Scientific Integrity and Ethical Issues In Research

Barbara E. Bierer, M.D.

Senior Vice President, Research, BWH Director, Center for Faculty Development & Diversity BWH Surgery Resident Research Conference November 1, 2010 [email protected]

(617) 732-8990

• • • How many of you have participated in an IRB approved study?

Have you been the PI on clinical investigative studies?

Have you done basic research?

• • • Have you ever had an scientific integrity issue?

How often have you witnessed your study team members get informed consent? Have you ever had an authorship dispute? Authorship question?

Do you know: • • Your Research Compliance Officer?

Allison Moriarty [email protected]

(617) 732-8335 • • • Do you know your Research Integrity Officer (RIO)?

me • Do you know the PHS Compliance Hotline?

(800) 856-1983 or http://www.partners.org/complianceline

•

Do you know:

• • •

Your Research Compliance Officer Your Integrity Officer (RIO) Your Authorship Dispute Officer (ADO)

•

At HMS

 Gretchen Brodnicki, JD   [email protected]

(617) 432-2496

Your RIO

• • • • • • • • • • BIDMC BWH CHB DFCI HMS HSPH Joslin Randy Mason Barbara E. Bierer Carleen Brunelli Beverly Ginsburg Cooper Gretchen Brodnicki Bernita Anderson Dianne McCarthy McLean Peter Paskevich MGH Richard Bringhurst Spaulding Roberta Nary

Federal regulations and guidance

• We must follow Federal regulations. We can’t make them up and we don’t have a choice.

• You must (not should)   also follow the Federal regulations.

Both ORI and NIH have regulations regarding integrity Each Harvard-affiliated institution also has a research integrity policy • Specific to human subjects: if you don’t know them, read the full text of the federal regulations (OHRP/FDA/OCR) regarding the main protections for human subjects:  IRB approval, Informed consent, Privacy • • or: Ask your friendly and knowledgeable IRB administrator Your IRB administrator will become, or is, your best friend.

• Your other best friend is your RIO.

Research integrity and research ethics

• Research with human participants raises particular ethical concerns: individuals accept risks to advance scientific knowledge, often without the promise of potential direct benefit.

• The success of, and future of, clinical research depends on public trust.

• The ethical principles that guide clinical research include:    Respect for persons (informed consent, impaired consent, confidentiality, welfare) Beneficence (scientific integrity and soundness, risks = benefits, clinical equipoise) Justice (even distribution of benefit and burden, access of benefits of research)

Research Integrity

• Research misconduct • Data integrity • Conflict of interest • Authorship • Confidentiality

Concerns,

among others •

Research misconduct

(as defined by 42 CFR Part 93) 

Falsification

(making up results and reporting them) 

Fabrication

(manipulating research materials, data, or processes; or manipulating data such that the research is not accurately represented in the record) 

Plagiarism

(the appropriation of another person’s ideas, results, or words without giving that person appropriate credit) • • An honest difference of opinion or an honest error can result in the occurrence of isolated non-compliance (including GCP noncompliance). This is NOT research misconduct.

Misconduct requires deliberate or repeated noncompliance

Ignorance

Unintentional

Sloppiness Falsification/Fraud

And it’s the right thing to do.

Intentional

And other concerns

• • • • • Data integrity Conflict of interest Confidentiality Coercion Clinical Research billing • The list goes on…..

Research integrity and clinical research

• Research with human participants raises particular ethical concerns: individuals accept risks to advance scientific knowledge, often without the promise of potential direct benefit.

• • The success and future of, clinical research depends on public trust.

The ethical principles that guide clinical research include:  Respect for persons (informed consent, impaired consent, confidentiality, welfare)  Beneficence (scientific integrity and soundness, risks = benefits, clinical equipoise)  Justice (even distribution of benefit and burden, access of benefits of research) • And If you don’t know them, read the full text of the federal regulations (OHRP, FDA) regarding the two main protections for human subjects:   IRB approval Informed consent And/or make sure your IRB administrator becomes your best friend.

