Weight loss among female health care workers- a 1-year workplace based randomized controlled trial in the FINALE-health study

The present study, termed FINALE-health is a cluster randomized single-blinded controlled trial conducted from May 2009 to the end of June 2010. The study consisted of 12 months intervention with tests performed at baseline, after three months and after one year. The intervention and effects after the first three months have previously been reported [14]. The present paper presents results after 12 months of the intervention. The project was ethically approved by The Central Denmark Region Committees on Biomedical Research Ethics (M-20090050), and qualified for registration in the International Standard Randomized Controlled Trial Number Registry (NCT01015716).

For detailed description of recruitment and randomization of participants, please see Christensen et al. 2011 [14]. In short, Randers municipality agreed to participate in the study, and one of nine possible areas of the municipality was found eligible to participate. Recruitment of participants was based on the payroll, and all employees were invited to participate in the project. Of 202 invited employees, 158 attended one of three introductory meetings. Among these employees, the predefined target group for the project were those fulfilling the following the inclusion criteria of being female, overweight (i.e. BMI > 25 or body fat % > 33 for age 18–40 or > 34 for age > 40), and being health care worker or primarily working with elderly care. 101 persons fulfilled the inclusion criteria, and 98 consented participation in the study.

All consenters filled out a screening questionnaire [15], were physically tested the following week, and enrolled and were divided in clusters which were randomly allocated to either intervention or reference group. Clusters were created based on information from the screening questionnaire and the management of working teams, day and evening/night shifts and close working relations. This approach was chosen to avoid contamination, and to benefit from the social support in work teams, thereby increasing compliance. The cluster-randomization was balanced on sex, age, job seniority and job type and done by an external research group, with no connection to the workplace or the participants [14].

In short, the intervention was carried out one hour a week during work time. The intervention lasted 12 months and consisted of two parts. The first part (0–3 months) focused on weight loss and included advice on dietary change based on the Danish Dietary recommendations, calorie counting, weight measurements, weight loss targets, strengthening exercises and initiating leisure time fitness exercise. The second part (3–12 months) focused on weight loss maintenance through further intervention with physical exercise and cognitive behavioral training, see Figure 1.

Before the intervention, all participants were encouraged to fill out nine days dietary records [16]. They were used to obtain information on dietary preferences and to create different exemplary courses so food intakes during the weight loss period matched to their normal preferences. Still, to change the participants usual food intake to a healthier diet, the courses were adjusted according to the Danish Dietary recommendations, e.g. reduction of refined sugar, reduction of fat – especially saturated fat, carbohydrates from primarily fiber-rich sources (e.g. whole-grain bread, whole grain pasta etc.), and 600 gram of fruit and vegetables per day. These guidelines are also consistent with new findings regarding weight loss [17]. These courses were proposed for every meal with specific calorie amounts adjusted to suit an individual calorie prescription. In advance, each individual's resting metabolism was calculated, based on gender, age and weight and multiplied by a Physical Activity Level factor (PAL) of 1.8 to estimate the daily energy requirement [18]. 1.000 – 1.200 calories was subtracted from the estimated daily energy requirements and the resulting value was used as individual calorie prescription. The calorie reduction in the present study was set to 1.000 – 1.200 calories to achieve an optimal weight reduction rate of about 1 kg per week [5]. This recommendation is based on a systematic review from 2007 by Franz MJ et al. “Weight-loss outcomes: a systematic review and meta-analysis of weight-loss clinical trials with a minimum 1-year follow-up”. The literature reviewed finds a 1.200 kcal reduction for women and a 1.500 kcal reduction for men in diet and exercise studies [5]. The American Dietetic Association’s Adult Weight Management Evidence-Based Nutrition Practice Guideline supports only a calorie deficit between 500 – 1000 calories [19], thus a calorie deficit between 1.000 – 1.200 calories was chosen”.

During the first three months of the intervention, if weight loss after two weeks was less than expected, the participants had their planned daily calorie intake lowered by further 300 kcal per day. The dietary advices and the weight check occupied approximately 30 min of the weekly session.

