Why PICO Frameworks Fail for Medical Device Literature Searches

Written by HATEM RABEH, MD, MSc Ing

Your Clinical Evaluation Expert And Partner

in
S

The problem is not that PICO is inherently wrong. The problem is that clinical evaluation teams apply it mechanically without adapting it to the reality of device evidence.

When a Notified Body asks why your literature search missed key studies on complications, or why your appraisal included irrelevant publications, the root cause often traces back to how you structured your search question. And in most cases, that structure was borrowed from clinical trial methodology without considering what makes device evidence different.

What PICO Was Built For

PICO stands for Population, Intervention, Comparator, and Outcome. It comes from evidence-based medicine and was designed to frame clinical questions for drug trials and systematic reviews.

For a pharmaceutical product, PICO works cleanly. The intervention is the molecule. The population is well-defined by disease stage or risk factors. The comparator is usually placebo or standard treatment. The outcomes are clinical endpoints measured in controlled trials.

Medical devices do not follow this pattern.

A surgical instrument is not an intervention in the same way as a drug dose. A diagnostic algorithm does not have a comparator in the traditional sense. A wearable monitor does not produce outcomes that fit into efficacy versus safety binaries.

This is where clinical evaluation teams run into trouble. They define the “I” in PICO as the device name or device category. Then they search for that term and retrieve studies that mention the device but do not evaluate its performance, safety, or clinical benefit in a meaningful way.

How PICO Breaks Down in Device Searches

I reviewed a clinical evaluation report for a neurostimulation device. The manufacturer defined their PICO as follows:

Population: Adults with chronic pain
Intervention: Neurostimulation device
Comparator: Standard analgesic therapy
Outcome: Pain reduction on VAS scale

The search returned over 300 articles. Most were excluded during screening because they described different neurostimulation technologies, different pain etiologies, or used the device as part of multimodal therapy where attributing the outcome to the device alone was impossible.

The manufacturer ended up with 12 relevant studies. The Notified Body pointed out that none of these studies addressed device-related adverse events in real-world settings, nor did they cover the specific patient subgroup most likely to use the device in clinical practice.

What went wrong?

The PICO framing was too broad in some dimensions and too narrow in others. “Neurostimulation device” captured too many device types. “Chronic pain” included conditions that were not relevant to the intended purpose. “Pain reduction” ignored safety, usability, and long-term performance issues that are critical for device evaluation.

The core issue is that PICO does not naturally accommodate the technical and contextual variables that define how a device works and how it produces outcomes.

What Changes When You Adapt PICO for Devices

The MDR requires that clinical evaluation be based on a systematic literature review. That review must cover both the device under evaluation and equivalent or similar devices. The search strategy must be transparent, reproducible, and justified.

MDCG 2020-6 on sufficient clinical evidence reinforces this. It states that the clinical evaluation must address the device’s safety and performance in the target population and under the conditions of use.

This means your search question cannot be framed as a simple intervention comparison. It must capture technical characteristics, clinical indications, user settings, and risk-relevant scenarios.

In practice, this requires modifying or expanding the PICO structure.

Population: Not Just Demographics

For devices, the population is not only defined by disease or age. It includes the clinical setting, the skill level of the operator, and the presence of comorbidities or contraindications that affect device safety.

A ventilator used in an ICU by trained intensivists is not the same as a portable ventilator used by home care patients. The population must reflect these contextual factors.

Intervention: Technical and Functional Specification

Instead of naming the device, define what the device does and how it does it. If your device is a laparoscopic trocar, the intervention is not “trocar.” It is “minimally invasive access with radially expanding mechanism.”

This allows you to capture studies on devices with the same functional mechanism even if they have different brand names or slight design variations.

Comparator: Often Not Relevant

Many devices do not have a direct comparator. Diagnostic devices are compared to a reference standard, not to another device. Surgical instruments may be compared to an open surgical approach, but that comparison is rarely the focus of the literature.

In these cases, the comparator can be omitted or redefined as the clinical standard of care without the device.

Outcome: Safety and Performance, Not Just Efficacy

Device outcomes must include adverse events, usability issues, technical failures, and long-term performance. A successful clinical outcome in one study may coincide with a high rate of device-related complications that are only reported in case series or post-market surveillance data.

Your search strategy must be sensitive to these outcomes even if they are not described in the abstract or keywords.

What I See Work in Real Submissions

The evaluations that pass Notified Body review without major objections on literature search do not strictly follow PICO. They follow a structured search strategy that is tailored to the device’s technical profile and risk class.

Here is what they do differently.

They start with the intended purpose and the technical characteristics that define the device. Then they map those characteristics to search terms that capture functional equivalence, not just brand names.

They include search terms for adverse events, complications, and failure modes. These terms are often not found in the title or abstract. They appear in methods sections, supplementary tables, or case reports.

They define the population in clinical and contextual terms. If the device is used by non-specialists, they search for studies that reflect that setting. If the device is used in pediatric patients, they explicitly include pediatric populations even if the intended use statement does not specify age.

They iterate the search. The first search is exploratory. They review the retrieved studies, identify additional relevant terms, and run the search again. This iterative process is documented and justified in the clinical evaluation report.

When PICO Becomes a Box-Checking Exercise

I have reviewed clinical evaluations where the PICO table was filled out in the appendix, but the search strategy itself bore no relationship to the PICO structure. The search was done using device names and synonyms. The PICO framework was added later to satisfy a perceived regulatory expectation.

This is worse than not using PICO at all. It signals to the reviewer that the team did not think critically about the search strategy. It suggests that the evaluation was done by template, not by analysis.

The MDR does not require PICO. It requires a systematic approach. MDCG 2020-13 on clinical evaluation provides examples of search strategies, but it does not prescribe PICO as the only acceptable method.

What the regulation requires is that the search strategy is appropriate for the device, transparent, reproducible, and capable of identifying relevant evidence on safety and performance.

If PICO helps you structure that strategy, use it. If it forces you into irrelevant constraints, adapt it or replace it.

What Matters More Than the Framework

The question you are answering is not “What studies exist about this device?”

The question is “What does the available evidence tell us about the safety and performance of this device and equivalent devices in the target population under the conditions of use?”

That question does not fit neatly into PICO. It requires you to define technical equivalence, clinical relevance, and contextual factors that influence how the device performs and how risks manifest.

Your search strategy must reflect that complexity. It must be broad enough to capture relevant evidence on similar devices and narrow enough to exclude studies that do not inform your evaluation.

This balance is not achieved by filling out a PICO table. It is achieved by understanding your device, understanding the clinical and technical landscape, and designing a search that captures the evidence you need to demonstrate conformity with the MDR.

Moving Forward

If you are using PICO, ask yourself whether it is helping you capture the right evidence or whether it is constraining your search in ways that miss critical studies.

If your Notified Body has questioned your search strategy, do not defend PICO as a regulatory standard. Instead, revisit the search logic. Document why each search term was included. Show how the search was tailored to the device’s technical and clinical profile. Demonstrate that the search is sensitive to safety and performance outcomes, not just efficacy.

The framework is less important than the reasoning. And the reasoning must be driven by the device, not by the template.