Effects of health education for migrant females with psychosomatic complaints treated by general practitioners: A randomised controlled evaluation study

2.1. Intervention 

The intervention strategies used followed the common methods used by migrant health educators active in general practices in Rotterdam. The characteristics of these methods include information transfer and counselling appropriate for the patient's culture and language, facilitation of communication between GP and patient, and transfer of information on relevant culture-specific matters aimed at the GP. Specific to the Bridge Project, the health educators helped the subjects understand the origins and symptoms of psychosomatic complaints and accompanied the patients to the GP to ensure that doctor and patient fully understood each other. The peer educators provided advice and taught skills the patients could use to change their beliefs about their psychosomatic complaints and promoted their self-efficacy to cope with these complaints. Because no intervention strategies for migrant patients with psychosomatic complaints were available, the Bridge Project was started. The positive experiences with lay health education by migrant health educators were an example for the education on stress-related symptoms [6]. General principles from stress reduction theory were translated to the intervention protocol. The promotion of the patients’ ability to restore balance when faced with stressors and to cope with their individual lives was the central principle for the patient centred education [7], [8]. The intervention protocol comprised the following phases of patient support: (1) clarification of physical and psychological complaints to the patient, (2) clearly formulating the request for help and optimising communication between patient and GP, (3) providing information to increase the patient's ability to restore balance in stressors and coping in personal life in eight sessions of group education, and (4) conclusion and evaluation.

A tailored counselling programme was devised for each patient. The mean number of counselling sessions between migrant health educator and patient averaged to 3.2 occasions. Consultations with GP and health educator present averaged 2.1 sessions per patient. A total of 25 patients participated in the group sessions and followed an average of 6.3 sessions. Not all patients participated in the group education sessions in phase 3, partly due to problems with program logistics and practical problems among the patients. Another possible reason for failing to attend the group sessions may have been the social stigma felt within the community if personal problems were discussed outside the immediate family. As a substitute, personal information sessions with the educator were provided to 11 patients in phase 3, averaging 4.6 occasions per patient; 11 patients failed to attend any phase 3 sessions. The mean intervention duration per patient was 12 months.

The education sessions and interventions were conducted by one Moroccan–Arabic health educator, one Moroccan–Berber health educator and two Turkish health educators. All the health educators had been trained to deal with the topic of psychosomatic disorders, had been given an introduction to mental health care facilities, and they had been trained to demonstrate relaxation exercises and to facilitate role-playing in the group education sessions.

2.2. Recruitment and data collection 

Patients were recruited from 39 GPs working in one or two-doctor practices situated in disadvantaged neighbourhoods of Rotterdam; 29 GPs actually referred female Turkish and Moroccan patients, who had visited their practices frequently over a 3-months period complaining of pain caused by muscular tension and of other stress-related symptoms as assessed by their GPs. Patients who, despite exhaustive physical examinations did not show abnormalities, or despite having received treatment were not happy with their progress, and for whom the GP's could not think of any further treatment, were recruited for the study. Exclusion criteria were, evidence of serious pathology, physical or mental, and any patient who resided in Turkey or Morocco for more than 5 months, consecutively.

Patients who met the inclusion criteria were divided into an intervention group and a control group who received regular treatment from their GPs, without the help of the migrant health educator. They were randomly assigned to the experimental or control situation once they had given informed consent. A blocked randomisation procedure was followed, randomising patients per general practice, and resulting in experimental and control patients in each general medical practice. This procedure was followed to meet the GPs requests to have access to the migrant health educators. The migrant health educators had not worked in the general practices before the evaluation study had started.

A baseline interview was conducted before randomisation of the patients into two groups. Post-test interviews took place after the conclusion of the intervention's phase 4. The interviews were held in the language of the patient using interviewers drawn from the Turkish and Moroccan ethnic groups. Approval was given for the trial by the Medical Ethics Committee of the Erasmus Medical Centre Rotterdam before the commencement of the trial.

2.3. Measures 

Given the aims of the Bridge Project, the measurement of perceived general and mental health was central in the interviews. Moreover, we measured intermediate outcomes that were aimed at in the individual counselling and group education sessions, i.e. perceived stressful situations, knowledge on pain and stress, social support, ability to cope with pain, and perceived disability due to pain. Perceived stressful situations were measured using 14 questions on stressful life events in the last year, adapted from the biographical probleminventorylist (BIOPRO) [9]. For example: Did you have problems in the relationship with your parents during the last 6 months? (1 no, 2 a little, 3 yes, α=0.69). Knowledge of the relationship between pain and stress was measured using five items, e.g.: Do you think that people who can easily put away their worries, experience less pain? (1 yes, 2 do not know, 3 no, α=0.63). Social support was measured using nine items from a validated questionnaire on perceived social support [10]. For example: Do you feel respected? (1 not at all, 2 hardly, 3 pretty much, 4 very much, α=0.84). A four-point scaled item on coping with pain was asked: How did you cope with you pain problems the past time? (1 bad to 4 good). General health and perceived pain were measured using two items from the short form health survey (SF36): one five-point scaled item on general health (1 very bad to 5 excellent) and one five-point scaled item on restriction of daily activities caused by pain during the last 4 weeks (1 not at all to 5 very much) [11]. Mental health was measured using the 90-items symptom checklist (SCL-90). The SCL-90 is a multidimensional symptom self-report inventory. It comprises 90 items, each measured on a 5-point scale of distress from ‘not at all’ to ‘extremely’. The SCL-90 quantifies psychopathology in terms of nine primary symptom constructs, however it is doubtful that these nine symptom constructs can be generalized cross-culturally. A unitary index of psychological discomfort or psycho-neurotism can also be calculated using all 90 items (1 not at all to 5 very much) [12], [13]. Because in this study the target population of the SCL-90 was Turkish and Moroccan, only the unitary index of psychological discomfort (psycho-neurotism) is presented.

