Essie Wilkie places the cherry tomato on the scale: too heavy. She picks up a knife and carefully trims a sliver from the offending orb. She weighs it again, verifies that the surgery has been successful, and drops the tinier tomato into a container of salad -- which is now exactly 68 grams.
Wilkie is not being obsessive about her diet. She is, however, being obsessive about someone else's. She works in the Johns Hopkins University ProHealth kitchen, which makes food for scientific research into the links between diet and human health.
By comparing the effects of different diets, the scientists hope to arrive at the healthiest way to eat. To ensure that results are accurate, participants must consume ruthlessly precise amounts of food -- hence Wilkie's tomato-paring.
Sometimes, even factory-sliced bread doesn't meet the ProHealth standard, says senior dietitian Phyllis McCarron, who oversees the kitchen. "It has to be 46 grams, and the manufacturer makes it 50 or 43," she complains.
A gram, by the way, is about 1 / 30th of an ounce.
Located in a two-story, red-brick building in suburban Woodlawn, the kitchen at first appears unremarkable -- until you notice the abundance of expensive precision digital scales.
Every weekday, Wilkie and nine other workers carefully weigh and assemble salads, sandwiches, snacks and entrees, each with an exact amount of calories, salt, fats and nutrients.
"We try to control all the factors," says Dr. Larry Appel, a Johns Hopkins medical professor who runs the research.
And he means all.
Those who participate in his "feeding studies" must agree that for the duration of the research, they'll subsist solely on food prepared by McCarron and her employees. That means no doughnuts, no fried chicken, no chili dogs -- not even an extra pinch of salt to add pop to a flavorless veggie burger. Subjects aren't even allowed to eat any additional healthy food because that would throw off the precise portioning.
There are a few exceptions: Participants may drink unlimited amounts of water and limited quantities of unsweetened tea and coffee, as well as diet soda and an occasional alcoholic beverage. And they can sprinkle their food with pepper, because, unlike salt, it doesn't raise the blood pressure.
Over the past decade, the kitchen has provided food for five feeding studies. The most famous is Appel's DASH study (Dietary Approaches to Stop Hypertension), which firmly established the cardiovascular benefits of eating a low-salt, low-fat diet. It is now a cornerstone in the treatment of high blood pressure.
The current trial, conducted with Brigham and Women's Hospital in Boston, is called OMNI-Heart (Optimal Macronutrient Intake and Heart Disease). The $6 million study will compare three diets for their effect on high blood pressure, cholesterol and overall cardiovascular health.
One regimen is high in "healthy" fats such as olive and fish oils; another offers lots of protein; the third emphasizes carbohydrates. But subjects don't gorge on tuna, cheeseburgers or pasta. As much as possible, ProHealth dieticians try to use vegetables and fruits in their recipes.
Researchers actually don't want participants to slim down. Losing weight -- or gaining it -- would add another variable to the experiment, which is supposed to evaluate the effect of diet alone. So every weekday at the center, subjects are weighed. If they've lost weight, the staff adds calories to their meals. Those who put on pounds get less food.
On completion, subjects receive the unprincely sum of $250. Appel says offering more might actually undermine the study. People who join to make money might be more likely to cheat, he suspects, because they wouldn't be inherently motivated to stick with the diet.
Over five months, subjects will spend five weeks on each of the three diets. The regimen can be demanding. "I never liked tomato juice. And suddenly I'm drinking three cans of tomato juice a day," says environmental engineer Ken Guttman, one of more than 160 subjects in the OMNI-Heart study.
The diet's constraints can vex researchers too. McCarron says it's not easy coming up with recipes that include the proper nutrients, vitamins and calories -- and also taste good. She throws tasting parties, using kitchen employees and former participants as guinea pigs. She and her colleagues tried 25 different versions of macaroni and cheese before hitting on an acceptable version.
The stringent requirements can lead to some strange concoctions. To meet the daily quota of olive oil, workers sometimes add it to yogurt. Otherwise, subjects would have to gulp it plain.
Most feeding-study fare, however, appears fairly normal. Spaghetti and meatballs, macaroni and cheese, salad and yogurt are all part of the menu. But the diet tends to have much less salt and saturated fat than the average American is used to.
"The food is so bland, it takes a while to get used to it," says 61-year-old Ron Jackson, another OMNI-Heart study subject.