During the last nine months of the intervention, every session still began with a weight check. Participants were clearly informed, that the aim of this period was to maintain their weight after the first 3 months loss of body weight, and that they no longer had to continue the reduced daily intake of calories. If participants during the first three months had problems quitting eating unhealthy food such as sweets and biscuits, they were advised to add smaller amounts of these products into their intake. If participants had problems with hunger, they were advised to eat larger and healthy meals. During the last nine months of the intervention, no further dietary counseling was provided unless questions were asked.

During the first 6 months of the intervention, about 15 minutes physical exercise training was included in the weekly session at the workplace (Figure 1). Focus was on strength training to maintain muscle mass in the lower extremities in order to maintain resting metabolism and physical capacity. These exercises were carried out in a meeting room at the workplace and consisted of one and two legged squats, with and without dumb bells and core balls, and lunges walking forward and to each side. Other exercises focused more on general strength, and included exercises for abdominal and back extension, shoulders and arms. Participants brought home a strength training program, picturing these exercises, and were encouraged to perform them twice a week at home. The dose of the instructed physical exercises in the sessions progressed in intensity throughout the weeks of the intervention, by increasing weights and repetitions. In addition to the brief training sessions, participants were encouraged to initiate leisure time aerobic exercises for two hours weekly such as biking, walking, running, swimming or attending different sports in the local area.

From the 6^th to the 9^th months of the intervention, a simple fitness gym was arranged with rubber bands, dumbbells, core-balls, skipping ropes and mattresses plus fitness machines for abdominal and back extension, leg curls and leg press. During this phase, the physical exercise was carried out in the fitness gym as circuit-training, both to progress muscle mass and aerobic fitness. Heart-rate monitors were lent to participants to learn about effective heart rate intensities for increasing aerobic fitness (>70% of maximal heart rate). In this phase, out-door running was also initiated.

During the 9^th to the 12^th month of the intervention, physical exercises primary took place as different local sports activities. It was the aim that all participants should attend sports at least twice a week (minimum of two hours in total), scheduled in the local area. Each of the seven training teams decided which sport activities to attend together during the sessions, and they ranged from attending fitness centers to zumba, spinning, swimming, water-gymnastics and line dance. Out-door running was still carried out during this phase. For motivation, participants were instructed to monitor leisure time exercises by training log books during the entire year.

Before the intervention, a specific CBT training tool based on Linton’s model for coping with chronic pain [20, 21] were modified to address discomfort during weight loss and to support a change to a more physically active lifestyle. The cognitive behavioral training tool was based on group discussions and a specifically tailored guideline, containing exercises such as pro-and-con schemes and positive thinking strategies with homework between each session.

During the first three months, about 15 minutes of the weekly sessions was used on CBT (Figure 1), helping the participants to make realistic weight loss targets and find personal strategies to ease hunger. From the 3^th to the 9^th month of the intervention, about 30 minutes were spent on CBT in the weekly sessions. Focus was to reflect on dysfunctional attitudes and coping behaviors with respect to the weight loss, and discuss functional alternatives and train the implementation of these in everyday life. During the 9^th to the 12^th months of the intervention, about 15 minutes of each session were spent on how to continue healthy behaviors, cope with social contexts and situations involving alcohol, food etc.

The reference group was offered a monthly two-hour oral presentation during working hours. The twelve presentations were based on the Danish National Board of Health and the Ministry of Food, Agriculture and Fisheries public websites and concerned the Danish Dietary recommendations and other health related topics.

