Global LIS Atlas

An open, crowdsourced world map of anatomic pathology: which AP LIS each operation runs, where it stands with digital pathology, and which bespoke tools its informatics team has built and would share. Published live on Observable and proposed to the Association.

The unit is the AP operation: the group that signs out surgical pathology and cytology for one or more sites. It may be a hospital department, an academic department, a contracted pathology group, or an independent or dermatopathology practice. Each one is counted on its own, never as part of the clinical lab that shares its building.

Source on GitHub: OpenPathologyFoundation/lis-atlas · Live page: openpathologyfoundation.github.io/lis-atlas · Observable notebook: @gershkovich/lis-atlas

Who runs what

Each country is colored by the most common AP LIS among responding operations. Blue, orange, and green mark the three largest vendors; gray groups everyone else. Hover a country for its top two systems. The small maps further down show where each major vendor is strongest. In the public view, a country is hatched instead of colored when no vendor there has at least operations, so no single operation’s system can be read off the map.

Installed base among responding operations.

Where each vendor is strongest. Share of responding operations in each country that run the vendor as their AP LIS.

Where digital pathology stands

operations () use digital slides in clinical sign-out, and sign out most cases on screen. The first bar covers all operations; each one after it is a type of operation, so a hospital department and a dermatopathology practice are compared side by side rather than pooled. Hover a segment for its count.

If every target holds. The share of operations that would be primary digital each year: those already there, plus every operation whose target year has come. Targets are plans, so they are only ever counted in totals.

Bespoke technology

Many AP operations run software their own informatics teams built to close gaps the core LIS leaves: case assignment, slide tracking, synoptic reporting, scan routing. We call it bespoke, not home-grown. Much of it is enterprise-grade software, built by embedded informatics teams with their own support and release cycles. The atlas asks two simple things: what these tools do, and whether the team would share them.

operations () run bespoke tools in clinical use, and more are building them. would share theirs or already do.

Tools on offer. Tools from operations shown by name that would share them or already do. In the live atlas, each entry would link to a contact or a code repository.

Who’s switching, and to what

operations () are evaluating, have selected, or are implementing a new AP LIS. Read the flow left to right: the AP LIS they run today, then the system they are moving to. Operations still evaluating have not picked a vendor yet. In the public view, moves made by fewer than operations are left out.

Main reason for switching

Upcoming go-lives

Every dot is one operation, placed in the quarter its new AP LIS goes live. go-lives are scheduled between now and 2030, and of them are moving to Sirius Cybernetics. Hover a dot for the operation and its move. In the public view, the dots show only plans that operations say are announced or under contract, from operations that agreed to be listed.

Share of responding operations on the three largest AP LIS vendors. Solid lines cover past switches; dashed lines project forward from scheduled go-lives. The remaining operations run other systems. Illustrative pattern, not a finding: the synthetic data was built to show one vendor gaining share.

Six degrees of the survey

Each response arrives through someone’s referral link, so the atlas also records how the survey spread. The Association sits at the center, linked to the board members who seeded the survey (dark dots with a ring). Every operation links to the colleague who invited it, and bigger dots brought in more operations. Hover an operation to trace its chain back to the board.

The median operation is ** hops** from the board, the longest chain has ** hops**, and **** of operations are within six degrees.

Where the map is still blank

Coverage drives the call to action. Each blank country needs a regional ambassador, and countries with only a handful of responses need a second push.

Hospitals covered in the US

In the US, the public CMS hospital list gives the atlas a denominator: hospitals in the 50 states and DC. Each operation names the hospitals it signs out for, so one answer from a contracted group that serves 20 hospitals fills in all 20 at once. Responding operations cover ** hospitals ()**.

Hospital totals per state are real; which hospitals are covered is synthetic. Hospitals are never hidden. Operations that agreed to be named count right away; those that asked not to be named count once of them in the same state have answered, so none can be picked out by the share it adds.

Rollout

Pilot (Aug 3–24). Association board and active members fill it in first, fix confusing questions, and seed the map so it isn’t empty at launch.
LinkedIn wave (Aug 25 – Sep 14). Leadership posts, members reshare, and everyone tags colleagues at other institutions.
Ambassadors (Sep 15 onward). Regional ambassadors take ownership of their countries, and the blank areas on the map become the call to action.

Named ( operations): name shown with systems, announced plans, and any tools offered for sharing.
Unnamed ( operations): systems and plans shown as, say, “a dermatopathology practice in Germany”. Where a country has fewer than responding operations, only the region is given.
Totals only ( operations): counted in maps, charts, and totals, but never listed individually.
Plans: a listed operation’s new vendor, go-live date, reason, and decision-maker appear only once it says they’re announced or under contract. Past go-live dates count only in totals. “Evaluating” is only ever counted in totals, and so are each operation’s digital pathology target, main barrier, and informatics staffing.
Bespoke tools: a tool’s name, and what it does, appear only when its operation is shown by name and offers to share it. For every other operation, what the tools do is counted only in totals.
Small-cell suppression: any number based on fewer than operations is hidden. Map areas and bar segments are hatched, and counts read "fewer than ". On the US hospital map, operations that asked not to be named count toward their state only once of them there have answered.

Switch Show as to Staff view at the top of the page to see the raw data these rules protect. In the live atlas, that view would sit behind a staff login.

From prototype to live

Survey. Any form tool that exports CSV or has an API works (REDCap, Qualtrics, Google Forms). The referral link is a URL parameter (?ref=CODE) stored with each response.
Finding operations. In the US, the public CMS list of CLIA-certified labs gives names and sites to suggest in the survey and a denominator for coverage. Other countries need their own lists.
Data. A scheduled data loader pulls the responses, removes names for operations that opted out, and publishes one cleaned CSV. The charts on this page read that file instead of the synthetic generator built into this notebook.
Governance. The Association owns the dataset. A small editorial group reviews new vendors and tools and merges duplicate operations, and an open license lets members reuse the data.