https://doi.org/10.5061/dryad.j6q573npb

Objective

The objective of this study was to determine if utilizing a contraceptive-specific electronic health records (EHR) system could improve potential barriers to the implementation of pharmacist-prescribed contraceptive services.

Materials and Methods

Study Design

• Type: Feasibility study
• Design: Crossover study
• Participants: 20 US-based pharmacists with active pharmacy licenses and patient care experience
• Encounters: Each pharmacist performed two standardized patient encounter simulations: one using the EHR and one using the paper-based workflow.
• Randomization: Pharmacists were assigned to odd and even groups via randomizer software (https://app.studyrandomizer.com). Odd-numbered participants began with the paper-based workflow; even-numbered participants began with the EHR workflow.

Data Collection

• Timing and Outputs: Encounters were timed, and contraceptive outputs were recorded by a third-party moderator in an Excel spreadsheet.
• Surveys:
  • Workload (NASA-TLX) and Usability (Health ITUES) Surveys: Completed after each encounter on paper forms, photographed, and mailed to the study team.
  • Perception, Attitude, and Satisfaction (PAS) Survey: Administered electronically after the final encounter.
• Data Recording: Data was recorded using the pharmacist's randomized number and the encounter designation (Patient 1 or Patient 2).

Data Processing

• Verification: Contraceptive outputs were verified on paper and electronic medical records.
• Transcription: Survey results were transcribed into the Excel spreadsheet.
• Statistical Analysis: Data from the Excel spreadsheet was uploaded to R software version 4.2.2 for statistical analysis.

Results

• Contraceptive Ineligibility: Pharmacists were more likely to identify contraceptive ineligibility using the EHR-based workflow compared to the paper workflow (p=0.003).
• Encounter Time: No significant difference in contraceptive encounter time between the two modalities (p=0.280).
• Mental Demand: Pharmacists reported lower mental demand with the EHR-based workflow (p=0.003).
• Perceived Usefulness: Greater perceived usefulness was reported for the EHR-based workflow (p=0.029).

Discussion and Conclusion

Pharmacist performance and acceptance of contraceptive service delivery were improved with the EHR workflow. The study suggests that pharmacist-specific contraceptive EHR workflows have potential to enhance the adoption and provision of appropriate contraceptive care.

Description of the data and file structure

• Dataset File: JAMIO_Phase_1_Data.zip
• Sheets and Descriptions:
  • Cohort Demographics: De-identified information about the pharmacists involved in the study including age range, sex, race, ethnicity, experience as a pharmacist (years), current practice setting, current state of practice, experience prescribing contraceptives, and experience with electronic medical records systems for patient-care activities. To decrease indirect identifiers the age and experience in years were converted to ranges. A question that was not answered is represented by NULL.
  • Encounter Completion Time: The time it took for the pharmacist to complete the contraceptive encounter by patient scenario (Patient 1 vs Patient 2) and modality (Oregon SSRAQ vs EHR). The pharmacists are listed by their randomized study number for deidentification. Each combination of the study variables is represented in a distinct table and ordered in a vertical orientation. The times were recorded by an independent third-party moderator using a standard digital stopwatch and recorded in MM:SS (MM = minutes, SS = seconds) and converted to seconds for analysis. 
  • Contraceptive Outputs: The clinical decision for a contraceptive and/or referral by each pharmacist under each condition. The pharmacists are listed by their randomized study number for deidentification. Each combination of the study variables is represented in a distinct table and ordered in a vertical orientation. The contraceptive selection is represented for both patient scenarios. The Patient 2 scenario has two additional fields, contraceptive category and progestin active pharmaceutical ingredient, which were recorded because the success criteria for the Patient 2 scenario depended on these factors. Where the contraceptive selection was a referral, the category and progestin selection fields were not applicable. The contraceptive outputs were recorded by an independent third-party moderator and were verified by the study team by reviewing the pharmacist documentation on the patient form or electronic health record.
  • NASA-TLX: The results of the NASA-TLX survey which was administered on paper by the independent third-party moderator after each of the patient simulations for each pharmacist. The pharmacists are listed by their randomized study number for deidentification. Each combination of the study variables is represented in a distinct table and ordered in a vertical orientation. The responses to the survey for each domain, the weight of each domain, and the weight-adjusted domain score was entered for each of the six domains followed by the overall raw and weight-adjusted scores in a horizontal orientation for each pharmacist. The numbers entered were obtained directly from the survey's and calculated from the Sources-of-Workload Tally Sheet (Appendix D) and Weighted Rating Worksheet (Appendix E) that can be found here: TLX_pappen_manual.pdf (nasa.gov)
  • Health ITUES: The results of the Health-ITUES survey which was administered on paper by the independent third-party moderator after each of the patient simulations for each pharmacist. The pharmacists are listed by their randomized study number for deidentification. Each combination of the study variables is represented in a distinct table and ordered in a vertical orientation. The responses to the survey for each question grouped by domain with the order the questions appeared on the survey completed by the pharmacists in the cohort displayed in parenthesis. The final column in each table is the overall score which is a sum of all the numbers for the individual responses to all questions.
  • PAS Survey: The results of the dichotomous and Likert on the PAS Survey which was administered electronically following the final patient simulation after the NASA-TLX and Health-ITUES were completed. The pharmacists are listed by their randomized study number for deidentification in order from 1-20. The statements are in row 1 and the direct responses from the participants are shown in rows 2-21.
  • PAS Survey Open-Ended: The results of the open-ended questions on the PAS Survey which was administered electronically following the final patient simulation after the NASA-TLX and Health-ITUES were completed. The pharmacists are listed by their randomized study number and broken into Odd and Even groups in a vertical orientation. The questions are in row 1 and the exact responses from the participants are shown in the subsequent rows. When a question was left blank by a participant, the word NULL has been inserted. 

Definitions and Units

• Randomized Number: Unique identifier assigned to each pharmacist for anonymization and de-identification.
• Patient 1 / Patient 2: Designation for the sequence of standardized patient encounters, corresponds to a specific standardized patient script.
• Encounter Time: Time taken for each contraceptive encounter, recorded in minutes/seconds.
• Contraceptive Outputs: Decision outcomes regarding contraceptive eligibility and recommendations by the pharmacist performing the standardized encounter.

External Resources

• Study Randomizer: Used for randomizing participant assignments.
• R Software: Used for statistical analysis.

Sharing/Access information

Creative Commons Attribution Non-Commercial license (OA-CC-BY-NC 4.0).  Please see the following website for instructions: http://creativecommons.org/licenses/by-nc/4.0/.