All participants were tested at baseline, after three months and after one year. Each test session lasted one hour and consisted of anthropometrical, health-related and physical capacity measures specified in the following. Height was measured without shoes. Body weight was measured wearing light clothes, but without socks and shoes. The weight measure was subtracted by 1 kg to compensate for clothing. Body Fat was measured by a bio impedance device (TANITA SC-330), which was set to 'standard' as body frame and the participant's age, height and gender were entered. Waist circumference was measured over the umbilicus and hip circumference on the hip part that gave the greatest circumference. Waist and hip circumference were measured standing up and with clothes on, using an ergonomic circumference measuring tape (Seco 203 Girth measuring tape). Blood pressure was measured in seated position after 10 minutes of rest with an electronic blood pressure monitoring device (Artsana CS 410). Three measurements were done one minute apart and the average value was used [22]. Cardiorespiratory fitness was measured using a Monark E327 bicycle ergometer and a heart rate monitor. Participants cycled for five minutes at 70 watts (60 rpm, 1 kp). Within these first five minutes, the test subject was asked to rate their fitness according to the Borg Scale. Hereafter load was increased by 35 watts (1/2 kp) every other minute until the test subject was forced to stop because of exhaustion. An algorithm was used to estimate maximal oxygen uptake (VO₂-max) [23].

Isometric maximal voluntary strength was obtained with a reproducible standardized setup [24], measuring maximal voluntary handgrip, sitting shoulder elevation, shoulder abduction and back flexion and extension force [25]. The participants performed a minimum of three attempts with steady increasing force to reach maximum within 3–5 seconds. The test was repeated until a maximal of 5 contractions if the last attempt showed a more than 5% increase from the previous maximum. The participant rested at least 30 seconds between each attempt. The maximal attempt was recorded for further analysis. Standardized verbal command and encouragement was given to maximize the effort. Handgrip was measured in both hands using a grip strength measurer (La Fayette) [26]. Shoulder elevation was measured with a Bofors dynamometer. The subject was seated erect on a chair with legs hanging freely, arms hanging along the side and head facing forward. The distance from pressure points to sternoclavicular joints were measured as the moment arms [27]. Shoulder abduction was measured with the Bofors dynamometers with the subject still seated erect on a chair with legs hanging freely, arms along the side and a 90 degrees flexion in the elbows and head facing forward. The distance from pressure points to the lateral epicondyles of the humerus were measured as the moment arms [27]. Back flexion and extension were measured with the subject standing, facing/backing a beam and support plate at the spina iliaca anterior superior. The Bofors dynamometer was fixed to pull horizontal with a belt positioned at the vertical level of m. deltoid insertion on the humerus. The distance from the belt to a line through the crista iliaca and lumbalcolumna (L4L5 level) was measured for the moment calculation [28].

Prior to the test session, participants were screened in accordance with the exclusion criteria for the test. The exclusion criteria for one or more of the tests were elevated blood pressure, defined as systolic values higher than 110 mmHg + persons age or diastolic values higher than 100 mmHg regardless of age [22], angina pectoris, heart or lung prescription medication, current or pervious illnesses and trauma, herniated disc, tennis elbow, golf elbow, Carpal Tunnel Syndrome, significant level of musculoskeletal pain at the time of the test, and pregnancy. An example is the bicycle-test. At baseline 35 in the intervention group and 22 in the reference group completed the test, the rest were excluded. After a year, only 23 and 16 in the two groups respectively could complete the re-test. A range of participants in the intervention and the reference group completing the different tests is given in Table 1. The test manager was blinded regarding the participants intervention status, and whenever possible the same test manager tested the subject at all three rounds of tests [14].

A questionnaire was completed three times, approximately one week before each test round. The questionnaire was developed for use in all workplaces participating in the FINALE program and consisted of 140 questions mainly of standardized and validated scales [15]. In this paper, responses to questions on musculoskeletal disorders are reported. Musculoskeletal disorders were measured with the Nordic questionnaire of musculoskeletal disorders [29], supplemented with questions about localized pain intensity [30].

A power calculation was carried out based on weight change to ensure a copious amount of participants in the intervention and the reference group [14]. Power was set to 0.8 with a significant level of 0.05. At least 30 participants in each group were needed to detect a difference in weight loss of at least 3 kg. With an estimated 30% drop out, 43 participants were needed in each group. Differences between intervention and reference group at baseline were tested with Pearson´s x² for distribution in sex, education (health care workers), current smoking status and the dichotomized parameter for musculoskeletal symptoms in neck, shoulders, upper- and lower back. All other parameters were tested with a Student´s t-test. When comparing intervention and reference groups over time, ANCOVA analysis were performed in accordance to the intention-to-treat principle, i.e. all randomized participants are included in the analyses with missing values substituted with carried forward or backwards measured variables. Clusters, age and the investigated value at baseline were included as covariates. All results are given as mean (SD). P < 0.05 are considered statistically significant.

