array(40) {
  ["request_overridden_res"]=>
  string(1) "3"
  ["project_status"]=>
  string(7) "ongoing"
  ["project_assoc_trials"]=>
  array(1) {
    [0]=>
    object(WP_Post)#4792 (24) {
      ["ID"]=>
      int(1256)
      ["post_author"]=>
      string(4) "1363"
      ["post_date"]=>
      string(19) "2014-10-20 16:11:00"
      ["post_date_gmt"]=>
      string(19) "2014-10-20 16:11:00"
      ["post_content"]=>
      string(0) ""
      ["post_title"]=>
      string(304) "NCT00642278 - A Randomized, Double-Blind, Placebo-Controlled, Double-Dummy, Parallel Group, Multicenter, Dose-Ranging Study in Subjects With Type 2 Diabetes Mellitus to Evaluate the Efficacy, Safety, and Tolerability of Orally Administered SGLT2 Inhibitor JNJ-28431754 With Sitagliptin as a Reference Arm"
      ["post_excerpt"]=>
      string(0) ""
      ["post_status"]=>
      string(7) "publish"
      ["comment_status"]=>
      string(6) "closed"
      ["ping_status"]=>
      string(6) "closed"
      ["post_password"]=>
      string(0) ""
      ["post_name"]=>
      string(189) "nct00642278-a-randomized-double-blind-placebo-controlled-double-dummy-parallel-group-multicenter-dose-ranging-study-in-subjects-with-type-2-diabetes-mellitus-to-evaluate-the-efficacy-safety"
      ["to_ping"]=>
      string(0) ""
      ["pinged"]=>
      string(0) ""
      ["post_modified"]=>
      string(19) "2025-10-24 15:20:52"
      ["post_modified_gmt"]=>
      string(19) "2025-10-24 19:20:52"
      ["post_content_filtered"]=>
      string(0) ""
      ["post_parent"]=>
      int(0)
      ["guid"]=>
      string(238) "https://dev-yoda.pantheonsite.io/clinical-trial/nct00642278-a-randomized-double-blind-placebo-controlled-double-dummy-parallel-group-multicenter-dose-ranging-study-in-subjects-with-type-2-diabetes-mellitus-to-evaluate-the-efficacy-safety/"
      ["menu_order"]=>
      int(0)
      ["post_type"]=>
      string(14) "clinical_trial"
      ["post_mime_type"]=>
      string(0) ""
      ["comment_count"]=>
      string(1) "0"
      ["filter"]=>
      string(3) "raw"
    }
  }
  ["project_title"]=>
  string(88) "Covariate-Adjusted Dose–Response Analysis With Incomplete Outcomes in a Phase II Trial"
  ["project_narrative_summary"]=>
  string(759) "Dose-ranging trials compare several doses to determine whether a treatment shows a dose-response relationship and which doses should be studied further. Standard MCP-Mod analyses may not fully use baseline patient information and can be affected by incomplete outcome data. We will use participant-level data from a phase II canagliflozin trial to evaluate MCP-ModX, a method that combines covariate adjustment with principled handling of missing outcomes while estimating the marginal dose-response curve. We will compare MCP-ModX with standard analyses and assess whether it improves the precision and reliability of proof-of-concept testing and dose-response estimation. This work may improve the analysis and design of future dose-ranging clinical trials."
  ["project_learn_source"]=>
  string(9) "colleague"
  ["principal_investigator"]=>
  array(7) {
    ["first_name"]=>
    string(9) "Cheng-Han"
    ["last_name"]=>
    string(4) "Yang"
    ["degree"]=>
    string(3) "PhD"
    ["primary_affiliation"]=>
    string(28) "Yale School of Public Health"
    ["email"]=>
    string(23) "cheng-han.yang@yale.edu"
    ["state_or_province"]=>
    string(2) "CT"
    ["country"]=>
    string(13) "United States"
  }
  ["project_key_personnel"]=>
  bool(false)
  ["project_ext_grants"]=>
  array(2) {
    ["value"]=>
    string(2) "no"
    ["label"]=>
    string(68) "No external grants or funds are being used to support this research."
  }
  ["project_date_type"]=>
  string(18) "full_crs_supp_docs"
  ["property_scientific_abstract"]=>
  string(1495) "Background: Phase II dose-ranging trials must detect a dose-response signal and characterize the dose-response curve. Prognostic baseline covariates can improve precision, whereas incomplete target-time outcomes can reduce information and compromise analyses that ignore the observation process. 

