The caregiver design gap

Design strategy
Written By Amruta Ghosalkar

👋🏼

The caregiver operates the device but wasn’t in the user model that shapes it.

In home cardiac monitoring, the caregiver charges the device, interprets the alerts and decides what needs immediate attention. Informal caregivers participated in only 37% of usability studies that identified them as a target user group, the lowest representation of any user category. That parttern is not unique to cardiac devices. It appears identically across home dialysis, pediatric ventilators, insulin pumps, and negative pressure wound therapy (NPWT) systems. The same user gets omitted from the brief for the same structural reasons.

Home dialysis: Designed for someone else’s hands

When Quanta Dialysis Technologies sought FDA clearance for home hemodialysis with a trained care partner, the human factors submission combined patients and caregivers into one user group. That regulatory decision reflects how caregivers are categorized in device design, functionally present, structurally invisible. In practice, caregivers manage device setup, respond to alarms, and troubleshoot failures. In the brief, they are afterthoughts. The interface problems in home dialysis cluster around predictable failures. Machines place control screens too far from the patient’s reach. Clamps require force levels that fail when blood pressure is low. Alarm messages surface thresholds crossed but not what to do. A UK study found caregivers disposing of blood due to equipment delays they could not resolve is the interface assumed clinical fluency they did not have.

Fresenius, Baxter and NX stage produce systems capable of sophisticated data management. Quanta dialysis technologies markets step-by-step touchscreen instructions “designed for patients and caregiver,” achieving >95% success in human factors testing. But the caregiver’s task set was designed after the patient interface was built, not as apart of the original user model. Caregiver burden data exists: 38-48% of dialysis caregivers report moderate to high burden using the Zarit Burden interview. The burden correlates with interface failures, but those pain points do not feed back into design process because caregivers were not in the room when the brief was written.

Pediatric Home Ventilators: Always awake, never acknowledged

Of the 9.4 million children in the US with special health care needs, those requiring chronic invasive ventilation at home present complex operational scenarios. American Thoracic Society guidelines explicitly requires “an awake and fully-trained caregiver” That operational requirement does not translate into a user requirement in device design. Philips Respironics and Res Med produce pediatric home ventilators optimized for clinical reliability. What they do no do is accomodate the parent managing a child’s ventilation mid-discharge, under time pressure, without clinical background. The interface does not differentiate between a respiratory therapist and a parent responding to a 2AM alarm.

Training program exist but there is no interface design for the trained non-clinical user. Remote monitoring systems highlight the gap: parents needed multiple caregiver logins, reminders, and individualized programming based on the child’s ventilation schedule feature that would have been specified upfront if the caregiver has been in the brief from the start. Shorter caregiver education duration is a risk factor for unscheduled hospital admissions. Caregiver quality of life deteriorates significantly over the first year, with declines in physical health and perceived health status. The device was not designed to prevent this.

Insulin Pumps: Remote monitoring as retrofit

Children under age seven with Type 1 diabetes are entirely dependent on caregivers. Pumps like Omnipod 5 (insulet), Medtronic 670G, and Tandem t:Slim X2 with control-IQ have made strides in automation. The problem is that caregiver support features were added as enhancements, not designed into the original user model. Omnipod DASH introduced remote monitoring through a mobile app in generation 4, years after platform launch. Automated bolus calculators exist because manufacturers recognized pot-deployment that grandparents and daycare workers needed decision support the original interface did not provide.

Caregiver distress is well documented. Mother report moderate but significant burden on the PAID-PR scale. Pumps reduce burden compared to multiple daily injections, but the task set, monitoring glucose, calculating doses, responding to alarms remain operationally complex for someone who was not the primary user in the original brief. The trajectory reflects a reactive design process: features added in response to observed pain points, not a user model that anticipated them.

NPWT: Training as interface substitute

Negative pressure wound therapy system and Pensar’s medical’s WoundPro are designed for homw use with caregiver operation. Caregivers manage dressing changes, monitor canister fluid levels, respond to alarms, and troubleshoot seal failures. A 2021 Delphi study identified 36 essential training components across 3 domains, pre-discharge education, treatment day education and ongoing at home support. The training exists becuase the interface does not. Alarm systems notify caregivers of blockages or leaks but do not guide resolution. Troubleshooting instructions are in manuals, not surfaced in the device intersection itself. WoundPro markets “24/7 support hotlines” and “expert consultation” an infrastructure that compensated for an interface assuming clinical fluency.

Home-based NPWT requires caregivers to be closely bound up with equipment 24 hours a day” The brief recognizing operational burden is not the same as designing to reduce it. If the training is comprehensive and the device still requires a 24/7 hotline, the interface has not been designed for the user operating it.

The upstream failure is structural, not accidental

The caregiver omission follows the same logic across product categories. Medical device regulatory pathways define user populations in intended use statements, which determine participates in human factors testing. Caregivers do not map cleanly onto FDA user categories: they are not the patient, not the clinician, and not always clearly defined as a “lay caregiver”. When Byonyks combined patients and caregivers into one user group for the X-1 APD cycler, that allowed FDA clearance but did not differentiate the caregiver’s task set from the patient’s.

Market dynamics reinforce the omission. Purchasing decisions for home medical devices are made by health systems, insurers, or prescribing physicians, not by family member who will operate the device. Manufacturers optimize for the decision maker whose satisfaction influences procurement. Caregiver experience is downstream and not tracked in outcome measures manufacturers are accountable to. There is no feedback loop connecting caregiver interface failures back to the design team.

What changes if the brief names them

The methodology to fix this already exists. Human factors engineering, participatory design, and contextual inquiry are standard in medtech. Research on co-design with family caregivers has found that engaging them from problem definition, not at usability testing, but when the brief is written, meaningfully changes what the design addresses. A brief that names the caregiver requires a user profile, the family member who takes over device management mid discharge, under time pressure, without clinical background, often coordinating across multiple caregivers with different involvement levels. Specific tasks: alert triage, device maintenance, data relay to care teams.Specific failure modes like “what does the caregiver do when an alarm fires and no guidance is available? What does correct caregiver behavior look like, and has anyone in the design process defined it?”

Cardiac monitoring devices from iRhythm (Zio XT, Zio AT), Boston Scientific (BodyGuardian MINI), Philips BioTelemetry, and Bardy Diagnostics have made strides in patient-centered design with adhesive patches that require no charging, remote data transmission, minimal disruption to daily life. The operational burden on caregivers in those systems is lower because the devices were designed to minimize intervention. But even those systems do not explicitly model the caregiver who interprets alert data, manages device troubleshooting, or coordinates care when the patient cannot. Emerging cardiac ultrasound patch technology from UC San Diego (Sheng Xu lab), Sonus Microsystems, and B-Secur's HeartKey platform will eventually move continuous structural heart monitoring into home settings. If those devices reach FDA clearance and enter home use, the same design question will arise: who is reading the ultrasound data, in what context, with what clinical background, and what does the interface need to do for that person?

The device cannot answer a question the brief never asked. Until caregivers are named in the user model upstream at the point where the intended use statement is written, not at the point where the interface is tested the third user will keep inheriting tools designed for someone else.

Amruta Ghosalkar
Next

The third user