A flowchart of the project is presented in Figure 2. From the employee list, 202 persons (8 men and 194 women), working at least 15 hours/week were invited to participate in the study. Among those, 144 consented to participate and were invited for baseline test and randomly allocated to either the intervention or the reference group. 98 met the full criteria to enter the target group, i.e. women, overweight based on BMI or fat percentage, health care workers or having similar education with daily patient care. After three months 91 were still participating and invited for test 2. At test 3 after 12 months, 83 completed the test.

Table 1 presents data from baseline (test 1) to 12 months (test 3) in both the intervention and the reference group. As a mean for all participants BMI was 30.6, aerobic fitness 26.3 mL O₂/kg/min, waist circumference 100.3 and blood pressure 131.8/83.7. No differences between the intervention group and the reference group were found at baseline.

There was a substantial variation in the weight loss between individuals as shown in Figure 3. In the intervention group the body weight changes over 12 months ranged from +15 to −42 kg, but the majority of the participants had weight losses between 0 and 10 kg. In order to test whether the weight loss depended on initial weight, a linear regression analysis was performed. In the intervention group a weak but significant linear correlation was found between weight change and initial weight (r = 0.30, P < 0.05). However this correlation relied on one outlier with a weight loss of 42 kg and an initial weight of 127 kg, when this individual was removed from the analysis there was no longer a significant correlation between weight loss and initial weight.

Regarding physical capacity measures, there were no significant Intervention group* Test round effects. Both the reference and the intervention group increased significantly over time in isometric muscle strength, for shoulder elevation and abduction and back extension. Only the intervention group significantly increased aerobic fitness over time, from 25.9 to 27.5 ml/min/kg. Regarding musculoskeletal pain, there was no significant Intervention group* Test round effect. In the intervention group, significant decreased pain level was found for both the neck and right shoulder. Except for the lower back, no changes in pain level in any of the body regions in the reference group were found.

Some of the factors that affect the risks for cardiovascular diseases are BMI, waist circumference and blood pressure. In both the intervention and the reference group, subjects in test 1and 3 were classified as above or below the recommended level for BMI (>24.9), waist circumference (>88 cm) and blood pressure (>139/89) in order to define the high risk fraction of the groups. The high risk group based on BMI, waist circumference and blood pressure showed a substantial decrease in the intervention group, in percentages from 90.7, 77.8 and 44.2 in test 1 to 64.8, 66.7 and 27.9 in test 3, respectively. The high risk fractions likewise decreased in the reference group but with much smaller changes.

This one year study was conducted as a cluster randomized single-blinded controlled trial, with data undergoing an intention-to-treat analysis (ITT). In spite of this rather conservative approach, we were able to reveal significant effects on the primary and secondary outcomes. The adherence rate of the study was very high, with a drop out of only 15% after 12 months. A limitation in the study is that in the integrated multiple interventions, the importance of each of the elements cannot be evaluated. Accordingly, we have not assessed the exact adherence to the diet or the other elements in the intervention. The diet was individually adjusted so participants were instructed that they could have a higher intake than calculated if they felt the need and could exercise more in order to keep the aimed calorie deficit. Even though the cognitive part of the sessions closely followed protocol, we could not entirely control what participants brought up for question and therefore we cannot state, that all participants got the precisely same guidance, which of course may affect the outcome. A well-defined protocol was followed concerning the brief physical exercises during sessions, but there was a limitation in the lack of quantitative registration of physical training doses in leisure time. The training logbook was primarily used to facilitate the individual coaching and serve as a motivating factor and no useful quantitative data on physical training could be obtained from these records.

The study was not designed for being very cost effective, but to investigate if it’s possible to generate a long-term weight loss among overweight health care workers in a workplace setting. After finding such positive results, a next step is to generate a cost-effective study with the same aim. Therefore, we did not perform a cost-effectiveness evaluation of this study.