Assessing Dietary Intake in the Management of Obesity
Abstract
O'NEIL, PATRICK M. Assessing dietary intake in the management of obesity. Obes Res. 2001;9:361S–366S.
This paper focuses on assessing the caloric-intake side of the energy balance equation in clinical settings. In the treatment of obesity, dietary assessment may have many purposes including the following: 1) establishing a baseline of eating patterns to determine targets of intervention and to gauge progress, 2) providing a means of monitoring change in the targeted dietary areas and behaviors, and 3) allowing for ongoing feedback to the patient. The types of data to be gathered in dietary assessment will depend on the purposes of the assessment. The nature, advantages, and disadvantages of the following dietary assessment methods are reviewed: 24-hour recall, diet history interview and questionnaire, and self-monitoring. When used on an ongoing basis in treatment, self-monitoring enhances weight-loss outcomes. However, compliance with self-monitoring varies widely across patients and over time. Possible methods of increasing compliance are discussed. Recent technological advances in software and hardware systems offer promise in improving compliance and effectiveness of self-monitoring.
Introduction
At its core, the goal of weight loss is to achieve a hypocaloric state, i.e., a state of caloric imbalance in which the calories consumed are fewer than the calories expended. Similarly, maintenance of body weight requires maintenance of a state of caloric balance. Efforts to help patients with these goals require that both the clinician and patient acquire information about both sides of the energy balance equation. This paper focuses on assessing the caloric-intake side of the equation in clinical settings. The intent is to examine assessment of caloric intake from a clinical perspective and not to provide a comprehensive and analytical review of the research literature.
Purposes of Dietary Assessment in Obesity Treatment
In the treatment of obesity, dietary assessment is useful for a number of reasons. Initial assessment establishes a baseline of eating patterns used to determine the most important targets of intervention and against which to gauge progress. Ongoing assessment of intake provides a means of monitoring change in the targeted dietary areas and behaviors. Furthermore, such assessments allow for ongoing feedback to the patient and can be a basis for problem-solving efforts to overcome challenges.
The process and results of dietary assessment permit an individual approach to dietary guidance that can avoid the “one size fits all” pitfall of many weight-loss diets. In this regard, whereas the above purposes primarily apply to the clinician, dietary assessment in weight-loss treatment should always have another important goal, that of increasing the patient's awareness of what he or she is actually consuming.
Dietary Assessment Information
Dietary assessment can provide different types of information that may be useful in different situations. For weight management, information about the total energy intake is vital. It may also be useful to know the macronutrient makeup of the diet to evaluate its overall healthfulness and to evaluate the extent to which the subject is following the nutrient distribution of the recommended food plan (e.g., low-fat or low-carbohydrate diet). When the nutritional adequacy of the diet is in question, it may be useful to know the micronutrient intake.
Dietary assessment can also provide information about the qualitative aspects of the patient's eating patterns. Determining which foods are most preferred by a patient can help in identifying intervention targets. Additional information about the patient's eating topography might include the temporal pattern of eating (e.g., whether calories are concentrated in the morning or evening), the distribution of calories (e.g., few large meals or frequent smaller feedings), and the speed of eating. Another aspect of eating topography is the location in which eating occurs. This might be viewed as one type of stimulus for overeating. Assessment may identify other types of stimuli for eating such as time of day, mood states, presence of other people, concurrent activities, etc. A final factor that can be revealed by dietary assessment is the degree of patient comprehension of dietary information: what foods are in which categories, what foods in what quantities are recommended in the weight-loss plan, etc.
Methods of Assessing Dietary Intake
There are many techniques for measuring dietary intake, including 24-hour recall interviews, diet history interviews and questionnaires, and self-monitoring. More detailed descriptions of the validity and reliability characteristics of these approaches and their use in research settings may be found elsewhere ((1-4)). This discussion will focus on the clinical applicability of these procedures with an emphasis on self-monitoring.
24-Hour Recall
This method consists of a guided interview by a trained interviewer, in which the patient is asked to describe all food and drink consumed during the previous 24 hours ((2)). The interviewer uses prompts and probes to attempt to trigger recall of all items consumed. Food models and photos, and careful questioning may be used to attempt to obtain accurate estimates of portion sizes. Occasionally, recalls may use a longer time period (e.g., 48 or 72 hours).
