ABSTRACT
The present study examined the relation between compliance with a web-based psychological self-assessment and continuation in obesity treatment. The study included 1827 female patients between the ages of 18 and 70 in treatment for obesity with a primary care physician. All patients were tested with an on- line psychological questionnaire at the start of treatment. Patients were urged to repeat the assessment at monthly intervals. It was suggested to patients that regular use of the test might prove to be beneficial to their treatment. We hypothesized that repeated use of the web-assessment would be associated with higher rates of continuation in treatment. Fifty-six % of patients completed the assessment at Time 2. We found that continuation rates at Time 3 and Time 4 were higher among those who took the test at Time 2 (72.4% at Time 3, 45.9% at Time 4) compared to those who did not (52.8% at Time 3, 29.9% at Time 4). Group comparisons on pre-treatment psychological variables and early weight loss outcome revealed few significant differences. These results suggest that repeated use of a web-based assessment and feedback program may improve continuation rates in obesity treatment.

Future research is needed to determine how the use of an Internet-based assessment influences treatment outcome, whether through direct effects on patient motivation, through effects on the doctor-patient relationship, or through other pathways.

INTRODUCTION
Obesity is now recognized as a most pressing public health issue (1). With rising prevalence and severity, the problem is now frequently referred to as "epidemic." Consensus statements from international bodies recommend moderate weight loss and greater efforts to improve weight loss maintenance (2).

Recently, an increasing number of researchers have suggested that among the most likely methods for improving long-term weight loss outcomes is to provide patients with continuing treatment (3,4,5). Indeed, the first requirement of successful obesity treatment is to keep patients in treatment, yet weight loss attempts are typically of short duration and long-term clinic-based maintenance programs are often poorly attended (3).

One modality that may offer an efficient means of increasing and extend ing treatment involvement is Internet-based intervention. Tate and colleagues recently reported positive results for behavior therapy for obesity delivered mostly via email (6). In the present study primary care patients receiving obesity treatment from general practitioners were given access to an on- line assessment procedure, which provided immediate feedback concerning eating and exercise behavior, as well as emotional factors and stress responses. Patients were encouraged to utilize the web-assessment on a regular basis as an important part of their treatment. Results of the assessment were to be used by the patient and her physician in developing behavioral goals and in evaluating treatment outcome.

While all patients had access to this system, predictably not all chose to utilize it beyond the required pretreatment assessment. In the present paper we examine the relation between repeating the self-assessment and continuation in obesity treatment. We.Web-based psychological assessment hypothesized that patients who voluntarily completed the assessment a second time would be more likely to continue in treatment than would patients who did not redo the assessment. We examined continuation rates at three time points (Time 2, Time 3, & Time 4), at 1½-month intervals.

METHOD
SUBJECTS
The study included 1827 female patients with a BMI of at least 25 (range = 25-65, mean = 32.2, SD = 6.0). Subjects ranged in age from 18 to 70 (mean = 41.7, SD = 11.5). All subjects were beginning treatment for obesity with a primary care physician. Physicians followed a consistent approach, involving low calorie diets and a brief form of behavior therapy. Sessions generally lasted 15 to 20 minutes, and involved a medical and behavioral evaluation and determination of behavior change goals until the next session. Sessions were held weekly for the first two months and then at two-week intervals. Treatment was open-ended and patients were urged to continue well into the weight loss maintenance phase. Patients were given the assurance that they could return to treatment at any time if they happened to stop for some reason. Thus, treatment combined the use of diets for short-term weight loss with behavior therapy for long-term behavior change and weight loss maintenance, in a continuous-care approach.

PROCEDURE
All subjects were tested with a behavioral/psychological questionnaire at the start of treatment. The assessment was administered on- line at our web-site (For information about the web-site and on-line assessment please contact the authors, or see www.mla.ca)

Subjects completing the on- line assessment received immediate feedback, indicating their score on a variety of behavioral and psychological dimensions, and providing behavior change recommendations concerning areas in need of improvement.

Patients were asked to repeat the test on a monthly basis. They were told that the assessment would: "Allow you to evaluate the habits, attitudes and behaviors that are responsible for your weight problem - to identify the possible causes." And that: "Repeating the test on a monthly basis will give you a tool that will prove indispensable in helping you get rid of bad habits once and for all." Thus, patients were given strong encouragement to repeat the assessment.

Subjects were classified into three groups at Time 2: Group 1 had dropped out of treatment, Group 2 was continuing but did not complete the psychological assessment at Time 2, and Group 3 was continuing and did complete the Time 2 assessment.

