# The Statistic That Asks If Groups Differ

By Nora Bissett, Pricing Editor (https://glpwatchdog.org/authors/nora-bissett) | Published: 2026-08-20 | Updated: 2026-10-27

Canonical page: https://glpwatchdog.org/research/the-statistic-that-asks-if-groups-differ

> A SOUL secondary analysis found oral semaglutide cutting heart-failure events in people who had heart failure and doing nothing in those who did not. The test of whether that split is real came out at P = .06.

## Key points

- HR 0.78 with heart failure at baseline; 1.01 without.
- P for interaction = .06, so the split is NOT established.
- Nearly all the benefit sits in preserved ejection fraction: HR 0.59 vs 0.98 for reduced.
- That subgroup carries extra weight because semaglutide has a dedicated HFpEF trial behind it.
- MACE showed no heterogeneity at all (P interaction = .77).

A secondary analysis of the SOUL trial found oral semaglutide cutting heart-failure events by 22% in people who already had heart failure, and doing nothing whatsoever in people who did not.[1] That looks like a clean split. The statistic that tests whether the split is real came out at **P = .06**. Once you understand what that number means, you will read every subgroup table differently.

## The analysis

SOUL enrolled 9,650 people with type 2 diabetes plus atherosclerotic cardiovascular disease, chronic kidney disease, or both, across 444 centers in 33 countries, and followed them for an average of just under four years. Roughly a quarter of them, 2,229 people, had a history of heart failure at entry. The analysis asked whether daily oral semaglutide changed a composite of heart-failure hospitalization, urgent heart-failure visit, or cardiovascular death.

*Composite heart-failure outcome with oral semaglutide against placebo, by baseline status.[1]*

| Group | Hazard ratio (95% CI) | Reading |
| --- | --- | --- |
| Heart failure at baseline | 0.78 (0.63–0.96) | 22% fewer events |
| No heart failure at baseline | 1.01 (0.84–1.20) | No effect |
| *P for interaction* | **.06** | **The groups are not shown to differ** |

Start with what the parent trial found. SOUL was designed as an event-driven superiority trial. Everyone in it was 50 or older and had type 2 diabetes plus established atherosclerotic disease, kidney disease, or both, and the median follow-up ran just past four years.[2] The primary outcome counted cardiovascular deaths together with heart attacks and strokes that patients survived. It struck **12.0%** of the oral semaglutide group and **13.8%** of the placebo group, a hazard ratio of 0.86 (95% CI 0.77–0.96, P = 0.006).

Two details belong with that. The confirmatory secondary outcomes, including a five-point kidney composite, **did not differ significantly** between groups, so the trial’s positive result is the cardiovascular one, not a general sweep. And serious adverse events were slightly *less* common on the drug, 47.9% against 50.3%, in a population sick enough that roughly half of everyone had one.

## What a P for interaction actually asks

Most people read a subgroup table by checking which rows are significant. That answers a different question from the one worth asking. Each row tests whether *that group’s* result differs from no effect. The interaction test asks whether *the groups differ from each other*.

> **Those come apart more often than people expect.** One subgroup can land just inside significance and another just outside without the difference between them being distinguishable from chance. A table can show “works here, does not work there” when the accurate summary is “we cannot tell whether it works differently in these two groups.”

At .06, this interaction test does not clear the conventional threshold. It is also close enough that dismissing the split entirely would be its own error, because .06 is weak evidence of a difference, not evidence of no difference. The correct reading is uncertainty, which is less satisfying than either headline.

> Checking which rows are significant answers a different question from the one you meant to ask.

## The split inside the split

Among the people who did have heart failure, the effect divided again by type of heart failure, and here the contrast is stark.

*Composite heart-failure outcome among participants with heart failure at baseline.[1]*

| Type | Hazard ratio (95% CI) |
| --- | --- |
| Preserved ejection fraction (HFpEF) | **0.59 (0.39–0.86)** |
| Reduced ejection fraction (HFrEF) | 0.98 (0.70–1.38) |

Essentially the entire benefit sits in preserved ejection fraction, the form where the heart muscle stiffens rather than weakens. That form is strongly associated with obesity, and treatment options for it have historically been thin.

