August 2017 Issue
Image-Based Dietary Assessments
By Jamie Santa Cruz
Vol. 19, No. 8, P. 40
Learn about their advantages and disadvantages and how to use them in practice.
Having accurate information about a client's diet is an important prerequisite for being able to counsel those clients on how to improve their diets. For this reason, dietitians often conduct dietary assessments in their first consultation with new clients as well as at regular intervals thereafter.
Traditionally, the most common methods of dietary assessment have been written dietary records, 24-hour oral recalls, and food frequency questionnaires. However, there are significant disadvantages to each of these methods; in particular, significant evidence suggests that clients underreport portion sizes and energy intake.1-3
"When we assess diet either for research or on an individual level, we know that we're missing a large portion of an individual's intake," says Luke Gemming, PhD, a lecturer in nutrition and dietetics at the University of Sydney in Australia.
In some cases, underreporting might be due to forgetfulness or to the difficulty of accurately estimating portions. But other factors such as social desirability bias also contribute to the inaccuracies, Gemming says. Misreporting is a particular problem among individuals who are overweight or obese. Written records and oral recalls in this population show a systematic bias toward small portions. In fact, these individuals' records often don't correlate to their body masses.4-6
Fortunately, there's a novel solution for many of the problems associated with traditional assessment methods: Dietitians can ask clients to photograph their meals.
Benefits of Images
The advantages of using photos in dietary assessments are numerous. To begin with, images provide better accuracy.7-9 Mary Cluskey, PhD, RD, an associate professor of nutrition at Oregon State University, says even clients with the best intentions frequently make mistakes in reporting their intake, such as erroneously recording they drank "100% juice" for breakfast when the beverage was actually Sunny Delight. Images leave little doubt about what foods clients are actually consuming.
In a similar vein, images can give dietitians a more accurate sense of a client's portion sizes. "Pictures don't lie," says Glen Tobias, MS, RD, CSSD, CDN, a sports dietitian in private practice in Connecticut and a long-time user of photo journaling with his clients. It's not that clients mislead intentionally, says Tobias, who's currently the team RD for the Boston Red Sox, but, in many cases, they simply interpret portions differently than a dietitian would. "You can journal that you had a small hamburger patty and some vegetables. Meanwhile, I see that it was a bigger burger … and you had one broccoli floret. [Having a photo] keeps us all on the same page," he says.
Image-based food journals also are less time-consuming than traditional written records.10 Clients frequently get overwhelmed by recording the minute details of meal after meal over a several-day period, which leads to a drop in the quality of written records over time.2,11,12 Snapping a photo, however, is relatively fast, so clients can more easily keep an accurate record over a sustained period. Assessing photos takes less time for dietitians, too. "You'd be surprised how quickly you can review meals and provide feedback using images. You can process information very quickly when you look at an image vs reading text of the dietary records," Gemming says.
Researchers in the field distinguish between two different ways to use images: image-based dietary assessment and image-assisted dietary assessment.9 According to Carol Boushey, PhD, MPH, RD, an associate professor at the University of Hawaii Cancer Center and a lead researcher in the field, an image-based assessment relies solely on images, not on any oral recalls or written records. By contrast, image-assisted assessments use images in combination with another assessment method. For instance, dietitians may ask for a traditional seven-day dietary record and also request pictures to supplement the written record, she notes. Or they could ask a client to take images of all foods consumed the day before a client meeting, then conduct an oral dietary recall in the office.
"You do the 24-hour recall, and then you have the person review the images to then say, 'Oh, I should have also remembered this or this,'" Boushey says. "That way those images can be used to assist with memory."
Although image-based assessments are a field of active research, images are unlikely to completely replace traditional assessment methods in the near future, according to Gemming. "In the near term, [photos] just provide a means to enhance traditional methods," he says.