The questionnaire was translated by certified interpreters in Turkish and Arabic and translated back by a native speaker, to ensure translation quality.

2.4. Data analysis 

Differences in background characteristics between the experimental and control group were analysed using the Chi-square test (p<0.05). Multiple regression analysis was used to analyse the effects of participation in the intervention on outcome measures. All scores were added to total scale scores meaning the higher score the better, except for the SCL-90 and BIOPRO. A higher SCL-90 total score indicates decreased psychological health. A higher score on the BIOPRO indicates more perceived stressful life events. Means of the differences between the intervention and the control group on the outcome measures post-test (T2) were tested (p<0.05) using multiple regression analysis with the outcomes at T2 as dependent variables and the baseline scores at T1 as co-variates. The statistical package SPSS 11.5 was used.

4.1. Discussion 

An evaluation study was conducted for the Rotterdam Bridge Project. Turkish and Moroccan migrant health educators helped female patients with muscular pain caused by tension and other stress-related symptoms with a probable psychosomatic cause, according to their GPs. The programme was aimed at improving the general and psychological health of Turkish and Moroccan immigrant women aged between 18 and 65 years. The evaluation study showed a positive effect on the patients’ mental health, general wellbeing and their reported ability to cope with pain. The increase in general and mental health is noteworthy, considering that the perception of stressful situations remained stable among the respondents during the study period. The intervention group reported a decrease in disability caused by bodily pain, however this effect cannot be ascribed to the counselling and information provided by the migrant health educators, as a decrease was also observed in the control group. Apparently patients experience the disabilities arising from pain to be less severe over time. An improvement in the patients’ knowledge of the relationship between stress and pain was not found, because many of the women already had high scores on the scale measuring knowledge. The peer education on the importance of obtaining support from the patients’ environment, i.e. family and friends did not result in effects on perceived social support. The rather diverse total of physical complaints and social complaints may have made it hard to attain effects on all outcome measures.

The non-response-rate of 25% of referred patients and loss-to-follow-up rate of 16% can be considered small, meaning that many women were in need of help to overcome their problems and were motivated to join the experiment. The reasons not to participate were in many cases practical of nature, e.g. lack of time due to work or absence due to a stay in the home country, however the recruitment team believed that some women were not allowed to participate by their families. Those who participated often were poorly educated housewives without a job who had been settled in the Netherlands relatively long time.

The study design used had some limitations. The intervention's effect size could be reduced because all the participating GPs had patients in both groups. The GPs could have used principles that they had learned from their contacts with a migrant health educator in their treatment of the control patients. The observed effects may be attributed mainly to the work of the adjunct migrant health educators, as this was the net exposure of the intervention group, compared to the control group. It is not expected that contamination occurred as a result of contacts between patients from the intervention group with members of the control group. Turkish and Morrocan families are reserved about talking to others about their problems and the help they seek; thus we expect that the migrant health educators’ intervention advise was not passed on within the communities.

The study design does not allow us to draw conclusions on what part of the intervention caused the effects that were found. Further research into the contribution made by the separate intervention elements to the effectiveness of migrant health education for patients with psychosomatic problems is recommended. Moreover a long-term follow-up evaluation study of the effects of counselling and education by the migrant health educators on the patients’ use of GPs should be carried out.

4.2. Conclusion 

Many patients referred by their GP showed a need for the migrant health education intervention. Their perceived general and psychological health and coping abilities improved through the intervention as a whole, despite the constant burden of stress from which the patients suffered. Not all outcome measures have been affected due to among others the diversity of physical and psychological complaints the patients suffered from, non-compliance to the intervention protocol, especially to the group education sessions, the patients’ relatively high level of knowledge on the relationship between stress and pain, and a perceived decrease of disability caused by the pain over time.

4.3. Practice implications 

The effects found and observed needs among the study group support a continuation of migrant health education in general medical practices. However, some barriers have to be overcome. The Bridge Project encountered difficulties in defining psychosomatic problems. At the start of the experiment, psychosomatic problems were defined as muscular pain caused by tension, i.e. headache, pain in neck and shoulders and lower back pain. Later the clause ‘other stress-related symptoms with a probable psychosomatic cause’ was added, because inclusion of patients lagged behind. Many patients suffered from a number of the pains mentioned, however 64% had other complaints, e.g. pains in the feet, knees, pain in the whole body, stomach trouble and vertigo. Often the women could not point to one main physical problem. All patients faced serious social problems, many of which were connected to family circumstances, e.g. maltreatment by the husband, problems in raising their children, disabled relatives and divorce. Moreover, many of the husbands were unemployed and some women had financial problems. The lay migrant health educators found it hard to work with this intervention group, despite the training and support they received from a community mental health service. The educators preferred to reduce their workload. They believed it would help to integrate the intervention methods with their daily activities for other patients with minor health problems. This they thought would make their workload bearable.

An improvement of social support could not be proven in this evaluation study. Social support from the women's environment may not have improved, considering the many social problems the patients endured with their relatives. It may be the direct support from the migrant health educator that contributed to the effects found. The absence of half the patients from the group education sessions may also explain the lack of an effect on social support. Reconsideration of the working methods seen as specific for the programme targets and target group using intervention mapping is advised [14]. Using other methods that are viable, taking into account the situation of Turkish and Moroccan women is to be recommended, e.g. an individual-directed approach aimed at acquiring sustainable social support from outside the family network might be considered.