> **This subgroup deserves more credence than most.** Semaglutide has a dedicated randomized trial in preserved ejection fraction behind it, so this is a pre-existing hypothesis being checked rather than a pattern found by looking. The difference between a prior expectation and a discovery in the table is the main thing separating a credible subgroup from a spurious one. It is still a subgroup.

## What did not split

The trial’s main cardiovascular endpoint behaved completely differently. Major adverse cardiovascular events were reduced similarly with and without a history of heart failure: hazard ratios of 0.83 and 0.86, with an interaction P of .77. That is about as clean a statement of “no difference between groups” as these analyses produce.

That contrast is informative. The drug’s general cardiovascular benefit appears not to care whether you have heart failure. Only the heart-failure-specific endpoint showed any suggestion of differing, and that suggestion did not reach significance.

Serious adverse events were similar in the heart-failure group on drug and on placebo (53.8% against 57.1%). That answers the safety question anyone would ask about giving a new agent to people with established heart failure.

## How to read the next subgroup table you see

- **Find the interaction P before reading the rows.** Without it, a table of subgroups is a list of separate questions, not a comparison.
- **Ask whether the subgroup was expected.** A group singled out because of prior evidence is far more believable than one that emerged from the data.
- **Notice which endpoints did not split.** A drug that behaves the same everywhere on its main endpoint, and differently on one specific one, is telling you something narrower than the headline.
- **Treat P = .06 as uncertainty, not as either answer.** It is neither a difference established nor a difference ruled out.

The companion problem, a subgroup with the biggest number and the least evidence behind it, is covered in [the biggest number is the weakest](https://glpwatchdog.org/research/the-biggest-number-is-the-weakest).

## Frequently asked questions

### Does oral semaglutide help heart failure?

In this secondary analysis, it reduced a composite of heart-failure hospitalization, urgent heart-failure visits and cardiovascular death by 22% among people who already had heart failure, with no effect in those who did not. The test of whether those two groups differ came out at P = .06, which does not establish the split.

### What is a P for interaction?

A test of whether two subgroups differ from each other, as opposed to whether each differs from no effect. One subgroup can be significant and another not while the difference between them remains indistinguishable from chance.

### Which type of heart failure benefited?

The preserved-ejection-fraction group: hazard ratio 0.59 (95% CI 0.39–0.86). People with reduced ejection fraction showed essentially nothing, at 0.98 (0.70–1.38).

### Should that HFpEF result be believed?

It carries more weight than a typical subgroup, because semaglutide has a dedicated randomized trial in preserved ejection fraction. This checks a prior hypothesis rather than reporting a pattern found by looking. It remains a subgroup analysis.

### Was it safe in people with heart failure?

Serious adverse events occurred in 53.8% of the semaglutide group and 57.1% of the placebo group among participants with heart failure.

## References

1. Pop-Busui R, Rasmussen S, Deanfield JE, et al. Oral Semaglutide and Heart Failure Outcomes in Persons With Type 2 Diabetes: A Secondary Analysis of the SOUL Randomized Clinical Trial. JAMA Internal Medicine. 2026. PMID: 41627802. https://pubmed.ncbi.nlm.nih.gov/41627802/
2. McGuire DK, Marx N, Mulvagh SL, et al. Oral Semaglutide and Cardiovascular Outcomes in High-Risk Type 2 Diabetes. The New England Journal of Medicine. 2025. PMID: 40162642. https://pubmed.ncbi.nlm.nih.gov/40162642/

---

Cite as: GLP Watchdog, Nora Bissett, "The Statistic That Asks If Groups Differ", published 2026-08-20, updated 2026-10-27. https://glpwatchdog.org/research/the-statistic-that-asks-if-groups-differ

Canonical HTML page: https://glpwatchdog.org/research/the-statistic-that-asks-if-groups-differ

Medical disclaimer: Nothing here is medical advice. We report on this subject; we do not treat anyone. Whether to begin a medication, supplement, or treatment, whether to adjust it, and whether to stop are questions for a clinician licensed to answer them, who knows your history.
