IVF lab KPIs on autopilot: how automated dashboards catch problems months before pregnancy rates do
Published 27 October 2026 · IVF Automation · Lab data and quality
In one multicentre study, a change of culture medium pushed a key embryology KPI down within a week. The drop in clinical pregnancy rate was not detected for another three months. That gap is the whole argument for monitoring lab KPIs continuously and automatically, instead of in a spreadsheet at the end of the quarter.
This guide explains which indicators the international consensus documents recommend, what the published evidence says about tracking them with control charts, and where automation and AI tools now fit for Indian IVF labs.
The standard set: Vienna and Maribor
The Vienna consensus (ESHRE Special Interest Group of Embryology and Alpha Scientists, 2017) recommended 19 indicators for fresh IVF and ICSI cycles: 12 key performance indicators (KPIs), five performance indicators and two reference indicators. Each comes with a minimum "competency" value and an aspirational "benchmark" value. A few examples:
- ICSI damage rate: competency 10% or less, benchmark 5% or less.
- ICSI normal fertilisation rate: competency at least 65%, benchmark at least 80%.
- IVF normal fertilisation rate: competency at least 60%, benchmark at least 75%.
- Cleavage rate: competency at least 95%, benchmark at least 99%.
Two points from the paper are often missed. The KPIs are mainly based on expert opinion, and the authors say each lab should select its own set based on its organisation and processes.
In 2021 the Maribor consensus added six indicators for clinical practice, including cycle cancellation rate before oocyte pick-up, the proportion of mature (MII) oocytes at ICSI, complication rate after pick-up, clinical pregnancy rate and multiple pregnancy rate. It also set a review rhythm: every six months or per 100 cycles for most indicators, and every three months or per 50 cycles for pregnancy and multiple pregnancy rates.
Together, the two documents give a lab and its clinicians a shared vocabulary. The hard part is collecting the numbers reliably, every week, without an embryologist spending evenings in Excel.
The evidence: KPIs as an early-warning system
A 2019 Human Reproduction study from Fertility Associates in New Zealand (Hammond and Morbeck) tested whether KPIs can spot trouble before clinical outcomes move. The team analysed 1,971 fresh IVF and ICSI cycles across three consecutive five-month periods, changing the culture medium in the middle period. Monthly KPIs were plotted on Shewhart control charts with warning (2-sigma) and control (3-sigma) limits.
- The Day 5 usable blastocyst rate fell from 32% to 25% after the medium change, and the fall was detected within one week. It recovered after the lab switched back.
- Clinical pregnancy rate also fell, from 51% to 36%, but this was not detected until three months after the KPI shift.
- The total usable blastocyst rate stayed inside control limits throughout, so a single overall blastocyst number would have missed the problem.
The authors flagged limits: the method may be less sensitive in smaller clinics with fewer cycles, and live birth rate was not included. A separate UK quality-improvement report in BMJ Open Quality, which used statistical process control charts to track ICSI KPIs after a root cause analysis, made the same point about low-volume clinics. Small clinics need enough data to be confident about a trend, so they risk being too slow to react, or reacting to false alarms.
Where automation comes in
None of this needs AI to start. It needs clean, structured data and a routine. That is exactly what manual paper and spreadsheet workflows struggle with. Automation helps at three levels.
1. Capture once, at the bench. A lab information system (LIMS), electronic witnessing system or time-lapse platform that records oocyte counts, fertilisation checks and embryo grades as the work happens removes re-typing. It also makes every KPI traceable to an operator, a date and a batch of media.
2. Automatic control charts. Once data is structured, weekly KPI calculation and control limits can run by themselves, with alerts when a point crosses a warning line. Our free KPI calculator is a simple starting point if you are still working from paper.
3. AI-assisted benchmarking. Some vendors now sell AI-based KPI monitoring. Fairtility, for example, describes its CHLOE KPI product as live, automatic monitoring of parameters such as fertilisation and blastocyst formation rates, flagging deviations from expected benchmarks. That is the vendor's description; ask for independent validation before relying on it.
Research is also exploring AI for fairer per-embryologist KPIs. A 2025 single-centre case study in the Journal of Clinical Medicine compared one senior embryologist's 474 ICSI cycles with pregnancy rates predicted by a model trained on 1,294 institutional cycles. Overall, observed and predicted rates were similar (0.31 versus 0.33), but they diverged in patients over 40. The authors concluded that personal KPIs are measurable but influenced by patient and physician factors, and that multi-operator, multi-centre validation is needed. In short: promising for adjusting for case mix, not yet a verdict on anyone's skill.
Don't forget environmental data
Lab KPIs tell you something went wrong; equipment data often tells you why. A 2026 RBMO abstract described an IoT monitoring system deployed across 48 centres of an IVF chain to track temperature, CO2/O2 levels and power. Deviations lasting more than five minutes triggered SMS alerts, then email at seven minutes and a phone call at ten, with a five-tier escalation path and central dashboards. The authors reported that power failures, gas cylinder issues and sensor failures were caught through timely alerts. Linking these logs with your KPI charts makes root cause analysis far faster. See also our guide to automated cryo-tank monitoring and the free lab QC logs.
A practical starter plan for Indian labs
- Pick five to eight KPIs from Vienna and Maribor that match your case mix, rather than all 25.
- Write down definitions, including the numerator, denominator and which cycles are included, so numbers stay comparable month to month.
- Record at source in your LIMS or witnessing system. If you are choosing one, our witnessing guide covers the questions to ask.
- Chart weekly, review monthly. Use control limits, and log every change of media, consumable lot, equipment or staff next to the chart.
- Mind your volume. If you run few cycles, group data over longer windows, as the Maribor review rhythm suggests, before acting on a trend.
- Keep the records. India's ART (Regulation) Act 2021 requires records to be kept for at least ten years; see our ART Act checklist.
- Test AI tools on your own data before buying. Ask vendors for validation studies, data-protection terms under India's DPDP Act, and how the tool exports data if you leave.
The bottom line
The consensus KPIs have existed since 2017. What has changed is that LIMS, witnessing, time-lapse and IoT systems now generate the data automatically, so a lab can watch its numbers weekly rather than discover a problem in next quarter's pregnancy figures. AI benchmarking is an interesting next layer, but the proven win today is simpler: consistent definitions, automatic capture and a control chart someone actually looks at.
Sources
- ESHRE SIG Embryology and Alpha Scientists in Reproductive Medicine, The Vienna consensus: ART laboratory performance indicators, Human Reproduction Open, 2017
- The Maribor consensus: performance indicators for clinical practice in ART, Human Reproduction Open, 2021
- Hammond ER, Morbeck DE, Tracking quality: can embryology key performance indicators be used to identify clinically relevant shifts in pregnancy rate?, Human Reproduction, 2019
- Improving ICSI success rates following root cause analysis and use of system behaviour charts, BMJ Open Quality, 2022
- Personal KPIs in IVF Laboratory: Are They Measurable or Distortable? A Case Study Using AI-Based Benchmarking, Journal of Clinical Medicine, 2025
- Dave V et al., Architecting an IoT-based tiered real time monitoring ecosystem for quality control vigilance in IVF laboratories, RBMO, 2026
- Fairtility, CHLOE KPI product page
- Assisted Reproductive Technology (Regulation) Act, 2021, India Code
Your next step
Try our free IVF lab KPI calculator to see where your lab sits against the Vienna competency and benchmark values. Vendors of LIMS, witnessing or lab-monitoring systems can claim a free listing so Indian lab managers can find you.
General operational information only; not medical, clinical or legal advice. Product names are mentioned for information and are not endorsements. See our disclaimer.