Scientific Papers

Reach and effectiveness of a worksite health promotion program combining a preventive medical examination with motivational interviewing; a quasi-experimental study among workers in low socioeconomic position | BMC Public Health

Study design and population

The WHPP was evaluated by employing a quasi-experimental study carried out at 14 departments in a hospital and a production company in the Netherlands. Workers in lower paid jobs who often performed physically strenuous work were targeted. These workers included kitchen workers, telephone operators, nutrition assistants, logistic staff, mechanics, paint sprayer, production workers. Eligibility criteria for individual workers in the study included 1) being in paid employment and 2) working at least 12 h a week. Participants were included between June 2019 and March 2020 after providing informed consent. At baseline a web-based questionnaire with questions on health, health behaviour, work ability, sickness absence, working conditions, and demographics was administered, and anthropometric measurements were carried out as part of the preventive medical examination. Six months after the baseline measurements participants were asked to fill-out a second web-based questionnaire with questions on the primary and secondary outcome measures and additional questions on participation in intervention activities. The Medical Ethics Review Committee of the Erasmus University Medical Center in Rotterdam waived the requirement for formal ethical application as the Medical Research Involving Human Subjects Act did not apply for the current study (MEC2018-1717). Informed consent was obtained from all study participants. The study is registered in the Netherlands Trial Register as NL8178 on 22/11/2019 (

The worksite health promotion program

Within the hospital the study was announced through e-mail and in the production company workers were notified about the study via their supervisor. The intervention consisted of the following components:

  1. 1)

    Preventive Medical Examination (PME), consisting of a web-based questionnaire with questions on health, health behaviour, work ability, sickness absence, working conditions, and anthropometric measurements.

  2. 2)

    Tailored coaching based on MI, up to 7 sessions within 6 months.

All workers enrolled voluntarily in the study by completing the PME, which took approximately 45 min. The communication and wording of the questions in the web-based questionnaire was adapted to the language level of the workers. Workers were given the opportunity by their supervisors to participate in the PME during working hours. In the hospital, the questionnaire was self-administered and workers performed their own anthropometric measurements with a toolbox they received at home. In order to improve accessibility of the PME at the hospital, workers could perform their measurements at the occupational health service. In addition, computers were made available at the department to fill out the questionnaire. At the production company, PMEs were already frequently offered to workers before the current program. The current PME was integrated in their existing procedure. In accordance to their existing PME procedure, the company doctor’s assistant administered the questionnaire and performed the anthropometric measurements face-to-face. After completing the PME, participants received a computer-generated overview of their results on specific factors related to health, health behaviour and work, and their cardiovascular risk profile. Results were represented as no risk (green), low risk (orange), and high risk (red). Focus groups that were performed at the organizations as part of the development of the program indicated that when the results were presented this way the presence and magnitude of the health risks were conveyed well and that they would be more inclined to change their behaviour.

Participants with a low or high cardiovascular risk were given the opportunity to receive tailored face-to-face coaching based on MI (30 min per session). These coaching sessions were focused on change in physical activity, smoking, alcohol consumption, nutrition, or relaxation, depending on the goal setting of the workers. In the hospital, workers with a low cardiovascular risk could participate in a maximum of 2 coaching session with a lifestyle coach and those with a high risk could receive 1 coaching session with an occupational physician (OP) and a at least 1 session with a lifestyle coach. Workers with a low risk at the production company could participate in at least 1 coaching session with either a dietician, physical therapist, or social worker, and workers with a high cardiovascular risk could receive 1 session with an OP and at least 1 with the other professionals. These MI sessions had to take place between the PME and follow-up measurement after 6 months. Whereas coaches within the production company received 3 days of training in MI, coaches within the hospital received 2 days of training in MI. Most of the coaches in the hospital already completed a training or workshop in MI before development of the program.

