Is It Quality Improvement or Research?

July 24, 2008

This article first appeared in the July/August 2008 issue of the newsletter "Quality Matters."

Case Study: Is It Quality Improvement or Research? The Experiences of Intermountain Healthcare and Children's Hospital Boston


By Douglas McCarthy

Summary: Case studies of two leading health care organizations show that they share a common perspective of quality improvement (QI) as integral to excellence in clinical practice and, therefore, subject to the same ethical obligation to deliver the best possible care to patients. Research on human subjects, in contrast, requires independent ethical oversight because of the potential conflicts between researchers' and patients' interests. Both organizations recognize that the decision to publish results or to use a particular analytic method alone does not distinguish QI from research. Moreover, in some cases QI can take on attributes of research, thus eliciting the need for prospective ethical oversight.

Issue: Because health care involves both benefits and risks to patients, health care professionals have long recognized an ethical duty to serve and protect their patients' interests. Research can pose additional risks to participants, both from the unproven nature of interventions being studied and from the potential that the researchers' interests may diverge from those of patients. For this reason, federal regulations require that institutions engaged in federally sponsored research commit to ethical conduct and oversight of research involving human subjects, including review by an institutional review board (IRB) when required. [1]

Health care practitioners and organizations are increasingly engaging in quality improvement (QI) activities to systematically bring routine care into conformance with evidence on best treatments and practices. [2] A recent investigation by the federal Office of Human Research Protections (which enforces the federal research regulations) of a QI initiative in Michigan hospitals has raised concerns about whether and how QI differs from research, as well as what type of ethical oversight should apply to QI activities. [3] The following case studies describe how two leading health care organizations have defined their obligations in this regard.

Case Study #1: Intermountain Healthcare

Organization: Intermountain Healthcare is a not-for-profit, integrated delivery system that provides care and coverage in Utah and southeastern Idaho. The organization employs 28,000 staff, including 700 physicians in a multispecialty group practice, and operates 21 hospitals, 140 clinics and physician offices, 42 pharmacies, and a 500,000-member health plan. Intermountain has been a pioneer in developing and implementing electronic medical records and in applying the principles of quality measurement and improvement to health care.

Topic Expert: Brent James, M.D., M.Stat., is vice president for medical research and continuing medical education and executive director of the Institute for Health Care Delivery Research, which provides technical support and education for clinical research and process management within Intermountain.

Date of Implementation: Intermountain began addressing this issue 15 years ago as part of broader discussions about information security issues related to electronic medical records (James chaired the organization's information security committee while also leading its quality improvement work). James' participation on a recent Hastings Center ethics panel on this topic added additional conceptual rigor to the organization's application of these policies to QI. [2]

Definition of QI: Intermountain considers QI to be part of "health care operations," consistent with the federal HIPAA (Health Insurance Portability and Accountability Act) Privacy Rule. Specifically, Intermountain defines QI to mean activities that:
  • focus primarily on the performance of local patient-care delivery, rather than the generation of new scientific knowledge;
  • attempt to consistently implement established best practices based on existing evidence (including randomized controlled trials, observational studies, and consensus-expert opinion); and
  • involve "open-loop systems" in which clinicians are instructed to modify implementation protocols based on patient need (so that QI does not conflict with a clinician's primary ethical commitment to a patient's well-being).
Intermountain excludes from QI (and hence considers research) any activity:
  • that involves experimental or unproven therapies (not evidence-based treatment);
  • in which patients are randomly assigned to competing treatments, potentially conflicting with a clinician's primary ethical commitment to each patient's well-being; [4]
  • that imposes additional testing burdens that represent a risk to a patient, while not conveying a countervailing potential benefit to that patient; or
  • that is funded by external grants or awards with primary or secondary goals of knowledge generation, such that those managing the endeavor have potential conflicts of interest that could place patients' interests secondary to some other goal.
Oversight of QI: Given the accumulated evidence of gaps in quality of care nationwide, Intermountain views QI as an ethical obligation to close the performance gap and deliver the best possible care to patients consistent with evidence-based standards (insofar as they apply to a particular patient). "Forty years of research has shown that the idea of a physician as a standalone expert, ethically founded, does not guarantee best care to a patient," James says.

