 0.87 1.28
    357 1.06 0.87 1.29
    360 1.06 0.87 1.29
    364 1.06 0.86 1.30
    371 1.06 0.86 1.32
    375 1.07 0.86 1.32
    378 1.07 0.86 1.33
    380 1.07 0.86 1.34
    385 1.07 0.85 1.35
    390 1.07 0.85 1.36
    392 1.08 0.85 1.36
    399 1.08 0.85 1.38
    400 1.08 0.85 1.38
    405 1.08 0.84 1.39
    413 1.09 0.84 1.41
    414 1.09 0.84 1.41
    420 1.09 0.83 1.43
    425 1.09 0.83 1.44
    432 1.10 0.83 1.45
    434 1.10 0.83 1.46
    435 1.10 0.83 1.46
    440 1.10 0.82 1.47
    441 1.10 0.82 1.47
    450 1.11 0.82 1.50
    452 1.11 0.82 1.50
    455 1.11 0.82 1.51
    462 1.11 0.81 1.53
    475 1.12 0.81 1.56
    480 1.12 0.80 1.57
    483 1.13 0.80 1.58
    490 1.13 0.80 1.60
    495 1.13 0.80 1.61
    500 1.13 0.79 1.62
    510 1.14 0.79 1.65
    520 1.15 0.78 1.68
    525 1.15 0.78 1.69
    532 1.15 0.78 1.71
    540 1.16 0.77 1.73
    546 1.16 0.77 1.75
    550 1.16 0.77 1.76
    560 1.17 0.76 1.79
    570 1.18 0.76 1.82
    585 1.18 0.75 1.87
    595 1.19 0.75 1.90
    596 1.19 0.75 1.90
    600 1.19 0.74 1.91
    615 1.20 0.74 1.96
    625 1.21 0.73 1.99
    630 1.21 0.73 2.01
    640 1.22 0.73 2.04
    660 1.23 0.72 2.11
    665 1.23 0.71 2.13
    675 1.24 0.71 2.16
    690 1.25 0.70 2.21
    700 1.25 0.70 2.25
    720 1.27 0.69 2.33



    Click image for larger version

Name:	VPA.tif
Views:	1
Size:	26.6 KB
ID:	1693813





    Code:

    mkspline VPA_min_knots = VPA_min, cubic knots(10 150 270) displayknots
    mat knots= r(knots)
    svy linearized: stcox VPA_min_knots* sex age education_cat bmi srvy_yr alcohol smoking marital_cat race limitations condition strength MPA_min
    testparm VPA_min_knots1 VPA_min_knots2
    levelsof VPA_min
    xbrcspline VPA_min_knots, values(`r(levels)') ref(150) matknots(knots)eform
    predictnl xb= _b[VPA_min_knots1]*(VPA_min_knots1-150)+ _b[VPA_min_knots2]*(VPA_min_knots2-150), ci(lo hi)

    gen hr = exp(xb)
    gen lb = exp(lo)
    gen ub = exp(hi)
    sort VPA_min


    graph tw line hr VPA_min, color(navy) ylabel(0.75 1 1.5) ///
    || rarea lb ub VPA_min, color(navy%30) ///
    yline(1) yscale(log) ytitle("Hazard Ratio") xtitle("VPA (weekly minutes)") ///
    xlabel(0(100)720) title({bf:VPA and alzheimer mortality})
    1
    try lincom
    0%
    0
    use a user command
    100.00%
    1
    Last edited by Ruben Lopez Bueno; 16 Dec 2022, 05:57.
Working...
X