Clinical trial needs and regulations for medical devices and how best practices from the pharmaceutical and biotechnology industry can be helpful

1. HOW CAN
MEDICAL
DEVICE
LEARN FROM
PHARMA?
Timothy C. Grammer, PhD MS MBA PMP
Outsourcing in Clinical Trials West Coast: Medical Device Stream
Burlingame, CA | February 13th 2019

2. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Avoid
Common Trial
Pitfalls
Site selection
Training
Enrollment
Retention
Documentation
Randomized
Controlled Trials
(RCTs)
Pros and cons of RCTs
Trial Design
Avoid Bias
Randomization
Blinding
Embrace
Trends
(not fads)
Real World Data
Patient reported
outcomes (PROs)
Connected devices
/ biosensors
Machine learning /
data analytics

3. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Let’s first make
sure we are all on
the same page

4. ARE YOU A
MEDICAL DEVICE?
Know if your
product is a
medical device
4

5. WHAT IS A
MEDICAL DEVICE?
Section 201(h) of the Federal Food Drug & Cosmetic (FD&C) Act
regulated by the Food and Drug Administration (FDA) defines a medical device as:
• an instrument, apparatus, implement, machine, contrivance, implant, in vitro
reagent, or other similar or related article, including a component part, or
accessory which is:
• recognized in the official National Formulary, or the United States
Pharmacopoeia, or any supplement to them,
• intended for use in the diagnosis of disease or other conditions, or in the
cure, mitigation, treatment, or prevention of disease, in man or other
animals, or
• intended to affect the structure or any function of the body of man or other
animals, and
• which does not achieve its primary intended purposes through
chemical action within or on the body of man or other animals and
which is not dependent upon being metabolized for the achievement
of its primary intended purposes. The term "device" does not include
software functions excluded pursuant to section 520(o).
The above definition is provided on FDA’s website. This presentation is US/FDA focused
EU and Global Harmonization Task Force have similar definitions

6. Some Devices Are Easy to Recognize
• Artificial knee
• Surgical stent
Some Devices Are Less Easy to Recognize
• In vitro diagnostic
• Reagent / reagent product
• Calibrators
• Wearable (smart contact lens, Apple Watch, Fitbit)
• Software app for behavioral modification
• Digital therapeutics*
ARE YOU A
MEDICAL DEVICE?
*There are discussions later today about digital therapeutics

7. Know your
device’s
risk classification
7
DEVICE CLASSIFICATION

8. What determines the classification
• Risks and the regulatory controls necessary to provide a
reasonable assurance of safety and effectiveness
Some hazard and risk considerations include:
• Treatment area (local versus broad)
• Extent of exposure risk (local versus systemic)
• Invasiveness
• Duration of contact with the body
• Potential for misuse
• Implications of defects or performance issues
8
DEVICE CLASSIFICATION

9. DEVICE CLASSIFICATION
9
US Device
Classifications are
risk based
Class I
Class II
Class III
Pharmaceuticals
are all considered
high risk
Lowest risk
Medium risk
Highest risk
The above classifications are for the US.
The EU splits Class II into 2 groups: Class IIa and Class IIb
Canada has 4 Classifications: I, II, III, and IV

10. DEVICE CLASSIFICATION
10
US Device
Classifications are
risk based
Class I
Class II
Class III
Pharmaceuticals
are all considered
high risk
The above classifications are for the US.
The EU splits Class II into 2 groups: Class IIa and Class IIb
Canada has 4 Classifications: I, II, III, and IV
General Controls apply;
Most exempt from
Premarket Notification
General Controls apply;
Special Controls apply;
Premarket Notification (510(k))
is the most likely path to
market
General Controls apply;
Special Controls apply;
Premarket Approval (PMA) is
the most likely path to market

11. DEVICE CLASSIFICATION
11
US Device
Classifications are
risk based
Class I
Class II
Class III
Pharmaceuticals
are all considered
high risk
The above classifications are for the US.
The EU splits Class II into 2 groups: Class IIa and Class IIb
Canada has 4 Classifications: I, II, III, and IV
Clinical Trial data
Is not needed
Clinical Trial data
may be needed
Clinical Trial data
will be needed
Usually,…