MEASURES
The assessment included measures of age, BMI, and four psychological variables: depression, stress symptoms, perfectionism and uncontrolled eating. The four psychological variables were measured with the Larocque Obesity Questionnaire (LOQ; 7)

LOQ - Depression included 8 items, which demonstrated a fairly high level of internal consistency (alpha = .79). Items measured feelings of hopelessness, sadness, inferiority, worthlessness, and crying. Previous research (7) has shown this scale to be highly correlated with the Beck Depression Inventory (8).

LOQ - Stress Reactions comprised 6 items, describing various symptoms commonly associated with stress (e.g., headaches, gastrointestinal complaints, difficulty.Web-based psychological assessment concentrating, dizziness, trembling, profuse sweating and unusual fatigue). This scale had an alpha of .72.

LOQ - Perfectionism included 8 items, tapping aspects of perfectionism such as disappointme nt with self, high expectations, guilt, preoccupation with mistakes, fear of failure, and need to be the best in everything. Alpha was .73.

LOQ - Uncontrolled Eating contained 12 items measuring a variety of eating behaviors, including rapid eating, eating in front of television, eating impulsively, emotional eating, eating to relax, and eating sweet and fatty foods. Alpha was .74. In previous research (7), higher scores on uncontrolled eating were associated with lower cognitive restraint and higher emotional eating on the Dutch Eating Behavior Questionnaire (8).

STATISTICAL ANALYSIS
We compared Time 2 groups on pretreatment characteristics using analysis of variance. We used chi-square analysis to compare groups on the proportion of patients who were still in treatment at Time 3 and at Time 4.

RESULTS
Table 1 presents group means on pretreatment characteristics. The only differences of note were that Time 2 dropouts had a slightly lower pretreatment BMI than the groups who were continuing in treatment (F=6.02, df=2, 1822, p<.05), and the group that continued in treatment but did not repeat the test was slightly older than the other two groups (F=4.97, df=2,1822, p<.05). None of the other group comparisons showed significant differences.

Treatment continuation rates are presented in Figure 1. The primary hypothesis was that patients who completed the assessment at Time 2 would be more likely to.Web-based psychological assessment continue in treatment. This was supported by our analysis at Time 3, c 2 (1, N=1046) = 43.4, p < .001, and at Time 4, c 2 (1, N=1046) = 28.9, p < .001. These analyses show that patients who completed the Time 2 assessment were significantly more likely to still be in treatment at Time 3 and at Time 4.

DISCUSSION
This study represents a preliminary investigation of the usefulness of a new technology in the treatment of obesity. We found that patients who voluntarily participated in our web-based assessment procedure at Time 2 were more likely to continue in treatment than those patients who chose not to complete the assessment. Assessment completers had a 20% greater treatment continuation rate at Time 3 and a 16% greater rate at Time 4.

These results are limited by the correlational nature of the study. Patients who chose to complete the Time 2 assessment may have been different than those who did not - i.e., a self- selection bias may explain the results. However, we were not able to detect any group differences at pretreatment, nor in the amount of weight lost by Time 2. Variables not included in the present study might be more successful in differentiating between groups. For example, lower continuation rates among those who choose not to repeat the web-assessment may reflect their lower outcome expectancies and self- efficacy in relation to weight control (10,11). Not doing the test at Time 2 may be an early indication that the individual is losing hope and/or interest in weight control. Repeated evaluation of subjects' weight control-related cognitions in relation to treatment continuation vs. dropout is required to test this hypothesis.

It is premature to conclude that the Internet assessment can influence treatment continuation rates. Such a conclusion will require future research involving random assignment to conditions. We are, however, encouraged by the results, given the possibility that the web-based assessment and feedback system may address the crucial maintenance problem in obesity treatment. This cost-effective and "always available" treatment tool may help keep patients connected and coming for treatment who might otherwise be vulnerable to dropping out.

Should the web-assessment prove capable of improving treatment adherence, additional studies would be required to explain its effects. The procedure may influence treatment outcomes in a number of ways. Receiving specific and quantitative feedback about behavioral and psychological factors may potentiate self- management processes such as (1) goal-setting, (2) self-observation, and (3) self-reinforcement. Alternatively, the assessment may improve doctor-patient communication and the therapeutic alliance. The assessment may promote a broader focus for both patient and doctor, with greater attention to emotional and behavioral factors. This type of time-efficient intervention may prove especially useful in primary care settings, where a high patient volume may sometimes inhibit an in-depth assessment of psychological and behavioral variables.