State of the Technology
Currently, the most common type of camera used in both image-based and image-assisted assessments is a smartphone. The cameras in mobile devices are fairly sophisticated, and a high percentage of Americans carry smartphones throughout the day, meaning the camera is readily available to snap a meal photo anytime, anywhere, Cluskey says.
Cell phone cameras represent an active modality of image collection, but researchers are also working to develop passive modalities—namely, wearable cameras that take photos automatically without any action on the part of the user.13,14 The eButton, a small pin-sized device developed at the University of Pittsburgh and worn on the chest, is one such option. The SenseCam, developed by Microsoft Research and worn on a cord around the neck, is another. Similar devices also have been developed to be worn over the ear.
A significant advantage of wearable devices, according to Gemming, is that they capture information about context, revealing, for instance, that the user eats a high percentage of meals in the car. But wearable cameras aren't ready for primetime use outside of research settings. "The technology is not quite there yet—we don't have wearable cameras that can capture high-quality images frequently enough that last all day, and we don't have the processes that make it easy to analyze thousands of images per day," Gemming says.
In the same way that wearable devices are still a work in progress, so is automation of analysis. Ideally, the apps used in image-based assessments could both identify the food items in a photo and automatically calculate portion sizes.
On the one hand, certain commercially available apps, including Calorie Counter by FatSecret, are already starting to integrate image recognition. However, the current capabilities of such apps are limited. "If you capture an image of a banana, it's going to know that it's a banana," Gemming says. "But if I put some mixed nuts in my hand … it might say walnuts, but it's not going to identify that it's all mixed up. So it's not there in terms of precision."
Boushey, who is involved in the development of a mobile app for automated dietary analysis through the Technology Assisted Dietary Assessment project at Purdue University, says accurately identifying food is difficult. "We can identify faces, but faces actually aren't really that complex because we all have some consistent features—eyes, nose, mouth. Food doesn't have consistent features," she says.
Some foods (milk, for example) are straightforward, but distinguishing between tuna salad and chicken salad is significantly more difficult. At this point, image recognition systems typically have to guess the identity of a food or a dish then prompt the user to confirm whether the answer is correct. Thus, while early forays into automated analysis are important signals of future possibilities, completely automated image recognition isn't yet a reality.
Meanwhile, the issue of portion size estimation is an even thornier problem. According to Gemming, recognizing portion sizes is more difficult than determining the identity of the foods, adding that the technology to support food volume recognition is "quite far away."
Due to the limitations of automation, many studies on image-based dietary assessment still rely on analysis by a trained dietitian, who can calculate specific nutrients from images and translate the image data into exchanges. Outside of the research world, where dietitians often use images more informally, many practitioners don't even try to do a complete, formal nutritional calculation of each image. In this case, Cluskey says, the purpose of images is mostly to "assist [dietitians] in understanding patterns of eating" and help them estimate total intake—but without trying to generate detailed information about specific nutrients in the individual's diet.
Apps: The Range of Possibilities
Because there are still relatively few commercially available apps that support image-based food logs, some dietitians who currently use photos in their practice have done so in an ad-hoc style, without actually using an app designed for that specific purpose.
For example, Beth Warren, MS, RD, CDN, founder of Beth Warren Nutrition and author of Living a Real Life With Real Food: How to Get Healthy, Lose Weight, and Stay Energized — the Kosher Way, invites clients to text images to her cell phone or tag her in social media postings, a method she has been using in conjunction with written logs for the last several years.
Eliza Whetzel Savage, RD, who works in private practice at Middleburg Nutrition in New York City, likewise doesn't use an app specifically aimed at photo journaling. For the last six months, she has begun inviting clients to post images in a shared iCloud photo album in lieu of keeping a traditional written log. "They upload photos of everything they eat to the shared album and add comments to clarify what the food is," Whetzel Savage says.