12. IS YOUR DEVICE
SR OR NSR?
Know your
device’s
risk significance
12

13. SR VERSUS NSR
DEVICES
• Significant Risk (SR) devices pose potential for
serious risk to the health, safety, or welfare of a
subject
• Non-Significant Risk (NSR) devices are defined
as those not being categorized as significant
risk (SR)
• This determination is important:
• SR devices require both IRB and FDA approval
before starting a clinical trial
• NSR devices require only IRB approval before
starting a clinical trial
13

14. DEFINITION OF
SIGNIFICANT RISK (SR) DEVICE
Under 21 CFR 812.3(m), an SR device means
an investigational device that:
• Is intended as an implant and presents a potential for serious
risk to the health, safety, or welfare of a subject;
• Is purported or represented to be for use supporting or
sustaining human life and presents a potential for serious risk
to the health, safety, or welfare of a subject;
• Is for a use of substantial importance in diagnosing, curing,
mitigating, or treating disease, or otherwise preventing
impairment of human health and presents a potential for
serious risk to the health, safety, or welfare of a subject; or
• Otherwise presents a potential for serious risk to the health,
safety, or welfare of a subject.
14
An NSR device is one that does not meet the definition above for an SR device

15. WHO DETERMINES
SR VERSUS NSR
Your Institutional Review Board (IRB) does (21 CFR 812.66)
If you want to start a clinical trial and you feel that your
device does not pose a significant risk to study subjects:
• You (the Sponsor) submit your protocol to your IRB along
with a statement of why your device trial does not pose a
significant risk to study subjects
• If the IRB agrees that the clinical trial does not pose a
significant risk and determines the trial device as an NSR,
then, no FDA approval is needed before you start the trial.
Your clinical trial can start under Abbreviated Investigational
Device Exemption (Abbreviated IDE) requirements.
• If the IRB disagrees and they deem your trial to pose a
significant risk, then your device is deemed SR and you must
submit an Investigational Device Exemption (IDE)
application to FDA for approval before starting your trial
15

16. WHAT IS YOUR
REGULATORY PATH?
Know your
regulatory
path
16

17. Pre-IDE IDE PMA PMA
supplement
STAGES OF FDA INTERACTION
Pre-IND IND
NDA /
BLA
sNDA /
sBLA
Typical
Pharma Path
Is There a
Typical
Device Path?

18. WHEN DO YOU INTERACT
WITH THE FDA?
Pre-
Submission
Meeting
IDE
510(k)
De novo
PMA
Optional meeting to discuss:
device design,
bench testing,
animal testing,
performance testing,
clinical trial design (if needed)
Request approval to:
start a clinical trial
for a significant risk
(SR) device not
approved for the
indication being
studied
File:
Pre-market notification;
Pre-market approval
If you are building an NSR device and choose not to have a pre-submission (“pre-sub”)
meeting, you may be only interacting with the FDA at the time of pre-market application

19. PREMARKET NOTIFICATION
VERSUS PREMARKET APPROVAL
2 main FDA Processes for Medical Devices:
• Premarket Notification process
• often referred to as a “510(k)”
• FDA grants Clearance
• Premarket Approval process
• referred to as a “PMA”
• FDA grants Approval
19

20. PREMARKET NOTIFICATION
510(K)
You (the Sponsor) must prove Substantial
equivalence to a predicate device(s) already
legally on the market through:
• Performance testing
• bench testing
• animal testing (if applicable)
Depending on the difference between your device and the
proposed predicate device(s), you may not need clinical trial
data. If you do, it’s usually a single arm unblinded study to
demonstrate substantial equivalence to the predicate device.

21. PREMARKET APPROVAL
(PMA)
You (the Sponsor) must demonstrate safety
and efficacy for your device
• Performance testing
• WILL require clinical trial support
If you cannot find a predicate device to apply for the 510(k)
process but feel your device does not have the safety risks
warranting a full PMA, look into the FDA’s “De Novo” path

22. MEDICAL DEVICE
CLINICAL TRIALS
Know if you
need a
clinical trial
22

23. DO YOU NEED A
CLINICAL TRIAL?
Pharma:
• When studying new drugs or biologics, a clinical trial is
always required
Medical Devices:
• clinical trials are not always required
• depends on a risk assessment
• Class I devices do not require clinical trials
• Class II devices sometimes require clinical trials
• Class III devices always require clinical trials

