The 11 cross-sectional studies are presented in Table 1. The cross-sectional studies showed; Firstly, a relation between emotion-oriented and avoidance coping strategies and the prevalence and symptom severity of depression. Secondly, the cross sectional studies cannot show whether these strategies were associated with first episode or with relapse of a depressive or manic episode.

The seven short-term studies are presented in table 2. Short-term studies are here defined as studies with a three- to six-month follow-up period. The short-term studies indicated: firstly, in four of the seven studies that emotion-oriented and ruminative coping style were state dependent 27293132, secondly, in four studies these coping styles were related to the severity of the depressive symptoms 27293031, thirdly, three studies showed that recovery from depression were associated with changes in coping style 272931, accordingly, recovered patients relied less on inappropriate emotion-focused coping strategies and finally, two studies showed that emotion-oriented coping predicted the presence of depression at follow-up 2830. Conversely one study showed that ruminative coping might be associated with a good outcome in depressed patients 33.

The 10 long-term studies are presented in table 3. Long-term studies are here defined as studies with a follow-up period longer than six months. The long-term studies indicated: firstly, that adaptive coping strategies predicted increased remission and decreased risk of relapse in half of the studies 3537394142. One study concluded that depressed persons did not differ in amount of problem-focused coping compared to control persons 34; another study did not find that problem-focused coping was a predictor of post-treatment symptoms 38. Secondly, the use of passive and ruminative coping strategies and coping by avoiding other people was found to be related to a higher risk of relapse and of greater severity of depressive symptoms at follow-up in six of the ten studies 343537394042. Thirdly, one study showed that a high level of stressors was connected with higher depression scores at follow-up 37. Finally, the study on bipolar disorder showed that patients that used stimulating coping strategies more often had a manic episode 42

Regarding the two hypotheses none of the studies were capable of answering whether certain coping strategies are predictive of affective disorder. Most studies showed that in depressive disorder avoidance and emotion-oriented coping strategies seem to be related to recurrence/relapse of depressive episodes. These coping strategies might also be associated with an increased time to recovery. Only a few studies on bipolar disorder were identified, limiting the findings to depressive disorder mainly. In the cross-sectional and short-term studies it is difficult to make a clear distinction between coping processes and symptoms, and avoidance and emotion-oriented coping styles seemed to be state related, so the results of these studies are not able to answer our hypothesis.

Regarding the trait or state effect, the use of cross-sectional designs and designs with a short follow-up period can be problematically because of the possibility of measuring residual depressive symptoms. The coping style of an individual may change during a depresive or manic episode, as illustrated in the short-term studies. Possibly, the emotion-oriented coping style found especially in the cross-sectional studies could be explained by a state-dependent consequence of a depressed mood. One way to determine whether the anomalies are caused by the depressive or manic episode (state model) is to include a control group. Among the 28 studies, 12 included a control group. In the follow-up studies, the lack of a control group makes it difficult to distinguish whether the changes that occurred reflect normalization of coping over time or whether the changes reflect changes in affective symptoms. The best impression of the relation between coping and state comes from the follow-up studies that used repeated measures of affective symptoms, life events and coping behaviour and include a control group 343536374043. The studies with a long follow-up time 37404143 illustrate the time-relation between coping and affective episodes: these studies support the hypothesis that emotion-oriented and avoidance coping strategies seem to be associated with increased risk of relapse. So, the controlled follow-up studies provided the most valid study design to comply with methodological problems.

It is difficult to compare studies involving heterogeneous samples and it is likely that selection bias influenced the results unpredictable. Most studies used clinical samples of unipolar patients. A general limitation is the lack of attention to confounding variables such as the duration and severity of the mood episodes and the number of previous episodes. Some studies investigated these parameters 21232830394041 and these results did not deviate substantially from those of other studies. The use of passive coping strategies predicted relapse of depression and stimulating coping strategies increased the relapse risk of mania but from the limited number of studies concerning bipolar disorder 2042 it is not possible to elucidate the relation between coping and bipolar disorder. The little attention to research into coping mechanisms and mania was already stressed in a case report in 1990 45 and latest by the authors in a research report from 1999 where the development and validation of the coping inventory for prodomes of mania was described 46.

Participants were registered differently according to comorbid psychopathology. In one study the most common comorbid disorders were panic disorder, dysthymia and social phobia 26. One study concentrated on comorbid agoraphobia 39 and one on comorbid personality disorder 44. In some studies patients with comorbid psychopathology were excluded 1621222829. Although the relation of coping processes and risk for substance abuse is a defined area of coping research 47, only two studies 3943 adjusted for the possible effect of comorbid substance use as a part of a coping strategy. This may have affected the results so that some inappropriate coping strategies are not identified.

As seen from the tables, the studies measured coping differently, which makes it difficult to compare across studies. The coping measures were based on self-reports and a variety of questionnaires were used without any golden standard. It is important to develop reliable and valid tools that assess how people cope with stressful situations and negative events. Therefore developing structured and easily scorable scales started in the 1960s 48 and during the 1980 and 1990s, – much research was conducted on self-reported measures of coping. Facing particularly stressful events, individuals are asked about the kind of coping behaviour(s) in which they have engaged 49. This research used widely divergent strategies, techniques and measures; additionally some of the scales possess a variety of psychometrically drawbacks 13. When examining the coping questionnaires, it is clear that substance use as a coping behaviour has virtually been ignored 13275051. The two studies concerning abuse showed that depressive symptoms were significantly associated with more drinking problems 43 and that improvement in functioning and well-being was associated with less alcohol consumption 39.

As discussed, in a critical review concerning the gap between coping theory and clinical intervention research 52, when using a coping checklist and thereby reducing coping to a summary score, a lot of information will be lost. The questions seem to be to general for the results to be valid and crucial aspects of timing, sequencing and appropriateness may be lost.

