Measuring the three shifts, with an equity modifier

One index for analogue to digital, hospital to community and treatment to prevention — adjusted for who the gains are reaching, wired to signals, actions, a board pack and commissioner sign-off.

1 The shifts, the number, and why it moves

The 10 Year Health Plan asks for three shifts: analogue to digital, hospital to community, treatment to prevention. Most boards I talk to measure them in three separate papers, on three separate cycles, with three separate definitions of progress. LeftShift Intelligence© puts all three on one page, scores each of them, rolls them into one index, and then adjusts that index for who the gains are reaching.

Three views of the same thing, on purpose. The 0–100 index and pillar scores band into Immature, Emerging, Developing, Established and Leading, so a non-executive can read the direction in one glance. The maturity rings score each axis out of 5.0 with red, amber and green indicator counts underneath, which is the view a transformation director works from. The System Transformation Index averages the three axes into one number for the annual plan, and comes out at 3.23 out of 5.0 here.

2 The equity modifier, and why it subtracts

An index that only ever goes up is an index nobody believes. The equity modifier is an adjustment to the headline score that reflects whether the gains are reaching everybody. It can be positive. In the demonstration it’s −3, and the reason sits on the same line as the number.

The reason given is digital exclusion rising in the 75+ cohort and in the highest deprivation quintile, and it sits on the front page immediately under the score. The failure mode of every transformation programme I have watched is the same one: the people who were already easiest to reach get reached first, the average improves, and the gap widens underneath it. The modifier stops the average from doing the talking.

3 What sits under a pillar score

A composite score is worthless if you can’t open it. Each pillar tab carries its own eight-quarter trend, nine key indicators with direction of travel, and a table of ten operational metrics that names the module each figure came from.

The right-hand column of that table is the part that matters to a buyer. Every operational metric names the module it’s drawn from, and those modules are real parts of the platform with their own APIs: prescribing, referrals, virtual ward, screening, social prescribing, flow. That’s the production path. Today the numbers are seeded and the page says so at the top of every tab, but the wiring is no data warehouse project. The metrics are already produced by modules this platform runs, which is the argument for measuring the shift inside the operational system instead of beside it.

The Community Shift tab adds one more thing: community nursing capacity fed from the scheduling partner that optimises the roster, so the capacity signal here and the roster the service runs are the same number.

4 Signals, and the actions they drive

A score tells you where you’re. It doesn’t tell you what to do on Monday. Shift Signals is the early-warning layer: each signal carries a severity, a domain, a named cohort, the evidence behind it and a recommended intervention, and it has a state that moves from New to Acknowledged to Action created.

Two of the five signals are positive, and that’s a design choice. A dashboard that only ever surfaces red teaches an organisation to bury things. The frailty virtual ward signal reads 8.4% thirty-day readmission against a 13.1% baseline and recommends scale-up; the hypertension signal reads an 8.2% improvement in blood pressure control over three months and recommends sharing the method. Those two would never reach a risk register, and they’re the ones a commissioner should act on hardest.

The urgent one is the opposite. Post-surgical discharge readmissions at 14.2%, up from 9.8% the previous quarter, three sentinel cases under root cause analysis, recommendation to pause expansion of the pathway. The signal is what stops a system scaling something that has started to hurt people.

5 The board pack writes itself, and admits what it can’t say

One button on the overview produces a six-section pack: executive summary, the index and its four components, priority risks and signals, interventions with delivery confidence, decisions required, and data confidence and limitations.

Section six is the one I would read first. It states the confidence level, the refresh interval for each pillar, and the fact that equity data is incomplete for ethnicity segmentation pending a data sharing agreement. A pack that tells a board what its own numbers can’t carry is worth more than one that doesn’t, because the board can then ask the right question instead of accepting the chart.

Section five is the reason the pack exists. Three decisions, each with a figure, a named recipient and a date: approve the recruitment funding, commission the independent review of the discharge pathway and report by the end of July, note the equity modifier deterioration and approve the inclusion plan. A transformation report that ends in three decisions has done its job.

6 Sign-off, and which jurisdictions it runs in

A period report that goes out unsigned is a leaflet. The Governance tab holds commissioner sign-off for the period, and a history of who signed what, in which jurisdiction and role, on which date, with their note and whether it was exported.

Two columns in that table earn their place. Jurisdiction, because the same index is signed off by different people under different accountability in different places. And notes, because the note from the ICS executive director says prevention metrics are below target and have been escalated to the system board. A dissenting sign-off is still a sign-off, and it belongs in the history, not in a minute nobody reads.

Scope is set by jurisdiction. LeftShift Intelligence© is enabled by default for England and for the first live jurisdiction, and other jurisdictions opt in. That matters because the three shifts are an English policy framing. The measurement model travels; the mandate it’s measured against doesn’t, and the platform treats that as configuration instead of as a separate build.

7 What is real today

The page carries an amber banner at the top of every tab that reads “Seeded demonstration data — not a live system feed”. It’s in every screenshot above because I left it there.

Here is the position without decoration. The index, the pillar scores, the maturity rings, the equity modifier, the signals, the actions, the cohorts, the capacity layer, the board pack and the sign-off history are all in-page demonstration content today, with no LeftShift endpoint in the platform behind them. Acknowledging a signal or creating an action changes what you see for the session and isn’t persisted. The export buttons on the board pack are a demonstrated control.

What is real is the measurement model and the modules underneath it. The operational metrics tables name the source of every figure, and those sources are live modules in the same platform with their own APIs. The production path is to compute each indicator from that live operational data on the refresh interval the tab already declares, and not to collect it by return. The same view is also embedded inside commvita Outcome Intelligence as its Shift Maturity tab, where the workforce experience partner layer beside it’s API-backed and live. Nothing is duplicated between them.

I would rather show a buyer a working measurement model on seeded numbers and label it, than show them a live-looking dashboard built on figures that came out of a spreadsheet last month.

Where it lives in commvita