This started as a near-miss, not a directive.
After Dorian in 2019 we lost one office entirely. What stayed hidden was damage to the neighbouring unit's plumbing. The inspector never flagged it because the vulnerability wasn't on our side of the building. When it surfaced the following summer, the physical files were safe. But the shock was real.
I suggested the transition while that feeling was still fresh. First because the government's digitisation schemes had just opened up, and also because five cabinets of records dating back to 2008 are one bad day away from being gone.
The charter was informal. I'm applying the structured framing now, after completing my Google PM cert. But the thinking was there at the time.
The providers were the hardest stakeholder to manage. Their time was clinical. Patients in chairs, instruments, schedules. Asking them to review how they'd been filling out forms felt like an administrative inconvenience with nothing to do with their actual work.
Direct correction requests didn't land. So I changed the approach. Instead of asking them to fix mistakes, I asked them to help me design a blank template showing what a correctly completed form looked like. One-time ask, and then my colleague and I could handle the corrections ourselves using the template as a guide.
It was a small reframe. But it got us what we needed without making anyone feel like they were being audited.
I didn't use a formal WBS at the time. My actual tools were Todoist for task assignment and Notion for logging what happened and when.
What I did have was a clear instinct about sequencing: insurance records first, health records later. Insurance records were the highest-risk item for the practice. Delayed or incomplete digitisation could mean unpaid claims. That made them the obvious first priority.
The WBS here is a retrospective mapping of what I actually did onto a formal framework. The thinking was there. The documentation wasn't. That's one of the honest lessons I took from this project.
| ID | Risk | Like. | Impact | Mitigation / Response | Status |
|---|---|---|---|---|---|
| R-001 | Time constraint — roughly 3 months in-person only, alongside existing duties | High | High | Prioritise insurance records; formally defer Phase 2 to a separate brief | ✓ Managed |
| R-002 | Provider resistance — clinical time seen as higher priority than admin validation | Medium | Medium | Reframed from correction requests to blank-template validation; delegated follow-up to office staff | ✓ Resolved |
| R-003 | Post-departure change management failure — new manager may revert to paper | Medium | High | SOPs written; staff trained; usability test passed. No formal internal champion designated. | ⚠ Materialised |
| R-004 | Insurance company rejects digital submission format | Low | High | Pre-check acceptance format per insurer; maintain fax line as fallback (→ CR-001) | ✓ Managed |
| R-005 | Data loss or misfiling during scanning and transfer | Low | V.High | Retain physical originals until digital version is validated; double-check filing protocol | ✓ No incident |
| R-006 | MS Access integration fails due to vendor-locked server backend | Medium | Low | Attempt; abandon if incompatible. Acceptable given it was outside Phase 1 core scope. | ✗ Abandoned |
R-003 is the one I wish I'd taken more seriously.
I knew I was leaving. I even have a note from that time about needing someone to own this after I left. But flagging a risk is not the same as managing it.
I wrote SOPs. I trained my colleague. The office manager passed the usability test. But I didn't get a formal commitment to designate an internal champion, and I didn't build enough pressure into the system to make the digital workflow obviously easier than paper.
A few months after I left, COVID was also complicating everything, and a less tech-comfortable manager went back to paper. I still get asked about it. The practice is partly family, so I stay in the loop.
Healthcare PM sits between administrative efficiency and patient dignity, and the two pull against each other constantly. Every system I design gets run against one question: who needs to use this under pressure, with minimal time, and possibly with a patient sitting right in front of them?
