"First, Do No Harm": Using Health Information Technology to Reduce Use of Preventive Services with Potential Harms

Background

The United States Preventive Services Task Force (USPSTF) issues recommendations for clinical preventive services based on critical appraisals of the evidence regarding the potential benefits and harms of the services. When there is no net benefit of the service, or the potential harm of the service outweighs the potential benefit (at the population level), the USPSTF issues a recommendation with a grade of D. This means that the USPSTF recommends that the service not be provided.

Communicating information about health care services that should not be provided – whether these are preventive services or treatment – is an essential element of reducing overuse and protecting patients from harm within the health care system. As electronic health records are implemented across the health care system, there is a clear expectation that providers and patients will have decision support available that effectively communicates “don’t do” recommendations as well as other types of health care information. However, formative research and feedback from practicing clinicians indicates that it is extremely difficult to communicate “don’t do” recommendations effectively.

The purpose of this project was to hold a working meeting; AHRQ brought together experts in clinical preventive services, health information technology, bioinformatics, health communication, and medical ethics to address the issue of how health information technology can be used to improve communication and clinical decision support regarding preventive services with potential harms. Meeting participants were asked also to identify key areas for future research. The meeting was held in March 2010 and chaired by Virginia Moyer, MD.

Results

Two critical questions were addressed by presentations and facilitated discussion:

• What have we learned from previous experience with clinical decision support – successes, failures, information needs, and links with performance measures – that can guide the development of effective clinical decision support regarding preventive services with potential harms?
• What are the critical differences in clinician and consumer decisionmaking regarding “don’t do” recommendations and “do” recommendations that must be considered in clinical decision support design and implementation in electronic health records (EHRs) and personal health records (PHRs)?

Topics for future research were identified by participants in six key areas:

1. Improve communication about the USPSTF recommendations
   • Determine whether educating about the potential harms of a service or recommending alternative approaches are effective in reducing unnecessary use.
   • Develop community-based communication interventions.
   • Evaluate the quality and level of patient–clinician discussion around health care decisions and compare the discussions with the correlating decisions.
2. Engage Health IT
   • Develop tailored communication programs tied to EHRs that help clinicians make personalized recommendations.
   • Link PHRs with information and decision-support tools that help patients and clinicians understand risk.
   • Use health IT to direct patients to good sources of reliable health information.
   • Use health IT to prompt an alternative course of action.
   • Move the decision points away from physicians and toward clinicians who may be better able to assist the patient.
   • Focus decision-support on consumers.
   • Create alerts regarding unnecessary services or add exclusion criteria to reminders.
3. Incorporate quality measurement
   • Develop a negative performance measure that tracks use of an unnecessary service.
4. Address clinician and patient behavior
   • Determine why clinicians ignore reminders and/or provide unnecessary services.
   • Evaluate the effects of individual preferences on health care decisionmaking to better understand what outcomes are most important to individuals.
   • Evaluate how clinicians interpret “don’t do” recommendations in terms of obligation (i.e., must/must not, should/should not, and may/may not).
   • Further evaluate the impact of feedback on practice variations in a manner that explains the potential harms. Research different interventions with clinicians.
   • Evaluate how target populations interpret “don’t do” recommendations.
5. Address financial incentives
   • Demonstrate how current reimbursement structures and liability affect clinicians’ recommendations and decisionmaking.
   • Evaluate whether saving time is an effective financial incentive for a clinician.
   • Evaluate the effects of financial disincentives on clinicians’ recommendations.
6. Consider other areas of research
   • Evaluate attempts to change clinical practice and behavior, such as efforts to reduce antibiotic use for viral infections, that may be applicable to the delivery of clinical preventive services.
   • Conduct root cause analyses to determine how often poor clinical outcomes can be linked to the injudicious use of screening tests.