In this close to community-based study we investigated self-rated health, symptoms of depression and general symptoms in survivors of a first-ever stroke 3 and 12 months after the event. After 3 months, 62% of the patients rated their health as very good or rather good, and this was the case for 78% after 12 months. In another community-based follow-up study on first-ever stroke patients' self-rated health, 60% rated their health as excellent, very good or good 12 months after stroke . However, these figures may not be directly comparable with ours, since the wording in our five-step scale for self-rating of health is symmetrical (middle value "neither good nor bad"), while their five-step scale is asymmetrical (middle value "good") which conforms to the wording in the first question in the quality of life assessment instrument SF-36. When self-rated health among people aged 75 years or older was assessed by use of the GQLI in a Swedish epidemiological study, 87% of the responders rated their health as very good or rather good, which is significantly better than in our study .
Although the majority of patients were totally, or close to totally independent regarding ADL both before and after the stroke, 21% had a higher degree of dependence 12 months after stroke. Self-rated health was strongly associated with ADL, which corresponds to the findings in other studies, where, for example, the impact of ADL-dependency on quality of life for post-stroke patients was shown in a 2-year follow-up study where quality of life was assessed by a VAS-scale , and in a 3-year follow-up study where it was assessed by the SIP-instrument (Sickness Impact Profile) .
More than half of the patients in our study had symptoms of depression at some time, and there was no change in frequency between 3 and 12 months. In a recent meta-analysis study on frequency of depression after stroke that is based on 51 studies, six of which were population-based, the pooled average of the population-based studies was 33% (95% CI, 0% to 72%) in the early phase (1–5 months after stroke) and 34% (95% CI, 24% to 43%) 6 months or more after stroke . Instruments for assessing depression differ, especially regarding patients classified as having symptoms of mild depression, and these patients may be included because of symptoms due to old age or physical diseases, and not depression. One observation from this meta-analysis study is that the highest frequencies of depression were found in the studies that used MADRS as the assessment instrument (pooled average 41%; 95% CI, 23% to 60%) . Compared to other depression rating scales, MADRS-S has relatively fewer somatic items, and is thus less sensitive to concomitant physical illness . This would be favourable in the present context, but it is probable that neurological and cognitive sequelae of stroke would nevertheless overlap with items in the MADRS-S scale. In a Swedish population-based study of depression in the oldest old (aged 85, 90 and 95+ years), the prevalence of any degree of depression, measured by the Geriatric Depression Scale-15 in combination with MADRS, was 26.9%. The 85-year-olds had a significantly lower prevalence than the two older groups (16.8% compared with 34.1% and 32.3%) .
Fatigue and sadness were among the most common symptoms at both 3 and 12 months after stroke. In another study, fatigue was reported by 68% of the patients 3 to 13 months after stroke, which corresponds well with our results . In a recent two-year follow-up study of post-stroke fatigue in Sweden, 39% of the patients responded that they always or often felt tired, and this was associated with feelings of depression and deterioration in several aspects of everyday life . The prevalences of general symptoms 12 months after stroke were surprisingly similar to the prevalences in a Swedish population (age range 16–92 years) . The stroke patients had higher prevalences of anorexia, loss of weight, dizziness, impaired hearing, constipation, pain in the legs, sweating and breathlessness, a lower prevalence regarding headache, and about the same prevalences of fatigue and sadness. According to another Swedish study, some symptoms decrease with age (fatigue, abdominal pain, nausea, diarrhoea, cough and headache), while others increase (sleeping disturbances, pain in the joints, pain in the legs, breathlessness and impaired hearing) . Most of the differences could thus be explained by the high age of the stroke patients.
Strengths of this study include the thorough, prospective search for patients, the absence of exclusion criteria, the use of a number of relevant outcome measures, and the use of everyday clinical records. It is thereby likely that these data provide a realistic picture of current practice in the area, which is sociodemographically representative of Sweden. However, it should be kept in mind that retrospective data from records only give information about what was actually recorded.
A limitation of the study is that the response rate for the interviews was rather low. One reason for this might be that patients with cognitive deficits, aphasia or severe handicaps were not excluded a priori in our study, and this may have influenced both the rate and quality of the responses. However, cognitive impairment or dementia were not more common among the patients who did not take part in the interviews and assessments at both 3 and 12 months after stroke than among those who did take part. The response rate was somewhat lower for the MADRS-S than for the GQLI. Our interpretation is that it was easier for patients to assess the presence or absence of symptoms (GQLI) than to use a rating scale to evaluate rather complex descriptions of symptoms and/or feelings (MADRS-S).
A potential limitation could be that despite our multiple search strategy, the search for cases might have been insufficient. However, there is a very strong tradition in Sweden and the other Nordic countries to refer patients suspected of having suffered a stroke to hospital for CT or MR imaging and, if possible, antiplatelet treatment. National, regional and local guidelines stress the importance of referring all patients with recent (less than a week) symptoms compatible with stroke or transient ischaemic attacks directly to hospital for further investigations and evaluation . Recent Swedish studies using community-based stroke registers show that only about five percent of first-ever stroke patients have not been in contact with hospital, and these comprise mainly patients living in nursing homes [34–37]. General practitioners' intensified efforts over a period of time to refer all cases of stroke did not significantly increase the incidence rates . Stroke patients whom we might have failed to include could thus be nursing home patients with concomitant severe diseases, or patients who did not see a doctor for their symptoms. However, because of the multiple overlaps in our search strategies, we are convinced that the study reflects the stroke situation in the community.
This study of first-ever stroke in a municipality provides a comprehensive picture of the well-being of these patients, who are often fragile in many respects and in need of different kinds of support. The clinical picture of these patients is complex and includes both neurological and less specific symptoms, and psychological symptoms may consequently be overlooked. To facilitate the continuity of patient care across different levels of care, it is important that doctors and other health care personnel not only communicate information about patients' neurological handicaps, treatments and co-morbidity, but also information about their moods, general symptoms, and worries and concerns about their own health. Many stroke patients suffer from fatigue and sadness at some time during the first year, and it is important to be observant in this regard and to investigate these findings further, since some patients may benefit from pharmacological treatment for depression. Further research is needed on the usefulness of these rating instruments in clinical settings.