I am trying to create a matched sample using hazard ratios by observation by id. I am conducting a study on the influence of a bill (treatment in dollars) to students who withdraw (exposure event) from a university. However, the treatment occurs too rarely to be measured and the model is not sufficient as I have confounders in my variable set. Further, the exposure event of withdraw is related (precedes) degree attainment (dependent outcome of interest).
I am attempting to mimic a method used by Dr. Bo Lu in "Propensity Score Matching with Time-Dependent Covariates" to use hazards as a propensity score to create a matched sample of subjects who experience the exposure event and those who never experience this event. I have searched heavily to find help for this but have not had luck in securing it. I am hoping someone here may have some suggestions/direction.
My dataset consists of 10559 subjects with 123719 records/ periods of observations. An example of my data is below - I used cube roots for dollar variables. My treatment is the amount of bill, but this occurs very infrequently. It requires a subject withdraw from the university to have a chance of experiencing a bill, thus I am interested in balancing my covariates to this event (WD_event). Nearly 600 subjects experienced at least once (some 2, 3 or 6 times during the period of observation). I used Cox regression to create my model, but when I generate the hazard variable =h, I only see about 17,000 hazards generated.
My question: How do I create hazard ratios for each observation in my dataset? How do I get Stata to use these hazards to create a matched sample of treated and untreated (WD_event) subjects for my final model which looks at a terminal outcome (degree attainment)? Any guidance would be appreciated.
Reference: Lu, B. (2005). Propensity Score Matching with Time-Dependent Covariates. Biometrics, 61(3), 721–728. https://doi.org/10.1111/j.1541-0420.2005.00356.x
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input str10 SID float(WD_event curt_bill) byte(FEMALE URETHNICITY) double HSGPA byte(IND RESIDENCY Housing) float(cum_GPA curt_TI curt_networth curt_semp curt_giftfuo curt_stuloan curt_plus) "12345" 0 0 0 0 3.63 0 1 0 1.36 54.92795 37.91602 11.84761 0 0 0 "12345" 0 0 0 0 3.63 0 1 0 2.23 54.92795 37.91602 11.84761 0 0 0 "12345" 0 0 0 0 3.63 0 1 0 2.3566666 54.92795 37.91602 11.84761 0 0 0 "12345" 0 0 0 0 3.63 0 1 1 2.5675 54.06861 37.44793 16.375246 0 11.408857 0 "12345" 0 0 0 0 3.63 0 1 1 2.786 54.06861 37.44793 16.375246 0 11.408857 0 "12345" 1 8.440975 0 0 3.63 0 1 1 2.3216667 54.06861 37.44793 16.375246 0 11.408857 0 "23456" 0 0 0 0 3.18 0 0 0 3.66 42.74466 12.59921 0 13.836535 0 0 "23456" 0 0 0 0 3.18 0 0 0 3.11 42.74466 12.59921 0 13.836535 0 0 "23456" 0 0 0 0 3.18 0 0 0 2.85 42.74466 12.59921 0 13.836535 0 0 "23456" 0 0 0 0 3.18 0 0 0 2.575 48.16408 7.937005 10.322802 10.558255 0 0 "23456" 0 0 0 0 3.18 0 0 0 2.72 48.16408 7.937005 10.322802 10.558255 0 0 "23456" 0 0 0 0 3.18 0 0 0 2.6333334 48.16408 7.937005 10.322802 10.564232 0 0 "23456" 0 0 0 0 3.18 0 0 1 2.685714 49.46087 0 18.395565 0 12.92062 18.147957 "23456" 0 0 0 0 3.18 0 0 1 2.35 49.46087 0 18.395565 0 12.92062 18.147957 "23456" 0 0 0 0 3.18 0 0 1 2.426667 49.46087 