Objective: To evaluate a proposed extension of generalized MCP-Mod that estimates covariate-adjusted marginal outcomes at each dose and accommodates incomplete outcomes through augmented inverse-probability-weighted standardization; we refer to the integrated procedure as MCP-ModX. 

Study Design: Methodological research using secondary analysis of a randomized phase II dose-ranging trial. 

Participants: The 322 participants randomized to placebo or canagliflozin 50, 100, 200, or 300 mg once daily in NCT00642278. 

Primary and Secondary Outcome Measure(s): Primary: change in HbA1c from baseline to Week 12. Secondary: Week-12 HbA1c <7%, change in fasting plasma glucose, and change in body weight. 

Statistical Analysis: We will compare generalized MCP-Mod, MCP-ModX, complete-case analysis, and the original last-observation-carried-forward approach. MCP-ModX will estimate marginal dose-specific outcomes and their joint covariance, test prespecified candidate dose-response shapes with multiplicity control, and fit supported models by generalized least squares. Sensitivity analyses will vary prespecified outcome and observation-model specifications."
  ["project_brief_bg"]=>
  string(2759) "Phase II dose-ranging trials support two linked decisions: whether there is evidence of a dose-response signal and how the response changes across doses. MCP-Mod addresses these goals by testing prespecified candidate dose-response shapes with multiplicity control and then fitting supported dose-response models [1]. Generalized MCP-Mod extends this framework to non-normal outcomes and to analyses in which suitable first-stage estimates and their covariance are supplied to the MCP and Mod stages [2].



Two features motivate the proposed work. First, randomized trials routinely collect prognostic baseline characteristics that can improve the precision of treatment-effect estimation when used appropriately [3]. Second, target-time outcomes may be incomplete. Complete-case analysis discards information and can be biased when outcome availability depends on measured baseline characteristics. Under a missing-at-random assumption, augmented inverse-probability-weighted estimation can combine an outcome model with a model for outcome availability and retain consistency when either working model is correctly specified [4].



We will evaluate a proposed MCP-Mod extension that integrates these ideas at the level of the marginal dose-response curve. For each randomized dose, the procedure estimates the marginal outcome in the trial population by combining standardized outcome predictions with an inverse-probability-weighted residual correction. A joint sandwich covariance accounts for estimation of the working outcome and observation models and is carried into both MCP-Mod stages. We refer to this integrated procedure as MCP-ModX.



NCT00642278 is a randomized, double-blind, phase II dose-ranging trial of canagliflozin in adults with type 2 diabetes. The trial randomized 451 participants among placebo, canagliflozin 50, 100, 200, or 300 mg once daily, canagliflozin 300 mg twice daily, and sitagliptin 100 mg once daily; the primary endpoint was change in HbA1c from baseline to Week 12 [5]. The original efficacy analysis used ANCOVA with baseline HbA1c and a randomization stratification factor and imputed missing efficacy outcomes using last observation carried forward [5].



This trial therefore provides a relevant empirical setting for evaluating whether a marginal, covariate-adjusted analysis with explicit treatment of outcome availability materially changes uncertainty, proof-of-concept testing, or dose-response estimation. The project is methodological; its purpose is not to revise the original clinical efficacy conclusions, but to evaluate a generalizable framework for future randomized dose-ranging trials with prognostic baseline information and incomplete outcomes."
  ["project_specific_aims"]=>
  string(1179) "Aim 1: Estimate the marginal Week-12 HbA1c dose-response relationship across placebo and canagliflozin 50, 100, 200, and 300 mg once daily using conventional generalized MCP-Mod and the proposed MCP-ModX procedure.