The 24-hour recall method has advantages of economy and expediency. Whereas a trained interviewer is required, the time necessary for a single interview is not excessive (under 0.5 hour). Little initiative, preparation, or self-direction is required of the patient, who must only answer questions by the interviewer. One issue with assessment of any behavioral variable, including food intake, is the reactivity of the assessment method, that is, the extent to which the assessment procedure influences the behaviors being assessed. If the patient does not know a recall interview is impending during the time interval assessed, the method is not reactive, i.e., it will not influence the intake being measured.
The main concerns regarding the 24-hour recall center on its validity. One concern is whether the 24-hour period studied is representative of a patient's normal intake. If this were an atypical day (e.g., holiday, infrequent travel day), the results will not reflect the patient's usual intake patterns, regardless of how accurately the patient reports the intake that occurred that day. Relying on memory and report, the accuracy of the intake data obtained in the 24-hour recall can be diminished by several patient and setting factors. Individuals’ memories of a previous day's intake can be imperfect, with the imperfections typically involving errors of omission that lead to underreporting. This may be enhanced by the demand characteristics of the situation, that is, the patients’ perceptions of the setting requirements that influence responses. Social desirability factors may discourage candor regarding foods the patient views as “bad” or undesirable. There is evidence that some obese people in particular may be less likely to disclose the full amount of their intake ((5)), a factor in the validity of all types of assessment relying on patient report.
A trained interviewer is required for this method, and must be scheduled with the patient. (If conducted over the phone, the opportunity to use food models and photos is lost, as is the interviewer's ability to detect nonverbal signals to guide the interview.) The experience and skill of the interviewer in setting a nonjudgmental tone and probing for clarification and additional information are important determinants of the method's validity. Finally, the method assumes that the patient is cognitively capable of recalling everything that was consumed the previous day.
Diet History Interview and Food Frequency Questionnaires
An alternative approach to the 24-hour recall is to ask for information about usual patterns of intake over a much longer period of time. This can be done either in person by a diet history interview or by paper and pencil self-report using a food frequency questionnaire ((1),(3)). In a diet history interview, information is recorded about the patient's customary intake patterns by a trained interviewer often using a standardized interview guide. Food frequency questionnaires present lists of specific foods in logical groups and ask the respondent to indicate how often each food has typically been consumed over the indicated time period (of weeks or months). Some instruments also elicit information about the amount of each item consumed.
Both of these procedures share the advantage of covering a rather long period of time, enhancing the representativeness of the sample period. Although the diet history interview entails considerable interviewer time, the food frequency questionnaire is very economical in that it is self-administered and may he computer-scored. However, staff participation may be necessary to advise about portion sizes for quantitative instruments and may be useful as a follow-up to enhance validity. These retrospective approaches are not reactive.
| Foods, quantities, and methods of preparation |
|---|
| Time and duration of consumption |
| Location of eating |
| Social, emotional antecedents, and concomitants |
| Hunger and satiety ratings |
| Use of behavioral strategies |
| Exercise/activity |
The primary limitations of these approaches relate to their validity. The subject is required not only to remember but also to condense that recollection into accurate frequency estimates. Particularly with the questionnaires, the subject is limited to reporting on the items that are presented, meaning that the procedure will miss less common (for the population) foods that may constitute significant portions of the individual's intake. These methods assume consistency of intake over the covered time period, whereas many individuals may demonstrate considerable changes in eating patterns over time. Unless quantities are elicited, absolute energy intake cannot be obtained with the questionnaires, a serious failing when used for weight management. Other factors affecting total energy may be missed by the questionnaires, such as food preparation methods and nonincluded food items. These retrospective reports are also susceptible to the influence of social desirability effects.
Self-Monitoring
Self-monitoring refers to the systematic observation and recording of one or more aspects of one's behavior ((6)). When applied to dietary assessment, the patient is asked to record all (caloric) food and drink intake over a specified period of time, often with forms that are provided ((1),(4)). In contrast to the methods discussed previously, self-monitoring (when done properly) is a concurrent, rather than retrospective, method of data gathering. Patients are instructed to record food and beverages at the time of their consumption if at all possible. If not, they are encouraged to record these later the same day. In practice, this method of assessing dietary intake for weight management is referred to by such terms as “food records,” “eating diaries,” and “food diaries.” When used clinically, self-monitoring permits assessment of a far greater variety of factors than is suggested by these informal names (Table ). Targets of self-monitoring most directly related to nutrition include types and amounts of foods and drinks consumed and methods of preparation. In addition, factors potentially relevant to increasing control of eating can be identified, such as the time, duration and place of consumption, and the social, emotional, and physical conditions in existence at the outset of an eating episode.
