Environmental Scan - Health Risk Appraisals in Primary Care
Findings by Research Topic
This section is organized around the six study topics. Under each topic, a summary of the environmental scan including the key informant interviews is presented, followed by a summary of the Expert Panel meeting. Conclusions by topic follow these summaries.
Between September 2000 and September 2001, the National Study of Physician Organizations and the Management of Chronic Illness used a telephone survey to measure organizational characteristics, care management processes, and health promotion practices of 1,104 of the 1,587 U.S. medical groups and independent practice associations (IPAs) with 20 or more physicians (a 70-percent response rate). The mean respondent had 71 clinic sites and 219 physicians. Thus, the study encompasses the practices of groups containing more than 220,000 physicians (including some who are part of multiple groups)—more than 25 percent of all physicians practicing in 2000. The survey asked "Does your group routinely administer a health risk assessment (HRA) protocol or questionnaire to identify patients who may benefit from counseling or other interventions to reduce their risk factors (do not include health history questionnaires)?" Of responding groups, 22.5 percent reported that they routinely administer HAs (Halpin et al., 2005). No information on HA content or format was collected.
Logistic regression revealed that a significantly larger percentage of medical groups than IPAs reported administering HAs (26.2 percent vs. 15.0 percent), and the odds of use increased as the number of physicians in the group increased. Multispecialty practices were less likely to use HAs, but this finding was marginally significant. Controlling for these factors, the odds of routine HA use were greater in organizations with external quality incentives and information technology capabilities. Practice age and ownership were not significantly associated with routine HA use. The report cautioned that these findings are correlational and may not imply causation.
Among the large groups using HAs, 88 percent stated they "give the questionnaire results to the patient's physician," and 52 percent stated they "use the results in a formal, organized process for contacting patients who are considered to be at risk" Meeting participants were extremely skeptical of these findings, which did not match their experience.
The report also stated that little else is known about routine U.S. physician use of HAs. Consistent with that view, the present review located no survey or population-based information on HA use among smaller practices. Halpin et al. speculate that "smaller organizations have a lesser capacity to administer HAs and are less likely to use them." (Halpin et al., 2005)
Beery and Greenwald (1996) suggested that in 1995 HAs largely were designed for people aged 18 to 64. For patients over age 75 they believed additional risks needed to be probed, patients were not greatly influenced by scores showing their probable remaining lifespan and ways to extend it, and written feedback was undesirable. More recently, Rubinstein et al. (2003) identified three HAs for the elderly that had been used in physician office settings but provided no information on how extensively they have been used. They described the HAs as follows:
- HRA—Elderly. Developed by John Beck, Lester Breslow, and colleagues at UCLA. Items in the questionnaire cover... comprehensive... content... relevant to health promotion in the elderly. Reports are generated for participants and their physicians. The instrument was tested... in senior centers, in a medical practice, and in random community samples.
- Interactive Multimedia HRA. Produced by the Oregon Center for Applied Science, Eugene, Oregon. This tool employs a kiosk system intended for use in medical facilities and senior centers. Based on touch-screen responses, the system creates a report designed to encourage specific behavior change in older adults.
- YOU FIRST Senior Health Assessment. Available from Greenstone Healthcare Solutions, Kalamazoo, Michigan. Detailed reports aid in increasing the speed with which the primary care provider identifies and acts on clients requiring care and targeted interventions. It includes a 15-item "readiness to change" scale.
The literature details a few project-specific protocols for HA use by physicians (Beery & Greenwald, 1996, Scariati & Williams, 2007, Levenkron, 1988, Greenwood, Ellis, & Gross, 1991, Stuck et al., 2007). The settings described include Group Health Cooperative of Puget Sound, a free rural clinic in Southwest VA, a general medical ambulatory clinic in Rochester NY, a small private practice in Minnesota, and a large randomized trial of an HA for the elderly in London, Hamburg, and Solothurn, Switzerland. Greenwood et al. (1991) found no comprehension differences between mail and nurse feedback; in contrast, Scariati & Williams (2007) found personal contact was superior. Harari et al. (2008) found minimal changes in health behavior or preventive services uptake in the 2,006-person London arm of the multi-national trial, where HAs were integrated into the electronic medical record, generating physician reminders as well as written feedback to participants.
An Austrian study (Eichler et al., 2007) assessed how often physicians who administered HAs to patients over age 70 followed up on risks identified. The likelihood of follow-up was 100 percent for hearing impairment, 93 percent for mobility/falls, 83 percent for depression, 65 percent for urinary incontinence, and only 18 percent for cognitive impairment.