Aim 2: Quantify how prespecified baseline prognostic covariate adjustment and explicit modeling of Week-12 outcome availability affect dose-specific estimates, their joint covariance, proof-of-concept testing, and fitted dose-response curves.

Aim 3: Assess robustness to prespecified alternative outcome and observation-model specifications and compare results with complete-case analysis and the trial’s original last-observation-carried-forward approach. Secondary analyses will examine Week-12 HbA1c <7%, fasting plasma glucose, and body weight.



We hypothesize that prognostic covariate adjustment will improve precision when the selected baseline variables are predictive of outcome, and that modeling outcome availability may alter estimates or uncertainty when missingness is associated with measured baseline information. These are empirical hypotheses; improved precision or changed conclusions are not assumed in advance."
  ["project_study_design"]=>
  array(2) {
    ["value"]=>
    string(8) "meth_res"
    ["label"]=>
    string(23) "Methodological research"
  }
  ["project_purposes"]=>
  array(2) {
    [0]=>
    array(2) {
      ["value"]=>
      string(37) "develop_or_refine_statistical_methods"
      ["label"]=>
      string(37) "Develop or refine statistical methods"
    }
    [1]=>
    array(2) {
      ["value"]=>
      string(34) "research_on_clinical_trial_methods"
      ["label"]=>
      string(34) "Research on clinical trial methods"
    }
  }
  ["project_research_methods"]=>
  string(1297) "Data source: De-identified individual participant-level data from NCT00642278 provided through the YODA Project secure platform.



The primary analytic sample will include all randomized participants assigned to placebo or canagliflozin 50, 100, 200, or 300 mg once daily (322 participants based on the published randomization counts [5]). These five arms define a common once-daily scalar dose range. The canagliflozin 300 mg twice-daily arm will be excluded because dosing frequency differs from the once-daily series, and the sitagliptin arm will be excluded because it is an active reference treatment rather than a canagliflozin dose.



No additional demographic or clinical exclusion criteria will be applied within the selected randomized arms beyond the original trial eligibility criteria. Participants will not be excluded because the Week-12 outcome is missing; outcome availability is part of the analysis. Analyses will follow randomized assignment.



No participant-level data from studies outside the YODA Project will be pooled with the requested trial data. Outcome, visit, and eligibility definitions will be verified against the protocol, statistical analysis plan, clinical study report, and data-definition documentation before analysis."
  ["project_main_outcome_measure"]=>
  string(1182) "Primary outcome: Change in HbA1c (%) from baseline to the scheduled Week-12 assessment, corresponding to the original trial’s primary efficacy endpoint [5]. For each participant, we will define an indicator for whether the Week-12 HbA1c required to construct this endpoint is observed. Participants with missing Week-12 HbA1c will remain in the target analytic population.



Secondary outcomes:

1. Binary glycemic response at Week 12, defined as HbA1c <7.0%.

2. Change in fasting plasma glucose from baseline to Week 12.

3. Change in body weight from baseline to Week 12.



These secondary outcomes were reported in the original trial [5]. Exact units, scheduled visit windows, derivation rules, and analysis flags will be verified from the protocol, statistical analysis plan, clinical study report, and shared data documentation before analysis.



The primary MCP-ModX analysis will not treat an unavailable Week-12 measurement as observed through last observation carried forward. LOCF will be reproduced, if supported by the shared data and documentation, only as a historical comparator to the original analysis."
  ["project_main_predictor_indep"]=>
  string(809) "The main independent variable is randomized treatment dose. The primary analysis will use five ordered once-daily dose levels: placebo (0 mg) and canagliflozin 50, 100, 200, and 300 mg once daily. Treatment will be analyzed according to randomized assignment.



