[
 {
  "trial": "T004",
  "source": "library",
  "arms": {
   "Coadministration group (propofol + ketamine)": {
    "randomized": "20",
    "analysed": "19"
   },
   "Propofol group (propofol alone)": {
    "randomized": "20",
    "analysed": "20"
   }
  },
  "depression": [],
  "pain": [],
  "adverse_events": [
   {
    "window": "intraoperative",
    "location": "Results, second paragraph",
    "values": {
     "Coadministration group (propofol + ketamine)": {
      "agitation": "0",
      "any_ae": "0"
     },
     "Propofol group (propofol alone)": {
      "agitation": "1",
      "any_ae": "1"
     }
    }
   },
   {
    "window": "emergence",
    "location": "Results, fourth paragraph (Propofol 'four'); abstract and Discussion (none with coadministration)",
    "values": {
     "Coadministration group (propofol + ketamine)": {
      "dissociation": "0"
     },
     "Propofol group (propofol alone)": {
      "dissociation": "4"
     }
    }
   }
  ],
  "blinding_check": null,
  "enrollment": null,
  "registration": null,
  "note": "Randomized 20 per arm (Methods). Analysed: Propofol 20, Coadministration 19 in the abstract, Table 1 and Figure 1 (the missing Coadministration patient is not explained); the Figure 2 caption gives them the other way round (Propofol n = 19, Coadministration n = 20). Depression: the 'depressed/elated' mood VAS (before surgery and 15 min after arrival in the PACU) is shown only in Figure 2 (median, 25th-75th percentiles) with no printed values, so no entries; the text says postoperative mood scores were higher with coadministration for all six mood states. Pain: no values at day 1 or 2; post-discharge pain at rest and during activity and hydrocodone use on the first postoperative day are given only as P values, and Table 3 gives time (days) to normalization, median (quartiles): Propofol pain at rest 5 (4, 5), during activity 4 (4, 4), hydrocodone 4 (4, 4); Coadministration 3 (2, 4) for each. Pain VAS on arrival to Phase 2 recovery (day 0, not extracted): Propofol 20 [2.0-20.0], Coadministration 0 [0.0-0.0], median [quartiles]; intraoperative fentanyl 1.1 ± 0.5 µg/kg in both (Table 1); PACU morphine not quantified. AEs: the only intraoperative adverse event was 'transient restlessness' in the Propofol group ('there was only one event'), coded agitation and any_ae; the Coadministration zeros follow from that statement rather than a printed 0. Mild perceptual changes (VAS scores ≤20) in surroundings, time, colors and sounds in 'four' Propofol patients (printed as a word), seen in recovery and absent 30 min after surgery, coded dissociation (derealization-type perceptual change; the source gives this specific count, not a combined psychiatric or psychotomimetic one, so not psychiatric_any), window emergence (it could also be read as early postoperative). The abstract ('without perceptual changes after surgery') and Discussion ('No psychotomimetic or perceptual symptoms were observed after surgery') give the Coadministration group's 0, stated in words. No changes in the suspicion, weird feeling or thought-process VAS in either group (scores, not counts). No enrollment period or registration given.",
  "reason": "The coders agreed on everything but the four Propofol patients with mild perceptual changes in surroundings, time, colors and sounds (Results, fourth paragraph), none with coadministration (abstract, Discussion). I kept coder B's dissociation in the emergence window: altered perception of surroundings, time, colors and sounds is the derealization the codebook puts under dissociation, and psychiatric_any is only for a source that gives just a combined psychiatric or psychotomimetic count, which this one does not. The changes were found on arrival in recovery and were gone 30 min after surgery, which is emergence from the sedation rather than the later postoperative course."
