[
  {
    "trial": "T017",
    "source": "library",
    "arms": {
      "group D": {
        "randomized": "99",
        "analysed": "95"
      },
      "group KD": {
        "randomized": "99",
        "analysed": "94"
      }
    },
    "depression": [],
    "pain": [
      {
        "measure": "100-mm VAS pain at 1, 6, 24, 48 h (rest or movement not stated)",
        "day": 1,
        "form": "mean_sd",
        "location": "Table 2, row VAS 24 h",
        "values": {
          "group D": {
            "mean": "34.9",
            "sd": "7.7",
            "n": "95"
          },
          "group KD": {
            "mean": "30.9",
            "sd": "6.7",
            "n": "94"
          }
        }
      },
      {
        "measure": "100-mm VAS pain at 1, 6, 24, 48 h (rest or movement not stated)",
        "day": 2,
        "form": "mean_sd",
        "location": "Table 2, row VAS 48 h",
        "values": {
          "group D": {
            "mean": "23.8",
            "sd": "6.3",
            "n": "95"
          },
          "group KD": {
            "mean": "22.8",
            "sd": "5.9",
            "n": "94"
          }
        }
      },
      {
        "measure": "PCA (morphine, ketorolac) consumption over 24 h",
        "day": 1,
        "form": "mean_sd",
        "location": "Table 2, row PCA consumption for 24 after surgery",
        "values": {
          "group D": {
            "mean": "65.3",
            "sd": "5.2",
            "n": "95"
          },
          "group KD": {
            "mean": "61.1",
            "sd": "7.2",
            "n": "94"
          }
        }
      }
    ],
    "adverse_events": [
      {
        "window": "overall",
        "location": "Table 2, row Hallucination",
        "values": {
          "group D": {
            "hallucination": "0"
          },
          "group KD": {
            "hallucination": "3"
          }
        }
      }
    ],
    "blinding_check": null,
    "enrollment": {
      "start": "2017-08",
      "end": "2018-01"
    },
    "registration": {
      "registry": "clinicaltrials.gov",
      "number": "NCT03194594",
      "date": null
    },
    "note": "Depression: PHQ-9 was measured before surgery and on postoperative days 1 and 3, but its values appear only in Fig. 2, and the text holds only the figure's caption ('Perioperative change of Patient Health Questionnaire (PHQ)-9 scores'), no printed values; the Results and abstract give only P values (groups K and KD lower on days 1 and 3 than before surgery and than group D, P < 0.01; group D unchanged; K and KD not different), so no depression entry is recorded. The 3.10, 3.50 and 2.95 (SD 0.67) in Methods are hypothesized means from a preliminary investigation for the sample size, not results. Arms: group K (ketamine alone, design role other) is left out. Randomized n = 99 per group (Methods and abstract); analysed n and every value's n are Table 2's column headers (group K 92, group D 95, group KD 94). The Results say 297 received study medication after randomization and 16 were withdrawn (conversion to open surgery, loss to follow-up, re-exploration for postoperative bleeding, refusal of postoperative assessment; Fig. 1, whose per-arm counts are not in the text). Pain: Table 2 heads the VAS rows 'Pain intensity with effort', so these are scores with effort, although the design's measure name says rest or movement not stated. VAS at 1 h and 6 h (day 0) are left out under the codebook's 24/48 h rule. Table 2 prints no unit for 'PCA consumption for 24 after surgery'; the PCA pump held morphine 60 mg, ketorolac 150 mg and ramosetron 0.6 mg in 100 ml saline (basal 2 ml/h, 0.5-ml boluses, 15-min lockout). Table 2 marks group KD's VAS at 1, 6 and 24 h with * (P < 0.01 vs group K) and its PCA value with ** (P < 0.05 vs the other two groups); the markers are dropped from the values. Adverse events: hallucination is printed as n (%) in Table 2 (group D 0 (0), group KD 3 (3.2)); the row alignment checks out, since genital pruritus (0, 4, 3) occurs only in the dexamethasone groups and hallucination (2, 0, 3) only in the ketamine groups. No time is stated for hallucination anywhere: the Methods list 'side effects of the study drugs' as a secondary outcome with no time, and the Results report them with age, weight and anesthesia duration, apart from PONV. In the extracted table its label lines up with the timed PONV rows (at 1, 6, 24, 48 h after surgery) but is not one of them. Ketamine was given under general anesthesia, so the count may include emergence reactions, which the codebook keeps apart from postoperative ones; it is coded overall. Table 2 also gives PONV at 1, 6, 24 and 48 h and genital pruritus, outside the codebook's categories; no other psychotomimetic event and no count of any adverse event is reported. No blinding check is reported (only 'double-blinded method'). Enrollment: the Results give no dates; these are from Methods ('The study was performed at University Hospital from August 2017 to January 2018'). Registration: NCT03194594 at clinicaltrials.gov, no registration date printed; 'Registration No. 3031' is the IRB approval (April 2016), not a trial registration. The batch text holds the Methods and Results twice (an excerpt and the full PDF text), and the copies agree.",
    "reason": "The coders differ only in the hallucination window (A overall, B postoperative); both counts (group D 0, group KD 3, Table 2) match the source. The source states no time for hallucination: Table 2's row, the Methods' 'side effects of the study drugs' and the Results' 'side effects (hallucination or genital pruritus)' carry none, and the Results report side effects apart from PONV, so B's inference from the row sitting under the PONV block is not what the source prints. Ketamine was given under general anesthesia, so the count may include emergence, which the codebook separates from postoperative; I kept A's overall. Everything else matches the source in both extractions; I kept B's row-level locations and A's registry field."
  }
]