Dose is the intervention defining the marginal dose-response curve. The first-stage estimation will treat randomized arm categorically to avoid imposing a dose-response shape before testing. The MCP stage will then evaluate prespecified candidate dose-response shapes across the numeric dose levels.



The 300 mg twice-daily canagliflozin regimen and sitagliptin reference arm will not be included in the primary scalar dose-response analysis because they do not belong to the common once-daily canagliflozin dose series."
  ["project_other_variables_interest"]=>
  string(1357) "Primary prognostic adjustment will use variables measured before randomization. The prespecified core set will include baseline HbA1c, baseline fasting plasma glucose, age as released in the de-identified dataset, sex, body mass index, duration of type 2 diabetes, and the randomization stratification indicator used in the original primary analysis, subject to confirmation of their representation in the shared data. Baseline body weight will be summarized descriptively; because it is closely related to BMI, the primary model will not automatically include both.



The same prespecified baseline information, together with randomized treatment arm, will be considered in the model for whether the Week-12 outcome is observed. Covariates are included for prognostic adjustment and modeling outcome availability, not to control confounding of randomized treatment assignment.



The core adjustment set will be specified before fitting outcome-association models and will not be selected by stepwise significance testing. Additional baseline variables documented in the protocol or data specification may be examined in sensitivity analyses only when their scientific role, coding, and completeness can be prespecified. Post-randomization measurements will not be treated as baseline adjustment covariates in the primary analysis."
  ["project_stat_analysis_plan"]=>
  string(4560) "The primary analysis will preserve randomized treatment assignment and target marginal Week-12 outcomes in the randomized trial population.



1. Data verification and descriptive analysis. We will verify treatment coding, visit definitions, baseline measurements, Week-12 outcome availability, treatment/study discontinuation variables, and relevant analysis flags against the protocol, SAP, CSR, and data-definition documentation. Baseline characteristics and Week-12 outcome availability will be summarized by randomized arm. Baseline characteristics will also be summarized by Week-12 observation status to describe the observed missing-data pattern; these comparisons will not be interpreted as tests of the missing-at-random assumption.



2. Estimands. For continuous outcomes, the targets are the marginal mean change from baseline at each selected dose and placebo-adjusted marginal mean differences at Week 12. For the binary HbA1c <7% endpoint, the targets are marginal response probabilities and prespecified placebo-adjusted contrasts. We will distinguish these marginal quantities from conditional regression coefficients.



3. Reference analyses. Generalized MCP-Mod will be applied using dose-specific estimates and their estimated covariance for the five selected once-daily dose levels [1,2]. A complete-case analysis will provide a reference for the effect of excluding missing target-time outcomes. When reproducible from the shared data, the original LOCF analysis will be included as a historical comparator [5].



4. MCP-ModX first stage. Under the primary assumption that Week-12 outcome availability is missing at random conditional on randomized arm and prespecified baseline information, we will estimate dose-specific marginal outcomes by augmented inverse-probability-weighted standardization [4]. A working outcome model will relate the endpoint to randomized arm and prespecified baseline prognostic covariates. A working observation model will estimate the probability that the target-time outcome is observed using randomized arm and prespecified baseline variables. The estimator combines standardized outcome predictions with inverse-probability-weighted observed-outcome residuals. Under the stated assumptions, it is consistent if either the working outcome model or the working observation model is correctly specified. For binary outcomes, fitted conditional probabilities will be standardized over the empirical baseline-covariate distribution to obtain marginal response probabilities.



5. Joint covariance and MCP step. The vector of marginal dose-specific estimates will be accompanied by a stacked M-estimation sandwich covariance that accounts for estimation of the working outcome and observation models. The estimates and covariance will be transformed to placebo-adjusted dose-response quantities. Before fitting the MCP tests, we will prespecify clinically plausible candidate shapes, including linear, Emax, and quadratic models, and fix any shape parameters needed to construct the candidate contrasts. Optimal contrasts and the joint distribution of the contrast statistics will be used for multiplicity-controlled proof-of-concept testing.