 },
 {
  "trial": "T005",
  "source": "library",
  "arms": {
   "Group A (propofol, fentanyl, ketamine)": {
    "randomized": "35",
    "analysed": "35"
   },
   "Group B (propofol, fentanyl)": {
    "randomized": "35",
    "analysed": "35"
   }
  },
  "depression": [
   {
    "measure": "Hamilton Depression Rating (HDR) score, 21 items",
    "day": -1,
    "form": "mean_sd",
    "location": "Results, third paragraph; abstract",
    "values": {
     "Group A (propofol, fentanyl, ketamine)": {
      "mean": "12.7",
      "sd": "5.4"
     },
     "Group B (propofol, fentanyl)": {
      "mean": "12.3",
      "sd": "6.0"
     }
    }
   },
   {
    "measure": "Hamilton Depression Rating (HDR) score, 21 items",
    "day": 1,
    "form": "mean_sd",
    "location": "Results, third paragraph; abstract",
    "values": {
     "Group A (propofol, fentanyl, ketamine)": {
      "mean": "9.9",
      "sd": "4.1"
     },
     "Group B (propofol, fentanyl)": {
      "mean": "14.4",
      "sd": "3.8"
     }
    }
   },
   {
    "measure": "Hamilton Depression Rating (HDR) score, 21 items",
    "day": 1,
    "form": "count",
    "definition": "HDR score up more than 5 points after surgery (a rise from the preoperative score, measured 2 days before surgery): a count of patients whose scores worsened, not of participants with depression (all had major depression); the source says only 'after surgery' (HDR was measured on days 1 and 3), and day 1 is the reading that fits its paragraph",
    "location": "Results, third paragraph",
    "values": {
     "Group A (propofol, fentanyl, ketamine)": {
      "events": "0",
      "n": "35"
     },
     "Group B (propofol, fentanyl)": {
      "events": "8",
      "n": "35"
     }
    }
   }
  ],
  "pain": [],
  "adverse_events": [
   {
    "window": "postoperative",
    "location": "Results, last paragraph",
    "values": {
     "Group A (propofol, fentanyl, ketamine)": {
      "psychiatric_any": "5"
     },
     "Group B (propofol, fentanyl)": {
      "psychiatric_any": "8"
     }
    }
   }
  ],
  "blinding_check": null,
  "enrollment": null,
  "registration": null,
  "note": "Group C (25 non-depressed controls, not randomized) left out. 35 per arm randomized and analysed (Methods, Table 1, outcome denominators 'of 35'); no per-time n is printed for the HDR means. Baseline HDR was measured 2 days before surgery (day -1). HDR means are from the Results text (also the abstract); Figure 1 (box plots) has no printed values, and day 3 has none (text: no significant difference between A and B). Count entry: 'Eight (23%) of 35 patients in Group B were up more than 5 points in HDR score after surgery, but there were no patients who were up more than 5 points in HDR score after surgery in Group A'; the time is not stated, so day 1 is a reading (the paragraph goes on to the day 1 means, the only time the arms differed), and the count could refer to day 3 or to either day; Group A's 0 is stated in words, Group B's 8 is printed as 'Eight', and Group A's n of 35 is the group's size, the sentence giving a denominator only for Group B. Table 2 gives HDR item scores (mean ± SD) 2 days before and the day after surgery (e.g., depressed mood 2.2 ± 1.2 vs 2.0 ± 1.0 before, 0.8 ± 0.8 vs 2.1 ± 1.0 after, A vs B; the text cites them as Table 1); items, not the listed total, so not extracted. Pain: VAS for A and B is printed only at 8 h (26.6 ± 8.7 vs 34.3 ± 12.0) and 16 h (24.9 ± 8.2 vs 31.1 ± 8.8) in the abstract, outside the 18-30 h window; the 24 h and later values for A and B are only in Figure 2 (only Group C's are printed). AEs: postoperative confusion by CAM (checked at least daily to day 7), 5 (14%) vs 8 (23%), coded psychiatric_any: it is the source's only psychiatric count, which the Methods analyse as 'psychosis emergence or confusion' and the Discussion treats as confusion 'such as hallucinations' and as emergence reactions; no excitement is described and it is not limited to emergence, so not agitation. Not coded: ventricular ectopic rhythm 1 vs 2; hypotension at induction 2 vs 2. No enrollment dates or registration given.",
  "reason": "Two differences. Postoperative confusion by CAM (5 in Group A, 8 in Group B): I kept coder B's psychiatric_any, not coder A's agitation, because it is the source's only psychiatric count, analysed as 'psychosis emergence or confusion' and likened to hallucinations and emergence reactions, while the codebook's agitation covers excitement on emergence, which the source doesn't describe. The patients whose HDR rose more than 5 points after surgery (0 of 35 vs 8 of 35): I kept coder A's count, since the source prints it and the codebook asks for an ambiguous time to be noted rather than the number dropped; it stays at day 1, the time its paragraph reports, with a definition saying it counts worsening, not depression, and that the day isn't stated."