Some examples of misconduct

Fabrication and/or Falsification in research:

• • • • • • • • • • Using historical data instead of real time data (“her tests were fine last week…”) Rigging medical tests Inventing subjects “Fudging” the data Creating documentation for visits/tests/interactions that didn’t exist Swapping samples Alteration of dates Falsifying consent form Eliminating outlier data Underreporting significant issues (ex. not reporting SAEs) • • • • Dropping “outlier” points Making up data or modifying data in a laboratory Reporting data as “representative of three independent runs” when only one done Using Photoshop to modify appearance of, or create data.

A word about “planned” protocol deviations • • • • • The subject was due for study visit on Monday (day 8) but Monday was Labor Day. The bilirubin is 2.2 and the subject is known to have Gilbert’s Disease.

Eligibility says “<60 years old” and the subject just turned 61 last week.

The screening hematocrit is supposed to be within 3 months, and the 10 year old subject had his blood drawn just 3 months and 2 days ago.

The patient forgot to sign the informed consent, so you have him backdate it to the date he verbally agreed.

Where do you draw the line?

How many “planned protocol deviations” compromise scientific validity?

How does the sponsor review?

How does the IRB review?

There are no “planned” protocol deviations that are permissible but none of these constitute scientific misconduct

Federal regulations on research misconduct

• • • First promulgated in 1990’s – 42 CRF Parts 50 and 93 Requires institutions that receive PHS research support to have policies and procedures for reporting and responding to allegations of research misconduct Have a number of definitive provisions and definitions:  Provisions:    Confidentiality Protections Stages of investigation  Definitions:

Complainant

means a person who in good faith makes an allegation of research misconduct. (Emphasis on “good faith.”)

Respondent

means the person against whom an allegation of research misconduct is directed or the person who is the subject of a research misconduct proceeding There can be more than one respondent in any inquiry or investigation.

• • • Both complainant and respondent have rights that must be respected.

The regulations are very general. The institutional policy is not. Any investigation must be timely and fair.

Federal Regulations on Research integrity investigations

Involvement of complainant and respondent in the (PHS) process: •

Inquiry stage:

 Institution must provide respondent with opportunity to comment on report;  Not the complainant •

Investigation stage:

 Institution

must

notify the respondent whether an investigation is warranted;   Institution

may

inform the complainant Institution

must

interview each complainant Institution

must

provide respondent with copy of the draft report; Institution

may

provide complainant with a copy of the draft report.

• First step: contact the institutional research integrity officer (RIO) • Our inquiries and investigations are performed in cooperation with Harvard Medical School for all individuals with appointments at HMS.

Institutional Policy Provisions

•

The first step

: contact your friendly research integrity officer (RIO) •

Partners policy - reporting

:  All individuals “should report observed or apparent research misconduct, or where there are reasonable grounds to suspect research misconduct, to RIO.”  If individual is unsure, can call upon RIO to discuss situation informally, get consult and “will be counseled about appropriate procedures for reporting allegations.” • • • • Confidentiality maintained and respected: Don’t talk to anyone else. Fair process: Do maintain evidence if any such evidence exists.

Protection against retaliation for complainant.

Reputation, livelihood and professional life should be restored.

It can happen anywhere

40 research misconduct cases involving clinical research coordinators reported to ORI in 10 years. (and this is only a fraction of actual issues) • • • Cases generally created by:  Productivity pressures     Excessive workload Lack of oversight Inappropriate delegation Pay incentives Most cases of reported clinical research misconduct involving CRCs occurred in academic medical centers.

In two national surveys, 17-18% of investigators report having witnessed misconduct in last 10 years.

Most CRCs

DO NOT

report misconduct of PIs and others.

And anyone can do it

• Anyone in the investigation team who has access to data • In clinical research, for instance:       Principal Investigator Co- or Sub-Investigators Study Coordinators Research Nurses Students, trainees, volunteers Monitors sent by the Sponsor  And, very occasionally, the Subject (“I just wanted to get into the trial. It was my only hope.”)