6. Mod step. Candidate models supported by the MCP step will be fitted to the estimated marginal dose-response vector by generalized least squares using its estimated covariance. Model-specific fitted curves and uncertainty will be reported. If multiple models are supported, any model-selection or model-averaging rule will be prespecified before the final analysis.



7. Sensitivity and secondary analyses. We will examine prespecified alternative functional forms and covariate sets for the outcome and observation models, while keeping the target estimands fixed. We will compare MCP-ModX with complete-case and LOCF analyses and report whether proof-of-concept conclusions, dose-specific estimates, and fitted curves materially change. Secondary analyses will repeat the framework for Week-12 HbA1c <7%, fasting plasma glucose, and body weight. No confirmatory subgroup analysis is planned.



The empirical trial analysis is intended to evaluate the operating implications and practical implementation of the proposed methodology in a genuine dose-ranging setting. It will not be used to claim that the missing-at-random assumption is empirically verified, and no benefit of covariate adjustment or missing-outcome modeling will be assumed in advance."
  ["project_software_used"]=>
  array(2) {
    [0]=>
    array(2) {
      ["value"]=>
      string(1) "r"
      ["label"]=>
      string(1) "R"
    }
    [1]=>
    array(2) {
      ["value"]=>
      string(7) "rstudio"
      ["label"]=>
      string(7) "RStudio"
    }
  }
  ["project_timeline"]=>
  string(1087) "Project start: upon approval and activation of data access.

Months 1–2: Review protocol, SAP, CSR, and data definitions; verify treatment, endpoint, visit, covariate, and missingness variables; construct the analytic dataset; finalize prespecified candidate dose-response models and analysis specifications.

Months 3–5: Conduct descriptive analyses and primary generalized MCP-Mod and MCP-ModX analyses; complete secondary and sensitivity analyses; verify numerical results and reproducibility within the secure platform.

Months 6–7: Integrate the empirical results with the methodological manuscript and complete internal scientific review.

Months 8–9: Prepare and submit the manuscript to a peer-reviewed journal.

Months 10–12: Address initial editorial or reviewer requests as feasible, report study status/results to the YODA Project, and prepare permitted supporting materials.

All analyses will be completed within the authorized data-access period; milestone dates will shift accordingly if access begins later than anticipated."
  ["project_dissemination_plan"]=>
  string(1078) "We plan to disseminate the work as a peer-reviewed statistical-methodology manuscript describing a covariate-adjusted extension of MCP-Mod for randomized dose-ranging trials with incomplete outcomes and demonstrating the approach using NCT00642278.



The primary audiences are biostatisticians, clinical-trial methodologists, statisticians in pharmaceutical development, and investigators designing or analyzing phase II dose-ranging trials. Potential journals include Biometrics, Biometrical Journal, Statistics in Medicine, and Pharmaceutical Statistics.



Results will be reported regardless of whether the proposed method materially changes the original trial conclusions or improves precision in this dataset. Publications and presentations will report only aggregate results permitted under the YODA Data Use Agreement and applicable disclosure requirements. No attempt will be made to identify participants. Methodological code that contains no participant-level data may be made publicly available subject to YODA and data-partner requirements."
  ["project_bibliography"]=>
  string(1140) "

Bretz F, Pinheiro JC, Branson M. Combining multiple comparisons and modeling techniques in dose-response studies. Biometrics. 2005;61(3):738–748. doi:10.1111/j.1541-0420.2005.00344.x.
Pinheiro J, Bornkamp B, Glimm E, Bretz F. Model-based dose finding under model uncertainty using general parametric models. Statistics in Medicine. 2014;33(10):1646–1661. doi:10.1002/sim.6052.
U.S. Food and Drug Administration. Adjusting for Covariates in Randomized Clinical Trials for Drugs and Biological Products: Guidance for Industry. May 2023.
Bang H, Robins JM. Doubly robust estimation in missing data and causal inference models. Biometrics. 2005;61(4):962–973. doi:10.1111/j.1541-0420.2005.00377.x.
Rosenstock J, Aggarwal N, Polidori D, et al. Dose-ranging effects of canagliflozin, a sodium-glucose cotransporter 2 inhibitor, as add-on to metformin in subjects with type 2 diabetes. Diabetes Care. 2012;35(6):1232–1238. doi:10.2337/dc11-1926.

