Page 114 - JCTR-11-5

Journal of Clinical and
Translational Research Predicting coordinator workload with adapted OPAL

error (MAE) and mean difference–were calculated, The MAE between the adapted OPAL-based estimated
consistent with early-stage predictive model validation workload and the actual coordinator hours logged for
practices. A paired Student’s t-test was used to assess the 12-month study period was 167.0 h, equivalent to
16
differences between estimated and actual hours, as both approximately 22.3 workdays (4.5 weeks or 1.0 month)
values were generated from the same coordinator-study using a 7.5-h institutional workday and 20.9 workdays
pairing. Unpaired Student’s t-tests were used for subgroup (4.2 weeks or 1.0 month) using an 8-h industry-standard
analyses (i.e., sponsor type and intervention type) to workday. Despite this variability, the average difference
compare mean values between independent groups. 17 between estimated and actual hours across all trials
These methods enable both absolute and relative was relatively modest at 24.1 h and was not statistically
evaluations of prediction accuracy, highlighting areas significant (t = 0.32, p=0.761). This difference represents
for potential refinements of the adapted OPAL score. approximately 7–8% of a full-time coordinator’s annual
Workday equivalencies were computed using 7.5-h effort. A detailed summary is provided in Table 2 and
institutional and 8-h industry standards. This adjustment Figure 1.
ensures consistency when comparing internal workloads When analyzed by sponsor type, industry-sponsored
to external benchmarks. p-values were not calculated for trials required more coordinator time, with an average of
individual trials because each trial’s estimated and actual 422.8 actual hours compared to 246.0 h for federally funded
values represent a single paired observation for the entire trials. This difference indicated a trend toward significance
12-month period, making statistical significance testing at (t = −2.06, p=0.095), suggesting that sponsor type may be
the trial level mathematically inappropriate. The reported influential in predicting coordinator burden. Although
p-value for the “estimated versus actual comparison” was industry trials demonstrated slightly higher average
calculated using the aggregated paired dataset across all adapted OPAL scores, the difference was not statistically
trials, allowing for appropriate variance estimation. significant (Figure 2). No substantial differences between
This study did not require Institutional Review Board drug and behavioral intervention trials were observed
approval, as it was classified as a quality improvement and
operational research initiative aimed at enhancing internal
clinical trial management processes. No identifiable
private information was collected, and the project was not
considered human subjects research.

3. Results
The seven interventional trials included in this study
consisted of five Phase 3 trials, one Phase 2/3 hybrid trial,
and one Phase 2 trial, spanning a range of therapeutic areas.

Table 1. Reference table for the adapted Ontario Protocol
Assessment Level score Figure 1. Estimated versus actual hours over a 12-month period
Adapted Ontario Protocol Estimated hours Estimated hours
Assessment Level score (6‑month period) per month
5.5 30.7 5.1
6.0 69.3 11.5
6.5 107.9 18.0
7.0 146.5 24.4
7.5 185.1 30.9
8.0 223.7 37.3
8.5 262.3 43.7
9.0 301.0 50.2
9.5 339.6 56.6
Note: Estimated coordinator workload hours were calculated using the
adapted Ontario Protocol Assessment Level reference table previously
published by Tyson et al. 9 Figure 2. Comparison of actual coordinator hours by sponsor type

Volume 11 Issue 5 (2025) 108 doi: 10.36922/JCTR025260032