Dealing with Falsification

• Prevention  Identify and eliminate risk factors • Detection  Monitor and recognize signs of falsification  Note that the system is designed to catch errors, not fraud • Correction  Investigate and report suspected falsification

Believe anyone who may come forward with complaints or suspicions

(and know that there are two sides to every story, and blame may be shifted)

So, what to do?

• • • • •

Approach Policies and Procedures A few ideas re:data Educational Directives Culture of Compliance

The single most important tool is

Source Documentation

   Even if you’re a pain in the … Even if inconvenient Even if you need support

Policies and Procedures

• • • • • • Clear Complete Readable and understandable Transparent Communicated Monitored But, some policies aren’t written.

If policies are written, some can’t be found.

If policies are written, individuals rarely (if ever) And, if read, individuals rarely remember.

read them. Updated HMS website/policy: http://hms.harvard.edu/public/coi/policy/integritypolicy.html

HMS Integrity in Science website

http://hms.harvard.edu/public/coi/policy/integritypolicy.html

Data Management

• Many institutions and laboratories establish standards for accurate collection and recording of data and for storing data.

 If your institution or department has established standards, know what they are.

 If your institution or department has not, consider establishing your own standards and expectations.

 Communicate these standards.

 Ensure consistency of application.

 Establish some system of oversight.

Data Management

• Data records:  Methodology notebook  Written  Web (Intranet/Internet) application  Privacy  Experimental notebook • Some laboratories keep a laboratory master log annotating individual research projects. • Some laboratories maintain a data selection file containing data selected for publication and documents for each publication and paper resulting from a study. • Many institutions and laboratories establish standards for accurate collection and recording of data and for storing data. There are lots of ways to manage data, and no right way to manage data.

Data Management: clinical research

• Data records:  Master notebook for each clinical trial  Master data record sheets • How to maintain “inventories” • How to utilize Partners web site • How to utilize sponsor study monitors, QA/QI program, auditors • Good clinical practices

Critical Record Keeping Practices

• • • • • • • • Common language (usu English unless discussed) An accurate reflection of process:    Ideally in ink, no skipped pages, no deletions Deleted or crossed out information initialed Not six months later: “I think I can. I think I can” Legible and understandable by others All experiments recorded, not just those that work Primary data kept in or adjacent to the notebook or log As applicable:   Mice breeding and housing, and reagent sources recorded Reagent source and location recorded, especially for items that aren’t commercially available.

Each notebook/log with front index and dates Life is long

Data Management

• Data: information, observations, experimental specimens, technologies, and products. • Data ownership: control and rights over the data, data management, and use.  Data is not (usually) owned by the PI but by the sponsoring institution (and sometimes funding agency) • Consistent, comprehensive, reliable and valid data collection and record keeping essential • Necessary and sufficient data storage protects research; access should be limited and data retention expectations known.

• Data analysis appropriate and data sharing policies known.

Data security and integrity: Three important considerations

• Physical security    Malicious hacker Unauthorized individual and someone not part of research team Part of research team but with no need to access data • Logical integrity  Maintain content of file  Prevent data from being “scrambled”  Protect against loss of data • Audit integrity  Verifiably know  Who entered data    When data were entered Who changed data When data were changed

PI Responsibility

• • • • Ensuring data validity Maintaining and reviewing notebooks/data files and, if they exist, laboratory master logs Ensuring that “notebooks” are clear, interpretable, and maintained Meeting with the research and staff associates to review progress and data at some frequency and periodicity   PIs should establish clear practices for maintaining research records Some PIs initial and date research records or reviewed data Look at the data Look at the research written record Question what you see And, if you don’t have time… Assign the responsibility to someone who does

Data Integrity and clinical research

• Note:    You can’t do everything yourself and You can’t watch everyone all the time You are dependent on your study team and collaborators • Thus:       Establish your expectations for your study team Model the behavior your wish to see Take your supervisory responsibilities seriously Ensure the environment is one in which you would wish to live Use every opportunity to educate Remember life is long (and science is small)

Authorship

Substantial contribution to: • Conception, design, data analysis or interpretation, and • Drafting or revising of the article, and • Final approval of manuscript Generally: provision of funding or of reagents not sufficient, being the big boss not sufficient, etc.

