reading data\raw\inspire-1.4.2\labs.csv.gz

==============================================================================
A. REPRODUCTION CHECK AGAINST THE PUBLISHED PEAK
==============================================================================
operations with an exposure: 2365 of 2365
max |recomputed peak - published peak| = 0.000000
max |recomputed draw count - published count| = 0

==============================================================================
B. HOW MUCH OF THE WINDOW IS RECONSTRUCTED RATHER THAN OBSERVED
==============================================================================
draws in window  patients  median_carried_h  median_span_h  median_step_vs_trapezoid_pct  max_step_vs_trapezoid_pct
              1       277             24.00           0.00                          0.00                       0.00
              2       337             23.08           0.92                          0.26                      36.62
            3-4       242             21.83           2.17                          0.60                     204.50
            5-9       967              5.58          18.42                          5.27                     106.73
            10+       542              2.83          21.17                          3.50                      29.38

patients with a single draw: 277 (11.7%). For them the time-weighted average is that one value.

==============================================================================
C. ENTANGLEMENT WITH SAMPLING INTENSITY
==============================================================================
Spearman correlation between the number of draws and the exposure.

                     exposure  spearman_rho_with_draw_count             p  median        iqr
     peak lactate (published)                         0.532 5.241804e-173    3.00  1.70-5.90
        time-weighted average                         0.363  1.900145e-74    2.04  1.30-3.30
lactate load above 2.0 mmol/L                         0.427 2.435193e-105    4.93 0.00-33.12

==============================================================================
D. SAMPLING INTENSITY BY EXPOSURE QUARTILE
==============================================================================
exposure  median draws G1  median draws G2  median draws G3  median draws G4  Q4/Q1 ratio
    peak                2                6                8               10         5.00
     TWA                3                6                7                9         2.57
    load                3                7                7                9         3.00

==============================================================================
E. ASSOCIATION WITH NEWLY DETECTED AKI
==============================================================================
                           exposure    n  events reference  HR[2.1-3.0] HR[2.1-3.0]_ci  HR[3.8-5.9] HR[3.8-5.9]_ci  HR[8.6 capped] HR[8.6 capped]_ci  per_stratum_HR per_stratum_ci  per_stratum_p reference_range  reference_n  HR[G2] HR[G2]_ci  HR[G3] HR[G3]_ci  HR[G4] HR[G4]_ci  per_SD_HR per_SD_ci  per_SD_p
peak lactate, published four strata 1834     196     <=1.9        1.159      0.79-1.69        1.482      1.04-2.10           2.135         1.22-3.74           1.251      1.08-1.45         0.0034             NaN          NaN     NaN       NaN     NaN       NaN     NaN       NaN        NaN       NaN       NaN
            peak lactate, quartiles 1834     196       NaN          NaN            NaN          NaN            NaN             NaN               NaN             NaN            NaN            NaN       0.50-1.50        501.0   1.065 0.70-1.62   1.238 0.82-1.87   1.923 1.29-2.87      1.257 1.10-1.44    0.0008
   time-weighted average, quartiles 1834     196       NaN          NaN            NaN          NaN            NaN             NaN               NaN             NaN            NaN            NaN       0.50-1.21        459.0   1.099 0.71-1.71   1.311 0.86-2.00   1.879 1.25-2.82      1.280 1.13-1.46    0.0002
  lactate load above 2.0, quartiles 1834     196       NaN          NaN            NaN          NaN            NaN             NaN               NaN             NaN            NaN            NaN       0.00-0.00        714.0   1.146 0.77-1.71   1.276 0.86-1.90   1.812 1.24-2.64      1.264 1.12-1.43    0.0001

==============================================================================
F. RESTRICTED TO PATIENTS WITH DENSER SAMPLING
==============================================================================
The carried-value assumption is weakest when draws are few. Restricting
to at least four draws removes the patients for whom the window is mostly
reconstruction, at the cost of removing the sparsely sampled comparison group.

                                    exposure    n  events reference_range  reference_n  HR[G2] HR[G2]_ci  HR[G3] HR[G3]_ci  HR[G4] HR[G4]_ci  per_SD_HR per_SD_ci  per_SD_p
          peak lactate, quartiles, >=4 draws 1250     121       0.50-1.90          335   1.790 0.97-3.30   2.634 1.50-4.63   3.893 1.89-8.02      1.489 1.26-1.76       0.0
 time-weighted average, quartiles, >=4 draws 1250     121       0.50-1.34          313   1.551 0.83-2.88   2.122 1.17-3.84   2.976 1.66-5.32      1.438 1.23-1.68       0.0
lactate load above 2.0, quartiles, >=4 draws 1250     121       0.00-0.00          335   1.883 1.02-3.48   2.243 1.22-4.12   3.255 1.80-5.88      1.383 1.19-1.60       0.0

written to results\manuscript_v3