Intermountain considers oversight for QI in the broader context of its ethical obligations for every aspect of patient interactions and care delivery. "It is not a question of whether we are going to oversee the ethics of it. . . but how we oversee it," James says. The level of oversight is determined based on the potential for conflicts of interest in comparison to those encountered in everyday clinical practice. When there is a potential that some factor other than the patient's best interest may take primacy in clinical decision-making, then a more intensive level of oversight is triggered.
  • Detection controls. All health care operations and research activities are subject to detection-oriented oversight designed to identify violations of policy. This means establishing a standard for ethical behavior (including the reporting of potential unethical behavior), training staff on the standard (such as data confidentiality), obtaining their commitment to the standard (such as through signed access and confidentiality agreements), monitoring for and investigating potential violations of the standard (such as systems for detecting inappropriate access to electronic medical records), and taking appropriate action against violators.
  • Prevention controls. Lower-volume activities with a higher risk of ethical conflict—such as clinical research activities—are subject to prospective ethical oversight designed to prevent conflicts of interest, such as review by an IRB or privacy board as required by the federal regulations.
Implementing QI so that it does not interfere with (but rather promotes) physicians' ethical duty to patients makes detection-oriented oversight an appropriate strategy for routine QI. Likewise, evaluating QI activities to learn whether they are effective is a function of organizational performance measurement. In circumstances that increase the potential for ethical conflict, QI activities are reviewed by an IRB or privacy board (as appropriate) to ensure patients' interests are protected. Such oversight is geared to the type of risk posed by the activity:
  • Risks to patient privacy: When someone decides to publish summary QI results outside the organization (analogous to the situation when a physician publishes a case-series report on his or her patients' clinical experience), ethical review (typically by a privacy board) is applied at that point in time to ensure the protection of patient confidentiality (not to approve the activity itself).
  • Risks to patient health: If patients will be randomly assigned to different QI interventions to determine which is most effective for providing evidence-based care (as distinguished from a nonrandomized pilot test comparing the experience of different organizational subunits), the project is prospectively reviewed by an IRB (typically one with expertise in QI-related research). In some cases, a project may qualify for "expedited" IRB review or for a waiver of informed consent.
Insights and Lessons Learned: Viewing QI within a framework of risk for conflict of interest is the clearest way to determine appropriate oversight, James says. This approach is more defensible than asking whether results will be published or a particular analytic method will be used. In the simplest analysis, research is trying to understand what is the best treatment, while QI is trying to implement what is known to be best treatment, he says. He acknowledges that there are some gray areas between the two, such as when randomly testing the relative merits of different QI methods ("comparative QI research" as distinct from routine QI practice). Intermountain takes a conservative approach in such cases.

Oversight through rigorous application of detection controls has proven to be an effective approach at Intermountain (although violations are regularly detected in routine clinical areas, none have been detected in QI activities). Moreover, this type of oversight is widely applied to health care delivery in the U.S.—such as by accreditation agencies and regulators including the OHRP in its oversight of IRBs. James points out that prospective ethical review is not required before obtaining patient consent for treatment, even though physicians have a financial interest in providing care. So requiring prospective review of routine QI would make little sense considering that it is designed to decrease the risk of poor quality.

For More Information: See B. James (2007) Quality-Improvement Policy at Intermountain Healthcare, in B. Jennings et al. (eds.) Health Care Quality Improvement: Ethical and Regulatory Issues, Garrison, N.Y.: The Hastings Center.

Case Study #2: Children's Hospital Boston

Organization: Children's Hospital Boston is a 397-bed comprehensive center for pediatric health care. Each year the institution admits more than 22,000 inpatients and serves more than 527,000 patients through specialized clinical programs. Children's is the primary pediatric teaching hospital of Harvard Medical School and is one of the nation's foremost pediatric research institutions, with more than 600 research scientists and $176 million in current research funding.

Topic Expert: Susan Kornetsky, M.P.H., is director of clinical research compliance at Children's, where she oversees the IRB administrative office, educates investigators on IRB regulations, and establishes policies to ensure institutional compliance with regulatory requirements.

Date of Implementation: Children's policy guidance on the distinction between research and QI activities was developed more than 10 years ago, with some recent refinements based on a recent Hastings Center report on this topic. [2] The policy was reviewed as part of Children's application for accreditation by the Association for the Accreditation of Human Research Protection Programs.

Definition of QI: Conceptually, the institution views QI as "activities based on existing knowledge about the enduring nature and function of people and their environment, rather than to develop new knowledge." QI includes projects designed to improve clinical care so that it better conforms to established or accepted standards, such as data-guided efforts to ensure the adoption of evidence based on practice guidelines or to introduce procedures to reduce medical errors. The institution is often required to conduct such activities to meet accreditation and regulatory requirements.