 },
 {
  "trial": "T011",
  "source": "library",
  "arms": {
   "K group": {
    "randomized": "60",
    "analysed": "60"
   },
   "C group (0.9% saline)": {
    "randomized": "60",
    "analysed": "60"
   }
  },
  "depression": [
   {
    "measure": "PHQ-9",
    "day": -1,
    "form": "mean_sd",
    "location": "Table 2",
    "values": {
     "K group": {
      "mean": "4.02",
      "sd": "0.13"
     },
     "C group (0.9% saline)": {
      "mean": "3.63",
      "sd": "0.14"
     }
    }
   },
   {
    "measure": "PHQ-9",
    "day": 1,
    "form": "mean_sd",
    "location": "Table 2",
    "values": {
     "K group": {
      "mean": "2.68",
      "sd": "0.73"
     },
     "C group (0.9% saline)": {
      "mean": "3.37",
      "sd": "0.91"
     }
    }
   },
   {
    "measure": "PHQ-9",
    "day": 5,
    "form": "mean_sd",
    "location": "Table 2",
    "values": {
     "K group": {
      "mean": "2.28",
      "sd": "0.61"
     },
     "C group (0.9% saline)": {
      "mean": "2.92",
      "sd": "0.63"
     }
    }
   },
   {
    "measure": "PHQ-9",
    "day": -1,
    "form": "count",
    "definition": "PHQ-9 ≥ 5 (the source's 'depressed mood')",
    "location": "Discussion, second paragraph (Table 3: 5–10 row 23 K and 22 C, >10 row 1 K and 0 C)",
    "values": {
     "K group": {
      "events": "24",
      "n": "60"
     },
     "C group (0.9% saline)": {
      "events": "22",
      "n": "60"
     }
    }
   },
   {
    "measure": "PHQ-9",
    "day": 1,
    "form": "count",
    "definition": "PHQ-9 ≥ 5 (the source's 'depressed mood')",
    "location": "Table 3, 5–10 row (>10 row 0.00 in both groups); Discussion, third paragraph (≥5: 7 K, 17 C)",
    "values": {
     "K group": {
      "events": "7",
      "n": "60"
     },
     "C group (0.9% saline)": {
      "events": "17",
      "n": "60"
     }
    }
   },
   {
    "measure": "PHQ-9",
    "day": 5,
    "form": "count",
    "definition": "PHQ-9 ≥ 5 (the source's 'depressed mood')",
    "location": "Table 3, 5–10 row (>10 row 0.00 in both groups)",
    "values": {
     "K group": {
      "events": "6",
      "n": "60"
     },
     "C group (0.9% saline)": {
      "events": "11",
      "n": "60"
     }
    }
   }
  ],
  "pain": [
   {
    "measure": "VAS score",
    "day": 1,
    "form": "mean_sd",
    "location": "Table 1",
    "values": {
     "K group": {
      "mean": "2.17",
      "sd": "0.24",
      "n": "60"
     },
     "C group (0.9% saline)": {
      "mean": "3.75",
      "sd": "0.42",
      "n": "60"
     }
    }
   }
  ],
  "adverse_events": [],
  "blinding_check": null,
  "enrollment": {
   "start": "2014-03",
   "end": "2014-09"
  },
  "registration": {
   "number": "ChicTR-OCC-14004280",
   "date": null
  },
  "note": "Table 2 prints PHQ-9 as 'mean + SD' with a plus sign (Methods: data are mean ± SD); its rows are C then K, as the per-group F values in the text show (C F = 2.17, K F = 14.67). The printed dispersions look too small to be SDs (0.13 and 0.14 before surgery, while Table 3 has scores from 0-4 to >10), so they may be SEs or misprints; similar small values appear in Table 1 (e.g., age 41.40 ± 0.16; day 5 VAS 1.88 ± 0.01). The preoperative assessment was 1 day before surgery (day -1). Table 2 prints no n (60 per