0 18.395565 0 12.918624 18.14897 "23456" 0 0 0 0 3.18 0 0 1 2.394 44.9258 8.2228985 22.722565 6.073178 15.511265 17.289171 "23456" 0 0 0 0 3.18 0 0 1 2.421818 44.9258 8.2228985 22.722565 0 15.511265 0 "23456" 0 0 0 0 3.18 0 0 1 2.386667 45.29989 11.362574 22.722565 8.5726185 0 0 "23456" 1 0 0 0 3.18 0 0 1 2.2030768 45.29989 11.362574 22.722565 0 0 0 "23456" 0 0 1 0 3.11 0 0 0 2.54 49.21026 18.171206 16.130444 0 9.517052 0 "23456" 0 0 1 0 3.11 0 0 0 2.77 49.21026 18.171206 16.130444 0 9.517052 0 "23456" 1 0 1 0 3.11 0 0 0 1.8466667 49.21026 18.171206 16.130444 0 9.517052 0 "23456" 1 7.945991 1 0 3.11 0 0 1 1.385 52.71893 22.2398 16.06484 0 10.495085 0 "34567" 0 0 1 0 3.11 0 0 1 1.62 54.92121 27.144176 12.15313 0 10.51321 0 "34567" 0 0 1 0 3.11 0 0 1 1.6833333 54.92121 27.144176 12.15313 0 10.51321 0 "34567" 1 0 1 0 3.11 0 0 1 1.442857 54.92121 27.144176 12.15313 0 10.510194 0 "34567" 0 0 1 0 3.11 0 0 1 1.43125 59.9198 21.68703 12.385623 0 20.50783 0 "34567" 0 0 1 0 3.11 0 0 1 1.438889 59.9198 21.68703 12.385623 0 12.92062 0 "34567" 0 0 1 0 3.11 0 0 1 1.295 59.9198 21.68703 12.385623 0 12.918624 0 "34567" 1 9.782434 1 0 3.11 0 0 1 1.1772727 62.48339 32.710663 17.213007 0 12.92062 0 "34567" 0 0 1 0 3.11 1 0 1 1.3041667 12.848336 0 18.197416 18.955574 15.78622 0 "34567" 0 0 1 0 3.11 1 0 1 1.473077 12.848336 0 18.197416 18.955574 15.78622 0 "34567" 0 0 1 0 3.11 1 0 1 1.367857 12.848336 0 18.197416 18.955574 14.760645 0 "34567" 0 0 1 0 3.11 1 0 1 1.4766667 12.848336 0 18.197416 12.050712 10.877427 0 "34567" 0 0 1 0 3.11 1 0 1 1.590625 18.197416 0 22.413 18.13783 16.186617 0 "34567" 0 0 1 0 3.11 1 0 1 1.6941177 18.197416 0 22.413 17.726313 14.762176 0 "34567" 1 10.02507 1 0 3.11 1 0 1 1.6 18.197416 0 22.413 17.200623 15.457044 0 "34567" 0 0 1 0 3.11 1 0 1 1.7105263 18.197416 0 22.413 15.38697 14.4257 0 "34567" 1 0 1 0 3.11 1 0 1 1.625 22.349113 0 22.413 16.423582 16.026 0 "34567" 0 0 1 0 3.11 1 0 1 1.704762 22.349113 0 22.413 17.001154 16.026 0 "34567" 0 0 1 0 3.11 1 0 1 1.7636364 22.349113 0 22.413 16.423582 14.206134 0 "45678" 0 0 1 1 3 0 1 0 2.87 28.165264 39.48592 19.55431 16.76493 12.146356 0 "45678" 0 0 1 1 3 0 1 0 2.785 28.165264 39.48592 19.55431 16.76493 12.146356 0 "45678" 0 0 1 1 3 0 1 0 2.8066666 28.165264 39.48592 19.55431 16.767302 12.148615 0 "45678" 0 0 1 1 3 0 0 0 2.9125 28.32238 0 14.422496 12.48692 12.064467 0 "45678" 0 0 1 1 3 0 0 0 3.13 28.32238 0 14.422496 12.48692 12.918624 0 "45678" 0 0 1 1 3 0 0 0 3.176667 28.32238 0 14.422496 12.491194 12.04612 0 "45678" 0 0 1 1 3 0 0 1 3.088571 26.686825 0 16.120188 12.6097 11.429308 0 "45678" 0 0 1 1 3 0 0 1 3.0775 26.686825 0 16.120188 12.6097 11.429308 0 "45678" 0 0 1 1 3 0 0 1 3.0944445 26.686825 0 16.120188 13.274253 11.429308 0 "45678" 0 0 1 1 3 0 0 1 3.06 28.84499 35.56893 14.666724 12.066757 13.550338 0 "45678" 0 0 1 1 3 0 0 1 2.972727 28.84499 35.56893 14.666724 12.066757 13.550338 0 "45678" 0 0 1 1 3 0 0 1 2.916667 28.84499 35.56893 14.666724 12.069046 13.550338 0 "45678" 0 0 1 1 3 0 0 1 2.730769 28.86301 30 14.666724 12.164404 19.52025 0 "45678" 1 12.053924 1 1 3 0 0 1 2.5357144 28.86301 30 14.666724 12.164404 0 0 "45678" 0 0 1 1 3 0 0 1 2.3666666 28.86301 30 14.666724 0 0 0 "56789" 0 0 1 0 3.88 0 0 0 3.43 42.01813 36.840317 11.6545 