For participants who normally dine with others, mealtimes can be a bit odd. Guttman, for instance, eats his ProHealth meal alongside his wife and two young children, who enjoy their usual dinner. His kids don't quite understand why they can't share their food with Daddy, as they did before the study. And when the family goes out for a meal, he dines beforehand and limits himself to a cool, refreshing glass of water at the restaurant.
Every weekday afternoon during the study, he drives to the ProHealth center for a supervised meal. To encourage proper consumption, subjects must eat either lunch or dinner at the site. (They also pick up their next two meals, and, on Fridays, take home weekend provisions. Everyone gets an insulated pack to carry food.)
Diners are overseen by Millie Laws, who is known as the "food police." Friendly but firm, Laws patrols the small, brightly lit dining room, making sure that all plates are perfectly clean.
"Everybody eats every single thing they're given," she says. "Every crumb. Every morsel."
This rule has no exceptions. Subjects are told to save a bit of bread so they can sop up any stray gravy or salad dressing. Those who end a meal with extra containers of margarine must scoop it out with a spoon and eat it plain. If someone drops food on the floor, even a raisin, Laws gets a replacement from the kitchen.
"She's strict," says Jackson, a retired federal worker who lives in Owings Mills.
But Laws can't oversee meals away from the center, and participants sometimes lapse. One recent evening, Jackson's wife, who is not taking part in the study, was frying up some chicken wings. The aroma overwhelmed her husband.
"I said to myself, 'I bought those chicken wings, I'm gonna eat one,' " he recalls. Afterward, his wife pointed out that this was his first slip, to which Jackson replied, "I didn't slip. I purposely had that chicken wing."
The next day, he confessed his sin to researchers, and was allowed to stay in the study. Despite the intense temptation, cheating seems to be rare, partly because the study uses a thorough screening process to weed out people who are unlikely to comply.
The researchers go to extraordinary lengths to keep their charges on track. During a recent study, a participant had to attend a funeral hundreds of miles away. So the kitchen prepared a week's worth of meals, packed them in a cooler filled with dry ice, and sent it via overnight delivery.
Science prevailed. The participant stayed on the diet.
Why would anyone submit to this rigorous discipline? Guttman, who lives in Mount Washington, joined in hopes of lowering his elevated cholesterol. There was also a certain convenience. "Having somebody make all your meals isn't too bad," he says.
Like many participants, Candace Green-Turner joined because her family has a history of cardiovascular problems. Her mother has had several strokes; her husband has high cholesterol; her father had a triple bypass; and five siblings have high blood pressure. So far, Green-Turner has been lucky, but she wants to help researchers understand these problems.
Some join for the free food. Statistical analyst Edwin Quiambao estimates that he will save about $1,000 over the course of the study (he also wants to reduce his blood pressure). He and his wife are putting themselves through a parallel sacrifice -- she's doing the Nutrisystem plan at the same time he's munching ProHealth meals.
But even the most disciplined diners deserve a break. All OMNI-Heart participants are now on a two-week holiday break, during which they can eat anything -- anything! -- they want.
Last week, as a reward, ProHealth threw a holiday party to reward its steadfast subjects. The food was far outside the bounds of normal research fare, and about 25 people feasted on baby back ribs, spicy shrimp, roast beef and biscuits.
As he waited in line for his plate, Guttman eyed the offerings eagerly. "Man, am I gonna enjoy this!" he said. "I hope there's enough shrimp left."
A Sample Menu
Here are typical menus for the OMNI-Heart study. Almost everything is low in salt, sugar and saturated fat (the kind that comes from animals). The meals contain no unhealthful trans fats.
BREAKFAST
Orange juice, with calcium
Thin white bread, toasted
Margarine made from olive oil
Banana
Milk, 2 percent or lower
Jelly
LUNCH
Cranberry juice cocktail
Chili with tofu
Baked tortilla chips
Omni bulgur wheat salad
Fig bars
Pear halves in juice, canned, not drained
Milk, 2 percent or lower
DINNER
Breaded Atlantic cod fillet, drizzled with canola oil
Lemon juice, canned (for the fish)
Spinach, fresh, steamed
Carrots, frozen, cooked without salt
Olive oil margarine
One peppermint candy
SNACK
Cranberry juice cocktail
Peanuts, dry roasted, unsalted
Raisins