group, Table 1; Table 3's categories total 60 per arm at each time). Both forms are given for PHQ-9 at each time. Count entries: the source calls PHQ-9 ≥ 5 a depressed mood; Table 3 gives categories 0–4, 5–10 and >10, its percentages being shares of each row, not of each arm. The baseline count is printed in the Discussion (24 K, 22 C with PHQ-9 ≥ 5); days 1 and 5 are Table 3's 5–10 row, its >10 row being 0.00 in both groups (day 1 matches the Discussion: 7 in K, 17 in C). Pain: VAS day 1 from Table 1 (scale range not stated); the text's t value for day 1 VAS (t1 = −2.051, the same as PHQ-9's) differs from Table 1's (−6.048); day 5 VAS (1.12 ± 0.87 vs 1.88 ± 0.01) not extracted. No psychotomimetic adverse events reported; Table 1 gives nausea (day 1: 12 vs 13; day 5: 0 vs 0) and insomnia (day 1: 10 vs 9; day 5: 8 vs 8), K vs C. Enrollment from Methods ('between March 2014 and September 2014'). The registration number is printed 'ChicTR-OCC-14004280' beside ethics approval Ky2012044, with no date.",
  "reason": "The coders differed only on whether to enter the PHQ-9 counts beside Table 2's means. I kept coder A's count entries: the codebook asks for every listed measure at every time the source reports it, and its count form is for participants with depression by the source's definition, here PHQ-9 ≥ 5 ('depressed mood'). Baseline is the Discussion's printed ≥5 count (24 K vs 22 C); days 1 and 5 are Table 3's 5–10 row (7 vs 17, 6 vs 11), its >10 row being 0.00, and day 1 matches the Discussion. n is 60 per arm. The means are unchanged."
 },
 {
  "trial": "T015",
  "source": "library",
  "arms": {
   "Ketamine group": {
    "randomized": null,
    "analysed": "162"
   },
   "Saline group": {
    "randomized": null,
    "analysed": "163"
   }
  },
  "depression": [
   {
    "measure": "EPDS",
    "day": 3,
    "form": "mean_sd",
    "location": "Table 4",
    "values": {
     "Ketamine group": {
      "mean": "7.2",
      "sd": "3.9",
      "n": "162"
     },
     "Saline group": {
      "mean": "7.2",
      "sd": "4.2",
      "n": "163"
     }
    }
   },
   {
    "measure": "EPDS",
    "day": 3,
    "form": "count",
    "definition": "EPDS ≥ 10 ('occurrence of postpartum depression', the primary outcome's threshold)",
    "location": "Table 4",
    "values": {
     "Ketamine group": {
      "events": "41",
      "n": "162"
     },
     "Saline group": {
      "events": "46",
      "n": "163"
     }
    }
   },
   {
    "measure": "EPDS",
    "day": 42,
    "form": "mean_sd",
    "location": "Table 4",
    "values": {
     "Ketamine group": {
      "mean": "5.6",
      "sd": "3.9",
      "n": "162"
     },
     "Saline group": {
      "mean": "5.7",
      "sd": "4.3",
      "n": "163"
     }
    }
   },
   {
    "measure": "EPDS",
    "day": 42,
    "form": "count",
    "definition": "EPDS ≥ 10 ('occurrence of postpartum depression', the primary outcome's threshold)",
    "location": "Table 4",
    "values": {
     "Ketamine group": {
      "events": "26",
      "n": "162"
     },
     "Saline group": {
      "events": "29",
      "n": "163"
     }
    }
   }
  ],
  "pain": [],
  "adverse_events": [
   {
    "window": "intraoperative",
    "time": "5 min after dosing",
    "location": "Table 3, 5 min",
