Skip to main content
Back to selected work

VisitSync

Clinical operations, made easier to run.

A commercial full-stack platform for independent research sites to coordinate patients, visit windows, protocol workflows and operational reporting.

View live product
VISITSYNC / OPERATIONSWEEK 20
MON 12TUE 13WED 14THU 15FRI 16
VIS-021Window confirmedVIS-034Procedures readyVIS-055Follow-upVIS-089Reminder sent
12 visits coordinated4 active sites
Ownership
Founder · product strategy · clinical workflow design · full-stack engineering
Platform
Next.js · TypeScript · PostgreSQL · Supabase
Domain
Regulated clinical research operations
Status
Commercial product · Production
01

Context and operational problem

Research sites often coordinate enrollment dates, protocol windows, procedures and follow-up across spreadsheets, calendars and messages. Every manual handoff creates another opportunity for a missed window, inconsistent record or preventable protocol deviation.

VisitSync grew from direct experience managing clinical data, source documentation, queries and audit trails across sponsor and CRO systems. It converts those domain constraints into one traceable operational workflow.

02

System architecture

CLINICAL INPUTSSites & rolesContextual accessProtocol templatesVisits · windows · procedures
OPERATIONS CORENext.js workspacePatients · studies · CRMVisit engineWindow calculation & validation
TRUST & OUTPUTPostgreSQL + RLSSite and role isolationAudit & notificationsReports · SMS · email
03

What was built

  • Patient and study operations
  • Protocol templates and automated visit-window calculations
  • Procedure and follow-up workflows
  • CRM and operational reporting
  • SMS and email reminders
  • Role-based administration and auditability
04

Engineering decisions

01

Server-owned workflow logic

Visit calculations and sensitive workflow rules remain on the server so the interface never becomes the source of truth.

02

Access at the data boundary

Row Level Security and explicit roles isolate site and study data while versioned migrations preserve controlled schema evolution.

03

Compliance before convenience

External AI is optional and gated until provider, privacy and data-handling requirements are approved.

05

Evidence and validation

105 generated routesBreadth of the implemented product surfacePlaywright production smoke suiteCritical routes and release behavior checked in productionVersioned database migrationsAuditable, repeatable schema evolution
06

What I learned

The central lesson was that clinical software cannot simply digitize a spreadsheet. Protocol rules, permissions and traceability must become explicit product behavior. Domain knowledge made it possible to design the workflow before writing the interface.

Have a complex workflow to simplify?

Start a conversation