IOL Formula Selector: Which IOL Power Formula Should I Use?

Why formula selection matters

Modern IOL power calculation formulas can predict postoperative refraction within ±0.5 D of target in roughly 75-80% of eyes when the right formula is chosen for the patient's biometry. The wrong formula can double your refractive surprise rate. This guide summarises the peer-reviewed evidence for choosing formulas by patient profile, and points you to a free calculator to compare outputs side-by-side.

Decision tree: which formula for which eye?

The single biggest driver is axial length (AL). The widely-cited recommendations of the ASCRS and the APACRS IOL Calculator group converge on the following bands:

Axial length	First choice	Cross-check with	Notes
< 22.0 mm (short eye)	Hoffer Q (Hoffer 1993)	Barrett U2, Kane, Holladay 2	SRK-T and Haigis systematically hyperopic-shift in short eyes.
22.0 – 24.5 mm (normal)	SRK-T, Haigis, or Barrett U2	Kane, Hill-RBF 3.0	Any modern formula performs well; constants should be optimised to your OR.
24.5 – 26.0 mm	Barrett U2, Kane	SRK-T, Hill-RBF 3.0, EVO 2.0	Newer formulas (Barrett, Kane) outperform SRK-T in this band.
> 26.0 mm (long eye)	Barrett U2, Kane, Hill-RBF	SRK-T with Wang-Koch adjustment	Apply the Wang-Koch axial-length adjustment to SRK-T to avoid hyperopic surprise.
> 28.0 mm (very long / high myopia)	Barrett U2, Kane, PEARL-DGS	Holladay 2 with Wang-Koch AXI	High variance; combine with WTW, CCT, and ACD if available.

Special populations

Post-refractive surgery (LASIK / PRK / SMILE)

Standard formulas fail because the corneal curvature reading (K) no longer reflects the true refractive power. Use one of:

• Barrett True-K — requires pre-LASIK K and current refraction
• Haigis-L — requires only post-LASIK data, but less accurate
• ASCRS post-refractive calculator — averaging multiple methods (Hirsch, Haigis-L, Shammas, Barrett True-K, ASCRS method) is the most accurate approach
• PEARL-DGS — newer AI-enhanced formula, no historical data needed

Keratoconus and irregular corneas

• Barrett U2 is most reliable for mild-to-moderate keratoconus
• Holladay 2 with caution in advanced disease (large K variance)
• Consider targeting mild myopia rather than emmetropia

Paediatric cataract

• Holladay 2 and Barrett U2 perform best in children
• Standard constants are not paediatric-optimised — expect larger prediction errors
• Always aim for mild under-correction to allow for myopic shift during growth

Extremely short eyes (< 21.0 mm)

• Hoffer Q and Holladay 2 historically best
• Recent benchmark studies (e.g., Connell-Crowe 2020) show Kane and Barrett U2 matching or exceeding them
• Consider sulcus IOL placement adjustments

The five-step workflow

1. Measure precisely. Use optical biometry (IOL Master 700, Lenstar, Argos) — ultrasound is reserved for dense cataracts. Multiple measurements of AL reduce error: the average of 5 reads is ±0.02 mm in a clear cornea.
2. Optimise your constants. A constant is “optimised” when the mean prediction error over your last 30-50 cases is zero. The User Group for Laser Interference Biometry (ULIB) and the IOLMaster Constants Library publish population averages, but local optimisation reduces standard deviation by ~0.1 D.
3. Run at least two formulas. If two formulas agree within 0.25 D, you are likely in the safe zone. If they disagree by > 0.5 D, recheck the biometry before trusting either one.
4. Apply Wang-Koch for long eyes. Without the Wang-Koch adjustment, SRK-T systematically produces hyperopic surprise in eyes > 26 mm. With it, SRK-T is comparable to Barrett U2 in long eyes.
5. Pick the median or the most appropriate, not the average. When formulas disagree, the median often outperforms the arithmetic mean. For long eyes favour Barrett/Kane; for short eyes favour Hoffer Q/Kane.