Self-monitoring can be used both to determine baseline status and to track intake patterns during and after treatment. In either instance, a decision must be made about the duration of the monitoring period. One-day records are of questionable validity, because of issues of representativeness and reactivity. Often, 3-, 4-, or 7-day periods are used. In the case of periods less than a week, the patient should be asked to include one weekend day. Longer (e.g., 7 days) periods are generally preferable, but compliance may be easier to obtain with the shorter durations. Within-treatment assessment may be continuous or pulsed using the above-described periods at specified intervals.
Patients are often provided printed forms (sheets, booklets, cards) on which to record the desired information. This can prompt the patient for the specific types of information desired and can help to organize the arrangement of the data for easier review and analysis. However, to enhance compliance, it is useful to assure patients that records will be accepted on any paper if they have misplaced the record forms. Other formats for recording intake have recently emerged, including software for computers and the more portable Palm-type devices (personal digital assistants [PDAs]).
Self-monitoring of dietary intake offers numerous advantages. It permits recording of all foods and drinks, in whatever quantities are consumed. There is wide flexibility regarding the duration and frequency of assessments. If factors such as quantities, methods of preparation, and brand names are included, the resulting information is suitable for nutrition analysis of energy, macronutrient, and micronutrient content. When additional information is elicited, data of behaviorally diagnostic value can be obtained, such as stimuli for eating or eating pace, and a fuller picture of the patient's overall eating patterns can emerge. Self-monitoring permits ongoing assessment of response to treatment for clinical and research purposes and continuous feedback regarding current eating patterns. Ideally, the recording is contemporaneous and not reliant on memory. Very little staff time is required to instruct the patient in monitoring requirements and no staff time is required to acquire the data. The primary demand on staff time is the review and analysis of completed records. As will be discussed in more detail below, self-monitoring can have a beneficial reactivity effect and is therefore a treatment tool as well as an assessment method.
One major disadvantage to self-monitoring is that the responsibility and control of the data acquisition are entirely in the hands of the patient: the method is completely reliant on patient compliance. As will be noted later, however. there are ways to improve compliance. Nonetheless, the method does pose a real demand on the patient, although the anticipated time and effort required are often greatly in excess of that actually demanded. From clinical experience, gaps in monitoring are likely to involve eating episodes that are viewed by the patient as excessive or undesirable. Social desirability and demand characteristics impact self-monitoring as they do other methods, highlighting the importance of emphasizing to the patient that the primary goal is accurate reporting regardless of the desirability of the actual amount or type of intake. The accuracy of the patient's portion estimates in the field should never be assumed; instructions for monitoring should include suggestions for determining portion sizes. Finally, self-monitoring is the most reactive dietary assessment technique, which may be problematic in circumstances where accurate measurement of naturalistic patterns is most important.
Role of Self-Monitoring in Enhancing Weight Loss
Self-monitoring often changes the monitored behavior in the desired direction ((6)). Evidence suggests that this is also the case in weight-management efforts, where self-monitoring of dietary intake seems to contribute to greater weight loss. An early study of a 19-week behavioral weight-control program found that the extent of dietary self-monitoring was significantly correlated with weight loss (r = 0.57, p < 0.05) ((7)).
The following studies by Kirschenbaum and colleagues convincingly demonstrate the contribution that self-monitoring can play in enhancing weight loss. Baker and Kirschenbaum ((8)) followed 56 patients during 12 weeks after initial treatment participation (average of 40 weeks), during which they had lost an average of 9.5 kg (21 lbs). They assessed the consistency of subjects’ self-monitoring on a weekly basis during this time and found that it varied widely across and within subjects. Several analyses showed the strong relation of self-monitoring consistency to weight loss during this time (r = 0.41 to 0.44, p < 0.01). When subjects were divided into quartiles based on extent of monitoring, the top quartile lost far more weight during the 12-week period (10.0 kg; 22.1 lbs) than did the bottom quartile, which gained 0.7 kg (1.6 lbs). Furthermore, when each subject's weekly weight changes were compared against the extent of monitoring in the corresponding weeks, subjects lost significantly more weight in their best week of monitoring than in their worst week of monitoring.