Table 1: Pretreatment characteristics based on Time 2 classification

                           Time 2 Classification
                ___________________________________________
                 Dropout     Continuing,      Continuing,
                             Did not          Completed
                             Complete Test    Test           F (2, 1822)
________________________________________________________________________
N                779         456              590
(% of sample)    (42.7)      (25.0)           (32.3)

Age              41.14a      43.15b           41.29a         4.97*
(SD)             (11.29)     (11.66)          (11.55)

Time 1 BMI       31.60a      32.50b           32.65b         6.03 *
(SD)             (5.51)      (6.48)           (6.20)

Time 2 BMI        ----       30.77            30.96          0.24
(SD)                         (6.32)           (5.94)

Depression 1     14.24       13.86            14.01          0.22
(SD)             (4.01)      (3.72)           (3.92)

Stress
Reactions 1      12.16       12.23            12.05          0.70
(SD)             (3.78)      (3.49)           (3.59)

Perfectionism 1  19.89       19.60            20.05          0.21
(SD)             (4.04)      (4.21)           (3.97)

Eating
Habits 1         25.06       25.16            24.85          0.44
(SD)             (5.86)      (5.42)           (5.71)

Note. Means in the same row with different subscripts indicate significant
      group differences (p < .05).
      * p < .01
      1 Higher scores indicate more symptoms.

Figure 1: Treatment continuation vs. dropout and participation in web-based assessment

Time 1                           N=1825
                                 /     \
                                /       \
                               /         \
Time 2                   Dropout          Continuing
                      (N=779,42.7%)     (N=1046, 57.3%)
                                        /          \
                                       /            \
                                      /              \
                                Did not            Completed test
                            complete test              (N=590)
                               (N=456)                 /  |  \
                                /|\                   /   |   \
                               / | \                 /    |    \
                              /  |  \               /     |     \
Time 3                   Dropout | Continuing  Dropout    |    Continuing
                         (N=216  |   (N=240    (N=163     |      (N=427
                         47.2%)  |    52.8%)   27.6%)     |       72.4%)
                                 |                        |
                                / \                      / \
                               /   \                    /   \
                              /     \                  /     \
                             /       \                /       \
                            /         \              /         \
Time 4                 Dropout     Continuing    Dropout     Continuing
                       (N=321        (N=135      (N=319        (N=271
                        70.4%)        29.6%)      54.1%)        45.9%)

ACKNOWLEDGEMENTS
The authors would like to acknowledge the help of Drs. Peter Forbes and Harry J. Lefebre and Mr. Paul Connolly for their support of the MLA research program.

Portions of this paper were presented at the annual convention of the North American Association for the Study of Obesity, Quebec City, October 2001.

REFERENCES
1. World Health Organization. Obesity: preventing and managing the global epidemic. Geneva, Switzerland: World Health Organization; 1998. WHO/NUT/NCD/98.1.

2. National Heart, Lung, and Blood Institute. The practical guide to identification, evaluation, and treatment of overweight and obesity in adults. Bethesda: NHLBI; 2000.

3. Jeffery, R.W., Drewnowski, A., Epstein, L.H., et al. Long-term maintenance of weight loss: Current Status. Health Psychol. 2000; 19 No. 1(Suppl.): 5-16.

4. Latner, J.D., Stunkard., A.J., Wilson, G.T. et al. Effective long-term treatment of obesity: a continuous care model. Int J Obes Relat Metab Disord. 2000; 24: 893-898.

5. Perri, M.G. The maintenance of treatment effects in the long-term management of obesit y. Clin Psychol: Science and Pract. 1998; 5: 526-543.

6. Tate, D.F., Wing, R.R. & Winett, R.A. Using internet technology to deliver a behavioral weight loss program. J Am Med Assoc. 2001; 285: 1172-1177.

7. Larocque, M. & Stotland, S. The Larocque Obesity Questionnaire: A new measure of the psychological factors in weight control. Am J Bariatric Med. 2000; 15: 12-14.

8. Beck, A., Ward, C., Mendelson, M., et al. An inventory for measuring depression. Arch Gen Psychiat. 1961; 4: 561-571.

9. Van Strein, T., Frijters, J.E.R., Bergers, G.P.A., & Defares, P.B. The Dutch Eating Behavior Questionnaire (DEBQ) for assessment of restrained, emotional eating and external eating behavior. Int J Eating Disord. 1986; 5: 295-315.

10. Schwarzer, R. & Renner, B. Social-cognitive predictors of health behavior: Action self-efficacy and coping self-efficacy. Health Psychol. 2000; 19: 487-495.

11. Stotland, S. & Zuroff, D. Relations between multiple measures of dieting self-efficacy and weight change in a behavioral weight control program. Beh Ther. 1991; 22: 47-59.