Counsellors contact dementia caregivers - predictors of utilisation in a longitudinal study

The concept of CCC is part of the intervention in the cluster-randomised controlled study "IDA". IDA is the German abbreviation of "Dementia Care Initiative in Primary Practice". The study protocol was published [16]. The study region was Middle Franconia, which is a mixed urban-rural area around Nuremberg in South-East Germany [16]. The IDA-study started in June 2006 and ended in December 2008. General practitioners were randomized in three intervention groups and all patients and family caregivers of these general practitioners were therefore gathered in one of these three groups in cluster-randomization. In two intervention groups, two non-medical interventions were compared to a control group with respect to cost and outcome. Control group A (171 patients) got usual care whereas in the two intervention groups B (109 patients) and C (110 patients), GPs recommended support groups (B and C) and offered family counselling beginning at baseline (only in C). The primary endpoint in the IDA-study was the patient's death or institutionalization. The special feature of IDA is that the general practitioner (GP) arranged additional support initiatives for the family caregivers. Usage of the additional offers was, however, voluntary and not requisite for study participation. Four counsellors offered family counselling following a newly-developed guideline ("Counsellor Contact Caregivers" - see Additional file 1, Table S1). The results presented here are taken from intervention group C, since CCC was applied beginning at baseline in this group.

The inclusion criteria for patients were: MMSE [17] score between 10 and 24 points, age ≥ 65 years, support by a family caregiver, living at home and member of a specific health insurance company (AOK). Patients with nursing home placement planned in the short-term or a shortened life expectancy of less than six months and those unable to give written informed consent were excluded. The data used here are medical examination data and data from telephone interviews with family members at baseline. The counsellors provided anonymous information about the general counselling contents. In order to protect the personal confidential relationship between the counsellor and caregiver, the following case-related data were not available for analysis: Resources and deficiencies within the families as well as detailed content and aims of counselling; individual risk factors for terminating care at home and the measures individually recommended by the counsellor (termed "protected data" in the informed consent).

The study was reviewed by the Ethics Committee of the Bavarian Chamber of Physicians (Date of approval: 30/05/2005, Reference number 05029) and is in compliance with the Helsinki Declaration. Signed informed consent was obtained from all patients and caregivers.

In order to meet the goals of patients and family members, who usually wish the patient to live at home for as long as possible; the family counsellors were trained with respect to the risk factors for terminating home care.

The major risk factors for institutionalisation were identified based on scientific literature. Afterwards, guidelines were defined in an expert group (three researchers in cooperation with the counsellors) on how to react if certain risk factors became evident (see Additional file 1, Table S1).

In absence of widespread specialised dementia care services in Germany, the GP is the primary contact person and source of information about further support services for patients with dementia and their families. Therefore it is important that he/she recommends counselling. The main intention of the counselling concept is to give the family caregivers information about the patient's situation, availability of help as well as psycho-educational support.

The four counsellors in the IDA project, who actively approached the family every 6 to 8 weeks by telephone, used elements of case management, when personal contact was established. The counselling process consisted of assessment, target agreement, intervention, monitoring and evaluation. The counsellors were available during normal working hours and offered calls to their clients' homes. They were state-registered nurses or nurses trained in the care of the elderly and all have several years experience in psychogeriatric care. At the beginning of the project, the counsellors participated in training sessions on communication and social insurance code. The counsellors were to attempt to make at least one house call or one personal contact in order to assess the care situation. As customary in other counselling processes, the first personal contact served primarily to get to know one another and to record health status and care needs [14, 18, 19]. Further personal contacts followed when, during the first contact, a sort of "counselling order" was established so that more extensive counselling contacts could be made. With regard to the contents of the counselling sessions, special attention was to be paid to possible risk factors, as described in the CCC guideline, (see Additional file 1, Table S1).The contents were documented in very generalized form for the scientists. The topic areas included: "Patient's physical situation" (General physical status, development of the disease, comorbidities of the patient), "Caregiver's physical situation" (General physical status of the caregiver), "Emotional situation of the patient" (Emotional issues concerning the patient), "Emotional situation of the caregiver" (Emotional issues concerning the caregiver), "General framework" (i.e. provision of medical and other supportive aids, suitable adaptation of housing, financial situation), "Caregiving activities" (i.e. information on ADL, IADL, supervision), "Social support" (i.e. Contact with friends and family or other supportive measures) and "Additional topics of the caregiver" (i.e. Further commitments of the caregiver, for example caregiving for further persons).

The unique characteristic of CCC is that approaching family members should enable earlier planning of assistance than is the case in the "usual" family counselling programs, which can only become active once the family members themselves seek to establish contact.

The patient's medical file contains the following data: Sociodemographic data, severity of dementia, measured with the Mini-Mental Status Examination (MMSE) [17], secondary symptoms of dementia, and comorbidities.

Statements were obtained from family members using a computer-assisted telephone interview. These included: Sociodemographic data, Barthel Index [20], to measure patient's functional independence, two subscales of the Nurses' Observation Scale for Geriatric Patients NOSGER (subscale IADL, instrumental activities of daily living, and subscale disturbing behaviour) [21], the Burden Scale for Family Caregivers (BSFC) [3] to assess caregiver's subjective burden and an extended version of the questions on informal care time of the Resource Utilization in Dementia instrument (RUD) [22, 23].

The following data were provided by the counsellors: date, duration and method of contact (telephone call, personal visit, home visit, written contact, contact with institutions, contact with GP) and up to three main conversation partners (i.e. caregiver, patient). In addition, up to eight much generalized topic areas were documented which were addressed in the contacts between counsellors and family caregivers (see section Intervention). These topic areas were divided into: "Patient's physical situation", "Caregiver's physical situation", "Emotional situation of the patient", "Emotional situation of the caregiver", "General framework, "Caregiving activities", "Social support" and "Additional topics of the caregiver".

For the statistical analysis, Statistical Package for Social Sciences (SPSS^® 16.0) was used. T-tests were used to analyse group differences in the case of continuous variables; Chi-square tests in the case of categorical variables. In order to reveal the relational structure of the 17 potential predictor variables and thus reduce collinearity-related problems during data analysis in a bivariate logistic multiple regression analysis, a factor analysis was first performed using the main component method with subsequent orthogonal rotation.

Five factors were extracted, which were then entered in a binary-logistic regression analysis as independent variables to determine the predictors for usage of extensive counselling. The dichotomous dependent variable was taking advantage of more than one personal contact ("extensive counselling", code 1), or no or only one personal contact (code 0). A p-value < 0.05 was defined as level of significance.

A sensitivity analysis including several interaction terms in the original model was performed. The interaction terms turned out to be statistically insignificant: Factor-1 × Factor-2 (p = .909); Factor 2 × Factor 5 (p = .500), Factor 3 × Factor 4 (p = .161). Therefore the main-effect model seems to be adequate.