About the Data
This page contains information about the data and how it was collected. All data was collected by Consortium members at our two Multi-site Clinical Centers (MCCs) located in Chicago and Michigan according to A2CPS procedures approved by our central Internal Review Board (IRB) at the University of Iowa. A comprehensive list of the questionnaires used to remotely collect self-reported outcomes is found in the tables below. Clicking "Source" will open a PDF containing the actual questionnaires, or Case Report Forms (CRF's), participants responded to. Clicking on "HEAL CDE" will download a .csv file containing information about the Common Data Elements used by HEAL to harmonize each variable collected with data from other studies.
Demographics
Self-reported demographic information is collected electronically through a variety of surveys using Research Electronic Data Capture (REDCap™) and/or MyDataHelps™ (RKStudio™, CareEvolution, LLC, Ann Arbor, MI). A2CPS includes the core demographic forms and associated Helping End Addiction Long-term (HEAL) Common Data Elements (CDE) for harmonization. Remote assessments have been implemented for both English and Spanish-speaking participants.
Psychosocial Data
Self-reported assessments are collected electronically through a variety of surveys using Research Electronic Data Capture (REDCap™) and/or MyDataHelps™ (RKStudio™, CareEvolution, LLC, Ann Arbor, MI) (see below for more details). A2CPS includes the core psychosocial assessments and associated HEAL CDEs for harmonization. Remote assessments have been implemented for both English and Spanish-speaking participants. Of note: two psychosocial items were also collected during brain-imaging sessions, a pain assessement and a mood assessment. These items can be found below in the "Imaging Source" section. [add a link to jump down]
Functional Testing (FT)
In the knee arthroplasty cohort, physical function and MEP are assessed with the Five-Times Sit-to-Stand (5TSTS) Test (145, 146) and the 10-meter Walk Test (10MWT) (147, 148) following standard protocols. Time to completion is recorded for both tasks as the primary measure of function. Pain is assessed prior to and immediately following each test to assess MEP using an 11-point numerical pain rating scale (NPRS) with anchors of 0 (no pain) to 10 (worst pain imaginable). In the thoracic surgery cohort, participants are asked to take 3 deep breaths and perform 3 forceful coughs. Pain is rated prior to and immediately following each task. MEP is defined as the difference in pain with activity minus the initial resting pain.
QST
Pressure Pain Thresholds (PPT)
Pressure Pain Thresholds (PPT) are assessed with a pressure algometer (Wagner Pain Test FPX25, Wagner Instruments, USA) applying a 1-cm2 rubber tip at a rate of 0.5 kgf/sec. Three repetitions are performed at the surgical site (knee or chest) and three at a standard remote site (mid-deltoid of the shoulder contralateral to surgical site) (149). The average of the repetitions at each site are used as the PPT values, where higher PPTs indicate less pressure sensitivity.
Mechanical Temporal Summation (TS)
Temporal Summation (TS) is thought to reflect ascending pain facilitation (150); it is assessed using a punctate stimulus (Neuropen®, Owen Mumford, United Kingdom) applied to the skin. A single repetition is applied, followed by 10 repetitions at a rate of 1 per second, at the surgical site (knee or chest) and a standard remote site (mid-deltoid of the arm contralateral to surgical site). Pain ratings from the single stimuli and maximal pain ratings following the 10 stimuli are recorded. TS is indicated as an increase in pain ratings from repetitive noxious stimuli compared to the single stimulus (150-152).
Conditioned Pain Modulation (CPM)
Conditioned Pain Modulation (CPM) is a psychophysical test paradigm which indirectly measures an individual's endogenous analgesia capacity (153-155). CPM assessment involves a test stimulus (PPT at the deltoid – 3 repetitions) before and immediately after application of a noxious conditioning stimulus (hand immersion in a 10°C cold water bath) for 60 seconds (or to tolerance). CPM is indicated as an increase in PPT (less pain sensitivity) following the conditioning stimulus (154, 155).
Dynamic Mechanical Allodynia
Allodynia is assessed in the thoracic cohort only using a standardized brush (Somedic SENSELab Brush-05, Somedic, Sweden or equivalent) (152). Five, 4-cm brush strokes are applied at the surgical site (chest wall) and the contralateral chest wall with pain ratings recorded after each stimulation. The brush is applied with sufficient force to slightly bend the bristles 45 degrees (~200 +/- 100 mN) to standardize the stimulus. The average pain ratings of the 5 repetitions at each site are used, where pain > 0 indicates allodynia.
