From epidemiology to intervention for depressive disorders in the general population: the ODIN study

Depression is a common, often chronic condition leading to personal disability and significant socioeconomic costs. Published European studies indicate a point prevalence of depression in adults ranging from 4.6 to 8.8% (1-4). Relapse is common, occurring in up to 75% of cases within 10 years (5). Depression has an impact on the community greater than that of many chronic diseases (6,7). Most research on depression in Europe has been undertaken in urban settings. The few available studies assessing differences in prevalence of depression between urban and rural areas vary strongly in their findings (1,3,8,9). The variety of outcomes could be attributed to differences in the measures used and in selection and sampling of rural and urban areas. The Office of Population Censuses and Surveys (OPCS) survey of psychiatric morbidity in Great Britain (3) reported higher rates of depression in urban than in rural areas but relied on the interviewers' opinion of whether subjects lived in an urban, semi-rural or rural area. For more valid conclusions on the impact of the urban:rural ratio, a comparative European study is needed using the same instruments and research design in different European regions.

Few studies have attempted to influence people with depressive disorders using an epidemiological framework. In addition to direct help by general practitioners (10), there is a need for strategies to enhance self awareness and encourage early help seeking, with the intention of reducing the duration and severity of depressive episodes. Problem solving approaches can be taught easily and have been shown to be effective in health care settings (11-13). Psychoeducational group approaches have been found effective (14), for example as a preventive measure with adolescents (15), and people who are unemployed (16).

In 1996 the European Commission funded the Outcome of Depression International Network (ODIN) study. This project had essentially two aims: a) to provide data on the prevalence, risk factors and outcome of depressive disorders in rural and urban settings within the European Union (EU) based on an epidemiological sampling frame; and b) to assess the impact of two psychological interventions on the outcome of depression, in terms of service utilisation and cost. The ODIN project is the first population-based study that incorporated into its design a randomised controlled trial of individual problem-solving treatment and a group psychoeducation programme. Subjects identified as cases in the epidemiological phase were offered the chance to take part in the controlled trial. This paper will describe the methods of the ODIN study and present a summary of the main findings (see also 17,18).

METHODS

Identification of centres and sites

The following criteria were used to identify suitable centres: expertise in mental health epidemiological research and/or research into strategies for the prevention of depression; access to urban and rural populations; a geographical and cultural spread across Europe. On this basis the project brought together five independent partners in Liverpool, Dublin, Oslo, Turku and Santander. Each centre identified a rural and urban setting in which to conduct the research. The urban areas divide into three large coastal cities (Dublin, Liverpool and Oslo) and two medium size towns (Santander and Turku). The rural areas were defined as having no centre of population greater than 15,000 people, and having at least 20% of economically active citizens engaged in occupations directly related to agriculture, fishing or forestry.

Sampling

Target populations and sampling frames

Adults aged 18-64 were the primary survey population. Community study samples were identified through census registers or lists of patients registered with primary-care physicians. These sampling frames may be considered equally valid in terms of the reliability of the data sets (19). Census registers may be highly reliable at the time of collection, but suffer from attrition over time, and the quality of electoral registers in Britain has been vitiated by their use for local taxation purposes. Primary-care registers, accessed conjointly across a locality, offer a slightly different but equally accurate representation of the population. In the present study, the different research teams involved made the choice to use one sampling procedure or the other - i.e. either census records or lists of patients registered with primary-care physicians - based on their previous experience in community surveys. Three centres (Oslo, Turku and Santander) had previously achieved high response rates through population register surveys, and therefore used this method for first-phase screening in the ODIN study. Subjects in Britain and Ireland were identified through primary healthcare registers, a selection process similar to the one used in the EURODEP study that assessed the prevalence of depression among those aged ??65 at the Dublin and Liverpool centres (20). The Irish research team had to reduce the scope of its intended sampling and interviewing procedures due to operational problems which arose during the study. At the rural site in Ireland, the registers of five general practitioners were involved, out of a total of 27; in Dublin, the registers of two general practitioners were involved, from a total of 390. At the British rural site, seven of the nine practices that covered the population area took part in the study; in Liverpool, 32 practices of the 106 that covered the population area participated.

In Oslo, Turku and Santander, the sample was randomly drawn from the population registers of the five sites (two in Norway and Finland, one in Spain) involved in the study. In Liverpool, a random set of patient names was obtained from health authorities, and interviewers contacted the practices with which the patients were listed. In Northern Wales and Ireland, the procedure was to identify relevant practices, and obtain random sets of names from their patient lists. The entire sample was stratified by sex and age in all the centres.