There is a substantial and partially causal link between life-events and depression 53 and the occurrence of major life events signals a period of increased risk of developing a depressive episode 54; several studies measured life events 171821232426272934353637394344. A general problem in measuring coping in connection with life events is the time relation, as participants should recall a stressful situation and then reconstruct how they dealt with it. Firstly, there is recall bias and secondly, coping has a temporal aspect: an individual can cope before a stressful event occurs, while it is happening and afterwards 55. The issue of timing is complex as there presumably may be an interaction between individual vulnerability, coping, life events and the risk of developing depression. Thus it is difficult to measure the coping process given the typical brief period between a severe life event and depression. Capturing the coping process represents a challenge 56 and it may require a more frequent monitoring of coping than is usually done in studies.

All coping measures are based on self-reports and participants are often asked in connection with hypothetical stressors e.g. the Coping Inventory for Stressful Situations 13 and it is difficult to know what kind of stressors were thought of. Many scales for coping are developed to assess overall coping strategies but it reasonable that human cope differently according to the nature of the event. Linking coping measures to recent actual stressors and attempting objective assessment of the stressful life event to avoid circularity is of main importance. The procedure of semi-structured interviews is time- and labour consuming and might be a reason for the consistent use of self-reports. The lacking use of coping in clinical assessment and treatment may be due to the rather confusing divergent coping scales. It can be rather meaningless to detect coping not knowing what kind of stressor the coping process relates to and the influence of personality traits cannot be ignored 56. For advancing research and clinical focus on coping there is a need of developing an integrated semi-structured interview detecting personality traits, life events and coping together. Beside, a specific development of coping schedules capable of identifying prodomal symptoms of bipolar episodes as described by Wong et al. 46 would be of clinical use.

Coping is closely related to personality and personality can affect coping measures and in coping as well as in personality research there is a distinction between trait and state variables. In the concept of coping responses 57, coping is seen as a response to specific stressful situations rather than a stable feature of personality. Another approach emphasizes the method of coping; whether a response entails primal cognitive or behavioural efforts 58. One study 26 showed that less-adaptive coping strategies (emotion-oriented coping) were associated with neuroticism and depression, whereas the reverse association was found regarding adaptive coping strategies (task-orientated coping). These findings are replicated in studies of the general population 5960. Finally, a study 61 from Japan found that task-oriented and avoidance coping were related to extraversion and that emotion-oriented coping was related to neuroticism. Personality traits are important in the assessment of coping and it is advisable to measure these traits when dealing with coping measures; however, only three studies have done so 264144.

In most of the present studies more women were included reflecting the gender differences in the prevalence of depressive disorder. In some studies gender differences were found and the general tendency was that men tend to distract themselves using active coping strategies, whereas women use strategies involving expressing emotion 162125. Other studies found no gender differences 223033; however, most studies did not take gender into consideration in the analyses. According to the hypothesis of Nolen-Hoeksema 62 the increased vulnerability of women to developing depression is related to gender differences in coping; men's response to their dysphoria is more behavioural and dampens their depressive episodes, whereas women's response to their dysphoria is more ruminative and amplifies them. Accordable to a review 63 of gender differences in depression, it is possible that men tend to distract themselves from their mood by engaging in physical or instrumental activities, whereas women are less active and ruminate over the possible causes and implications of their depression. These hypotheses are compatible with findings from other studies 646566 the latter study involving data covering representative population samples from six European countries. Conversely, an older prospective one-year study 67 of 100 healthy community-residing men and women, in which participants were interviewed seven times at four-week intervals, found no gender differences in emotion-focused coping. Another study 68 of couples that recently had experienced at least one threatening life event that was potentially depressonigenic for both showed that women had a greater risk than men of depressive episodes following the life event. The greater risk was restricted to episodes that followed events involving children, housing or reproductive problems. Women's greater risk was only present among those couples for whom there were clear gender differences in associated roles.

Age differs between the studies but most studies had a broader age range. How age influences the stress and coping process is not clear: The Normative Aging Study 69, a longitudinal study examined stress, appraisal and coping in three groups' middle-aged, young-old and old-old men. The study followed 2280 men for more than 30 years. A significant overall effect of age on coping strategies: instrumental action, cognitive reframing, social support and interpersonal hostile strategies were found and all coping strategies showed linear decrease with age. The relation between age and coping is complex and there is no clear answer as to whether persons cope better or worse as they age. In a recent study, the association between life events and onset of depression and mania was not found to change throughout life 70

The episodic course of affective disorders presents a challenge to the clinician and the patient, in making and fulfilling a treatment plan. Effective treatment depends on medication adherence also when treating medical patients with comorbid depression. When treating affective disorder, doctors and patients have to deal with different problems: lack of insight, symptom-free intervals, residual affective symptoms and poor social support, which can complicate long-term treatment 71. Identification of these factors may add to increase medication adherence as seen in a study 72 that investigated medication adherence in 32 patients with bipolar I disorder. Consistent with the hypothesis that acceptance coping would be positively associated and denial coping negatively associated with adherence to mood-stabilizing medication, it was found that low levels of acceptance and a high level of denial undermined medication adherence.

Coping strategies could be a target for selective prevention targeting subgroups of the population whose risk of developing an affective disorder is higher than average. Further, a way to ameliorate the course in affective disorder is to help patients to identify prodomal symptoms and individual maladaptive coping strategies and try to change these, e.g. problem training as a treatment for depression 467374 or intervention designed to improve coping. More recent literature supports the utility of individual cognitive behavioural and psycho-educational approaches, particular in enhancing medication adherence, so medication and psychotherapy are not only compatible but also synergistic 75 and a strong focus in psychological treatment involves structured attempts to teach patients new coping skills.