When I built the folder naming conventions, I was thinking about the locum clinician who'd never seen our system before. When I kept the fax line, I was thinking about the patient whose insurer would otherwise miss a payment. Accessibility isn't a compliance checkbox. It's what makes systems work for everyone, including the people who weren't in the room when you designed it.
| Stakeholder | Information need | Channel | Frequency | Owner |
|---|---|---|---|---|
| Practice Owner | Progress vs. scope; any budget or change issues; Phase 2 readiness | Shared Google Sheet + in-person | Weekly summary; ad hoc on issues | Gabriel |
| Office Colleague | Daily task priorities; filing standards; quality checks | Todoist + end-of-day 1:1 | Daily | Gabriel |
| Dental Providers | Template validation requests only; minimal impact on clinical time | In-person (brief) + blank template | As needed; batched | Gabriel |
| Insurance Companies | Format acceptance; any submission requirement changes | Phone + email + fax (one insurer) | As needed | Gabriel |
| Incoming Manager Gap — identified post-project | System overview; SOPs; where things are and why | Written SOP + walkthrough | One-time handoff | Gabriel |
There were no formal meetings on this project. The practice didn't work that way. The pace was set by patient appointments, not project schedules.
Most of my communication happened in the margins: quick asks between patients, end-of-day check-ins with my colleague, phone calls with insurers at lunch. The shared Google Sheet meant the owner could check progress without needing to ask me directly.
This worked fine for execution, but it created a gap at handoff. The incoming manager row in that table was added in retrospect. I had planned a walkthrough. I just didn't build enough redundancy to make the system work without me being there to explain it.
| CR-ID | Description | Raised by | Impact | Resolution | Status |
|---|---|---|---|---|---|
| CR-001 | Family Guardian requires fax-only submission — digital format not accepted under any condition | Gabriel (compliance check) |
Medium | Fax line maintained as exception. Fax workflow documented as separate SOP entry. Insurer flagged for Phase 2 renegotiation. | ✓ Resolved |
| CR-002 | Provider validation method changed: direct correction requests replaced with blank-template approach | Gabriel (process change) |
Low | Providers create one reference template each. Office staff handle all corrections independently using template as guide. Scope unchanged. | ✓ Resolved |
| CR-003 | Patient health data migration formally de-scoped from Phase 1; moved to Phase 2 | Gabriel & Owner (scope review) |
High | Agreed as correct prioritisation. Phase 2 requirements documented and handed to owner for future resourcing. | ✓ Deferred |
| CR-004 | MS Access integration attempted for records cross-reference; abandoned due to server constraints | Gabriel (technical) |
Low | Vendor-locked backend incompatible. Dropped without Phase 1 scope impact. Noted for future consideration if infrastructure changes. | ✗ Abandoned |
CR-001 caught me off guard. By 2021 I'd assumed every insurer was digital. Most were. But Family Guardian is a major local Nassau carrier and their submission process was simply fax-only, full stop. The size of the practice had nothing to do with it. EDC is the largest single dental practice in the Bahamas and that still wasn't the conversation.
So we built around it. The fax line stayed, the workflow was documented as an SOP exception, my colleague knew the process, and I flagged it for Phase 2 renegotiation. That's a policy conversation that needs a longer runway than a digitisation sprint.
What I took from it: compliance checks belong at the start of planning, not mid-execution.
My primary KPI was qualitative, and I stand by it. I'd sit with the office manager or the EA and ask them to find a specific patient file. If they could do it without asking me, we were on track.
For quantitative tracking I used a Google Sheet. A running count of what was done versus what remained, reviewed weekly with my colleague. My Notion journal was the informal "what happened today" log.
I'd build something more structured now. The dashboard on the left is the framework I'd use on a comparable project. The placeholders are an honest acknowledgment that real figures are confidential. Structure matters more than the numbers anyway.
The project succeeded. Phase 1 was delivered, the system worked, and the office could find things. But the change didn't stick, and that's the gap I think about most.
There's a difference between delivering a project and making a change last. On-time and within scope isn't enough if the system reverts when you leave. Lasting change needs an owner, a plan, and enough friction removed that the new way is genuinely easier than the old way.
I didn't have the vocabulary for that at the time. I do now. And that's probably the most useful thing I took from completing my Google PM cert: being able to name exactly what happened and understand where it broke.