When participants had the intention to change their behaviour and were willing to undertake actions during the coaching sessions, they received suggestions on health promotion activities based on their personal preference. Participants could either independently undertake action to change their behaviour or they could take part in health promotion activities offered by the organisations. The latter activities include discount on gym membership, (e-health) interventions for quitting smoking, improving vitality, and mindfulness, and the use of mobile apps to track and improve physical activity. Reach was defined as participation in the MI after the PME compared to persons only participating in the PME.


Primary outcome measures

Participation in health promotion activities

Participation in health promotion activities was measured after 6 months using the question ‘Since you participated in the PME 6 months ago, have you taken action in order to improve your health behaviour (physical activity, smoking, alcohol use, nutrition or relaxation). Workers could indicate whether they individually attempted to improve their health behaviour or participated in health promotion activities offered by the organisation (e.g. courses on smoking cessation, mindfulness or vitality, or discount on gym membership). If workers indicated that they had taken action on at least one health behaviour, this was considered as having participated in health promotion activities.

Bodyweight and BMI

During the PME at baseline bodyweight and height were measured. Whereas workers in the hospital measured their own bodyweight and height, at the production company these measurements were performed by the doctor’s assistant. After 6 months self-reported data on bodyweight were collected. BMI was derived from the weight and baseline height of participants and expressed in kg/m2. BMI was categorised into workers with a healthy weight (18.5 kg/m2 to 25 kg/m2), overweight (25 kg/m2 to 30 kg/m2) workers and obese workers (≥ 30 kg/m2).

Secondary outcome measures

Health behaviour and self-rated health

At baseline and after 6 months respondents were asked to indicate their current health behaviour with respect to the number of days a week (0–7 days) they performed vigorous physical activity (activities which cause persons to sweat) during leisure time, smoking status (yes/no), alcohol consumption in alcohol-units per week (‘1 = less than 1 glass per week’ to ‘6 = more than 28 glasses per week’), daily intake of fruit (‘1 = none’ to ‘6 = 3 or more pieces of fruit a day’) and vegetable consumption (‘1 = none’ to ‘6 = 4 or more spoons a day’). Dichotomous variables were made based on adherence to Dutch guidelines (yes/no) for performing leisure time vigorous physical activity for at least 2 days a week [18], not smoking, not drinking more than 7 glasses of alcohol per week [19], eating at least 2 pieces of fruit a day, and at least 4 servings of vegetables (200 g) per day [19]. Self-rated health was measured at baseline and after 6 months by asking respondents to rate their generate health on a scale from 1 (very good) to 5 (very bad) [20]. Self-rated health was categorized into workers having good or very good self-rated health, and those having less than good self-rated health.

Work ability and sickness absence

Self-reported work ability was measured at baseline and after 6 months using the first dimension of the Work Ability Index (WAI) [21]. Respondents were asked to indicate their current work ability as compared to their lifetime best on a scale from 0 (not able to work) and 10 (lifetime best). At baseline and after 6 months it was examined how many days in the past year participants were not able to work due to sickness, admission to hospital or research on a 5-point scale (‘1 = 0 days’ to ‘5 = 100 to 365 days’). Dichotomous variables were made for work ability (work ability higher than 6 versus 6 or lower) and sickness absence (less than 10 days sickness absence versus at least 10 days).


At baseline, socio-demographic information was collected on gender, age in years, highest completed educational level, and marital status. Age was categorized in 4 groups; lower than 30 years, between 30 and 40 years, between 40 and 50 years, and 50 years and older. Education was divided in three groups; low (no education, primary school, lower and intermediate secondary schooling, or lower vocational schooling), intermediate (higher secondary schooling, or intermediate vocational schooling), or high education (higher vocational schooling, or university). For marital status being married or cohabiting was compared with unmarried, divorced or single workers. Additionally, experience of financial pressure at baseline was assessed and dichotomized (intermediate or high financial pressure versus little or no financial pressure). Several working conditions were assessed at baseline. Autonomy (Cronbach’s alpha = 0.81) and work pressure (Cronbach’s alpha = 0.67) each were measured with 3 items derived from the Job Content Questionnaire (JCQ) with the answer categories 1 (always), 2 (often), 3 (regularly), 4 (sometimes), and 5 (never) [22]. Based on mean scale scores, low autonomy (scores 4 to 5) was compared with high autonomy (scores 1 to 3) and high work pressure (scores 1 to 3) with low work pressure (scores 4 to 5). Dichotomous variables (yes/no) were made for other work-related aspects including having a high physical workload (whether or not workers had to lift, push, pull or turn heavy loads) during daily work, performing shift work, performing mainly physically demanding tasks or both mentally and physically demanding tasks, less than good work-life balance, and working at least 36 h a week.