Authorship issues

• Authorship is complicated:    Criteria for authorship unclear Disagreements common No established standard for order of authors (first, middle, 7 of 11, last, *co-authors) • Among many problems with authorship are:  Guest authors (persons who have made trivial contributions to the paper)  In study: 21% of articles 1  Ghost authors (individuals who have made substantial contributons but are not listed as authors)  In study: 13% of articles 1 • The revised federal regulations include authorship disputes as research misconduct.

Copyright Law differs from Principles of Authorship 1 Flanagin A et al. JAMA 1998; 280: 22-224.

HMS Authorship Guidelines

• • • • • • Everyone who is listed as an author should have made a substantial, direct, intellectual contribution to the work. For example (in the case of a research report) they should have contributed to the conception, design, analysis and/or interpretation of data. Honorary or guest authorship is not acceptable. Acquisition of funding and provision of technical services, patients, or materials, while they may be essential to the work, are not in themselves sufficient contributions to justify authorship. Everyone who has made substantial intellectual contributions to the work should be an author. Everyone who has made other substantial contributions should be acknowledged. When research is done by teams whose members are highly specialized, individual's contributions and responsibility may be limited to specific aspects of the work. All authors should participate in writing the manuscript by reviewing drafts and approving the final version. One author should take primary responsibility for the work as a whole even if he or she does not have an in-depth understanding of every part of the work. This primary author should assure that all authors meet basic standards for authorship and should prepare a concise, written description of their contributions to the work, which has been approved by all authors. This record should remain with the sponsoring department.

http://hms.harvard.edu/public/coi/policy/authorship.html

Order of Authorship

Many different ways of determining order of authorship exist across disciplines, research groups, and countries. Examples of authorship policies include descending order of contribution, placing the person who took the lead in writing the manuscript or doing the research first and the most experienced contributor last, and alphabetical or random order. While the significance of a particular order may be understood in a given setting, order of authorship has no generally agreed upon meaning.

As a result, it is not possible to interpret from order of authorship the respective contributions of individual authors. Promotion committees, granting agencies, readers, and others who seek to understand how individual authors have contributed to the work should not read into order of authorship their own meaning, which may not be shared by the authors themselves.

• • • The authors should decide the order of authorship together.

Authors should specify in their manuscript a description of the contributions of each author and how they have assigned the order in which they are listed so that readers can interpret their roles correctly.

The primary author should prepare a concise, written description of how order of authorship was decided.

http://hms.harvard.edu/public/coi/policy/authorship.html

HMS Authorship Disputes Nada

HMS Authorship “Implementation”

     Research teams should discuss authorship issues frankly early in the course of their work together. Disputes over authorship are best settled at the local level by the authors themselves or the laboratory chief. If local efforts fail, the Faculty of Medicine can assist in resolving grievances through its Ombuds Office. Laboratories, departments, educational programs, and other organizations sponsoring scholarly work should post, and also include in their procedure manuals, both this statement and a description of their own customary ways of deciding who should be an author and the order in which they are listed. They should include authorship policies in their orientation of new members. Authorship should be a component of the research ethics course that is required for all research fellows at Harvard Medical School. These policies should be reviewed periodically because both scientific investigation and authorship practices are changing. http://hms.harvard.edu/public/coi/policy/authorship.html

Resources

Partners Human Research Committee:

(617) 424-4100 http://healthcare.partners.org/phsirb/home.htm

Office of Research Compliance (ORC):

Allison Moriarty BWH Research Compliance Officer / Director, ORC http://bwhbri.partners.org/research_compliance (617) 732-8335

BWH Institutional Official:

Barbara Bierer, M.D. Senior Vice President, Research / BWH Research Integrity Officer (“RIO”) (617) 732-8990

Partners Compliance Helpline (anonymous):

1-800-856-1983 http://www.partners.org/complianceline

Questions?

Allison Moriarty [email protected]

(617) 732-8335 Barbara E. Bierer, MD [email protected]

(617) 732-8990