    "values": {
     "Ketamine group": {
      "hallucination": "37",
      "dizziness": "113",
      "visual": "31"
     },
     "Saline group": {
      "hallucination": "0",
      "dizziness": "16",
      "visual": "0"
     }
    }
   },
   {
    "window": "intraoperative",
    "time": "15 min after dosing",
    "location": "Table 3, 15 min",
    "values": {
     "Ketamine group": {
      "hallucination": "8",
      "dizziness": "79",
      "visual": "14"
     },
     "Saline group": {
      "hallucination": "0",
      "dizziness": "10",
      "visual": "1"
     }
    }
   },
   {
    "window": "intraoperative",
    "time": "immediately before leaving the operating room",
    "location": "Table 3, out of operation room",
    "values": {
     "Ketamine group": {
      "hallucination": "3",
      "dizziness": "17",
      "visual": "4"
     },
     "Saline group": {
      "hallucination": "0",
      "dizziness": "8",
      "visual": "3"
     }
    }
   }
  ],
  "blinding_check": null,
  "enrollment": {
   "start": "2015-10-08",
   "end": "2016-03-10"
  },
  "registration": {
   "number": "ChiCTR-IPR-15007159",
   "date": null
  },
  "note": "Randomized per arm not printed (330 enrolled and randomized 1:1; per-arm numbers are only in the flow chart, Figure 1, which has no values in the text). 1 subject missed the 3-day visit and 4 were lost by 6 weeks; 325 analysed (Ketamine 162, Saline 163, as in every table's header). Day 42 = '6 week postpartum'. Depression from Table 4: EPDS mean ± SD and 'occurrence of postpartum depression' (EPDS ≥ 10); both forms are given at each time. Pain: NRS only at 3 days (4 (0–7) vs 4 (0–8), median (range)) and 6 weeks (1 (0–8) vs 2 (0–9); NRS ≤3 129 vs 111, >3 33 vs 52), outside the day 1/2 windows, so not extracted. Adverse events: Table 3 ('Side effects after administration during the operation'; drug given within 5 min of cord clamping) counts patients at each of three checks (5 and 15 min after dosing, and just before leaving the operating room), one entry per check, named in 'time'; a patient can appear at more than one check, so the entries are not to be summed. Diplopia coded visual. Also in Table 3 (5 min, 15 min, leaving; Ketamine vs Saline), not codebook categories: vomiting 3/1, 7/1, 1/2; headache 8/1, 4/0, 3/0; drowsiness 71/4, 26/2, 5/1; Ramsay score >3 15/0, 0/0, 0/0. The Discussion says side effects were not fully documented after the operating room and that no adverse events associated with ketamine were found during postoperative follow-up (no counts). Enrollment dates are from Methods ('conducted from October 8, 2015 to March 10, 2016'); registration has no date.",
  "reason": "The coders differed only on the primary outcome, the occurrence of postpartum depression (EPDS ≥ 10) in Table 4: 41 of 162 vs 46 of 163 at 3 days and 26 vs 29 at 6 weeks. I kept coder A's count entries beside the EPDS means: the codebook asks for every listed measure at every time it is reported, and its count form is exactly a count of participants with depression by the source's definition, so dropping it would lose the trial's primary outcome. The rest is as both coders had it; each adverse-event entry names its check in 'time' and its place in 'location'."