While the options for apps specifically designed to support image-based food logs are limited, there are a few available. For instance, MealLogger allows users to create photo food journals and then share them socially with other users or with a nutrition professional. Alongside their photos, users can provide textual descriptions of each photo. Their dietitian can then make comments on individual meals. Although photo journaling is the app's main focus, clients also can record exercise, daily steps, and sleep activity.
Whereas MealLogger is expressly designed around allowing users to create image-based food logs, Calorie Counter is a slightly more comprehensive option. As the name implies, the app is a calorie counter and diet tracker that also allows users to log exercise sessions, record weight loss progress, and connect socially with other users in their weight-loss journey. The latest versions of the app offer a photo journaling option, including limited image recognition capacity. Like MealLogger, Calorie Counter allows users to share their data with a health professional of their choice, with the added bonus that both client and professional versions are free.
A completely different option is Healthie, the app of choice for Monica Heather Auslander, MS, RD, LD/N, founder of Essence Nutrition in Miami and team RD for the Florida Marlins. Technically, Healthie is a complete practice management tool and telehealth platform aimed at nutrition professionals offering everything from online scheduling and electronic charting to an integrated video-chat portal. However, the mobile app also allows clients to create an image-based food diary to share with their dietitian (with the option for written notes alongside the photos). Auslander particularly likes the fact that Healthie encourages users to record their emotional state with each photo.
Putting Images Into Practice
Experts say there are several key steps to using photos in a dietetics practice setting that will maximize the chances of success for both clients and nutrition professionals, including the following:
• Understand what photos can—and can't—do. Although the ability to see a client's portion sizes is one benefit, dietitians such as Warren and Whetzel Savage—both of whom have used photos in their own practices—say images aren't a perfect solution, noting that portions still are difficult to estimate. Research suggests that these two dietitians aren't alone in having difficulty: A February 2017 study of 114 dietetics students published in Nutrients found that most of the students were inaccurate in gauging portion size from images, especially with foods of high-energy density or with an amorphous shape such as ice cream.15 Not only are portions difficult to judge but images also can communicate only so much about a dish's ingredients; for instance, a photo of potato salad can't tell a dietitian whether the dish contains only one teaspoon of mayonnaise vs several tablespoons, and it won't indicate whether the mayonnaise is full fat or low fat. The strength of images, Cluskey says, is that they can open up conversation, not that they serve as an end-all solution.
Similarly, although images can improve objectivity, clients sometimes forget to take photos. In particular, Cluskey says, clients tend to miss "small foods—like if you just walk by your coworker's desk and they have candy on their desk." This limitation isn't unique to image-based assessment methods, she says, but it's helpful to be aware that photos can do only so much.
• Determine how to use images. While it's possible to use images as the basis for a detailed analysis of individual nutrients in a client's diet, many dietitians use photos simply as a way to understand clients' general eating patterns. For some dietitians, images are useful as a means to keep in touch with clients in between appointments. For example, Tobias uses oral recalls for formal dietary assessments both in his initial consultations with clients and in follow-ups, but he relies on his athlete clients' photo journals to stay in touch when they're on the road.
Auslander has a similar strategy. Like Tobias, she relies on oral recalls for formal dietary assessments during office appointments, and she uses photo journals in between appointments to keep tabs on how clients are doing. "It's a fabulous avenue to stay connected with clients and ensure they don't fall off the wagon," she explains.
• Start with younger clients. According to Auslander, while many clients believe photo logs create extra accountability and motivation, some find the idea intrusive and upsetting. Tobias and Whetzel Savage say the difference is largely generational, with photo journaling typically most appealing to younger, more tech-savvy clients. "This is what they know," Tobias says. "Pad and pen are old-fashioned for them. … It's caveman drawings."
• Give tips on photo quality. According to Boushey, an ideal photo journaling app would guide users to the proper angle for taking a photo, check image quality, and prompt users to retake dark or blurry photos. However, most commercial apps lack advanced quality controls, so clients need pointers on how to take photos that are easy to evaluate. In particular, Cluskey recommends reminding clients to take photos in good light and from a 45-degree angle above their plate.