24. DO YOU NEED A
CLINICAL TRIAL?
Medical Devices:
• 510(k) submissions sometimes require clinical trials, which
are usually confirmatory to demonstrate the new device
is as safe and effective as (i.e., substantially equivalent to)
a legally marketed predicate device
• PMA submissions always require clinical trials, which
provide reasonable assurance that a device is both safe
and effective for its intended use
Therefore, whether or not you need a clinical trial depends
on your risk classification (II or III) and whether you are
applying for a 510(k) or a PMA

25. MEDICAL DEVICE
CLINICAL TRIALS
Know your
clinical trial
regulations
25

26. US HUMAN RESEARCH
REGULATIONS
Code of Federal Regulations (CFR)
• Title 21- Food and Drugs (apply to medical devices)
• Part 11 Electronic Medical Records
• Part 50 Informed Consent
• Part 54 Financial Disclosure of clinical investigator
• Part 56 IRB
• Part 312 IND
• Part 314 NDA
• Part 600 Biologics
• Part 812 IDEs
• Part 814 Premarket Approval for Medical Devices
• Title 45 - Public Welfare (apply to medical devices)
• Part 46 DHHS Protection of human research subjects
• NIH Grant policies for Clinical Trials (apply to medical devices)
• Check the NIH website to confirm if your NIH funded study (e.g., SBIR grant) is
deemed a clinical trial and therefore must comply with NIH trial policies and
forms particularly around protection of vulnerable populations

27. SOME ADDITIONAL US
REGULATIONS TO KNOW
Code of Federal Regulations (CFR)
• Title 21- Food and Drugs (apply to medical devices)
• Part 58 Good Laboratory Practice for Nonclinical
Laboratory Studies
• Part 211 Current Good Manufacturing Practice for
Finished Pharmaceuticals
• Part 820 Quality System Regulation

28. SOME MEDICAL DEVICE
USAGE IS EXEMPT FROM
21 CFR PART 812 (IDE)
Medical Device Sponsors and Clinical Investigators may be exempt from
21 CFR Part 812 regulations regarding IDE requirements (meaning that
they need neither FDA nor IRB approval) if the medical device is:
• used in accordance with its legally marketed labeling,
• testing only consumer preference (i.e., no safety or efficacy tests),
• testing a modification that (1) is not for the purpose of determining
safety or efficacy and (2) does not put subjects at risk,
• intended solely for veterinary use, or
• a diagnostic that:
• complies with 809.10(c) labeling requirements and
• the testing
• is noninvasive;
• does not pose a significant risk;
• does not introduce energy into a subject; and
• is not used as a diagnostic procedure without confirmation by another
medically established diagnostic product or procedure;
28

29. IDE REGULATION
21 CFR PART 812
Device trials need IDE determination from an IRB (i.e., IRB
approval) when:
• Data is being collected on safety and effectiveness
regardless of whether or not the results are intended for
FDA submission
• New disease or condition
• New intended use
• New patient population
• Significant design changes
• Significant use changes
29

30. MEDICAL DEVICE
CLINICAL TRIALS
Know what
trial phase
you are in
30

31. Early
Feasibility Feasibility Pivotal
Post-Market
Surveillance
CLINICAL TRIAL STAGES
Phase
0
Phase
I
Phase
II
Phase
III
Phase
IV
Typical
Pharma Path
Typical
Device Path
Pre-
approval
Post-
approval
# Subjects
# Subjects <10
10-15 100-200
100-50010-50
20-100 100s-1000s 1000s
100s-1000s

32. Early
Feasibility Feasibility Pivotal
Post-Market
Surveillance
CLINICAL TRIAL STAGES
Phase
0
Phase
I
Phase
II
Phase
III
Phase
IV
Typical
Pharma Path
Typical
Device Path
Pre-
approval
Post-
approval
Health status
Health status patients
healthy patients
patientspatients
healthy patients patients
patients

33. Early
Feasibility Feasibility Pivotal
Post-Market
Surveillance
CLINICAL TRIAL STAGES
Phase
0
Phase
I
Phase
II
Phase
III
Phase
IV
Typical
Pharma Path
Typical
Device Path
Key Outputs
Key Outputs
Safety
Performance
Human Factors
Early PK/PD
MOA
Safety
Efficacy
Safety
Performance
Efficacy
Safety
Performance
Efficacy
Safety Safety
Efficacy
Safety
RWD
Safety
Performance
RWD
Registry Trials
Pre-
approval
Post-
approval
MOA = mechanism of action PK/PD = pharmacokinetics / pharmacodynamics
RWD = Real World Data; post-approval data relating to patient health status and/or the delivery of health care

34. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Now we should all be on the same page
and I can start my talk, because you know:
1. Your product is a medical device
2. Your risk class (I, II, or III)
3. Your regulatory path (510(k) or PMA)
4. Your need for clinical trial data
5. Your risk significance (NSR or SR)
6. Your clinical trial regulations
7. Your clinical trial stage (Early Feasibility, Feasibility, Pivotal)

35. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Avoid
Common Trial
Pitfalls
Site selection
Training
Enrollment
Retention
Documentation
Randomized
Controlled Trials
(RCTs)
Pros and cons of RCTs
Trial Design
Avoid Bias
Randomization
Blinding
Embrace
Trends
(not fads)
Real World Data
Patient reported
outcomes (PROs)
Connected devices
/ biosensors
Machine learning /
data analytics

36. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Avoid
Common Trial
Pitfalls
Site selection
Training
Enrollment
Retention
Documentation
Randomized
Controlled Trials
(RCTs)
Pros and cons of RCTs
Trial Design
Avoid Bias
Randomization
Blinding
Embrace
Trends
(not fads)
Real World Data
Patient reported
outcomes (PROs)
Connected devices
/ biosensors
Machine learning /
data analytics

37. COMMON CLINICAL TRIAL
SPONSOR PITFALLS
• Failure to follow the clinical trial regulations
• Lack of proper oversight of the trial stakeholders
(e.g., CRO, site, labs)
• Lack of proper study subject protection
• Inadequate reporting to the IRB
• Inadequate reporting to the FDA, when needed
• Poor study design
• Poorly written clinical trial documents
• Failure to maintain accurate, current, and
complete clinical trial documents
37

38. COMMON CLINICAL TRIAL
PEOPLE PROBLEMS
Poor Site
Selection
Difficult Principal
Investigator (PI)
PI and other site staff
conflicts of interest
Lack of PI and/or staff
experience
Lack of a dedicated Study
Coordinator
Competing studies
ongoing at the site
Frequent site staff turnover
Poor Study
Enrollment
Poor subject recruitment
efforts
Lack of incentive for
enrollment (e.g., small site
payments)
Overly restrictive Exclusion
Criteria
Poor site location (e.g., few
patients nearby)
Small or sparsely located
patient population
Poor Subject
Retention
Lack of incentive to stay
in the trial (e.g., subject
reimbursement)
Unaddressed safety
concerns
Demanding number or
types of site visits
Poor communication
from the site
Unnecessarily long or
complicated trial design

39. COMMON CLINICAL TRIAL
DOCUMENTATION PROBLEMS
Informed
Consent
Form (ICF)
issues
Poorly written /
confusing ICF
Unapproved ICF
versions used
Study
procedures
done prior to
signature of ICF
Protocol
Deviations
Poorly written /
confusing protocol
Failure to follow
protocol
Inclusion/exclusion
criteria violations
Improper Adverse
Event (AE) reporting
Source
Document
Errors
Poorly written /
confusing Case
Report Form (CRF)
Incorrect or missing
CRF data
Discrepancies
between CRF and
other source
documents (e.g.,
lab results)
Missing or lost
source documents
Inadequate
Investigational
Product
Accountability
Lack of Investigational
Product Accountability
records
Discrepancies
between subject CRF
and product
accountability logs
Improper
documentation of
return/destruction of
unused product

40. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Common
clinical trial
best practices
40

41. ENSURE CLEAR ROLES
• Who is the sponsor?
• Try to avoid bias when the physician is the device designer and
trial sponsor
• Use physician PIs for your study who are not involved in the
device design or funding
• Who is the monitor?
• Do not have the PIs or the site staff do the monitoring
• Use independent monitors to audit the site(s)*
• Who is the manufacturer?
• There can be only one manufacturer of record
• When the sponsor is not the manufacturer, there needs to be a
contract between the sponsor and manufacturer
• Make sure that you (the sponsor) have clinical trial insurance
*The ICH E6 Guideline for Good Clinical Practice (GCP) is an outstanding resource for understanding
the roles of the Sponsor, Investigator, monitors, and manufacturers.