Delivery of the intervention

Intervention exposure was indicated by the number of MI coaching sessions workers attended and the mean total duration of these sessions as reported by coaches. The number of health promotion activities workers participated in was acquired from the 6 month follow-up questionnaire. The quality of MI coaching was determined by audio recordings, which were made by the coaches. The OPs and dietician were instructed to record every 4th session they provided and the other professionals had to record every 3rd session according to an agreed protocol, to ensure random selection of the recordings. The frequency of recoding was set lower for the OPs and dietician as they were expected to meet more participants than the other professionals. Recorded sessions were analysed using the Motivational Interviewing Treatment Integrity code (MITI) version 4.2.1 [23], which represents a reliable instrument for assessing MI treatment integrity [24]. Coding was done by a trained researcher (DV). Quality of MI was expressed by the mean technical (average of global scores cultivating change talk and softening sustain talk)- and relational score (average of global scores partnership and empathy), both on a scale from 1 to 5. A technical score of 3 and relational score of 4 was indicative of acceptable quality. The proportion of recordings with at least 40% complex reflections and ratio of reflections versus questions ≥ 1 was considered as adequate MI. In addition, self-reported data from the 6-month follow-up questionnaire was used to indicate the percentage of workers (strongly) agreeing with the following statements on the contact with coaches; ‘I was treated in a pleasant way’, ‘the coach had expertise’, ‘my medical data was treated confidentially’, ‘I was satisfied with the contact in general’, ‘the contact lived up to my expectations’, ‘because of the contacts I know how to improve my health behaviour’, ‘the contact with the coach contributed to change in health behaviour’, and ‘the number of contact moments was sufficient’. Answers were given on a 5-point scale ranging from 1 (strongly disagree) to 5 (strongly agree).

Statistical analysis

Baseline characteristics were compared between 1) workers who completed the PME and attended subsequent MI coaching sessions, and 2) workers who completed the PME but did not attend MI coaching using chi-square tests. Logistic regression analysis was performed to determine whether baseline characteristics were associated with drop out after 6 months.

Associations between baseline characteristics and participation in the PME and follow-up MI coaching sessions were tested with multivariate logistic regression analyses. Backward elimination based on maximum likelihood estimates was used to select a combination of characteristics that contributed most to participation in the intervention group. For each individual, the propensity score was estimated, indicating the probability to participate in the PME as well as MI coaching compared to participation in the PME without additional MI coaching.

Multivariate logistic regression analysis was performed to test whether MI sessions were associated with participation in health promotion activities. This analysis was adjusted for sociodemographic characteristics (gender, age, educational level), organisation, and the propensity score to decrease the impact of selection bias and the voluntary choice to participate in the intervention due to the quasi-experimental study design [25]. Changes in health, health behaviour, work ability and sickness absence between baseline and 6 months within each group were evaluated using paired T-tests for continuous variables and McNemar’s test for dichotomous variables. The effectiveness of subsequent MI coaching after 6 months compared to workers who did not attend MI coaching was analysed with linear regression models. Analyses were performed among workers with complete data at baseline and 6 months follow-up. First, we employed models in which participation in MI coaching was included as independent variable, with adjustment for baseline values on outcome measurements, socio-demographic characteristics (gender, age, educational level), and organisation. Next, the propensity scores were additionally included as covariates to the models. All analyses were performed with IBM SPSS Statistics version 25.

Source link