Few orders are more reflexively written, or more quietly ineffective, than standalone sliding-scale insulin (SSI). A patient with type 2 diabetes is admitted, the order set offers “insulin sliding scale,” it gets checked, and for the rest of the stay their glucose is managed reactively a dose of rapid-acting insulin given after each finger-stick that’s already high. It feels like diabetes management. It mostly isn’t.
The practice
Standalone SSI means using only reactive, correction-dose insulin a graduated dose triggered by the current glucose reading with no scheduled basal or nutritional (prandial) insulin. Glucose is checked (typically before meals and at bedtime), and insulin is given in response to the number that’s already there. It is, by design, a system that treats hyperglycemia after it has occurred and does nothing to prevent the next episode.
Why we think we should
The practice persists for understandable reasons, which is why it’s so durable:
- It feels responsive and safe. You’re “doing something” about every high, and giving insulin only when glucose is elevated feels like it should minimize hypoglycemia.
- It’s cognitively easy. A single check-box on the order set, no need to estimate basal requirements or match insulin to meals, no titration burden.
- It’s familiar. Generations of clinicians learned it, order sets enshrine it, and it requires no calculation so it reproduces itself.
The intuition is that reactive correction is simpler and safer than proactive dosing. Both halves of that intuition turn out to be wrong.
Why there’s no good reason
Standalone SSI is reactive by construction it chases glucose after it rises and never addresses the underlying insulin deficiency that produces the rise. The result is predictable and well documented: patients on SSI alone spend more time hyperglycemic, ride a rollercoaster of highs-and-corrections, and derive no benefit over more physiologic dosing.
The pivotal evidence is the RABBIT 2 trial (Umpierrez and colleagues), which randomized non-critically-ill inpatients with type 2 diabetes to a basal-bolus regimen (scheduled long-acting basal + prandial insulin + correction) versus sliding-scale insulin alone. Basal-bolus achieved substantially better glycemic control, and contrary to the “SSI is safer” intuition without an unacceptable hypoglycemia penalty. A large fraction of the SSI-alone group failed to reach target and required rescue. The reactive approach didn’t protect patients; it left them hyperglycemic while feeling like action.
The mechanistic reason is simple: a patient with diabetes has a scheduled, ongoing insulin requirement (basal) and meal-related requirements (prandial). SSI addresses neither it only reacts to the failures those unmet requirements produce. You are, in effect, waiting for the fire and then reaching for the extinguisher, over and over, instead of not starting the fire. Correction-dose insulin is a useful component of a regimen; as the entire regimen, it’s a system for documenting hyperglycemia rather than preventing it.
Guidelines have reflected this for years: standalone SSI is discouraged for inpatient management of hyperglycemia, with scheduled basal (or basal-bolus) regimens preferred, and correction-dose insulin used as an adjunct rather than the sole strategy.
What to do instead
- Use a scheduled basal (or basal-bolus) regimen as the foundation for inpatients with type 2 diabetes and sustained hyperglycemia, rather than correction-only. Estimate a weight-based total daily dose, split into basal and prandial, and titrate.
- Keep correction-dose insulin as an adjunct a supplement layered on top of scheduled insulin to handle excursions, not the whole plan.
- Match nutritional insulin to intake hold or reduce prandial doses for patients who aren’t eating, which addresses the real hypoglycemia concern more precisely than abandoning scheduled insulin entirely.
- Reassess and titrate daily using the glucose pattern, rather than repeating the same reactive scale for the whole admission.
The takeaway
Standalone sliding-scale insulin is the archetype of a reactive practice that feels like management but chases the problem instead of preventing it. The evidence RABBIT 2 and the guidelines that followed is clear and long-standing: scheduled basal-bolus dosing controls inpatient hyperglycemia better, and correction-dose insulin belongs as an adjunct, not the entire regimen. The next time an order set offers “sliding scale” as a standalone option, treat that check-box as the prompt to write an actual regimen instead. Reacting to every high is not the same as preventing the next one.