| Frequent attenders in primary health care: A cross-sectional study of frequent attenders’ psychosocial and family factors, chronic diseases and reasons for encounter in a Finnish health centre | ||
|---|---|---|
| Prev | Next | |
The study population was a sample of the whole population aged 15 years or older in Oulainen, a small town situated in a rural area in Northern Ostrobothnia in Finland. The computerized population register in the health centre comprises a database of all the inhabitants in the municipalities of Oulainen, Merijärvi and Vihanti. The selection of frequent attenders was based on the computerized annual statistics, which are routinely collected in all health centres in Finland for administrative purposes. The age and sex-matched controls were selected from the same population register. The control group also included some subjects who had not visited a GP at all during the study year. Thus, the controls represented the whole non-frequently attending population in Oulainen.
Previously, most of the studies on frequent attendance have been based on patient populations (Table 2) and only a few on unselected population samples (Wamoscher 1966, Smedby 1974, vab & Zaletel-Krakelj 1993). Several studies have included no control group. In some studies, all non-frequent attenders or such patients divided into categories by utilization or the total population have been considered as controls (Table 2). In some studies, the control group or groups have been poorly defined, and only in some studies has an age- and sex-matched defined population been used as a control group (Hood & Farmer 1974, Westhead 1985).
The study population represents the population of a small Finnish town, and this fact must be borne in mind when generalizing the results to the whole Finnish population utilizing primary health services.
The study was a cross-sectional controlled study with equal numbers of frequent attenders and age- and sex-matched controls. The collection of study data can be divided into three phases, which were carried out during the years 1995–1996 (Figure 1 of Paper II). The examination of medical records and the collection of data of the study population from the annual statistics took the year 1995, during which the mail inquiry was also carried out. The personal interviews were carried out by two trained nurses during the year 1996.
The interviewees were selected from the original list in the date of birth order, so that every second person from the list of frequent attenders and from the list of controls was selected, to maintain the representativeness of the sample.
The cut-off point of eight face-to-face contacts with GPs in the health centre per year was used as a criterion of frequent attendance in this study. Previously, some researchers have used the upper quartile (Courtenay et al. 1974) or decile (Westhead 1985, Von Korff et al. 1992, vab & Zaletel-Krakelj 1993) of visits stratified by age and sex as the criterion of frequent attendance, although most of the previous studies have used cut-off points ranging from five to 20 consultations (Table 2). In Finland, Larivaara (1987) used eight visits to GPs as the cut-off point of frequent attendance in his study in a rural health centre in Northern Finland. The criterion of frequent attendance in the study of Karlsson (1996) was 11 visits (including the visits to physicians other than GPs in a health centre).
The method of selecting frequent attenders was feasible and reliable, but it assumes the use of a certain cut-off point of consultation frequency as the criterion of selection. Methodologically, a certain cut-off number of visits is an easier way to identify the frequent attenders from the statistics, but the method does not take into account the differences in the frequency of consultations between males and females and between various age groups. Females and the older age groups are over-represented in the study population, and the results are not fully comparable with all previous studies. In this study, the visits in the private medical sector were omitted, because the public health care facility is the primary way to access health care services in Oulainen. In addition to this, the gathering of reliable statistics from private health care would have been laborious or impossible.
The inclusion of non-attenders in the control group must be borne in mind when comparing the results with the results of the earlier studies. However, in the study of Browne et al. (1982), the non-attenders (zero users) differed significantly from the modal users only by having a higher internal locus of control.
The statistical information of the utilization of health services in primary and secondary health care can be regarded as reliable and free of systematic bias. The information of the medical records in the health centre was gathered and coded by the author himself, which means that there is a certain risk of subjectivity when coding the information on, for instance, the ICD diagnoses of chronic diseases or the main ICPC-based reasons for encounter. The author could not be “blind” to the thickness of the medical files of the frequent attenders, either.
The problems of reliability implicit in the use postal questionnaires in general have been previously discussed (McDowell & Newell 1987, Hyyppä & Kronholm 1994). Nevertheless, the postal questionnaire was used to gather the basic sociodemographic data and information about the housing, family and health matters. The response rate in the mail inquiry was satisfactory and equal in both study groups. Although the non-responders were younger and more likely men than responders, there were no significant differences between the non-responders and responders in the sociodemographic backgrounds.
The structured personal interview and the included questionnaires aimed to gather confidential and intimate information about family, social support, health behavior and psychological items, which could have been difficult to obtain otherwise. The reliability of the interview phase was enhanced by using two trained nurses to minimize both systematic and random errors. The participation rate of the interview phase was quite equal and high enough in both study groups. The non-participants were younger and more often men than the participants, but there were no significant differences between the non-participants and participants in the number of visits or the basic sociodemographic characteristics.
Validated measures were used to assess the health, health-related quality of life and psychosocial and family factors of the participant. The Finnish version of NHP has been validated among the Finnish population as a measure of health-related quality of life (Koivukangas et al. 1995). The CMI health questionnaire has been used earlier in Finland to determine the severity of mental disorders (Väisänen 1975, Lehtinen et al. 1993). The SCL-36 questionnaire is a reduced version of the widely used SCL-90 questionnaire, which has been validated in a Finnish community sample. In this validation study, the Finnish interviewees scored consistently higher on all subscales of SCL-90 than the American population or patient samples, and this difference must be borne in mind when comparing the results with international studies. (Holi et al. 1998.)
As a criterion of somatization, we used the cut-off point of eight symptoms out of the 12 listed on the SCL-36 somatization subscale. We did not use any psychiatric diagnostic interview to validate our cut-off point. Previously, various criteria have been used to define somatization. Escobar et al. (1987, 1998) used four symptoms for men and six symptoms for women in DIS to define “abridged somatization”. Kroenke et al. (1997,1998) defined the multisomatoform disorder as three or more medically unexplained physical symptoms and a history of two years of somatization. According to their study, the optimal threshold for pursuing a diagnosis of multisomatoform disorder, was seven or more symptoms (Kroenke et al. 1998). Portegijs et al. (1996) used five symptoms on the DSM III symptom list as a cut-off point for somatization among frequent attenders in his study.
The validity of TAS-20 has been shown to be relevant (Parker et al. 1993) and the Finnish translated version has also been validated (Joukamaa et al. 2001). WI is a reliable instrument for assessing hypochondriasis among primary health care patients (Wyshak et al. 1991, Speckens et al. 1996). Previously, DAS has been found to be a reliable measure of marital adjustment (Spanier 1979), and MCI has also been validated in a Finnish population (Kinnunen et al. 1975).