Mobile electronic versus paper case report forms in clinical trials: a randomized controlled trial

Abstract

Background: Regulations, study design complexity and amounts of collected and shared data in clinical trials render efficient data handling procedures inevitable. Recent research suggests that electronic data capture can be key in this context but evidence is insufficient. This randomized controlled parallel group study tested the hypothesis that time efficiency is superior when electronic (eCRF) instead of paper case report forms (pCRF) are used for data collection. We additionally investigated predictors of time saving effects and data integrity.

Methods: This study was conducted on top of a clinical weight loss trial performed at a clinical research facility over six months. All study nurses and patients participating in the clinical trial were eligible to participate and randomly allocated to enter cross-sectional data obtained during routine visits either through pCRF or eCRF. A balanced randomization list was generated before enrolment commenced. 90 and 30 records were gathered for the time that 27 patients and 2 study nurses required to report 2025 and 2037 field values, respectively. The primary hypothesis, that eCRF use is faster than pCRF use, was tested by a two-tailed t-test. Analysis of variance and covariance were used to evaluate predictors of entry performance. Data integrity was evaluated by descriptive statistics.

Results: All randomized patients were included in the study (eCRF group n = 13, pCRF group n = 14). eCRF, as compared to pCRF, data collection was associated with significant time savings across all conditions (8.29 ± 5.15 min vs. 10.54 ± 6.98 min, p = .047). This effect was not defined by participant type, i.e. patients or study nurses (F_(1,112) = .15, p = .699), CRF length (F_(2,112) = .49, p = .609) or patient age (Beta = .09, p = .534). Additional 5.16 ± 2.83 min per CRF were saved with eCRFs due to data transcription redundancy when patients answered questionnaires directly in eCRFs. Data integrity was superior in the eCRF condition (0 versus 3 data entry errors).

Conclusions: This is the first study to prove in direct comparison that using eCRFs instead of pCRFs increases time efficiency of data collection in clinical trials, irrespective of item quantity or patient age, and improves data quality.

Trial registration: Clinical Trials NCT02649907 .

Keywords: Data handling; Electronic case report form; REDCap; Time efficiency.

Fig. 1

Summary of the study design. A nutritional assessment study served as use case to test our hypotheses regarding data collection methods in clinical trials. In this study, patient information was collected in two ways. Patients either entered information themselves into standardized instruments such as questionnaires or information about the patient was obtained by a study nurse and then entered. Entering patient information could be done in two ways that included either direct access to the digital database through an electronic case report form (eCRF) or indirect access by filling a paper based case report form (pCRF) that was subsequently transferred to the database (pCRF to eCRF). Precise records were taken for times required for data entry by either CRF type, subject type performing data entry and instrument type being used

Fig. 2

CONSORT 2010 Flow Diagram. This diagram illustrates that all patients that were assessed for eligibility to participate in this study also agreed to participate. Non of the included patients withdrew consent or decided not to use the randomized data entry method. Given the cross-sectional design without follow-up there was also no loss to follow-up. Study nurses were not included in this diagram since they were not allocated one particular intervention, i.e. data entry method, but changed methods between visits

Fig. 3

Plot of average time consumption for data entry procedures. Patients and study nurses could enter data to the database in two ways; they either entered data directly through eCRF or first to a pCRF whose content was subsequently transferred from pCRF to eCRF resulting in a pCRF total time. Results show that entering data through eCRF is significantly faster than the complete pCRF procedure (‑2.25 ± .99 min, p = .047). Errors bars represent standard deviation