Assessment methods

First phase

The first-phase assessment identified possible cases of depression using the Beck Depression Inventory (BDI) (21), with a threshold score above 12 (22). The BDI was combined with a questionnaire on social support (23), the List of Threatening Experiences (24), and sociodemographic details. In Santander, the first phase was conducted by home-based personal interview. In all other centres, it was conducted using an initial postal survey, with postal, then telephone, then home-visit follow-ups. All refusals were accepted, and non-responders were contacted up to three times.

Second phase

All of those scoring at or above the BDI threshold and a random 5% of responders were offered detailed interviews with research workers trained in mental health, conducted in the subject's language. To date, most of the epidemiological studies on depression in the general population have used strict definitions of depression, according to DSM-III/IV and ICD-10 criteria, focusing on the prevalence of depressive episodes or major depression. This may lead to a tendency to consider severe/major depression as the only affective disorder worthy of intervention. In order to overcome such a prejudice, in our study we extended the definition of depressive disorders to include dysthymia and adjustment disorders with depressive mood. The Schedules for Clinical Assessment in Neuropsychiatry (SCAN) Version 2.0 (25) was used to generate diagnoses of depressive disorders on the basis of ICD-10 and DSM-IV categories. For ICD- 10, these include single and recurrent depressive episodes (F32, F33), bipolar and persistent affective disorders (F31, F34), and adjustment disorders with a depressive component (F43.2). For DSM-IV, these include depressive, bipolar and adjustment disorders with a depressive component (codes 293.83, 296.xx, 300.4, 309.xx, 311, V62.82). Disability was assessed using the short form of the Medical Outcome Study (MOS) General Health Survey (SF-36) (26,27). Social adjustment was assessed using a modified version of the Social Functioning Schedule (28), and quality of social network by a modified version of Miller and Ingham's 29 questions about close and diffuse support contacts. Cognitive style was assessed using the Problem-Solving Inventory (30) and the Automatic Thoughts Questionnaire (31). The Client Service Receipt Inventory (CSRI) (32) - a validated means of collecting data on a wide range of services, including health and social services and the voluntary and complementary sectors, and on the tasks undertaken by informal care givers - was used to estimate the direct costs of depression in the overall sample, drawing on the model proposed by Rice and Miller (33), and to assess the cost effectiveness for problem solving treatment versus control groups.

Follow-up

Each study subject was offered an interview 6 and 12 months after the initial diagnostic interview. These interviews followed a similar protocol to the initial structured interview to allow cross-checking of the possible effect of illness on the initial assessments. At the six month followup, the Personality Assessment Schedule (PAS) (34) was used to generate five categories of personality traits, if possible with the involvement of an informant other than the subject. The use of the PAS at this stage offered the best trade off between potentially confounding depressed mood and response rate attrition.

Intervention

Respondents identified as cases of depressive disorders were randomly allocated to one of three groups. A community mental health facilitator (MHF) based in each centre delivered the intervention.

1. Individual intervention: Problem Solving Treatment (PST). The MHF contacted each subject randomised to this group, described the diagnosis reached during the detailed interview, and offered one-to-one training in individual problem solving. The procedure has the following stages: identifying and clarifying the problem; setting clear achievable goals; brain-storming to generate solutions; selecting a preferred solution; clarifying the steps necessary to implement that solution; and evaluating progress. It involves six treatment sessions over three months, the first 60 minutes, the rest 30 minutes duration (11).

2. Group intervention: the Coping with Depression Course (CWD). The intervention for subjects randomised to this group was presented by the MHF in a group setting, following the educational model described by Lewinsohn et al (35) and Muñoz and Ying (36), with modifications to increase the content devoted to social support. Within each study site, up to six groups of 6-10 subjects were formed on a locality basis. Each group met for 8 two hour sessions.

3. Control group. No intervention from research team.

Training and quality control

The Diagnostic Interviewers (DIs) were psychiatrists, general practitioners or psychologists. All DIs received an initial weeklong training course at an approved SCAN training centre, and subsequently practised the full diagnostic interview schedule on at least 10 volunteer subjects. Inter-rater reliability over time was monitored by means of assessment and feedback using a standardised videotaped consultation. A videotape including a full SCAN interview was used for this exercise, supplied by the WHO-approved SCAN training centre that trained ODIN's first-phase DIs. Each DI was asked to rate and score the interview, then send his/her score sheets to a central analysis centre (Liverpool). Scores were compared with the 'official' set of ratings which accompanied the video. The videotaped interview contained 113 questions that could be rated, and all 13 of the DIs were included in this exercise. A 100% agreement was reached for overall diagnosis (moderate depressive episode) and for diagnostic category (F32.1). There was 70% inter-rater agreement on scores for individual questions.