Blood Draw Source
Blood samples are collected from all study participants prior to surgery. Medication intake within the 24 hours prior to the visit, and time of last food and stimulant intake are recorded prior to the blood draw. Two tubes, a 10mL BD K2EDTA vacutainer (Becton-Dickinson) and a 2.5mL BD PAXgene® Blood DNA Tube, are drawn at each visit. The samples in K2EDTA tubes are centrifuged within 30 minutes, processed and temporarily stored at -80 degrees within 1 hour locally at the clinic sites. The PAXgene samples are placed directly at -80 degrees. All samples are barcoded to link participant ID to the sample and are scanned into REDCap. Samples are shipped on dry ice from the clinical site to a central omics center (UC San Diego) for long-term storage and distribution. Sample aliquots are provided to each of the ODGC sites for genetic variant, exRNA, proteomic, lipidomic, and metabolomic analyses.
Imaging Data
The raw imaging data contains five kinds of images across four modalities. The MRI protocol is built upon the protocol used for the Adolescent Brain Cognitive Development study but has been adapted to the scanners in the A2CPS consortium. A detailed overview of the acquisition is presented by Sadil et al. (2024).
Anatomical (anat)
Structural MRI (sMRI) provides information related to the anatomic layout of the brain, and allows calculations of cortical and subcortical volumes and morphology. It provides a stable reference for cross-subject and cross-modality alignments.
Diffusion (dwi)
Diffusion-weighted MRI (dMRI) captures microstructural information on how tissues and structures hinder the diffusion of water molecules. Each scan measures diffusion in one direction, with more diffusion at a location leading to less signal there.
Fieldmap (fmap)
MRI scans are distorted by small magnetic field variations across the brain. To correct these, we acquire paired images with opposite distortions, which can be combined to create fieldmaps of the underlying variations.
Functional (func)
fMRI measures hemodynamically driven signal changes that result from localized changes in oxygen uptake and blood flow as a result of neural activity, and are typically analyzed by looking for signal variations associated with scan conditions (often controlled with a “task”) or at how spontaneous intrinsic variations in signal are coordinated across brain regions (“functional connectivity” of those areas). Note that in the dataset “resting state” is labeled as another type of task (task-rest).
Resting-state functional MRI (REST; task-rest) are six minutes long and are done with an inflatable limb cuff, of the type used for taking blood pressure readings, worn but uninflated. A REST1 scan precedes cuff inflation, and (assuming this occurs -- in some subjects inflation and thus cuff scans may be contraindicated) a REST2 occurs at the end.
Inflated cuff functional MRI (CUFF; task-cuff) are also six minutes long and have a user-calibrated inflation of the cuff (CUFF1; pressure to invoke a pain intensity rating of ~4/10) followed by a cuff inflation to a standard pressure (CUFF2). Not all subjects have either, but if at least one is present the protocol calls for a REST2 scan following them.
Each data release includes the raw imaging data, scan metadata, and an extensive set of derivatives produced from the scans. For details on the derivatives, please see the individual Release Pages.
Related, Non-Imaging data
Before, during, and following the functional MRI runs the patient is asked to rate any pain during the scanning procedure within their whole body, at their surgical site (knee or chest), and associated with the cuff pressure. Patients are also asked to complete a brief mood inventory. Note that because of the breadth and richness of the A2CPS dataset, information relevant to the imaging portion can also be found in other sections, e.g. caffeine use (Blood Draw CRF) or sleep (Psychosocial).
Multiomics Data
No omics data is included in release 1.0.0. Future releases will include data from blood plasma collected at three timepoints (pre-surgery, 6 weeks and 3 months post-surgery) to generate genomic measurements including proteins, lipids, metabolites and extra-cellular RNA molecules. DNA from blood will also be used to generate genetic variant data using a SNP microarray with custom content, with imputation to whole-genome coverage. Data generation protocols will include batching multiple timepoints from the same individual together and balancing of samples by surgery type, collection site, sex and race within each batch to alleviate technical and biological confounding. Relevant experimental meta-data will be provided with released data and normalization will be performed where possible, with all processing steps documented for each data type.
EHR Data
No EHR data is included in release 1.0.0. Future data releases will utilize Electronic Health Records (EHR) to extract clinical and procedural characteristics, including age, sex, height, weight, BMI, diagnoses, and surgical details. Surgical information will consist of procedure name, Current Procedural Terminology (CPT) code, surgeon, laterality, duration, and anesthesia details. Anesthesia data will include the American Society of Anesthesiologists (ASA) physical status classification, primary anesthesia type, duration, and medications administered during care. Postoperative data will include pain assessments, medication use, and length of stay, with opioid medications converted to morphine milligram equivalents (MME). For the knee arthroplasty cohort, implant name, manufacturer, and serial number will be recorded. For the thoracic surgery cohort, additional structured data on medical history and 30-day postoperative outcomes will be obtained from the Multicenter Perioperative Outcomes Group (MPOG) and Michigan Society of Thoracic and Cardiovascular Surgeons (MSTCVS) databases.