Participants discussed the scope and use of HAs. They agreed successful use of HAs in primary care requires linkage within the physician practice and between the practice and preventive service providers so that relevant information reaches the physician, is included in the patient's medical record, and stimulates prompt and appropriate follow-up. Efficient linkage to appropriate preventive follow-up care is a major challenge. Developing and promoting HA standards and protocols for follow-up referral could be very helpful.
Nevertheless, the challenges in implementing this approach can be formidable. A panelist reported that the Health Partners HRA generates a one-page clinic report for the patient to give the clinician as well as an individual feedback report. The clinic report also goes into the EMR, but the primary care provider has no reason to pick it up, has no prompt, is not trained for it, has no system to support it, and gets no reimbursement for reviewing and acting upon it. So the HA is disconnected from a systems perspective. Another panelist stated that physician discomfort with the unprecedented inclusion of patient-entered data in the EMR accentuates the problem.
Several potentially effective models for integrating HAs into primary care were discussed, including work place, primary care office, and independent/stand alone administration. HAs have been effectively used for more than 30 years in work sites and participation rates are rising. At many large companies, the benefit plan design and premium are linked to the participation rate, and 90 percent participation has become common. Linkage of HAs administered by employers or their health plans into primary care are discussed under topic 3. Stand-alone HA interventions such as Revolution Health and Health Vault have an independent method of collecting data and transferring it to physicians. Finally, in the clinician's office waiting room, patients often have time to complete health histories and assessments including HAs that can be integrated into the primary care visit.
Physicians lack both the time and training to provide the preventive follow-up that may be identified when assessing results from an HA. A community of resources and specialists that require referrals are often required. One method, therefore, would be PCP referral to community providers with a brief physician follow-up 3-4 months later. Another effective delivery system may be contracting out HA counseling, referral, and follow-up to assure compliance.
Participants pointed out that many physicians have little experience changing health risk behaviors (e.g., through smoking cessation, weight loss, stress management) and that they should focus on a few key areas that they can address. One suggestion was to limit HA questions to those behaviors that can result in death and which have valid measurements; i.e., smoking, alcohol abuse, poor nutrition, and lack of physical activity.
Halpin et al. (2005) found that 22.5 percent of large medical groups and IPAs routinely used HAs in 2000, with findings typically reported to the primary care physician. The panel was skeptical of these findings and suggested linkage and possibly even usage was more likely to be sporadic than routine. A more probing repeat survey and a survey of a sample of smaller practices seem desirable. These surveys also should investigate the content, format, and scoring of HAs used in primary care and the guidance/training that primary care providers receive about interpreting and acting on the HA information they are provided.
The review did not find any publicly available practice guidelines or standards of care for primary care physicians to follow regarding the administration of HAs, ways to process or evaluate HA data, or follow-up steps to take with patients who have completed an HA. Panel participants agreed that physicians have little experience with high risk behavior change and recommended developing standards for HAs and their use in primary care settings. Halpin et al. (2005) also suggest that facilitating information technology (IT) use related to HAs might accelerate use in primary care.
Other than one Austrian study (Eichler et al., 2007), minimal information was found on the care provided as a result of HA responses. Replicating that study at large U.S. IPAs or medical groups that use HAs seems desirable and may be completed with claims data.
Topic 2: How HAs might be used more effectively to improve the delivery of preventive services in primary care settings
A recent review concluded "Interventions that combine HRA feedback with the provision of health promotion programs are the interventions most likely to show beneficial effects. HRA questionnaires must be coupled with follow-up interventions (e.g., information, support and referrals) to be effective. The HRA questionnaire alone or with one-time feedback is not an effective health promotion strategy" (Rubinstein et al., 2003). The interview respondents echoed these conclusions, stating that HAs alone do not produce many beneficial results. With appropriate follow-up, they believed HAs would be cost-effective in primary care settings and could offer a sound return on investment, reduced medical costs, and improved productivity.
Pronk (2005) addresses how physicians should approach treating modifiable risk factors, and suggests using the 5 As, a method that has been used to address tobacco use and smoking in the primary care setting. With the 5 As method, providers assess, advise, agree, assist, and arrange for follow-up. The primary care physician plays a role in collecting risk assessment data and providing goal-oriented counseling and follow-up.
For follow-up, it is suggested that clinic personnel, including nursing staff and physicians, stay connected with the behavior change staff using interactive technology such as a secure Web site, telephone, and fax. With consent, Web sites can give physicians and their nursing teams access to data on individual patients. At the time of the next visit, the patient's records can be reviewed and advice can be tailored to progress in the treatment program. The Health Partners Health Behavior Group used this approach in their weight control program (Pronk, 2005).
