| The association between atopic disorders and depression: The Northern Finland 1966 Birth Cohort Study | ||
|---|---|---|
| Prev | Chapter 6. Discussion | Next |
The major strengths of this study was that it was based on a large, unbiased and genetically homogenous Northern Finland 1966 Birth Cohort database: In the provinces of Oulu and Lapland, 96% of all women (n = 12068), with an expected date of delivery falling between 1st January and 31st December, were evaluated. These women gave birth to 12058 liveborn infants. A variety of biological, socioeconomic and health conditions as well as living habits and family characteristics of the cohort members have been collected prospectively from pre-natal stages up to the age of 31 (Rantakallio 1969, 1988; Rantakallio et al. 1995, 1997).
Hospital diagnoses, with regard to atopic disorders and depression, were obtained from the national hospital discharge registers, which cover all treatment episodes in general, mental, military, prison, and private hospitals as well as the inpatient wards of local health centers nation-wide. These register data have been shown to be a reliable tool for epidemiological research (Poikolainen 1983, Keskimäki & Aro 1991).
In addition, the diagnostic accuracy concerning hospital-treated depression is good due to the validation process. All cohort members, having been admitted to a hospital during 1982-1997 for reasons of a psychiatric disorder, were identified from the FHDR and their hospital case notes were collected and scrutinised; The FHDR diagnoses were carefully validated by re-checking them against DSM-III-R criteria (Isohanni et al. 1997, Moilanen et al. 2003).
Skin prick tests are known to be one of the best objective assessments for demonstrating an IgE-mediated mechanism underlying clinical atopic symptoms (The European Academy of Allergology and Clinical Immunology 1993). In the original publications II–IV, the detection of a cohort member’s atopic disorder was based on positive responses from skin prick tests, and in this respect only a very small number of false positive cases may have been included in the analyses.
Finally, the HSCL-25 has been proved to be an acceptable screening scale in obtaining information on symptoms of depression in the normal population (Sandanger et al. 1998). In addition, the HSCL-25 was found to be a valid instrument in a Nordic multicentre investigation (Fink et al. 1995) as well as a Finnish one (Veijola et al. 2003).
There were several limitations that need to be commented on in this thesis. As in any large population sample, only part of the potential exposure or confounding variables could be taken into account. First, it was not explored whether the association between atopic disorders and depression is specific to atopic disorders, or more generally to virtually any chronic physical disorder (for a review, see Katon 2003). Among the major weaknesses in this study were also that solid data on immunoglobulin-E or cytokine and other immunomediator profiles were lacking.
Further, the overall number of the subjects suffering from hospital-treated atopic disorders was rather low. Likewise, the number of hospitalisations due to depression was also rather low in both genders and this imposes restrictions on statistical analyses. In general, in the database of the FHDR a small amount of missing or erroneous personal identification numbers are known to exist, involving the treatment periods from the 70s to the early 80s and amounting to 15% and 10%, respectively (Näyhä 1992). Since then, thanks to computerized hospital registers and automatic checking systems for the personal identification number, there are practically no erroneous data in the FHDR any longer (Mähönen et al. 2000). In any case, there is no reason to assume that the missing and erroneous data in the FHDR would have been unevenly distributed in the study populations of this thesis.
With regard to self-reported lifetime doctor-diagnosed depression, the definition was based on a single question obtained from postal questionnaires. However, treatment-seeking had earlier been found to be a good predictor of lifetime depression (Kendler et al. 2001b). Further, the findings of this thesis about the gender differences can be biased due to the fact that cases of depression are more common among females and that the validity of diagnoses through questionnaires are inferior to those of structured diagnostic rating scales; it is also possible that men have underreported their depressive disorders.
Self-report rating scales such as the HSCL-25 conducted at one time point (in this study at the age of 31) give a very limited approximation of lifetime depression of all cohort members. However, in the original publication III, in addition to HSCL-scores, the data on self-reported doctor-diagnosed lifetime depression could also be used. In spite of the fact that the exact diagnosis of the self-reported depression was not available and that further it was also not known when exactly a cohort member had been treated by a doctor for depression, the findings of the original publication III about the highly symptomatic cohort members (i.e., having HSCL-25 depression subscale scores in excess of 2.0), together with the doctor-diagnosed depression, mean that those cohort members have very likely had, before the age of 32, a major depression – either a single episode or a recurrent one. It must also be remembered that self-filled rating scales may include items such as “crying easily” and “lost interest in sex”, which have recently been suggested to give too high a score of depressiveness in females when screening depression (Salokangas et al. 2002).
It was unfortunate that the temporal sequence could not be taken into account in the statistical analyses of the original publications. The variables for “symptoms of allergies” and for “self-reported doctor-diagnosed clinical histories of allergic disorders” did not include the information on the onset age of allergic symptoms or disorders.
Finally, the definitions of parental atopy and a sibling’s atopy were also based upon postal questionnaires and objective measurements, such as skin prick tests, were not available. This can lead to a biased definition of familial atopic disorders. In general, using self-administered questionnaires may easily lead to an underestimation rather than overestimation of the phenomenon studied. However, the procedure does facilitate comparisons with earlier epidemiological studies of atopic disorders, for most of them have also used written questionnaires (Burr 1992, Remes 1998).