42. COMMON CLINICAL TRIAL
SPONSOR BEST PRACTICES
• Follow the clinical trial regulations
• Ensure proper study subject protection
• Put the proper thought into your study design
• Write clear, concise clinical trial documents
• Have an experienced monitoring team
• Maintain proper oversight of the trial stakeholders
(e.g., CRO, site, labs)*
• Maintain proper reporting to the IRB and FDA
• Maintain accurate, current, and complete clinical
trial documents and make sure that the sites do, too
42
*Hal Mann is hosting the roundtable discussion later today about vendor partnerships

43. DEVICE TRAINING
IS ESSENTIAL
In addition to the standard protocol training that
all study staff require, device trials differ from
pharma trials in that device trials require hands-
on device training for the physician and site staff
This is not only required by IDE rules, but also
essential because the efficacy and safety of the
device may be highly dependent upon the skill
and technique of the physician and site staff
who are using the device.
43

44. SELECT A QUALITY
PRINCIPAL INVESTIGATOR (PI)
44
PI qualified by experience,
education, training, and lack
of bias
• experienced in the
therapeutic area
• knowledgeable of Good
Clinical Practice (GCP)*
• Properly trained in the use of
the investigational product
• Transparent and disclosed
potential conflicts of interest
*Unlike pharma trials where the PI completes a Form 1572, medical device trials don’t have a
specific Statement of Investigator form but the PI does need to sign a document committing them
to follow the protocol and IRB/FDA rules (see 21 CFR 812.110)

45. SELECT A QUALITY
STUDY SITE
45
Study Site qualified by experience, knowledge,
training, lack of COIs, and history
• Study Staff, especially a dedicated Study
Coordinator, experienced in the therapeutic area
• Study Staff, especially a dedicated Study
Coordinator, knowledgeable of Good Clinical
Practice (GCP)
• Properly trained in the use of the investigational
product
• Lack of potential conflicts of interest (COIs) such as
competing concurrent studies in the same
therapeutic area that can compete for time,
subjects, attention, and other resources
• Strong history of good subject enrollment/retention
and good prior regulatory compliance records

46. ENSURE
SUBJECT PROTECTIONS
46
Protect Study Subject’s Welfare
• Informed Consent with risks
clearly explained
• Proper Study Design with
proper Incl/Excl criteria
• Proper reporting to IRB and
regulatory agency
• Diligent monitoring of site
documentation, AE/SAE
reporting, and Subject
follow-up
AE = Adverse Event; SAE = Serious Adverse Event; The ICH E6 GCP Guideline gives a good explanation of
the definitions of AE, SAE, AR, SAR, SUSAR, etc. though the document is for pharma/drug events.
Know the regulations and important differences about AE and SAE reporting for medical devices

47. PREPARE SUBJECT
RECRUITMENT STRATEGIES
Site patient
databases
Physician
referrals
Community
networking
Newspaper and radio
advertisements
Internet websites
and social media
Speaking to patient
advocacy groups
Ensure that all subject recruitment messages are properly IRB approved

48. KNOW WHERE YOU ARE
LOSING SUBJECTS
Enrollment Allocation Follow up Analysis
# assessed for eligibility;
# excluded (reason);
# randomized
# allocated to intervention;
# did not receive intervention (reason);
# received intervention
# lost to follow up (reason);
# discontinued intervention (reason);
# completed intervention
# excluded from analysis (reason);
# analyzed

49. UNDERSTAND THE ROOT CAUSES
OF YOUR ADVERSE EVENTS
When an Adverse Event occurs in your trial,
you need to consider:
• Is it treatment related or not (AR versus AE)
• If treatment related (i.e., an AR), is it due to:
• User error (e.g., improper physician
technique)
• Device defect (e.g., broken part)
• Device error (e.g., improper calibration)
• Device hazard not due to error or defect
(e.g., allergic reaction)
49
AE = Adverse Event due to any cause; AR = Adverse Reaction deemed possibly due to the treatment
Reminder to know the regulations about AE/SAE reporting for medical devices to your IRB and FDA

50. ENSURE INVESTIGATIONAL
PRODUCT ACCOUNTABILITY
50
Have the site keep an accurate, complete,
and current Device Accountability Log that
records device:
• Receipt
• Date
• Quantity
• Batch or other identifiers
• Signature of receiver
• Use and disposition
• Return
• Reason (e.g., defect, repair, end of study)
• Disposal
• date, quantity, batch, signature, etc.