Most physicians need to be convinced to use HAs in the clinical care setting (B. Bagley). Many family practice doctors, for example, assume that they already are catching most risk factors in their practices (e.g., blood pressure, sugar level, cholesterol, and smoking) and need to be shown the gaps in what they are collecting. To persuade clinicians to use HAs more widely in routine practice, convincing research needs to show that critical items are not being addressed (B. Bagley).
One interviewee aptly summed up HA potential (R. Harmon). It is likely that HAs can strongly influence the delivery of health care services, especially if they become part of the electronic health record (EHR), but this does not happen often. The HA combined with the health record would be shared with the patient and the clinician. An important focus would be on preventive services, which could affect outcomes in an integrated system. Outcomes also might improve with an annual assessment of the status of risk factors, progress made, and clinical services needed. Integrating HAs/IT with the process of care (such as kiosk computers in clinics) could help improve patient-provider communication, services coordination, and quality of care. Rubinstein et al. (2003) provide a more detailed vision of the potential roles of technology in HA administration and follow-up.
Strengths of Using HAs in Primary Care
According to many key informants, use of HAs with supporting systems in place can produce good outcomes. Specific strengths of HAs used in primary care include:
- Ideal HAs are systematic, standardized, and user friendly. They engage patients at the point of entry and can be administered to large numbers of healthy people.
- HAs are preventive screening tools. They provide a method of identifying health problems and risks, especially psychosocial ones.
- HAs are oriented around behavior change and capture information most medically oriented history forms do not. HAs can link health history with outcomes as they process information, and capture health outcomes. They are developed to be actionable and result in follow-up.
- HAs can facilitate the relationship between patient and provider, empowering and preparing patients. They can connect the right people with the right services and create a "teachable moment" with the patient.
- HAs provide meaningful feedback to patients and help motivate them to change behavior. At the same time, they allow clinicians to target interventions and programs to be more cost effective.
- HAs are predictive of outcomes and health status. It is hard for clinicians to focus on prevention without data; with HAs, they can see the progress they are making.
Additionally, the better HA tools allow clinicians to prioritize several interventions (J. Mold). Otherwise, physicians follow disease-oriented guidelines, which are often too long lists of actions. HAs can make the experience more person-centered, with goal-oriented recommendations by converting long lists of possible interventions and tailoring them to the person.
Increasing Response Rates
Response rates for HAs vary. In the workforce setting, meaningful incentives are a key participation driver, experts say. "If we offer a health risk appraisal with no incentive, even if it's communicated well, on average we'll see only 15 to 20 percent participation," according to David Anderson of Staywell. "Offering even a small incentive raises that rate to more than 50 percent. Some companies get virtually 100 percent participation" (Aase, 2006).
Response rates in primary care settings also generally increase with the use of incentives (R. Goetzel, others). In general, the better the incentive, the better the participation (R. Harmon, N. Pronk). Examples of patient incentives to complete HAs are co-pay reduction or waiver; reduced deductible; completing the HA as a condition of coverage; paying $100 for participation; a $25 gift card (a weaker incentive); or T-shirts or trinkets (also weaker incentives).
More provider-oriented approaches to increasing response rates include a request from the primary care physician to the patient to complete the HA; organizational support; improved communication within the setting; and better understanding of the clinical setting, such as promoting programs that support the bottom line (D. Grossman). Physician requests resulted in the highest completion rate at Group Health Cooperative.
Respondents also indicated response rates are sensitive to mode of administration. HAs may be administered through personal interview, telephone interview, paper-and-pencil tools (on-site completion/submission), or online. Computerized forms offer several advantages. They are more anonymous, and thus people often are willing to disclose more. The skip patterns, which can be quite complex and confusing on paper, are automatically calculated for the respondent. Disadvantages of online administration are that some patients lack access to computers and some lack computer skills and need assistance to use an automated form. Some medical offices and clinics have set up kiosks with computers so patients can complete HAs before their appointments. Multiple respondents stated that health care or administrative staff members may be available to assist with this process.
Doctors Are Busy and May Lack the Time and Resources to Use HAs
Pronk (2005) states that busy clinic schedules, limited time for a physician to spend with the patient, the issue of financial reimbursement, and other factors result in missed opportunities to talk to patients about taking action on their modifiable adverse health risks. Kottke, Edwards, & Hagen (1999) consider it unlikely that physicians will ever have more time to devote to counseling for health behaviors due to the high cost of physician time and the time pressures under which they do their work.
