Table 3 Summary of input and output results for designing the Kerry Town ETC health information system

 I1: Data use priorities

High priority: patient care; mandatory reporting to the government and SCI leadership; Medium priority: longer-term medical research; Low priority: assessing data collection quality control, actively informing treatment practices.

 I2: Necessary patient data

Patient care: baseline (intake) information and daily medical records; Reporting: daily status updates by various breakdowns (e.g. demographics, outcome); Research: detailed medical records, epidemiological factors.

 I3: Data collection challenges

Data collection restricted by PPE, limited ETC equipment/facilities, high variance in clinical skill levels, language differences between patients/staff and amongst staff.

 I4: Patient flow

See Fig. 1.

 I5: Clinical workflow

Strictly designated rotations for patient rounds, medication administration, food, and other care during morning, afternoon, evening, and night shifts. Patient data were collected and reviewed in the red and green zones depending on the task. Similar to patients, clinicians also had one-directional flow in ETC. See section A5 of Additional file 1 for more specific details.

 I6: Data access – who, which, where, how often

Red zone: bedside patient records for clinician rounds; green zone: daily individual patient data for clinical, pharmacy, laboratory, HIS, and patient care (e.g. psychosocial and community) staff; aggregate daily patient numbers for operations management team; weekly aggregate updates for leadership.

 I7: Activities interfering with HIS

WASH staff incinerating red zone materials could destroy patient records; clinical protocol changes could alter HIS workflow; staff turnover and limited handover time could affect system maintenance; infection control means monitoring system is difficult in red zone.

 I8: Available resources

Equipment: generators, printer/copier, scanners, laptops, office supplies (toner, paper). Additional for EHR - waterproof tablets, server, Wi-Fi routers with uninterruptable power sources (UPSs). Money: minimal needed for PHR, additional £25 k pounds from SCI for EHR development/equipment. Personnel: 1–5 person site-based HIS staff and 1–2 overseas advisors for PHR; additional off-site software development team for EHR [11]. Time: requirement for functional (but later modifiable) HIS needed for site opening.

 O1: Which data and where

We categorized possible patient data as essential versus desirable, aiming to start with essential only. We mapped data flow across the ETC including how to split data across forms/modules and rooms (Additional file 1; section A6).

 O2: Data platform

We used 1) Microsoft Word to adapt ISARIC’s case record form and develop other forms for the PHR, 2) EpiInfo and EpiData databases for some PHR data entry, and 3) OpenMRS for the EHR. We converted final paper forms to PDFs.

 O3: Data collection design decisions

For the forms/modules, we aimed for minimal information, large font sizes (no smaller than size 14 on forms) for entry/review in red zone, checkboxes or buttons for faster and clearer entry/review, ordering information in intuitive ways (e.g. symptoms from top to bottom of body). Examples of these decisions are demonstrated in the PHR baseline assessment form (Fig. 2) and the EHR IV fluid ordering/monitoring module (Fig. 3).

 O4: Who collects data

Selecting which staff should record patient data varied based on which data were being collected, data collection platform being used, preferences of medical lead, and staff member availability and skills (e.g. language, computing, writing).

 O5: How data are communicated

Site layout and infection control dictated communication methods. Key approaches were 1) radio transmission (typically for complicated information i.e. drug orders, vital signs), 2) “duplicate” charts in green zone from “collective memory” of clinicians returning from red zone (typically general information i.e. patient status, symptoms), and 3) red zone Wi-Fi scanner (only retrospectively due to logistical problems). Clinicians brought in patient record copies from green zone to review in red zone. EHR automatically transmitted information over the Wi-Fi router to on-site server for access by Wifi-enabled devices.

 O6: How errors are minimized

We 1) assigned patient ID numbers with a check digit (instead of sequential) final number (Additional file 1; section A7), 2) carefully handwrote patient wristbands in advance, 3) reviewed log books when individuals patient files had errors or unusual values, 4) performed retrospective accuracy checks on subset of digitized data and data re-entry.

 O7: How records are digitized / processed

HIS staff used simple EpiData database to digitize data needed for daily and weekly reporting from PHR forms. De-identified data were exported and converted into Stata format for analysis and report production. Additional patient data were retrospectively digitized by securely scanning PHR forms and entering selected data into EpiInfo database. EHR exported data into a CSV file.

 O8: How and where records are stored/secured

Patient records were stored in: 1) red zone until patient visit was over; 2) green zone clinicians’ station for log books and duplicate charts of current patients, and 3) longer-term storage in a locked HIS office cabinet for discharged patient files. Scans and databases were secured using Safehouse Explorer Encryption software on password-protected laptops stored in a locked cupboard. Some departments stored own additional patient records, including pharmacy and labs. EHR data were on a secure green zone server with nightly backups. Clinical staff logged in to access patient files but only HIS team lead could access downloadable files on server.