51. SELECT A PROPER
STUDY DESIGN
51
Design a study that is well suited for safety,
statistical relevance, and lack of bias
• Appropriate mitigation of safety risks
• Proper inclusion / exclusion criteria
• Proper AE monitoring and reporting
• Appropriate sample size considering study
phase and amount of prior safety data
• Well designed Statistical Analysis Plan
• Data handling (e.g., imputing missing data)
• Minimize Bias
• Proper primary and secondary endpoints
• Appropriate control
• Randomization
• Blinding

52. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Avoid
Common Trial
Pitfalls
Site selection
Training
Enrollment
Retention
Documentation
Randomized
Controlled Trials
(RCTs)
Pros and cons of RCTs
Trial Design
Avoid Bias
Randomization
Blinding
Embrace
Trends
(not fads)
Real World Data
Patient reported
outcomes (PROs)
Connected devices
/ biosensors
Machine learning /
data analytics

53. CAN DEVICE TRIALS USE
RANDOMIZED CONTROLLED TRIAL
(RCT) DESIGN?
RCT is the gold-standard for pharma
clinical trials, which typically are:
• Randomized Multi-arm
• Placebo- or comparator-controlled
• Double-blinded (Subject and Physician)
53

54. BUT WAIT A MINUTE…
Device trials
are unique
54

55. DEVICE TRIALS ARE DIFFERENT
THAN PHARMA TRIALS
Minor device design changes are common to
enhance reliability and ease of use:
• These often do not require clinical trial evaluation
• Bench and/or animal testing may be enough to validate the
design change prior to market launch
When a clinical trial is required, evidence usually
comes from sources other than RCT clinical
studies, such as:
• Partially controlled or single-arm clinical trials
• Published or Real World Data with an already marketed device
55

56. DEVICE TRIALS ARE DIFFERENT
THAN PHARMA TRIALS
When a device trial is needed, it tends to be:
• smaller than Pharma trials
• Single Arm without a control comparison
• Heavily dependent on the skill level of the
user and doctor’s technique
• Prone to have frequent design
modifications during the lifetime of the trial
• Difficult to blind
• Difficult to randomize

57. CAN DEVICE TRIALS USE
RANDOMIZED CONTROLLED TRIAL
(RCT) DESIGN?
It may seem like RCT need not be used if it is…
• Difficult or impractical to randomize if there are:
• published predicate data already available to allow
“substantial equivalence” comparison for a 510(k)
clearance
• no standard of care comparators
• ethical and practical issues about placebo controls
(e.g., sham surgery)
• Hard to blind if:
• Subjects cannot receive a sham control or alternate
therapy comparator
• Doctor is intimately involved in the procedure (e.g.,
invasive surgery)
57

58. CAN DEVICE TRIALS USE
RANDOMIZED CONTROLLED TRIAL
(RCT) DESIGN?
But let’s take a closer look:
Comparator control
Design types
Randomization
Blinding
58

59. COMPARATOR CONTROL
If an effective standard-of-care therapy or device exists,
it can be used as a comparator control:
• Randomization to multiple arms can be done
• Can choose Non-inferiority or Superiority design
• Demonstration of superiority can positively impact
regulatory and reimbursement decisions
If no effective standard-of-care therapy or device exists,
consider a sham control if:
• There is high potential for placebo effects
• Sham treatment is both possible and ethical

60. DESIGN TYPES
Parallel arm design:
• Separate arms for Control and Treatment(s)
• Easier to implement and analyze than cross-over designs
• Better for acute conditions or devices with long lasting effects (ex.
surgically inserted device)
Cross-over design:
• Subjects “cross-over” from one study arm to the other
• Allows intra-subject evaluation in addition to inter-subject evaluation
• May be used for chronic conditions when the device effects are not long lasting
and can be removed via a “washout” period
• ISSUE: Cross-over can lead to confounding issues if the washout is not effective at
removing carry-over effects
• ISSUE: Studies can be long, which is not ideal for devices that undergo frequent
design changes and have short lifespans

61. RANDOMIZATION
When there is a control arm,
randomization of subject
allocations to each study arm
is necessary to avoid bias