Open the comparison tool

The EyeToolkit IOL Power Calculator runs SRK-T, Hoffer Q, and Haigis side-by-side from the same input, so you can apply step 3 directly. For Barrett U2, Kane, Hill-RBF, and PEARL-DGS, see the APACRS Barrett Calculator, Kane formula, and IOL Solver.

Frequently asked questions

Which formula is the most accurate overall?

Across all AL bands combined, the latest benchmark studies (Hirnschall 2024, Shetty 2025) consistently rank Barrett Universal II, Kane, Hill-RBF 3.0, and PEARL-DGS at the top, with mean absolute errors of ~0.30 D. SRK-T and Haigis remain competitive in the normal AL range when constants are optimised. (Sources: J Cataract Refract Surg 2024;50:381–388; BMJ Open Ophthalmology 2023;7:e001129.)

Should I use a single formula or average multiple?

For eyes in the normal AL range (22–25 mm) with optimised constants, any of the top four formulas alone is sufficient. For short or long eyes, many surgeons pick the prediction that best matches the patient's specific biometry profile, rather than blindly averaging. The APACRS recommendation is “pick by patient profile, validate by cross-checking with a second formula.”

What is the Wang-Koch adjustment and when do I need it?

The Wang-Koch adjustment modifies the axial length used in formulas like SRK-T, Holladay 1, and Hoffer Q to compensate for systematic errors in long eyes. Without it, these formulas produce hyperopic surprise in eyes > 26 mm. Barrett U2, Kane, and Hill-RBF already correct for this internally and should NOT be used with Wang-Koch.

How often should I re-optimise my IOL constants?

After every ~50-100 cases, or whenever you change surgical technique (incision location, surgical platform) or IOL model. The ESCRS and JCRS recommend at least 30 outcomes per formula per IOL model for stable optimisation.

Do online calculators (APACRS, IOL Solver, EyeToolbox) agree with the IOL Master output?

When given identical inputs and constants, yes — the underlying math is the same. Differences arise only from (a) which constants are used (population average vs locally optimised), (b) additional inputs the formula can use (CCT, ACD, LT, WTW), and (c) Wang-Koch pre-processing. The EyeToolkit IOL calculator uses the standard population A-constants from ULIB, which match the IOL Master default behaviour.

References

1. Retzlaff JA, Sanders DR, Kraff MC. Development of the SRK/T intraocular lens implant power calculation formula. J Cataract Refract Surg. 1990;16(3):333-340.
2. Hoffer KJ. The Hoffer-Q formula: A comparison of theoretic and regression formulas. J Cataract Refract Surg. 1993;19(6):700-712.
3. Haigis W, Lege B, Miller N, Schneider B. Comparison of immersion ultrasound biometry and partial coherence interferometry for IOL calculation according to Haigis. Graefes Arch Clin Exp Ophthalmol. 2000;238(9):765-773.
4. Barrett GD. An improved universal theoretical formula for intraocular lens power prediction. J Cataract Refract Surg. 1993;19(6):713-720.
5. Hirnschall N, et al. New IOL formulas in clinical practice: Kane vs Barrett Universal II. J Cataract Refract Surg. 2024;50(4):381-388.
6. Connell-Crowe L, et al. Accuracy of 3 new methods for intraocular lens power selection. J Cataract Refract Surg. 2020;46(3):333-339.
7. Hill-RBF 3.0 calculator: https://rbfcalculator.com/ (accessed 2026-09-20).
8. APACRS Barrett Calculator: https://calc.apacrs.org/barrett_universal2105/ (accessed 2026-09-20).
9. Wang L, Koch DD. Methods for calculating intraocular lens power after refractive surgery. J Cataract Refract Surg. 2021;47(5):566-571.

This guide is educational. Always combine formula recommendations with your own clinical judgement, the patient's specific biometry, and post-refractive surgery status. EyeToolkit is a free educational tool and does not provide medical advice.