Clinical Evidence

NMRA · CER for general devices · Performance evaluation for IVDs

Why Clinical Evidence Is Required

The NMRA requires clinical evidence to establish that a device performs safely and as intended when used in patients in clinical practice. Clinical evidence is distinct from bench testing — it addresses the device's actual clinical behaviour, not just its physical or chemical properties.

General Devices — Clinical Evaluation Report (CER)

For general medical devices, the clinical evidence is compiled into a Clinical Evaluation Report (CER) following a systematic methodology:

CER Process

1. Define scope — device description, intended purpose, clinical claims
2. Identify applicable clinical data — from three possible sources:
   • Clinical investigations of the subject device
   • Equivalent device data (with justified equivalence)
   • Published clinical literature (systematic literature review)
3. Appraise data quality — assess relevance, methodological quality, and risk of bias
4. Analyse and conclude — evaluate whether the data supports the safety and performance claims
5. Document residual uncertainties — and how they are addressed (e.g., by PMCF)

CER Depth by Class

Class	CER Depth
Listed/I	Abbreviated review may suffice for well-established technologies
IIa	Literature-based CER with systematic review methodology
IIb	Comprehensive CER; clinical data from investigations may be required for novel technologies
III	Robust clinical data including clinical investigations where no equivalent device exists

IVDs — Performance Evaluation

IVD registration requires performance evaluation data rather than a clinical evaluation report:

Analytical Performance

Parameter	Description
Analytical sensitivity	Limit of detection and limit of quantification
Analytical specificity	Cross-reactivity and interference testing
Precision	Repeatability and reproducibility
Accuracy / trueness	Comparison to reference method
Linearity/measuring range	For quantitative assays
Stability	Shelf life and open-vial stability

Clinical Performance

• Sensitivity and specificity compared to reference method or gold standard
• Positive and negative predictive values in the intended use population
• Results from clinical studies or specimen bank evaluations

Overseas Clinical Data

NMRA accepts clinical data generated outside Sri Lanka, provided:

• Device used in the overseas study is identical (or equivalence is properly justified)
• Study population is representative of the Sri Lankan patient population
• Clinical conditions and intended use are comparable

When using the reliance pathway, the reference NRA's clinical assessment and conclusions form a key component of the reliance evidence package. The NMRA may accept these conclusions as supporting evidence, though the NMRA retains the right to conduct additional review if needed.

Post-Market Clinical Follow-Up (PMCF)

For Class IIb and III devices, the NMRA may require a PMCF plan as part of the registration dossier. PMCF collects clinical data after registration to:

• Confirm long-term safety and performance
• Detect rare or late-emerging adverse events
• Address residual uncertainties identified in the pre-market CER

In Overseas Clinical Data section, add: 'Where clinical data is from significantly different geographic, climatic, or epidemiological regions, provide justification that the data remains applicable to the Sri Lankan patient population, or plan for supplementary local clinical data collection.'