62. CHOOSE THE RANDOMIZATION
STRATEGY THAT WORKS BEST FOR
YOUR NEEDS
Several strategies for random subject allocation to control and treatment arms:
• Simple Randomization uses a single sequence of random assignments
• Can be computer generated or use a random numbers table
• ISSUE: Can lead to unequal number of subjects in each arm
• Block Randomization uses small balanced randomization blocks with
predetermined group assignments
• Prevents unequal number of subjects in study arms
• Prevents distribution differences over time
• ISSUE: Covariates and confounding traits may still be imbalanced
• Stratified Randomization uses separate randomization schedules for
subjects with and without covariate factor(s) that may negatively impact
the primary outcome analysis
• Balances covariate factor(s) between study arms
• Particularly useful in small trials
• ISSUE: Works best when all participants have been identified and
covariate factor(s) known before assignment to a study arm

63. BLINDING
Use Blinding to avoid bias; consider:
• Subject Blind: When possible, have the subject
unaware of which treatment arm they are on
• Physician Evaluator Blind: Have one physician
(unblinded) apply the treatment device while
another physician (blinded) evaluates the
safety and efficacy measures
• Lab and Site Staff Blind: If you are using a
central laboratory, have the lab staff blinded
to which samples are from control versus
treatment arms. Same for all study staff that
can be blinded (ex. site nurses)

64. MEDICAL DEVICE
CLINICAL TRIALS
Some
additional trial
considerations
64

65. CONSIDER ALL THE
SAFETY REQUIREMENTS
Devices have many requirements to show safety:
•Subject safety
•User safety
•Biocompatibility
•Electrical safety
•Toxicity
•Traceability
•Sterilization (if needed)

66. PLAN FOR
PHYSICIAN USER VARIABILITY
Account for Physician variability:
• Technique Variability: It may be very difficult to convince a
physician to change their technique if your protocol
procedure is significantly different from their experience. May
want to cluster randomization by Investigator or clinical site
based on technique differences (ex. surgical technique or
post-operative physical therapy) and take this into account in
your Statistical Analysis Plan (SAP)
• Experience Variability: You don’t want your safety and efficacy
outcomes highly dependent on the skill of the individual physician.
In addition to good training, consider:
• Mitigation Option 1: Have a run-in period prior to the study
analysis phase to allow for appropriate physician and site staff
learning curve in using the device
• Mitigation Option 2: Split the data into two subgroups: those
subjects randomized during the learning phase compared to
those randomized after the learning phase

67. CONSIDER INTERIM ANALYSIS
Incorporate Interim Analysis milestones to test:
• Safety
• Should the study be ended because of low benefit/risk
• Best to use unbiased assessors that go by many names
such as Independent Data Monitoring Committee (IDMC)
or Data Safety Monitoring Board (DSMB)
• Efficacy
• Should the study be ended because the primary objective
has been reached and it would be more ethical to move
the control group to the treatment, if possible
• Should the study be ended because the primary objective
will not be reached and the study should be ethically
ended because of futility
• Should the study be modified if part of an Adaptive Design
Trial
67

68. CONSIDER ADAPTIVE DESIGNS
• The FDA has released guidance for Adaptive Design
for Clinical Trials of Drugs and Biologics
• Adaptive Design Trials use a prospectively planned
opportunity for modification of one or more
specified aspects of the trial design based on
analysis of data
• Adaptive Designs make clinical trials more flexible
and cost effective
• EXAMPLES: Data from planned interim analysis used
to decide early stopping or sample size expansion
• IMPORTANT: The modifications must be pre-planned
and the decision points clearly specified in advance
68
*The FDA Guidance document titled Adaptive Designs for Clinical Trials of Drugs and Biologics is a
great resource for understanding Adaptive Trial Designs, which can be considered for med device, too

69. DIAGNOSTIC DEVICES NEED
SPECIAL CONSIDERATION
Unlike devices used for treatment:
Diagnostic devices are typically evaluated on indirect benefits of
clinical utility rather than direct benefit to the patient’s clinical
condition
Diagnostic devices need to be evaluated on:
• diagnostic accuracy
• sensitivity (does it detect disease when it is present; true positive)
• specificity (does it exclude disease when it is absent; true negative)
• clinical utility
• therapeutic impact
• patient outcomes
Essentially, if a subject or patient has a positive test result:
• how likely is it that they have the condition being tested
• how useful is it for the physician making a clinical decision
69

70. PARTNER WITH
PHYSICIAN AND ACADEMIC
INVESTIGATORS
Consider partnerships with the physicians and
academic researchers in your field of study:
• Academic research collaborations
• Investigator Initiated Trials (IITs)
• Human factors analysis and user feedback
• Design change ideas
• Advisory Boards
• Co-development of clinical trial design
• Participation on Steering Committees or other
committees involved in assessing clinical data
• Publications in impactful journals
• Presentations at impactful conferences