0 19.10925 0 "56789" 0 0 1 0 3.88 0 0 0 3.295 42.01813 36.840317 11.6545 0 19.10925 0 "56789" 0 0 1 0 3.88 0 0 0 3.446667 42.01813 36.840317 11.6545 0 19.111076 0 "56789" 0 0 1 0 3.88 0 0 1 3.1925 46.74969 31.072325 0 0 19.566505 0 "56789" 1 0 1 0 3.88 0 0 1 2.554 46.74969 31.072325 0 0 19.566505 0 "67890" 0 0 1 1 3.46 0 0 0 2.96 35.1104 10 7.937005 24.139236 0 0 "67890" 0 0 1 1 3.46 0 0 0 2.645 35.1104 10 7.937005 9.337017 0 0 "67890" 0 0 1 1 3.46 0 0 0 2.3133333 35.1104 10 7.937005 9.344658 0 0 "67890" 1 18.703476 1 1 3.46 0 0 0 1.735 35.787067 10 7.937005 20.18986 0 0 "67890" 0 0 1 1 3.46 0 0 0 1.688 35.787067 10 7.937005 20.553696 0 0 "67890" 0 0 1 1 3.46 0 0 0 1.4066666 35.787067 10 7.937005 20.554485 0 0 "67890" 0 0 1 1 3.46 0 0 1 1.6485714 35.56893 0 12.59921 7.889095 0 0 "67890" 0 0 1 1 3.46 0 0 1 1.875 35.56893 0 12.59921 0 0 0 "67890" 0 0 1 1 3.46 0 0 1 2.0366666 35.56893 0 12.59921 0 0 0 "67890" 0 0 1 1 3.46 0 0 0 1.833 37.977135 7.937005 0 13.09205 13.550338 0 "67890" 0 0 1 1 3.46 0 0 0 1.939091 37.977135 7.937005 0 13.09205 13.550338 0 "67890" 0 0 1 1 3.46 0 0 0 1.9633334 37.977135 7.937005 0 13.095937 13.550338 0 "67890" 0 0 1 1 3.46 0 0 0 2.043077 37.977135 7.937005 0 13.661972 0 0 "67890" 0 0 1 1 3.46 0 0 0 2.04 37.84354 12.59921 0 17.005766 13.55941 0 "67890" 0 0 1 1 3.46 0 0 0 2.144 37.84354 12.59921 0 16.362804 14.75146 0 "67890" 0 0 1 1 3.46 0 0 0 2.19125 37.84354 12.59921 0 16.606897 13.062813 0 "78901" 0 0 1 0 3.66 0 0 0 3.46 39.44162 32.39612 14.49901 10.409656 11.365154 9.333192 "78901" 0 0 1 0 3.66 0 0 0 3.155 39.44162 32.39612 14.49901 10.409656 11.365154 9.333192 "78901" 0 0 1 0 3.66 0 0 0 3.2133334 39.44162 32.39612 14.49901 10.409656 11.367735 9.337017 "78901" 0 0 1 0 3.66 0 0 0 3.41 39.44162 32.39612 14.49901 0 0 0 "78901" 0 0 1 0 3.66 0 0 1 3.416 41.41964 32.075344 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 0 1 3.368333 41.41964 32.075344 11.049365 0 0 0 "78901" 1 0 1 0 3.66 0 0 1 2.887143 41.41964 32.075344 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 0 1 2.9175 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 0 1 2.965556 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 0 1 3.004 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 0 1 3.030909 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 1 1 3.055833 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 1 1 3.1015384 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 1 1 3.144286 48.89537 69.58943 11.049365 0 0 0 "78901" 0 0 1 0 3.66 0 1 1 3.188 48.89537 69.58943 11.049365 0 0 0 "89012" 0 0 0 0 4 0 0 0 3.76 50.49286 35.982243 5.593445 16.090631 0 0 "89012" 0 0 0 0 4 0 0 0 3.88 50.49286 35.982243 5.593445 16.090631 0 0 "89012" 0 0 0 0 4 0 0 0 3.92 50.49286 35.982243 5.593445 16.091919 0 0 "89012" 0 0 0 0 4 0 0 0 3.83 50.49286 35.982243 5.593445 12.59921 0 0 "89012" 0 0 0 0 4 0 0 0 3.824 50.49286 35.982243 5.593445 12.59921 0 0 "89012" 1 0 0 0 4 0 0 0 3.186667 50.49286 35.982243 5.593445 12.59921 0 0 "89012" 0 0 0 0 4 0 0 1 3.2857144 50.49286 35.982243 5.593445 12.59921 0 0 "89012" 0 0 0 0 4 0 0 1 3.36375 50.49286 35.982243 5.593445 12.59921 0 0 end
0 Response to Hazard as a Propensity Score for Matching - Propensity Score Matching with Time-Dependent Covariates
Post a Comment