Connected Health and Wearables · Study deck

Medication IoT: Adherence Evidence and Verification

A lid-open event is not the same as a swallowed dose.

Ada is your guide for this deck.

casesmedication
iotclass.org

After studying this chapter

Learning objectives

You will be able to:

  • Explain: It may send events through BLE to a phone, Wi-Fi to a home router, cellular for independent operation, or MQTT/HTTPS over TLS to a cloud service.
  • Explain: A smart dispenser can combine load cells, lid switches, beam-break sensors, RFID/NFC tray identity, motor encoders, real-time clocks, tamper switches, local storage, and battery state.
  • Explain: It still does not prove that the medicine was absorbed or had the intended effect, and every relay adds consent, privacy, usability, and workflow duties.
  • Explain: The family was worried, so they got him a Smart Pill Bottle.
Smart dispenser evidence diagram separating door and tray observations, weight change, and reminders from downstream cloud inferences about medication events.
Smart dispenser evidence diagram separating door and tray observations, weight change, and reminders from downstream cloud inferences about medication events.
iotclass.org

Major section

Adherence Data Provenance

The system should record those states rather than silently converting them into adherence or non-adherence.

  • A smart dispenser can combine load cells, lid switches, beam-break sensors, RFID/NFC tray identity, motor encoders, real-time clocks, tamper switches, local storage, and battery state.
  • EHR integration is a reconciliation problem.
  • The event store should be append-only and idempotent.
Periodic checkups capture about 30 minutes of monthly Parkinson observation, while continuous IoT monitoring captures 43,200 minutes per month, about 1440 times more evidence for earlier detection and medication optimization.
Periodic checkups capture about 30 minutes of monthly Parkinson observation, while continuous IoT monitoring captures 43,200 minutes per month, about 1440 times more evidence for earlier detection and medication optimization.
iotclass.org

Major section

Adherence Data Provenance (continued)

It may send events through BLE to a phone, Wi-Fi to a home router, cellular for independent operation, or MQTT/HTTPS over TLS to a cloud service.

  • Each event needs a device id, compartment id, medication id, schedule version, timestamp source, sequence number, confidence, and offline-sync status.
  • Missing patch contact, delayed activation, phone offline state, patient refusal, duplicate ingestion signals, and portal acknowledgement are all different states.
  • A FHIR or HL7 interface must match device events to the current order, patient identity, encounter context, and clinician workflow.
  • Security and privacy controls sit in the data model, not only at the login screen.
An ingested pill leaves a sensor activated by stomach fluids. It signals a skin patch, which confirms the pill to a phone. The figure ends at a care dashboard that displays the relayed event rather than depicting the sensor’s later physical path.
An ingestible sensor activates in the stomach, signals a wearable patch, and relays a timestamp through a phone and cloud to the care dashboard.
iotclass.org

Major section

For Kids: Meet the Sensor Squad!

Medication adherence means taking your medicine at the right time, in the right dose, every single day.

  • It sounds simple, but millions of people forget -- and smart sensors are here to help!
  • The family was worried, so they got him a Smart Pill Bottle.
  • If the weight doesn't change when it should, I know he forgot!".
  • If you gave 6 out of 7 days, that is about 86% in this story!
A dissolving smart pill powers a brief body signal that a patch and phone relay to the provider's adherence record.
A dissolving smart pill powers a brief body signal that a patch and phone relay to the provider's adherence record.
iotclass.org

Major section

Smart Medication Dispenser Architecture

The low-power body signal reaches a wearable patch, then a phone relays the event through the cloud to a provider dashboard.

  • That chain can add a precise ingestion timestamp, stronger evidence than a reminder or opened bottle.
  • It still does not prove that the medicine was absorbed or had the intended effect, and every relay adds consent, privacy, usability, and workflow duties.
  • The dispenser at the top holds four kinds of hardware, and each one produces a different fact: which tray was opened, how much weight left it, whether the carousel moved, and whether a prompt was given.
An ingestible sensor activates in the stomach, signals a wearable patch, and relays a timestamp through a phone and cloud to the care dashboard.
An ingestible sensor activates in the stomach, signals a wearable patch, and relays a timestamp through a phone and cloud to the care dashboard.
iotclass.org

Major section

Chronic Monitoring Value

Patients with chronic neurological conditions like Parkinson's disease experience daily symptom fluctuations that monthly doctor visits cannot capture.

