Early intervention has become the most repeated phrase in group risk.
Every insurer talks about it. Every group income protection (GIP) provider has a programme for it. Rehabilitation services, vocational support, fast-track physiotherapy, mental health pathways.
The intent is good. The infrastructure is not.
The problem is not a lack of services. It is a lack of connection between the services that exist and the people who need them, at the point when intervention would actually change the outcome.
Most early intervention programmes activate after a claim is made or after an absence is reported.
By that point, the employee has already crossed a threshold. They are already off work, already in the claims process, already costing the scheme.
The intervention is early relative to long-term absence, but late relative to when the conditions that produced the absence were first visible.
That distinction matters and the group risk market has largely ignored it.
The lead time nobody measures
Every group income protection claim has a lead time.
The employee who submits a mental health claim in September was showing signs of unsustainable pressure in May.
The musculoskeletal claim that arrives in Q4 started as discomfort in Q1 that was never addressed because the employee didn’t know their cash plan covered physiotherapy.
This is not speculation. Occupational health research has been quantifying these trajectories for more than 30 years.
When workplace demands (workload, time pressure, role conflict, emotional labour) consistently exceed the resources available to cope (autonomy, support, development, feedback), the progression is well-documented: disengagement, then presenteeism, then absence, then claim.
The research tells us two things.
First, the trajectory is predictable. Second, it is interruptible, but only if someone is measuring the conditions while there is still time to act.
The group risk market currently measures claims. It does not measure the conditions that produce them.
Why insurer early intervention falls short
Insurer-provided early intervention services are valuable. Fast-track access to cognitive behavioural therapy (CBT), physiotherapy pathways, vocational rehabilitation, second medical opinions.
These services genuinely improve outcomes for employees who access them. The limitation is structural.
These services activate when the insurer becomes aware of a case, which typically means when a claim is submitted, a referral is made, or an absence hits a threshold.
The insurer does not have visibility of the employee’s risk profile before that point.
They cannot see that a team’s demands have been rising for three months. They cannot see that an individual’s self-reported anxiety has been climbing since January.
They see the claim when it arrives.
This is not a criticism of insurer early intervention. It is an observation about where the data sit.
The insurer has the services. The employer has the proximity to the employee. The adviser sits between them. And nobody is measuring the upstream conditions that would make early intervention genuinely early.
Connecting measurement to the benefits infrastructure
The occupational health research has always known what to measure. What it has never had is a direct line to the services that can act on the finding.
When an employee’s check-in responses cross established thresholds for anxiety, burnout risk, or musculoskeletal discomfort, the question is not whether they need support. The research answers that.
The question is whether anything happens before they deteriorate further. In most organisations, nothing does. The employee either self-refers if they know how, or they don’t.
What changes when measurement is embedded inside the benefits infrastructure rather than sitting outside it is that the finding triggers a specific action.
Not a report. Not a recommendation to speak to HR.
A direct route to the service the employer already provides: the PMI mental health pathway, the EAP, the cash plan physiotherapy benefit, with access instructions at the point the risk is identified.
That is what genuinely early intervention looks like.; the insurer’s services reach the employee before absence, before the referral, before the claim.
At team level, aggregated patterns show where demands are rising and resources are falling weeks before the conditions become visible in absence data.
This is what we built at Alltoogether. And it sits where it should always have sat: upstream of the claim, inside the broker relationship that connects the employer, the insurer, and the employee.
What this means for the renewal conversation
At team level, managers see anonymised, aggregated patterns with strict privacy thresholds showing where demands are rising and resources are falling.
This gives the broker something that claims analytics alone cannot provide: a leading indicator of where claims pressure is building.
An adviser who can walk into a renewal meeting and show that risk indicators in a particular department improved after targeted interventions, that EAP utilisation rose because employees were actively routed to it, and that the conditions that typically precede GIP claims were identified and addressed before absence occurred, has a fundamentally different conversation with both the employer and the insurer.
That is not a wellbeing pitch. It is a claims prevention strategy with measurable data and it sits where it should always have sat: inside the broker relationship that connects the employer, the insurer, and the employee.
The next step for group risk
Early intervention is not a new idea, but genuinely early intervention, measuring conditions before they become claims rather than intervening after they do, requires infrastructure that the group risk market has not yet built.
The insurer provides the services. The employer provides the proximity. The broker provides the relationship. What has been missing is the measurement layer that connects all three before the claim is made.
That layer now exists. The question for the group risk market is whether it is willing to move upstream of the claims data it has spent five years perfecting, and start measuring the conditions that the data are telling them about after the fact.