"
  ["project_suppl_material"]=>
  bool(false)
  ["project_coi"]=>
  array(1) {
    [0]=>
    array(1) {
      ["file_coi"]=>
      array(21) {
        ["ID"]=>
        int(20188)
        ["id"]=>
        int(20188)
        ["title"]=>
        string(40) "SV_57KskaKADT3U9Aq-R_1YGioHXT7JHnKrZ.pdf"
        ["filename"]=>
        string(40) "SV_57KskaKADT3U9Aq-R_1YGioHXT7JHnKrZ.pdf"
        ["filesize"]=>
        int(34542)
        ["url"]=>
        string(89) "https://yoda.yale.edu/wp-content/uploads/2026/10/SV_57KskaKADT3U9Aq-R_1YGioHXT7JHnKrZ.pdf"
        ["link"]=>
        string(86) "https://yoda.yale.edu/data-request/2026-1072/sv_57kskakadt3u9aq-r_1ygiohxt7jhnkrz-pdf/"
        ["alt"]=>
        string(0) ""
        ["author"]=>
        string(4) "2611"
        ["description"]=>
        string(0) ""
        ["caption"]=>
        string(0) ""
        ["name"]=>
        string(40) "sv_57kskakadt3u9aq-r_1ygiohxt7jhnkrz-pdf"
        ["status"]=>
        string(7) "inherit"
        ["uploaded_to"]=>
        int(20187)
        ["date"]=>
        string(19) "2026-10-06 06:02:22"
        ["modified"]=>
        string(19) "2026-10-06 06:02:24"
        ["menu_order"]=>
        int(0)
        ["mime_type"]=>
        string(15) "application/pdf"
        ["type"]=>
        string(11) "application"
        ["subtype"]=>
        string(3) "pdf"
        ["icon"]=>
        string(62) "https://yoda.yale.edu/wp/wp-includes/images/media/document.png"
      }
    }
  }
  ["data_use_agreement_training"]=>
  bool(true)
  ["human_research_protection_training"]=>
  bool(true)
  ["certification"]=>
  bool(true)
  ["search_order"]=>
  string(1) "0"
  ["project_send_email_updates"]=>
  bool(false)
  ["project_publ_available"]=>
  bool(true)
  ["project_year_access"]=>
  string(4) "2026"
  ["project_rep_publ"]=>
  bool(false)
  ["project_assoc_data"]=>
  array(0) {
  }
  ["project_due_dil_assessment"]=>
  array(21) {
    ["ID"]=>
    int(20297)
    ["id"]=>
    int(20297)
    ["title"]=>
    string(47) "YODA Project Due Diligence Assessment 2026-1072"
    ["filename"]=>
    string(51) "YODA-Project-Due-Diligence-Assessment-2026-1072.pdf"
    ["filesize"]=>
    int(108839)
    ["url"]=>
    string(100) "https://yoda.yale.edu/wp-content/uploads/2026/10/YODA-Project-Due-Diligence-Assessment-2026-1072.pdf"
    ["link"]=>
    string(93) "https://yoda.yale.edu/data-request/2026-1072/yoda-project-due-diligence-assessment-2026-1072/"
    ["alt"]=>
    string(0) ""
    ["author"]=>
    string(4) "1885"
    ["description"]=>
    string(0) ""
    ["caption"]=>
    string(0) ""
    ["name"]=>
    string(47) "yoda-project-due-diligence-assessment-2026-1072"
    ["status"]=>
    string(7) "inherit"
    ["uploaded_to"]=>
    int(20187)
    ["date"]=>
    string(19) "2026-10-21 19:12:18"
    ["modified"]=>
    string(19) "2026-10-21 19:12:18"
    ["menu_order"]=>
    int(0)
    ["mime_type"]=>