In addition, Children's considers surveys that have a primary purpose of gauging the opinion and perceptions of internal and external customers (e.g., trainees, staff, patients, referring physicians) to be an integral component of organizational quality assessment. This includes, for example, surveys to determine users' satisfaction with a service and to gather information on how to improve the service.

Oversight of QI: For the purposes of determining appropriate oversight, Children's distinguishes among pure QI activities that are exempt from IRB review, research that is subject to IRB review under federal regulations, and QI activities that are also research and therefore require ethical review as such. The following examples are indicative of the types of studies that may qualify for IRB review:
  • Studies in which subjects or groups of subjects may be randomized to different interventions or treatments (the IRB can consider a waiver or alteration of informed consent requirements when the intervention involves minimal risk to patients and it would be impractical to obtain individual consent).
  • Studies in which the anonymity of participants cannot be ensured (participants are defined as individuals who are being asked to provide feedback on a QI initiative, not individuals or services evaluated as part of the QI process).
  • Studies involving care practices, interventions, or treatments that are not standard (i.e., neither consensus- nor evidence-based).
  • Studies that involve more than minimal risk to patients.
External publication of findings from a QI project does not by itself trigger IRB review at Children's, provided that the project is clearly described as a QI activity and is not represented as research. (The hospital has not experienced any difficulty with journals accepting QI articles for publication under this framework.) QI projects are also subject to the organization's privacy and security policies.

Insights and Lessons Learned: Kornetsky has found that written guidance on this topic educates the institution's staff and helps them frame their projects appropriately as QI or research. "You have to be very clear up front whether you are considering a project research and apply the appropriate criteria" such as obtaining a waiver of informed consent, she says. On the other hand, classifying something as a QI activity "doesn't mean that it can be done sloppily or without attention to ethical issues." To qualify for exemption from IRB review, for example, a satisfaction survey must be completely voluntary and its results completely anonymous.

Before issuing written guidance on this topic, the hospital's IRB was becoming overwhelmed with requests to review surveys and QI projects. The hospital's staff automatically assumed these activities required IRB approval because they used research-like methods or involved the publication of results. According to Kornetsky, "Our IRB would look at these and say, 'this isn't research. This is to help the institution figure out a better way of doing things, so why are we looking at this?'"

Kornetsky finds that the institution's guidance is adequate on its own about half the time. Some projects fall into a gray area, or staff have questions that are not answered by the guidance. For example, randomization may be an indication that a QI project is "beginning to cross the line into research," but it does not definitively classify as research. Kornetsky talks these issues over with staff and encourages them to form a defensible rationale. The final determination about whether to submit the project for IRB review is made by the individual or group responsible for the activity.

For More Information: Download the hospital's Guidance on What Quality Improvement and Education/Competency Evaluation Activities Are Considered Research and Subject to Committee on Clinical Investigation Review?

Limitations: The organizations described above conduct both QI and clinical research and have established their own IRBs. Their experience may be more relevant to similar types of organizations rather than to community hospitals and other nonacademic institutions that do not conduct clinical research.

References
[1] The Federal Policy for the Protection of Human Subjects, as adopted by the Department of Health and Human Services (Code of Federal Regulations, Title 45, Chapter 46), defines research as "a systematic investigation… designed to develop or contribute to generalizable knowledge." See http://ohsr.od.nih.gov/guidelines/45cfr46.html.

[2] J. Lynn et al. (2007) The Ethics of Using Quality Improvement Methods in Health Care, Annals of Internal Medicine 146, 666–673. This report defines QI as "systematic, data-guided activities designed to bring about immediate improvements in health care delivery in particular settings."

[3] B. M. Kuehn (2008) Update: HHS Reverses Decision to Halt Quality Improvement Study, Journal of the American Medical Association 299, 1416.

[4] Intermountain distinguishes between the use of randomization for purposes of experimental research and other circumstances in which randomization is used for purposes of equity to allocate a scarce resource, or when a new treatment or service is pilot tested in a randomly selected organizational subunit to see if claims for its utility (as described in the published literature) hold up in practice.

back to top>


This study was based on publicly available information and self-reported data provided by the case study institution(s). The aim of Fund-sponsored case studies of this type is to identify institutions that have achieved results indicating high performance in a particular area, have undertaken innovations designed to reach higher performance, or exemplify attributes that can foster high performance. The studies are intended to enable other institutions to draw lessons from the studied organizations' experiences in ways that may aid their own efforts to become high performers. The Commonwealth Fund is not an accreditor of health care organizations or systems, and the inclusion of an institution in the Fund's case studies series is not an endorsement by the Fund for receipt of health care from the institution.