 },
 {
  "trial": "T025",
  "source": "library",
  "arms": {
   "Ketamine group": {
    "randomized": "41",
    "analysed": "41"
   },
   "Placebo group (normal saline)": {
    "randomized": "43",
    "analysed": "43"
   }
  },
  "depression": [
   {
    "measure": "MADRS (also at discharge)",
    "day": -1,
    "form": "median_iqr",
    "location": "Table 1",
    "values": {
     "Ketamine group": {
      "median": "28",
      "q1": "26",
      "q3": "31",
      "n": "41"
     },
     "Placebo group (normal saline)": {
      "median": "27",
      "q1": "25",
      "q3": "31",
      "n": "43"
     }
    }
   },
   {
    "measure": "MADRS (also at discharge)",
    "day": 2,
    "form": "mean_sd",
    "location": "Results, Primary and Secondary Outcomes, second paragraph",
    "values": {
     "Ketamine group": {
      "mean": "16",
      "sd": "9"
     },
     "Placebo group (normal saline)": {
      "mean": "21",
      "sd": "9"
     }
    }
   },
   {
    "measure": "MADRS (also at discharge)",
    "day": 3,
    "form": "count",
    "definition": "responders: MADRS reduction ≥ 50% from baseline (the primary outcome; a count of responders, not of participants with depression)",
    "location": "Table 3 (also abstract: 17/41 vs 7/43)",
    "values": {
     "Ketamine group": {
      "events": "17",
      "n": "41"
     },
     "Placebo group (normal saline)": {
      "events": "7",
      "n": "43"
     }
    }
   },
   {
    "measure": "PHQ-9 (also at discharge)",
    "day": -1,
    "form": "median_iqr",
    "location": "Table 1",
    "values": {
     "Ketamine group": {
      "median": "13",
      "q1": "11",
      "q3": "15",
      "n": "41"
     },
     "Placebo group (normal saline)": {
      "median": "13",
      "q1": "12",
      "q3": "15",
      "n": "43"
     }
    }
   }
  ],
  "pain": [],
  "adverse_events": [
   {
    "window": "postoperative",
    "location": "Table 3, Psychiatric outcomes (postoperative day 3)",
    "values": {
     "Ketamine group": {
      "dissociation": "3"
     },
     "Placebo group (normal saline)": {
      "dissociation": "1"
     }
    }
   }
  ],
  "blinding_check": null,
  "enrollment": {
   "start": "2017-07-05",
   "end": "2018-08-21"
  },
  "registration": {
   "number": "NCT03086148",
   "date": null
  },
  "note": "ITT: all 84 analysed after imputing missing postoperative scores (baseline values for 9 subjects at day 1, 7 at day 2 and 2 at day 3; last observation carried forward for 3 at discharge), so no per-time n is printed for the day 2 means. Baseline (Table 1) was 1 day before surgery (day -1). Day 2 MADRS '16 [SD = 9] vs 21 [SD = 9]' are means (Figure 2 plots means with SD). Day 1 and day 3 MADRS and all postoperative PHQ-9 values are only in Figure 2 (no printed values; PHQ-9 P = .036 at day 2 and .015 at day 3). The day 3 count is the primary outcome; its RR is 2.55 (1.18–5.50) in Table 3, 2.25 in the text and 2.51 in the abstract, with the same CI in each. Not entered: response at day 2, given only as percentages (43.9% vs 20.0%); the MADRS reduction from baseline at day 3, '11 vs 7, respectively; 95% CI, −0.51 to 7.96' (no estimate of the difference printed; whether means or medians not stated); outcomes at discharge, which has no fixed day (hospital stay 7 (6–9) vs 8 (6–10) days): response 16 (39.0) vs 9 (20.9), RR 1.86 (0.93–3.74), and remission (MADRS ≤ 10) 12 (29.3) vs 3 (7.0), RR 4.20 (1.28–13.80). Pain: only cumulative severe pain (NRS ≥5) over the first 3 postoperative days, 19 (47.5) vs 17 (39.5), not at day 1 or 2, so not entered (47.5% does not match 19/41). AEs: Table 3's psychiatric outcomes are on postoperative day 3; dissociative symptoms (CADSS ≠ 0) coded dissociation. Not coded: delirium on day 1 by CAM-ICU, 1 (2.4) vs 3 (7.0), RR 0.35 (0.04–3.23), a secondary outcome not described as agitation or excitement and not on emergence (psychiatric_any does not apply, as specific psychiatric counts are given); manic symptoms (YMRS ≥5) 2 (4.9) vs 1 (2.3); psychotic symptoms (4 BPRS items incl. hallucinations, ≥5) 3 (7.3) vs 4 (9.3), not specific to hallucinations; complications at discharge: aphasia 5 vs 3, epilepsy 1 vs 1, hemorrhage 5 vs 9.",
  "reason": "The coders differed only on delirium on postoperative day 1 by CAM-ICU, 1 (2.4%) vs 3 (7.0%) in Table 3, which coder A coded as agitation. I left it out, as coder B did: the codebook's agitation is agitation, including delirium-like excitement on emergence, and this is delirium of any kind screened the day after surgery, with no excitement described; psychiatric_any doesn't apply because the source gives specific psychiatric counts, including the dissociation kept here. The rest is as both coders had it."
 }
]