The public health need for patient-reported measures and health behaviors in electronic health records: a policy statement of the Society of Behavioral Medicine

WIDESPREAD USE OF EHRS IS WELCOME—BUT THERE IS A PROBLEM

The HITECH Act and the Patient Protection Affordable Care Act place new emphasis on the widespread and meaningful use of electronic health records (EHRs). While we welcome this emphasis and its likely attendant advances in patient care and clinical health research, there is a significant problem: Currently, EHRs fail to capture data reflecting crucial social and behavioral determinants of health.

EHRS MUST CAPTURE BEHAVIORAL AND SOCIAL DATA

EHRs now capture such measures as weight, blood pressure, and health history, and results from tests and procedures. These measures help healthcare providers make informed diagnoses and treatment decisions, which, in turn, can produce desirable health outcomes. That being said, studies show socioeconomic status, anxiety, depression, and such health habits as tobacco and alcohol use, diet, and physical activity, often play an equal or greater role in determining health outcomes. Cancer, heart disease, obesity, HIV, and diabetes are among the serious chronic diseases inextricably linked to behavior. EHRs represent a valuable opportunity to capture standardized, patient-reported behavioral and social data and outcomes from hundreds of millions of patient encounters.

EHRS REPRESENT THE BEST OPPORTUNITY FOR CAPTURING BEHAVIORAL DATA

Decision makers at all levels—from health care providers to local, state, and federal health administrators—need sufficient data. But the health care quality data reported to the National Committee on Quality Assurance (NCQA) and recorded in the Healthcare Effectiveness and Information Set (HEDIS) include few data on personal characteristics, mental health, social environment, or behavioral issues—all of which have profound impact on healthcare and health outcomes. Further, the current HEDIS measures primarily emphasize process of care, but do not capture information on outcomes from the patient perspective. There is no standard way to compare performance of different healthcare organizations on behavioral or psychosocial issues, nor is there a way to reliably estimate future healthcare demands that depend on broader determinants of health.

Most important, three critical national efforts will be hindered without the inclusion of patient reported factors

The patient-centered medical home and patient-centered comparative outcomes research clearly cannot achieve their full promise without the inclusion of the patient perspective and PROs. Unless changes are made, personalized or precision medicine will also be operating without the most important tailoring tool–patient preferences and perspective.

SBM PROPOSES A STANDARDIZED, PRACTICAL TOOLKIT OF MEASURES FOR EHR INCLUSION

We propose the development of a standardized, practical toolkit for measuring behavioral and psychosocial patient report variables to be routinely included and confidentially protected in the EHR. These measures should also be included as part of quality and pay for performance measures for preventive and chronic illness care, such as HEDIS, and the primary care medical home. Tools for the kit will be chosen because they are practical, actionable, and feasible, and will build on important existing efforts (e.g., PROMIS, NQF, and NCQA) that capture some, but not most of the proposed domains. Selection criteria include reliability, validity, sensitivity to change, feasibility, importance to clinicians and to public health, and user-friendliness. Unlike 10 years ago, such practical measures do exist but are not being widely used.

A harmonized set of such measures should include: (1) well-established health behavioral risk factors (smoking/tobacco use, physical activity, eating patterns, risky drinking, and medication taking); (2) socioeconomic determinants (education, age, zip code); (3) psychosocial problems (anxiety, distress, depression, stress; (4) patient-reported outcomes (functional status, health-related quality of life);and (5) patient goals and preferences for care and communication. These domains and examples of practical measures within each of these categories are available and listed in Table 1.

Table 1.

Recommended Domains and Example Measures^*

Domain	Example measure(s)
Health behaviors
Smoking/tobacco use	SRNT items; one Fagerstrom item for smokers
Physical activity	BRFSS, IPAQ or pedometer readings
Eating patterns	Starting the conversation or NCI fat and fruit/vegetable screeners
Risky drinking	2 items from AUDIT or BRFSS
Medication taking	Hill-bone adherence scale
Optional items	Customized to site priorities—e.g., salt intake, sleep patterns
Psychosocial and patient/environmental characteristics
Depression/anxiety	PHQ 2 or 4
Quality of life	PROMIS questions
Stress/distress	Distress scale or distress thermometer
Health literacy/numeracy	Chin and Fagerlin health literacy and numeracy items
Patient goal(s)	Free text on specific measurable goal and goal attainment
Demographic characteristics	Race, ethnicity, zip code for GIS coding
Optional characteristics	Customized to setting: patient priorities and preferences (e.g., preferred level of participation in medical decision making; mode of contact-e-mail vs. phone)
Issue patient most wants to discuss during next contact:

^*Note: We propose these specific measures to make clear that validated, practical measures to effect these recommendations exist. Final choice of measures will be determined by future review and rapid consensus methods.

While opinions may differ on the specific measures to be included, such differences can be bridged through modern online techniques for achieving such data harmonization. With today's electronic tools, the increasing prevalence of patient portals (PHRs), automated telephone calls, waiting room data collection, and cell phone technologies, PROs can indeed be routinely collected. Carefully selected PROs are feasible and are more valid, reliable, less expensive, and time-consuming than are office measurements of blood pressure and weight—which no one would think of excluding from the EHR.

THE TIME IS NOW

As noted in a recent Institute of Medicine report, “...the United States lacks both a cohesive national strategy and the appropriate measurement tools to track and respond to the social and environmental factors that affect health outcomes.” The time is now to standardize practical behavioral and psychosocial measures to be included as HEDIS indices, and to have them built into confidential data capture for EHRs. Doing so should improve human health and healthcare, reduce suffering, and develop better information for addressing significant gaps in care and population health management. The consequences of failing to develop standardized patient report data elements are lost opportunities to enhance patient care and understanding of population health.