A later study by Boutelle and Kirschenbaum ((9)) closely replicated the above findings. In this study, conducted with different participants from the same ongoing treatment program, subjects had already lost an average of 15 kg (33 lbs) over varying periods of time (mean = 17 months). Again, varying rates of monitoring were found, and consistency of monitoring was correlated significantly with weight loss.
In addition, another study by Baker and Kirschenbaum ((10)) found that, among participants in cognitive-behavioral treatment of obesity, weight control during the end-of-the-year holiday season was better for patients who monitored more consistently than for those who did not. Boutelle et al. ((11)) subsequently demonstrated that weight-loss participants who received an additional intervention to encourage self-monitoring during the holidays demonstrated more consistent self-monitoring and better weight outcomes during that period than did participants who received usual treatment.
| Percentage of days self-monitored | Percentage of subjects: 12-week study ((8)) | Percentage of sujects: 8-week study ((9)) |
|---|---|---|
| 76 to 100% | 39% | 46% |
| 51 to 75% | 27% | 28% |
| 26 to 50% | 16% | 16% |
| 0 to 25% | 18% | 10% |
| Percentage of days self-monitored represents percentage of days in study period that subject recorded intake for entire day. | ||
Levels of Compliance with Self-Monitoring
Taken together, the above findings point to the substantial effect on weight loss that self-monitoring can offer when conducted over an extended period of time. However, they also support the clinical impression that compliance with self-monitoring varies greatly across people and within the same person, over time, particularly when it is expected on a continuous basis. Table shows rates of compliance with self-monitoring (defined as monitoring intake for the entire day) observed in two of the previously described studies, in one case over 12 weeks and in the other over 8 weeks ((8),(9)). As is shown, compliance varies widely across subjects, and is far from perfect. One-fourth to one-third of the subjects kept records on fewer than one-half of the days. Compliance rates were lower with increased durations of observation.
Thus, it is not realistic to expect perfect compliance with self-monitoring when it is required on an ongoing basis. Fortunately, near-perfect adherence is not necessary. Boutelle and Kirschenbaum ((9)) concluded that a realistic goal for success within a comprehensive weight-control program might be recording all intake on 75% of days. They further suggested that monitoring on less than one-half of the days might lead to failure.
Increasing Compliance with Self-Monitoring
Given the importance of achieving a moderately high rate of compliance, the clinician should employ all possible means to assist and encourage patients with their monitoring goals. The goals themselves should be realistic, e.g., record for 5 days out of the week. Initial instructions should include discussion of the importance of both the process and products of self-monitoring for the weight-control effort. It should be made clear that the top priority is recording intake regardless of whether the intake is in line with assumed or explicit dietary recommendations. In other words, the clinician should emphasize that turning in the records is important regardless of the dietary mishaps they may depict. Needless to say, staff responses to actual records must be consistent with this message. They should be nonjudgmental and supportive and should reinforce the behavior of recording.
Attention to the mechanics of self-monitoring may also improve compliance. Instruction in portion estimation may increase the accuracy of recording and make the task seem less daunting. Recording forms are more likely to be used if they are user-friendly, e.g., conveniently sized and organized with adequate space for required information. Comprehensible abbreviations and symbols should be encouraged when they can facilitate quicker and easier recording.
Demands on the patient should be as few as possible, and only data that are very important at the time should be elicited. Not all data are needed throughout treatment, and the self-monitoring targets may be reduced in number and may vary over time. For example, as treatment progresses, it might be most important to record all food and caloric drink intake, but not water intake, location of eating, or the presence of others. Alternatively, the goal may be to record only between-meal items if meals seem stable, or record only dinner or meals out if those are problematic.
An important, but sometimes unmentioned, way to increase compliance is to ensure that the patient sees that keeping and submitting diet records will directly improve his or her treatment. Feedback should be provided regularly and promptly in a nonjudgmental, supportive manner, so that provision of self-monitoring data is seen as a means to obtain more focused and individualized dietary and behavioral advice. If reported dietary lapses are treated as challenges and opportunities for problem solving, the reluctance to report such slips should decrease, and the patient's ability to conquer problematic situations should improve.