Eye Toward the Future
Although the use of images in dietary assessment is still a relatively novel methodology, it's one dietitians should take seriously, according to Lauri Wright, PhD, RDN, LD, an assistant professor in public health at the University of South Florida and a spokesperson for the Academy of Nutrition and Dietetics. "[Image-based dietary assessment] is the wave of the future," she says. "We are in the early stages of its application to practice, but I do believe dietitians will be using this method more and more in counseling."
— Jamie Santa Cruz is a freelance writer of health and medical topics based in Parker, Colorado.
1. Poslusna K, Ruprich J, de Vries JH, Jakubikova M, van't Veer P. Misreporting of energy and micronutrient intake estimated by food records and 24 hour recalls, control and adjustment methods in practice. Br J Nutr. 2009;101(Suppl 2):S73-S85.
2. Livingstone MB, Robson PJ, Wallace JM. Issues in dietary intake assessment of children and adolescents. Br J Nutr. 2004;92(Suppl 2):S213-S222.
3. Neuhouser ML, Tinker L, Shaw PA, et al. Use of recovery biomarkers to calibrate nutrient consumption self-reports in the Women's Health Initiative. Am J Epidemiol. 2008;167(10):1247-1259.
4. Bartholome LT, Peterson RE, Raatz SK, Raymond NC. A comparison of the accuracy of self-reported intake with measured intake of a laboratory overeating episode in overweight and obese women with and without binge eating disorder. Eur J Nutr. 2013;52(1):193-202.
5. Vance VA, Woodruff SJ, McCargar LJ, Husted J, Hanning RM. Self-reported dietary energy intake of normal weight, overweight and obese adolescents. Public Health Nutr. 2009;12(2):222-227.
6. Karelis AD, Lavoie ME, Fontaine J, et al. Anthropometric, metabolic, dietary and psychosocial profiles of underreporters of energy intake: A doubly labeled water study among overweight/obese postmenopausal women — a Montreal Ottawa New Emerging Team study. Eur J Clin Nutr. 2010;64(1):68-74.
7. Gemming L, Utter J, Ni Mhurchu C. Image-assisted dietary assessment: a systematic review of the evidence. J Acad Nutr Diet. 2015;115(1):64-77.
8. Stumbo PJ. New technology in dietary assessment: a review of digital methods in improving food record accuracy. Proc Nutr Soc. 2013;72(1):70-76.
9. Boushey CJ, Spoden M, Zhu FM, Delp EJ, Kerr DA. New mobile methods for dietary assessment: review of image-assisted and image-based dietary assessment methods [published online December 12, 2016]. Proc Nutr Soc. doi: 10.1017/S0029665116002913.
10. Storey KE. A changing landscape: web-based methods for dietary assessment in adolescents. Curr Opin Clin Nutr Metab Care. 2015;18(5):437-445.
11. Gersovitz M, Madden JP, Smiciklas-Wright H. Validity of the 24-h. dietary recall and seven-day record for group comparisons. J Am Diet Assoc. 1978;73(1):48-55.
12. Rebro SM, Patterson RE, Kristal AR, Cheney CL. The effect of keeping food records on eating patterns. J Am Diet Assoc. 1998;98(10):1163-1165.
13. O'Loughlin G, Cullen SJ, McGoldrick A, et al. Using a wearable camera to increase the accuracy of dietary analysis. Am J Prev Med. 2013;44(3):297-301.
14. Gemming L, Rush E, Maddison R, et al. Wearable cameras can reduce dietary under-reporting: doubly labelled water validation of a camera-assisted 24 h recall. Br J Nutr. 2015;113(2):284-291.
15. Howes E, Boushey CJ, Kerr DA, Tomayko EJ, Cluskey M. Image-based dietary assessment ability of dietetics students and interns. Nutrients. 2017;9(2):E114.