The MHFs had qualifications in psychology, nursing and allied health professions. The MHFs providing the individual intervention received an initial two day training session from a team of approved trainers. Each MHF then practised with at least six volunteer subjects. All these sessions were audiotaped with the subjects' permission. Tapes of each second session, together with summaries of the other five sessions and translations if needed, were sent to one of the trainers for feedback. A final training was held in October 1996, during which each MHF's competency to deliver the intervention was assessed. During the trial the quality of the intervention was monitored by continued audiotaping of the study treatment sessions: a random 20% of the tapes of each second session (with translation if needed) were sent to a senior trainer for assessment. Each MHF providing the group intervention participated in an introductory two day course with an experienced trainer. This was followed by running at least one practice group with volunteer subjects. The second session of each group was videotaped and sent to the trainer for assessment and feedback. Quality control during the intervention phase followed a similar pattern.

Field trials

During Autumn 1996 each centre undertook field trials of the screening procedures and research instruments. 318 subjects across the five centres were invited to take part, of whom 172 (54%) agreed to do so. 36 (21%) had BDI scores above threshold. The response rates were highest in Santander, where a completely face-to-face methodology is being employed. Response was lowest in Dublin, where the trial was conducted in a part of the city with extremely high levels of socioeconomic deprivation, and in the rural Norwegian community, where no follow-up was made of trial subjects. The proportion of returned questionnaires successfully completed was uniformly very high. The proportion of subjects scoring at or above BDI threshold was highest in the British and Irish urban areas, and lowest in the British rural area and the Spanish urban area. Following feedback from these trials, modifications were made to methods of following up non-responders, with particular emphasis on techniques for telephone follow-up. The order of the instruments in the screening questionnaire was amended and final decisions were taken about the precise format of the Client Service Receipt Inventory.

Statistical analysis

Epidemiological study

Routine data management and description of the results were carried out using SPSS 7.5 for Windows. Prevalence estimates were carried out using STATA Release 6.0 after allowing for both the two-phase sampling procedure and different response rates across sites through the use of weights (37,38). Information arising from the first-phase screening results and the secondphase sampling mechanism was processed by assigning a 'sampling weight' to each individual subject, given by the inverse of the phase-two sampling fraction.

Randomised controlled trial

Six month and twelve month outcomes were analysed separately. In each case, relevant baseline scores were used as covariates and both Treatment Group and Centre were treated as qualitative factors. Centre effects were treated as either random or fixed. Qualitative outcomes (diagnosis of depression) were analysed using logistic regression. All analyses of treatment effects were carried out using the following procedures from STATA Version 6.0: reg, xtreg (random effects models), logit and xtlogit (random effects models). All analyses were carried out by intention to treat, but with alternative approaches to dealing with the effects of drop-outs (39).

RESULTS

Prevalence estimates

Figure 1 gives the weighted prevalence of depressive disorders (ICD-10 criteria) for survey responders on each site, together with the 95% confidence intervals. An analysis of the combined sample (n= 8,764) gave an overall prevalence of 8.56% (95% CI 7.05-10.37). The figures were 10.05% (95% CI 7.80-12.85) for women and 6.61% (95% CI 4.92-8.83) for men. Rates in Liverpool were more than six times higher, and in Oslo over three times higher, than those in Santander. There was relatively little variation among the four rural areas, with weighted prevalence ranging from 6.1% in Wales to 9.3% in rural Norway. In Britain and Ireland, urban rates were two to three times higher than in rural communities, but in Norway and Finland there was little difference between the urban and rural figures (18).

Figure 1.

ODIN prevalence of depressive disorders in participating centres: diagnostic cases for men and woman (95% confidence intervals) (from Ayuso-Mateos et al [18]).

Trial of psychological interventions

Following the diagnostic interviews, 453 subjects with depressive disorders were entered into a randomised controlled trial of the two psychological interventions (PST and CWD). Of these, 128 were randomised to PST, 120 to CWD and 205 to the control group. The trial outcomes are presented in terms of two domains: acceptability and effect on depression diagnosis.

Acceptability

PST subjects were significantly more likely than CWD subjects to complete the intervention offered to them. Eighty three (65%) of subjects randomised to PST completed the intervention, compared with 56 (47%) of those randomised to CWD (chi square, p=0.0059). There was little difference in the proportions of subjects discontinuing each intervention, once started. However, subjects were significantly more likely to refuse CWD than PST initially: 39% vs. 17%, chi square p=0.0002.

Effect on depression

SCAN data for caseness, on an intention to treat basis, showed that both PST and CWD subjects were less likely than controls to be cases at 6 months. Amongst PST subjects, 39/97 (40%) were still cases at six months, compared with 36/82 (44%) of CWD subjects and 92/164 (56%) of controls. On chi square, differences between PST and controls achieved conventional levels of significance (p=0.0216), whereas those between CWD and controls did not (p=0.1065). At 12 months, the proportion of PST subjects who were still cases was unchanged: 34/85 (40%). The proportion of CWD cases at 12 months had increased to 43/78 (55%), while the proportion of control cases had further reduced to 66/150 (44%). On chi square, the difference between PST and CWD at 12 months approached conventional levels of significance (p=0.075).