    string(15) "application/pdf"
    ["type"]=>
    string(11) "application"
    ["subtype"]=>
    string(3) "pdf"
    ["icon"]=>
    string(62) "https://yoda.yale.edu/wp/wp-includes/images/media/document.png"
  }
  ["project_title_link"]=>
  array(21) {
    ["ID"]=>
    int(20298)
    ["id"]=>
    int(20298)
    ["title"]=>
    string(44) "YODA Project Protocol - 2026-1072 - 26-10-06"
    ["filename"]=>
    string(44) "YODA-Project-Protocol-2026-1072-26-10-06.pdf"
    ["filesize"]=>
    int(129913)
    ["url"]=>
    string(93) "https://yoda.yale.edu/wp-content/uploads/2026/10/YODA-Project-Protocol-2026-1072-26-10-06.pdf"
    ["link"]=>
    string(86) "https://yoda.yale.edu/data-request/2026-1072/yoda-project-protocol-2026-1072-26-10-06/"
    ["alt"]=>
    string(0) ""
    ["author"]=>
    string(4) "1885"
    ["description"]=>
    string(0) ""
    ["caption"]=>
    string(0) ""
    ["name"]=>
    string(40) "yoda-project-protocol-2026-1072-26-10-06"
    ["status"]=>
    string(7) "inherit"
    ["uploaded_to"]=>
    int(20187)
    ["date"]=>
    string(19) "2026-10-21 19:12:34"
    ["modified"]=>
    string(19) "2026-10-21 19:12:34"
    ["menu_order"]=>
    int(0)
    ["mime_type"]=>
    string(15) "application/pdf"
    ["type"]=>
    string(11) "application"
    ["subtype"]=>
    string(3) "pdf"
    ["icon"]=>
    string(62) "https://yoda.yale.edu/wp/wp-includes/images/media/document.png"
  }
  ["project_review_link"]=>
  array(21) {
    ["ID"]=>
    int(20299)
    ["id"]=>
    int(20299)
    ["title"]=>
    string(36) "YODA Project Review - 2026-1072_site"
    ["filename"]=>
    string(38) "YODA-Project-Review-2026-1072_site.pdf"
    ["filesize"]=>
    int(1321265)
    ["url"]=>
    string(87) "https://yoda.yale.edu/wp-content/uploads/2026/10/YODA-Project-Review-2026-1072_site.pdf"
    ["link"]=>
    string(80) "https://yoda.yale.edu/data-request/2026-1072/yoda-project-review-2026-1072_site/"
    ["alt"]=>
    string(0) ""
    ["author"]=>
    string(4) "1885"
    ["description"]=>
    string(0) ""
    ["caption"]=>
    string(0) ""
    ["name"]=>
    string(34) "yoda-project-review-2026-1072_site"
    ["status"]=>
    string(7) "inherit"
    ["uploaded_to"]=>
    int(20187)
    ["date"]=>
    string(19) "2026-10-21 19:12:50"
    ["modified"]=>
    string(19) "2026-10-21 19:12:50"
    ["menu_order"]=>
    int(0)
    ["mime_type"]=>
    string(15) "application/pdf"
    ["type"]=>
    string(11) "application"
    ["subtype"]=>
    string(3) "pdf"
    ["icon"]=>
    string(62) "https://yoda.yale.edu/wp/wp-includes/images/media/document.png"
  }
  ["project_highlight_button"]=>
  string(0) ""
  ["request_data_partner"]=>
  string(0) ""
}
data partner
array(1) {
  [0]=>
  string(0) ""
}


pi country
array(0) {
}


pi affil
array(0) {
}


products
array(0) {
}


num of trials
array(1) {
  [0]=>
  string(1) "0"
}


res
array(1) {
  [0]=>
  string(1) "3"
}