Background
Family health history (FHH) has long been acknowledged as an important part of the medical examination [
1]. In the current age of genomics, the importance of FHH is becoming ever more apparent. According to Francis Collins, “Virtually every human illness has a hereditary component” [
2] and current professional guidelines for cardiovascular disease [
3], diabetes [
4], breast cancer [
5], and colorectal cancer [
6] among others strongly endorse FHH risk stratification to develop personalized prevention strategies. Despite this, collection and use of FHH for clinical decision making in primary care is underutilized.
Many barriers exist to the accurate and complete collection and application of FHH within the traditional primary care model. Patients are frequently unprepared to provide FHH, usually due to either lack of communication among family members or failure to appreciate its importance [
7,
8]. At the same time, physicians find it difficult to acquire and use FHH due to time constraints, lack of standardization, and difficulty synthesizing into actionable prevention strategies [
9‐
12].
Self-collection tools have been shown to be as good or better than the current practice of FHH collection by medical providers [
13,
14]. These factors make a patient-oriented FHH collection and risk stratification tool a compelling approach for overcoming barriers and improving patient care. In 2004 the Genomedical Connection, a consortium of Duke University, the University of North Carolina at Greensboro, and Cone Health System, developed the Genomic Medicine Model (GMM). The central component of the GMM was the creation of MeTree
©, a computerized FHH collection and decision support tool for integration into primary care clinics, a practice environment uniquely suited for widespread population impact [
8]. This manuscript describes the experiences of the providers and patients at intervention clinics who used MeTree
© as part of a Department of Defense (DoD) (grant # W81XWH-05 1-0383) funded hybrid implementation-effectiveness controlled study.
Discussion and conclusions
Using an implementation study process to integrate the GMM resulted in tremendous support from patients and providers. This was remarkable considering physicians had a number of very strong and potentially valid concerns about a negative impact on workflow and the possibility of “hijacking” patient-provider discussions. Further, physicians often felt they were already adequately capturing FHH and its implications for disease risk in their patients’ preventive care plans. Implementation of MeTree© using ongoing feedback and adaptation proved that the model could be successfully adopted within primary care, even among busy real world clinical practices. In fact these practices, as opposed to many clinical trial sites, were not early adopters who were strongly motivated to see the intervention succeed; instead they were chosen by the health system’s administration based upon their size and diversity. Clinic practices were compensated for participating, but had no pre-existing interest regarding the outcomes of the study.
While other electronic FHH and CDS tools exist, to our knowledge this is the first trial exploring direct integration of a FHH tool into real world primary care practices and the first to show that known barriers in the clinic can be successfully overcome [
33‐
39]. The finding that patients talked with their family members, acquired new knowledge about their FHH, and changed their perception of risk, awareness, and attitude towards health supports the idea that by educating patients on the importance of collecting their FHH and its impact, the model has the potential to empower patients to take more responsibility for their care and can improve the dynamic of the patient-provider relationship. A similar improvement in risk perception and awareness has also been seen in other family history studies [
33]. In addition, by providing risk stratification and actionable CDS in areas that require complex calculations and decision making with which most PCPs are not comfortable [
9,
10], MeTree
©, was able to provide a valuable and time saving resource.
We examined the patient experience taking important demographic factors into account, in particular recruitment and satisfaction among minorities and undereducated. There were no significant differences in patient satisfaction in these groups. In most analyses of patient experience, age showed a small but statistically significant difference in needing more time or assistance. Recruitment and satisfaction among minorities and the under-educated was the same as the underlying population, and though age was statistically significant, the effect sizes were not clinically significant. Survey results suggest that the positive patient experience could be attributable to the extensive education available at each step in the model: collecting FHH, entering FHH into MeTree©, and risk assessment actions. Nevertheless we can still improve aspects around collecting FHH and talking with relatives, especially since those patients who did talk with relatives were significantly more likely to feel prepared. Focusing upon expanding tools to further improve communication among family members could have a significantly positive impact on the quality of FHH provided by patients.
MeTree© may improve provider discussions on decision-making, as shown by the appropriate increase in discussions for patients at higher risk, though some MeTree© recommendations were not extensively discussed, particularly those related to tamoxifen, ovarian cancer, and thrombosis. Several possible explanations exist: failure of patients to recognize the survey item representing their discussions, providers not addressing the topic because they were uncomfortable with it, or providers’ assessment that the recommendation was inaccurate or inappropriate. The latter seems unlikely, since few providers disagreed with report recommendations. One caveat is that those who answered the patient-provider discussion questions on the survey were only a subset of the study population, 31%; however, they were statistically similar to the underlying population with the exception of age, which at a difference of 2 years is clinically negligible. The low response rate may have been due to the fact that there was no option to record that none of the topics were discussed, which may have been the case for some patients, and that it was the only question located on the reverse side of the survey making it easy to miss. Another limitation is that the reliability of patient self-report of discussions with their provider is unknown. Further research will be necessary to understand the disparity between recommendations and discussions during face-to-face time between patients and providers.
An important study limitation is that implementation study designs allow adaptation to promote GMM optimization for the current setting, thus there is no assurance of generalizability. Further study across a diversity of settings is necessary to better evaluate this. In addition, several aspects of the GMM are still under evaluation. In particular we are assessing the accuracy and quality of the FHH provided as described in several previous studies [
8,
12,
13,
40‐
43], the impact of education on FHH collection and quality, the impact of CDS on provider care plans and on patients’ primary prevention and lifestyle behaviors, and its cost-effectiveness. While study enrollment may seem low, 72% (N = 4,277) of those contacted agreed to participate. The greatest barrier was only being able to recruit one individual per clinic per one hour time slot due to kiosk access.
Further study will offer an opportunity to obtain real world outcomes data on the potential impact of MeTree© implementation on provider practice and patient behavior, both in terms of utilization of screening and genetic counseling and on lifestyle behavior of patients. Our model lays the groundwork for engaging community based practices in genomic research by outlining a model through which risk assessment and follow-up counseling and medical management occurs as a basis for implementing and evaluating a health services framework.
Competing interest
The authors declare that they have no competing interests.
Authors’ contributions
RRW participated in interpretation of data, drafted and critically revised the manuscript. LAO contributed to study design, interpretation of data, and critically revised the manuscript. TLH analyzed and assisted in interpretation of the data and assisted in drafting the manuscript. AHB contributed to concept and design of the study and critically revised the manuscript. KPP assisted in data acquisition and interpretation and critically revised the manuscript. ERH contributed to analysis and interpretation of the data and critically revised the manuscript. ABA contributed to conception and design of the study, assisted in data acquisition and critically revised the manuscript. VCH contributed to conception and design of the study and critically revised the manuscript. GSG contributed to conception and design of the study and critically revised the manuscript. All authors read and approved the final manuscript.