A logistic regression was performed to obtain estimates of treatment effects on the diagnosis of depression, using three computational methods: first with centre effects random; second with centre effects fixed; and third with observations weighted to compensate for missing data. Using odds ratios with 95% confidence limits, at six months the first method shows a significant treatment effect for PST alone, the second method shows a significant effect for both PST and CWD, and the third method shows a significant effect for CWD alone. At twelve months, none of these methods show any differences between the two interventions and controls.

DISCUSSION

One of the findings of the present study is the wide difference in the prevalence of depressive disorders across the study sites and between urban and rural centres. Taking the genders together, the centres fall into three categories: high prevalence (urban Ireland and urban UK, 12.3% and 17.1% respectively), low prevalence (urban Spain: 2.6%) and medium prevalence (the rest of the sites: 5.9-9.3%). The study found high proportions of depression among survey responders in some centres, particularly among the female population in urban areas. At seven of the nine study sites, the prevalence of depressive disorders was higher among women than among men, confirming the results of several previous studies (40,41). Over the last few decades, there has been growing evidence of significant inter-gender differences in the rates of specific mental disorders (2). A variety of social and medical factors have been considered in an attempt to explain the higher rate of depressive disorders in women (1,8,41). Further analysis of the data collected in the epidemiological arm of the ODIN study will enable us to test whether some of the gender differences in the depressive disorders prevalence estimates across sites could be explained by different levels of exposure to life events and other social factors.

The apparent variation in prevalence of depressive disorders between the nine centres needs careful assessment. The methodological decision to offer diagnostic interviews to only a 5% sample of subjects below BDI cut-off, led to considerable higher standard errors (and hence wider confidence intervals), than would have been the case if a larger proportion of BDI-negative subjects had been included in the second phase of the community survey. Secondly, there is evidence of systematic differences between survey responders and the populations from which they were drawn. Responders were more likely to be female, and older, than non-responders. Both of these factors may have introduced a bias towards higher prevalence rates amongst responders than amongst the survey populations. This trend may have been compounded in the British sites and in Dublin, where the primary care samples were older than their corresponding census populations. Conversely, the tendency for socio-economically deprived subjects to be less likely to respond - demonstrated for the Liverpool sample - may have introduced a bias towards lower rates of depression than in corresponding community as a whole.

Further analysis of the ODIN sample will enable us to study to what extent urban/rural differences in the prevalence of depressive disorders may be related to differential exposure to life events or differential levels of social support networks, as has been recently proposed (42). In addition, we are currently working on analyses of health care utilisation patterns, in relation to the initial sample and to the interventions, that will provide valuable information on the current treatment coverage for these conditions and its impact on the proposed interventions.

Reasons for apparent differences in acceptability between PST and CWD need to be explored further. The decision to offer different interventions in different sites may have introduced within-study biases, such as differential activity between research teams in organising therapeutic events, or persuading subjects to participate. However in Liverpool, the only site where both interventions were offered, the refusal rates between PST and CWD were markedly different. The ability to provide PST in subjects' homes, whereas CWD subjects had to travel to an alternative venue, may also have reduced the likelihood of participation amongst depressed subjects, who have intrinsically lowered motivation as a consequence of their condition. There were greater time delays in organising group events, which can decrease motivation further. Finally, depressed subjects may feel embarrassed or stigmatised by their condition, and consequently be less inclined to participate in public group activity than a private individual treatment process.

We conclude that PST and CWD may be recommended as effective interventions for people with depressive conditions in urban and rural community settings, on the basis that, at least in the short term, they reduce the severity and duration of depressive disorders, and improve subjective mental and social functioning. Our results should influence the focus of psychological services in primary care, by emphasising the benefits of treatments which are specific, brief, easy to learn and simple to implement. This may have specific implications for the training and employment of practice counsellors. Our findings should also encourage a broader vision for mental health strategies within the general population, providing effective help for people who may previously have been excluded from health care services.

Appendix

The ODIN Group is composed of the academic colleagues and research and administrative staff who have worked on this part of the ODIN Project. They include: Javier Ballesteros, Gail Birkbeck, Trygve Børve, Maura Costello, Pim Cuijpers, Ioana Davies, Juan Francisco Diez- Manrique, Nicholas Fenlon, Mette Finne, Fiona Ford, Luis Gaite, Andres Gomez del Barrio, Claire Hayes, Andres Herrán, Alfonso Higuera, Ann Horgan, Tarja Koffert, Nicola Jones, Marja Lehtilä, Catherine McDonough, Christine Murphy, Anna Nevra, Teija Nummelin, and Britta Sohlman.