  • By the time a patient sees their neurologist, the "good days" and "bad days" have averaged out, making medication optimization nearly impossible.
  • "Instead of patients actively performing certain tasks, could we monitor disease progression passively in the background?".
  • The gain is the days between visits, not a better measurement on the day itself.
Periodic checkups capture about 30 minutes of monthly Parkinson observation, while continuous IoT monitoring captures 43,200 minutes per month, about 1440 times more evidence for earlier detection and medication optimization.
Periodic checkups capture about 30 minutes of monthly Parkinson observation, while continuous IoT monitoring captures 43,200 minutes per month, about 1440 times more evidence for earlier detection and medication optimization.
iotclass.org

Deck summary

Key takeaways

The system should record those states rather than silently converting them into adherence or non-adherence.

  • It may send events through BLE to a phone, Wi-Fi to a home router, cellular for independent operation, or MQTT/HTTPS over TLS to a cloud service.
  • Medication adherence means taking your medicine at the right time, in the right dose, every single day.
  • The low-power body signal reaches a wearable patch, then a phone relays the event through the cloud to a provider dashboard.
  • Patients with chronic neurological conditions like Parkinson's disease experience daily symptom fluctuations that monthly doctor visits cannot capture.
iotclass.org

Retrieval practice

Recall check 1 of 5

Ada says: answer from memory, then check your reasoning.

Q1An ingestible-sensor patch loses contact and the portal shows no new signal. What should the care record say?

ARecord a certain missed dose, because lost patch contact establishes that ingestion did not occur.
BMark ingestion unverified because patch contact failed during signal collection.
CRecord a swallowed dose, because the portal's prior status confirms ingestion after the new signal is absent.
DRecord treatment refusal from missing patch contact, because contact loss is evidence of a deliberate decision.
Show answer

Answer: B Separate the device fault from the still-unknown medication event.

iotclass.org

Retrieval practice

Recall check 2 of 5

Ada says: answer from memory, then check your reasoning.

Q2A dispenser reconnects and uploads repeated door-open events. What should the event store preserve?

AProvenance and idempotent reconciliation of duplicates
BA new adherence event for each received upload
CA non-adherence label for the offline interval
DA corrected patient note without its source context
Show answer

Answer: A The chapter distinguishes sequence, schedule, and offline state rather than treating repeated uploads as new adherence.

iotclass.org

Retrieval practice

Recall check 3 of 5

Ada says: answer from memory, then check your reasoning.

Q3A pill bottle reports its lid opening to a care system. What is the device’s directly observed event?

AA prescribed dose being swallowed, inferred solely from opening the lid.
BA clinician's final care decision, inferred from the bottle's switch.
CThe bottle lid opening, directly reported by the device sensor.
DA patient's response to medication, inferred from a lid-opening event.
Show answer

Answer: C The basic example describes a sensor that detects lid opening and sends the event.

iotclass.org

Retrieval practice

Recall check 4 of 5

Ada says: answer from memory, then check your reasoning.

Q4A patch loses contact after a patient takes a sensor-enabled pill. Which claim can the app safely make?

ARecord a confirmed missed dose because the patch stopped reporting.
BRecord confirmed ingestion, because the prescribed dose schedule establishes that the dose was taken.
CMark ingestion uncertain and show the missing patch-contact interval.
DHide the gap and keep the last dose status as if it were current.
Show answer

Answer: C Lost contact is a device state; it cannot prove that a dose was taken or missed.

iotclass.org

Retrieval practice

Recall check 5 of 5

Ada says: answer from memory, then check your reasoning.

Q5A care team wants to compare a planned tapping test with background movement sensing. What is the key difference?

ATapping is prompted, while background sensing observes daily routines.
BTreat both as diagnostic proof, because prompted and observed movement provide equivalent clinical evidence.
CBackground sensing needs no consent because it is not a planned test.
DA planned tapping test records ordinary movement continuously all day.
Show answer

Answer: A Active tests need a deliberate task, while passive sensing can capture daily behavior.

iotclass.org

Print reference

Answers

Answer key.

  1. B · Separate the device fault from the still-unknown medication event.
  2. A · The chapter distinguishes sequence, schedule, and offline state rather than treating repeated uploads as new adherence.
  3. C · The basic example describes a sensor that detects lid opening and sends the event.
  4. C · Lost contact is a device state; it cannot prove that a dose was taken or missed.
  5. A · Active tests need a deliberate task, while passive sensing can capture daily behavior.
iotclass.org