Applications of Technology to Self-Monitoring for Weight Management
Technological advances offer possibilities for enhancing the process and effectiveness of self-monitoring. Available software products allow the user to enter dietary data into a computer, and more recently, a Palm-type PDA, for storage and analysis, either onboard or through uploading to an Internet site. These procedures offer the opportunity to provide automatic nutrient analysis without staff effort and to combine that analysis with a comparison to the individual's recommended food plan. Furthermore, in the case of Palm-type PDAs and other portable hardware, this capacity can be employed before the planned intake actually occurs, to aid in patient decision-making. The information and awareness functions of self-monitoring are thus enhanced with these applications. With ongoing storage of entered data, intake trends can also be depicted. These technologies also have far more capacity than paper formats for engaging user interest and providing additional information to the user.
An early, small study examined the effectiveness of a weight-loss program using a small portable computer for self-monitoring compared with a traditional paper-and-pencil format ((12)). Subjects in the computer-assisted group lost more weight at 8 weeks and at 8 months. Another more recent weight-loss trial examined the effects of a PDA-based monitoring and feedback system compared with standard paper and pencil monitoring. All subjects also received 12 weeks of standard behavioral treatment in a group setting at the beginning of the trial. Preliminary results suggest that subjects using the PDA system device lost more weight than the paper-and-pencil subjects at 12 and 18 months (J. Sanderson, J. Larsen, unpublished data presented at the National Nutrient Databank Conference, July 2000).
A recent comparison of two Internet-based weight-loss interventions demonstrates the possibilities of this medium, especially when it involves self-monitoring with feedback ((13)). One intervention was essentially educational, providing information and weight-related resources on the Internet, with unmonitored and otherwise unassisted encouragement to self-monitor. The behavior therapy program added the requirement of weekly submission of self-monitoring information for feedback through E-mail from a therapist. Subjects in the behavior therapy group lost more weight at both 3 and 6 months. Furthermore, as in the previously noted studies of more traditional treatment approaches, frequency of submission of self-monitoring records was significantly correlated with weight loss in the behavior therapy group.
Results of these preliminary studies suggest the potential of technologically enhanced self-monitoring to increase the amount and duration of weight loss in lifestyle-based treatment programs. Regardless of evidence for effectiveness, it is likely that the demand for such hardware and software will increase with growing consumer interest and increased pervasiveness of computers, PDAs, and related devices. Technology that can aid data collection and provide increasingly fast and personalized analysis, feedback, and guidance will hold great appeal. As wireless technology advances, the opportunities for instant transfer of patient-entered data to monitoring centers and nearly instantaneous responses from the center to the patient will permit more constant interaction between patient and the “treatment program,” be it human or digital. Challenges for researchers will include the careful, impartial evaluation of the effects and cost-effectiveness of such technological treatment additions and the determination of which components of these additions are most important. Clinicians will face challenges such as staying abreast of technological advances, learning to incorporate technologically sophisticated devices into treatment without losing focus of the overall goals for the individual patient, and maintaining a view of technology as a means to help humans with the problem of obesity, not as an end in itself.
Conclusions
The various methods of dietary assessment have differing advantages and disadvantages for the clinician. The choice of method should be determined by the clinical needs at a given point in treatment and by the available staff and patient resources. For example, the 24-hour recall may be most useful for diagnostic and follow-up purposes in situations where the staff and patient time is available, the patient is interviewed during a representative period of intake, both qualitative and quantitative information are desired, and the numbers of opportunities for patient contact are limited. Diet history and food frequency questionnaires may be employed for diagnostic purposes when only one opportunity for assessment exists and the emphasis is on qualitative patterns of food choices. Self-monitoring may be most valuable when analysis of behavioral patterns is desired, when specific and concurrent information is needed, or when the reactivity effect may be exploited to control intake. Self-monitoring is especially useful for ongoing assessment of progress as well as baseline assessment where reactivity is not a concern. Technologically assisted means of self-monitoring may be of particular value for enhancing compliance and for providing ongoing, rapid feedback to the patient.
In clinical settings, validity, though important, is not the only concern. Although underestimates of intake may be assumed to occur frequently, they may be taken into account in dietary prescriptions. Most important, the method(s) of dietary assessment should be acceptable to the patient and should provide information whose value is apparent to both patient and clinician.
Acknowledgment
Preparation of this paper was supported in part by HealtheTech.