Pronk (2005) discusses the significance of the physician's role in the primary care setting and how the physician can set the tone for a fluid transfer of patient information (intake, measurements, HAs). In a typical, busy clinic schedule, the physician has approximately one minute available to address prevention-related topics; thus, a referral to another member of the extended care team should be made to do the 5-A "assist" and "arrange" for follow-up.
Other Barriers to Using HAs in Primary Care
HAs have not been used a great deal in primary care settings for varied reasons, listed below. Some barriers were cited by one interviewee, as noted; the remainder were cited by many or nearly all respondents.
- Physicians commonly use a medically oriented history form, which does not focus on health behavior changes (P. Renner). It can be difficult to incorporate traditional HAs into clinical care; e.g., most insurance companies do not pay for a wellness visit (Z. Nagykaldi).
- Few links exist between health promoters and primary providers. The HA is too much of a stand-alone tool; it is not used enough as an integrated tool with the health care system (R. Soler). Effective HA use requires integrating the HA with follow-up programs, but few system have a feedback loop. The primary care setting lacks a mechanism to show that a particular problem has been addressed.
- Physicians do not always know what to do with HA results (P. Renner). Many physicians are not trained in health behavior change or do not often prescribe health education and health promotion (B. Bagley). Although clinicians need to be able to respond to HA output, HAs typically are not designed to guide their response.
- The ultimate goal is to motivate the patient, but changing health behavior is very difficult to accomplish (P. Renner, Z. Nagykaldi). Patient readiness to change, thus, is a key factor in the success of the HA process.
- HAs lack standardization—the tools vary widely in length (R. Goetzel) and in topical coverage. Patient burden becomes an issue with the longer HAs and the need for duplication when sharing information. Electronic forms can be costly and complex tools.
- HAs provide a self-report restricted to one moment in time. Repeat HAs to track outcomes generally have low response rates.
Participants discussed what might enhance HA effectiveness. They debated if HA questions should focus only on areas with evidence-based interventions. Some Panel members said not to make the HA intervention-based because all risk factors are modifiable, and it is important to ask a broad set of questions to understand what is possible. HAs lack enough clinical detail to inform timely primary care action, especially if staff and other resources in the care setting is limited. Also, current interventions may not be very effective, but that does not mean effective ones could not be developed in the future. One panelist therefore advocated increasing the range of topics that HAs cover and the depth of coverage. Others thought, instead, that with limited time, the need is to focus on interventions that are evidence-based and can be addressed effectively now. One said if the HA is limited to risks with proven interventions, it would be possible to design a 10-20 question tool, a very practical length. In primary care settings, this individual recommended a narrow HA focus, while HAs in work place or insurer settings could be much broader.
One study showed that HA questions about significant health impacts worked well. Conversely, some participants thought that patient-specific performance measures were premature, that "we're not there yet."
The lack of physician time for HAs, pointed out in the scan, also was discussed. Many primary care physicians never look at HAs, even if they have easy access to them. Providers often are uncomfortable addressing certain risk factors when they are not trained for it and have no supports for follow-up. The clinician needs to be involved in deciding appropriate parameters, and guidelines should be developed and incorporated into HAs or electronic tools, or both. The panel reiterated the value of the "5A's" model which recognizes that everything does not have to happen in the primary care setting.
The critical challenges in tying HAs into primary care effectively relate to linking information identified in the primary care setting to follow-up services. Physicians often lack both the time and the training needed to coordinate effective preventive health services in follow-up to HA findings. Standards and protocols might help to address this issue. In large practices, specialist staff might be hired or trained to handle preventive service coordination and delivery. This is a larger challenge for smaller practices.
Another question is the appropriate form and content for HAs used in primary care. Some feel that the ideal HA for primary care would focus on a small set of risky behaviors and suggest to the physician a prioritized report on what preventive care needs exist. A broader behavioral focus and outputs that include mortality or disease risk scores and related patient feedback, however, is the model with proven value in other settings. A meta-analysis on the effectiveness of risk communication in primary care (Edwards et al. 2000), while not focused on HAs, is relevant here. In general, it found risk communication by physicians changed patient decisions (treatment choice, screening uptake, or risk behavior) more when based on individual than population risk data.
Strategies to improve the delivery and coordination of health care services include HA use by providers based in work site clinics, the team approach to HA interpretation and follow-up at the primary care provider's office, systems redesign, addition of reimbursable CPT codes for HA administration and follow-up, and developing clear guidelines for the clinician and personalized wellness plans for the patient.