Effect of Preventive Messages Tailored to Family History on Health Behaviors: The Family Healthware Impact Trial

Our FM residency reviewed this article for our latest journal club. A few of the limitations of the study concerned us, especially concerning the lack of diversity in the study population. Our patients are rarely at goal with healthy behaviors and this study is not able to address if tailored messages would be beneficial for this group. Also concerning is the high level of literacy and computer access required. Hopefully there is more accessible information available and more low tech ways to assess family history. Finally, we find that many patients do not know family history. Perhaps just as important is trying to emphasize healthy habits whether family history indicates a specific risk. We are excited to see more research in this area for different populations.

Competing interests:   None declared

It might have been nice to have a 'usual care' or 'no intervention' arm, as both groups received messages, though I understand the hypothesis was specifically about using messages tailored to family history - the intervention group simply had messages tailored for their family history, but both groups got messages. We know that in current practice most patients receive limited prevention messages. It was interesting that the two areas where benefit was shown were ones about which patients already understand the need (improved diet content and exercise), as opposed to those which are more likely to require physician buy-in or assistance - BP, glucose & cholesterol screening.

Competing interests:   None declared

Ruffin et al.’s study “Effect of Preventive Messages Tailored to Family History on Health Behaviors: The Family Healthware Impact Trial” is a positive step in the right direction that family history gathering and interpretation really does matter with regard to patient behaviors. We as family physicians have intuitively known this but it is nice to have some hard data.

Personally, the small benefit shown does not bother me. A family history tool such as Family Healthware will ultimately be most effective by changing generations of behavior going forward. Meaning, the risk and prevention messages will have effects on the proband but these effects will be amplified in their children and hopefully their children's children. I am not sure the best way to study that effect, but intuitively it makes sense that children will adopt the healthy behaviors of their parents just as they adopt the unhealthy ones.

Also, I think much of the art of medicine comes into play in how a physician interprets the family history information. Sure we give standard messages about exercise and nutrition, but don't most of us tailor the message based on the patient's socio-cultural background, educational level and the patient's own context with regards to illness? Again, how do you study the subtle ways in which family physicians "tweak" the message. The benefit of Family Healthware is that it starts the process of discussing and recommending risk reductive behaviors. If we can implement this tool widely, it will force physcians and those of us who teach physicians to learn the process of effectively communicating risk to patients again hopefully magnifying the amount of change seen in the present study.

Ultimately, I think this family history tool and others like it, should lead us down the path of more individualization based on risk. I would love to have the graphic representation of family history already created by the time I meet the patient, especially if we can change it over time, add new elements, and refine it within the continuity relationship. Also, I would add the genogram elements and use it as an important historical reference as well as a road map for future care. Unfortunately, my experience thus far with Electronic Health Records is that they gravitate toward deindividualization, and much in the way of family history is either omitted or lost.

Competing interests:   None declared

The study by Mack Ruffin and colleagues is a useful contribution to the emerging literature on the effectiveness of web-applications to provide individualized patient education.

It is, however, disappointing to observe the limited effect of personalized health education messages based on individual lifestyle and family history.

As noted, the generalizability of this study is markedly contrained by the self-selection of more literate participants. Furthermore, most participants were already at higher levels of targets for healthy behaviours.

I would like to add that family risk assessment is probably not a sufficient motivating factor to alter lifestyle. The contribution of genetic determination to disease expression is low for most common conditions seen in primary care. Furthermore, risk of genetic susceptibility is not precisely assessed by simple historical data about family members and the age of onset of disease in them.

Further research is needed to assess risk more precisely and develop engaging methods of online health education.

Competing interests:   None declared

This is an important paper as it is the first RCT to examine the impact of family history screening, delivered through an electronic family history tool, on patient health behaviours. Family history remains an important predictor of disease risk even in the era of genomic medicine and the promise of DNA-based disease risk assessment. Until now we have had very little robust data to know whether formally screening an adult primary care population for familial disease risk alters health behaviours including both primary and secondary preventive activities.

The trial found small but statistically significant improvements in meeting recommended targets for fruit and vegetable consumption and physical activity. Between 25 and 33 people would need to be screened for one person to change their behaviour in one of these domains. Somewhat surprisingly the intervention was associated with fewer people having their cholesterol tested in the last 5 years. The authors rightly acknowledge important limitations of the study and its findings. Perhaps of most relevance is the recruitment bias and uptake of the FHITr. Only 17.8% of people invited actually entered the study and only 15.6% used the FHITr. This population were more likely to be white, female and from a high socioeconomic level and probably represented a healthier cohort. There was a resultant ceiling effect given that many participants were already achieving recommended targets in the specific health behaviours.

The authors also discuss the uncertainty about whether the family history information was actually used in their consultation with a primary care physician. It is interesting to surmise whether the intervention might had have a larger effect if family history screening were more directly linked to a preventive health visit with a physician or other member of the primary health care team. Using an electronic tool to support family history screening does seem therefore to have a small but statistically significant effect on some disease preventive behaviours, at least in a more advantaged population. The challenge as always in health promotion is reaching those most at risk because of their lifestyles. Whether computer-based family history screening is the most efficient approach to achieve this remains uncertain.

Competing interests:   None declared