PREDICTUS dataset (GREDOS deposit) — README
Version: v1.0 (processed files), generated on 2026-01-21
License: Creative Commons Attribution 4.0 International (CC BY 4.0)
Access: Embargoed for 6 months from the deposit/publication date in GREDOS.

1) Title
PREDICTUS_dataset_predictors_of_90_day_functional_outcome

2) Authors
Diana L. Tarruella-Hernández; Antonio García-Molina; Marta M. Dolcet-Negre; Alicia Aliena-Valero; Álvaro Lucero-Garófano; Manuel Pedrero-Prieto; Lluís Morales-Caba; Gerardo Fortea-Cabo; Fernando Aparici-Robles; María Jesús Rivas-López; Juan B Salom; José I Tembl; Irene Escudero-Martínez

3) Contact
- Marta M. Dolcet-Negre (Universidad de Salamanca): martamdn@usal.es
- Alicia Aliena-Valero (Universitat de València): aliena4@uv.es

4) Associated publication
Manuscript under review. Title (working): “Higher Lymphocyte-to-Monocyte Ratio Predicts Favourable 90-Day Functional Outcome After Reperfusion-Treated Ischaemic Stroke”.
This record should be updated with the paper DOI upon acceptance.

5) Study design and population
Retrospective observational cohort of adults (≥18 years) with acute ischaemic stroke treated with mechanical thrombectomy, with or without intravenous thrombolysis (rtPA). Baseline clinical data, laboratory blood counts, imaging scores and 90-day mRS were collected.

Spatial coverage: Valencia, Spain (Hospital Universitario y Politécnico La Fe / Instituto de Investigación Sanitaria La Fe).
Temporal coverage: start=2016-04; end=2024-07.

6) Ethics
Approved by CEIm Hospital Universitario y Politécnico La Fe (protocol #2024-0422-1).

7) Funding
- Spanish Ministerio de Ciencia e Innovación: PID2021-125211OB-I00
- Instituto de Salud Carlos III (RICORS-ICTUS): RD21/0006/0014
- Junta de Castilla y León (Regional Research Program): SA2117P23

8) Files included in this deposit
A) Raw input file
- PREDICTUS.xlsx
  Raw spreadsheet as provided for analysis (Spanish column names in sheet 'Trombos').

B) Processed datasets (English variable names + exclusions)
- PREDICTUS_processed_full.xlsx
  Cleaned/processed dataset with English variable names, exclusion criteria applied, and derived binary outcome.

- PREDICTUS_processed_models.xlsx
  Subset intended for modelling, removing variables not considered in regression models (see Section 10).

C) Documentation
- PREDICTUS_data_dictionary.csv
  Variable-level codebook including coding/allowed values and a flag indicating whether the variable was used in models.

- README_PREDICTUS.txt
  This file.

9) Unit of analysis
One row per patient.

10) Key processing steps
Processing was applied to generate the 'processed' datasets:
- Removed three patient records flagged for review: CLOT-LF70, CLOT-LF97, CLOT-LF181.
- Excluded patients with baseline disability: baseline mRS ≥ 2 (kept baseline mRS < 2).
- Removed rows with missing 90-day mRS.
- Created a binary 90-day outcome:
  * favourable_mRS<2  if mRS_90d < 2
  * unfavourable_mRS>=2 if mRS_90d ≥ 2

Variables NOT considered in regression models (removed in the 'models' dataset):
- id (identifier)
- non_contrast_CT, angio_CT, perfusion_CT (near-constant / not informative for modelling)
- atrium (high proportion of missing values)
- heidelberg (high proportion of missing values)
- TICI_baseline (near-constant)

11) Missing data
Missing values are represented as empty cells / NA. The processed datasets remove rows with missing mRS_90d only; other variables may still contain missing values.

12) Data quality notes
Some categorical fields may contain unexpected codes due to data entry or spreadsheet formatting; these are flagged in the codebook (see 'notes' column). Users should perform validation checks before analysis.

13) Recommended citation (to be updated with the dataset persistent identifier)
Dolcet-Negre, Marta M.; Tarruella-Hernández, Diana L.; García-Molina, Antonio; et al. (2026). PREDICTUS dataset: Predictors of 90-day functional outcome. GREDOS, University of Salamanca. Persistent identifier: [to be assigned by GREDOS].