71. STUDY DESIGN CONCLUSION
• Design your clinical trial thoughtfully to best
demonstrate benefit-to-risk advantage and avoid bias
• RCT features: Control Arm, Randomization, Blinding
• Consider trials to show superiority to standard-of-care
– Remember that you need to overcome both
regulatory and reimbursement hurdles
• Partner with physicians and academics
– Design assistance and usability testing
– Clinical trial advice and/or participation
– Marketing assistance: publications, talks, webinars

72. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Avoid
Common Trial
Pitfalls
Site selection
Training
Enrollment
Retention
Documentation
Randomized
Controlled Trials
(RCTs)
Pros and cons of RCTs
Trial Design
Avoid Bias
Randomization
Blinding
Embrace
Trends
(not fads)
Real World Data
Patient reported
outcomes (PROs)
Connected devices
/ biosensors
Machine learning /
data analytics

73. REAL WORLD DATA
The FDA uses real-world data (RWD) and real world evidence
(RWE) to monitor safety and AEs for regulatory decisions
Healthcare companies use RWD and RWE to make coverage
decisions
Consider creating a platform for your device to collect and
analyze RWD/RWE that can be helpful post-launch to evaluate:
• Safety
• Efficacy
• Performance
• Reliability
• New markets and uses
• Clinical trial design
• Observational / Registry studies*
73
*Diane Covington is speaking later today about Registry Studies

74. PATIENT REPORTED
OUTCOMES
• Patient reported outcomes (PROs)* are typically defined as “any report of the status
of a patient’s health condition that comes directly from the patient, without
interpretation of the patient’s response by a clinician or anyone else”
• PROs are playing an increasingly important role in regulatory and coverage
considerations
• PROs may be gathered using questionnaires or patient diaries during a study that
capture:
• patient satisfaction
• patient preference
• Quality of life (QoL)
• Impact on the patient’s family and caregivers
• acceptance
• compliance
• Choice of which PROs to capture and measure depends on:
• stage of device development (e.g., concept, feasibility, pivotal, post-approval registry)
• quality criteria (performance, reliability, durability)
• usability criteria (convenience, ease of use)
• safety and efficacy objectives
74
*Ritu Verma spoke earlier today about PROs

75. CONNECTED DEVICES /
BIOSENSORS
• There is a tidal wave of increased use of microchip
biosensors and connected devices both in and outside of
healthcare (hello Alexa, Google, and Siri)
• Digital health, wearables, sensors and remote monitoring
capabilities are providing “Big Data” that companies can
use to optimize their clinical trials and gather data on
safety/efficacy
• Consider this trend for use in your clinical trial:
• Remote monitoring to track:
• clinical supply chain logistics (e.g., chain of custody; damage)
• Real time device performance and reliability
• Real time patient health status
• Traceability and product recall streamlining
• Collecting RWD post-launch
75
*Ashleigh Dawley spoke today about patient-centric digital health

76. MACHINE LEARNING /
DATA ANALYTICS
How do you handle the “big data” from all
your clinical trials?
• Machine learning techniques can use data to
train algorithms to identify safety signals or
other key performance indicators (e.g., MRI
results; enrollment/retention issues)
• Data analytics can be used to identify:
• Optimal patient populations to recruit from
• Confounding variables to consider in your SAP
• Early signs of safety issues in your trial data
• Areas of future clinical trial improvement
76

77. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Now we should all be on the same page
and I can end my talk. Please consider:
1. Your product is a medical device
2. Your risk class (I, II, or III)
3. Your regulatory path (510(k) or PMA)
4. Your need for clinical trial data
5. Your risk significance (NSR or SR)
6. Your clinical trial regulations
7. Your clinical trial stage (Early Feasibility, Feasibility, Pivotal)
8. Your options to incorporate lessons from pharma…

78. HOW CAN MED DEVICE
LEARN FROM PHARMA?
Avoid
Common Trial
Pitfalls
Site selection
Training
Enrollment
Retention
Documentation
Consider
Randomized
Controlled Trials
(RCTs)
Avoid Bias
Demonstrate superiority
to standard-of-care
Overcome regulatory
and reimbursement
hurdles
Embrace
Trends
(not fads)
Real World Data
Patient reported
outcomes (PROs)
Connected devices
/ biosensors
Machine learning /
data analytics