The reliability of SQLape® indicators depend on the quality of data, which might have different impact. For instance, if the patients have different identifiers from one year to another, this will have strong impact on readmission measurement, but no effect on the other indicators. If all procedures dates are identical in each stays, it will not be possible to measure reoperations’ rate, but this will not bias other indicators. But if the exhaustiveness of diagnoses coding present gaps, it might impact all indicators.
To estimate the quality of data, we use four criterions : exhaustiveness, uniqueness, precision, conformity, which are detailed below.
– diagnoses: proportion of therapeutic procedures justified by a corresponding diagnosis: organ or tissue removal for transplant or graft, foreign body removal, incidental appendectomy, procedures on skin and lymph nodes, diagnostic procedures.
– procedures: proportion of operations coded for patients with at least one diagnosis requiring surgery: abortion, anal fistula, cataract, cesarean delivery, cox arthritis, benign neoplasm, goiter, femur fracture, hemorrhoids, hydrocele, inguinal hernia, meniscus disorders, retina detachment, skin cancer, tendon tear, uterine prolapse as main diagnosis; appendicitis as main or secondary diagnosis.
– generally associated (appendicitis, abortion, retina detachment, femur fracture, uterine prolapse, etc.)
– other variables: proportion of observations without missing values
– proportion of unique stay identifiers in the whole SQLape_input.txt file
– proportion of patients with a same identifier over two years
– diagnoses: proportion of codes with at least a specified pathology or organ (cf. list (1) below)
– procedures: proportion of codes with at least a specified operation or organ (cf. list (2) below)
– operation dates: proportion of cases with at least two different operation dates during the same stay
– diagnoses: official ICD-10 codes* ICD-10
– procedures: official ICD-9-CM codes* ICD-9-CM
– dates (yyyymmdd): day 1-31, month 1-12, year > 1900
– other: Swiss Federal Office of Statistics nomenclatures
*Swiss specific positions are automatically transcoded into international codes, including updated codes (e.g. I84 into K649 for hemorrhoids).
The minimal requirements are given in table below, for each variable with the concerned indicators
A minimal proportion is defined to assess each requirement.
Requirements are set specifically to each indicator. For instance, the reliability of the patients’ identifiers is required only for readmission or the discharge decision only to identify deaths (X = required required).
If a requirement is not met for an indicator, it is colored in blue and a warning will appear for the corresponding Indicator.
List (1) of diagnoses codes without specified pathology or organ:
A09, A499, A64, B349, B49, B889, B89, B99, C80, C969, D099, D369, D489, D739, D759, D849, E079, E215, E237, E259, E279, E329, E349, E569, E619, E639, E889, F349, F39, F489, F609, F69, F99, G319, G64, G729, G939, G959, G969, H029, H059, H279, H359, H439, H449, H539, H579, H619, H749, H839, H939, I279, I516, I519, I739, J069, J180, J189, J399, J989, K089, K229, K238, K272, K274, K276, K319, K629, K639, K669, K769, K829, K839, K869, K929, L989, M139, M199, M219, M249, M259, M489, M519, M629, M759, M799, M899, M959, N289, N329, N369, N399, N429, N489, N509, N649, N739, N949, O94, O95, P279, P969, Q289, Q899, Q999, R068, R104, R238, R258, R268, R298, R398, R418, R498, R568, R69, R829, R99, T149, T659, T789, T819, Z529, Z532, Z549, Z73, Z739, Z769.
List (2) of procedure codes without specified operation or organ:
0299, 0399, 0499, 0589, 059, 0698, 0699, 0749, 0759, 0779, 0799, 0899, 1299, 149, 1698, 1699, 189, 2099, 2199, 229, 2699, 2799, 2899, 2999, 3198, 3199, 3398, 3399, 3489, 3499, 3599, 3999, 409, 4198, 4199, 4299, 4499, 4699, 799, 4899, 4999, 5099, 5199, 5299, 5499, 5599, 5699, 5799, 5899, 5999, 6099, 6299, 6498, 6499, 6599, 6999, 7091, 718, 719, 7399, 7599, 7990, 8199, 8399, 8499, 8599, 8699.