Validation of the Edinburgh Postnatal Depression Scale- Previous findings in another sample of Greek women and review of the differences with the newer results.
Angeliki Leonardou, University of Athens, 1st Department of Psychiatry, Eginition Hospital
11 January 2010
Dear Sirs, In the interesting paper The Edinburgh Postnatal Depression Scale: translation and validation for a Greek sample, the authors mention that “the purpose of this study was to validate the Greek version of the EPDS…”. We would like to inform you that our group has already published a validation of the EPDS based on a different translation and with a different methodology, which has been published in the Journal of Reproductive and Infant Psychology [1]. The fact that our study was not mentioned in the paper by Vivilaki et al. (whereas a study using our translation was referenced therein) [2], might not be significant if the two studies did not produce such different results. We believe that it is important to review here the methodological differences between our study and the study by Vivilaki et al., so that the clinicians and researchers interested in utilising the Greek EPDS can decide on which cut-off they should use in evaluating scores.
Our study was conducted in a sample of new mothers who have given birth in both private and public large obstetric hospitals in Athens. We approached 109 women, of which 95 (87%) agreed to participate in the study. Our initial recruitment of the study subjects took place on the second postpartum day, where initial questionnaires, as well as demographic data were collected. The reassessment of the study subjects took place in the 8th postpartum week. At that time the researcher, blind to the results of the initial questionnaires, conducted a semi-structured interview (SCID), to ascertain the presence of major depression. The diagnosis of minor depression, based on DSM-IV research criteria, was added to the assessment, as it has also been done elsewhere, as 50% of the cases, if unnoticed and untreated, will later suffer from major depression [3]. In the validation study by Vivilaki et al. the recruitment of the study subjects took place between 4 days till 16 weeks postpartum. This poses a major issue regarding the diagnosis of postpartum depression. Newly delivered mothers at 4 days postpartum could be suffering from maternity blues, whereas the women assessed in the 16th week postpartum, could indeed be suffering from postpartum depression. “Postpartum blues”, also known as “5th day blues” is a condition affecting 30-80% of women during the first 3-15 postpartum days, it is significantly influenced by the hormonal fluctuation and it is usually self-limited [4]. Postpartum depression usually manifests itself at 6-8 weeks postpartum. This to our opinion poses an issue on the homogeneity of the group assessed for depressive symptoms in the Vivilaki et al study. Our results showed 5% major and 7.4% minor depression, which is consistent with the percentage found by other authors [5]. The high 31.7% diagnosis of mild depression in the BMC study could be accounted by the inclusion of maternity blues cases.
In our opinion this difference accounts for the rather unexpected difference in the cut-off points in the two Greek validation studies: our validation study shows a cut-off point of 11/12, whereas the newer validation study showed a cut-off point of 8/9, which in our opinion is too low and would include a lot of false positive cases. Unpublished data of our study have shown that the EPDS on the second postpartum day has poor psychometric properties, regarding the ‘prediction’ of postpartum depression in the 6-8th week postpartum. ROC analysis showed that at a cut-off point of 8/9 had a sensitivity of 90.00% and a specificity of 53.52%, which means that one of the two women identified by the questionnaire as possibly depressed would be falsly alarmed. So we have found that the use of the EPDS on the first postpartum days is not encouraged, as it has already been shown elsewhere [6]. In contrast to our procedure, the validation study by Vivilaki et al. has validated the questionnaire against another, already validated questionnaire, a procedure as well already adopted by others. However it is well known that using questionnaires to diagnose depression shows a higher prevalence of the disorder [7]. This fact along with the timing of the assessment could account according to our opinion for the cut-off point of the scale being found in this newer validation study a lot lower than in the initial one.
Finally, there is another point to be made, regarding the use of the scale solely as a screening instrument. In the current paper that we comment on it is mentioned that the Greek EPDS could be used by Greek health professional for diagnostic purposes. However as the National Screening Committee [8] has already clearly stated, as well as the constructors of this questionnaire [9], we would like to suggest that it is our responsibility to teach health professional willing to use this screening questionnaire that it is NOT a diagnostic tool, and thus it can only identify possibly depressed subjects, but can not and should not be used for making the diagnosis of depression.
We do not believe that the different translation of the EPDS used by Vivilaki et al. played a significant part in the results reported. The EPDS is a fairly easy scale to translate. However, it might be confusing to users to have many different versions of the same scale, however small differences might be. Our translation came from a previous version used in Greece for over a decade, and was in a study that was published in the same journal our validation study was published [10]. It is also the only one referenced by the originator of the EPDS scale, Prof. J. Cox [11]. Our translation has already been utilized by three other Greek studies [12, 13, 14]. All the above mentioned publications can be found through the Internet when the key words in Google search engine are ‘EPDS, Greek’.
Best Regards Angeliki A. Leonardou, Ioannis M. Zervas
References 1. Leonardou AA, Zervas YM, Papageorgiou CC, Marks MN, Tsartsara EC, Antsaklis A, Christodoulou GN, Soldatos CR: Validation of the Edinburgh Postnatal Depression Scale and prevalence of postnatal depression at two months postpartum in a sample of Greek mothers. Journal of Reproductive and Infant Psychology, 2009, 27 (1): 28-39. 2. Gonidakis F, Rabavilas AD, Varsou E, Kreatsas G, Christodoulou GN: A 6-month study of postpartum depression and related factors in Athens Greece. Comprehensive Psychiatry 2008, 49: 275-282. 3. Horwath E, Johnson J, Klerman GL: Depressive symptoms as relative and attributable risk factors for first-onset major depression. Archives of General Psychiatry, 1992, 49: 817-823. 4. Kennerley H., Gath D: Maternity blues. I. Detection and measurement by questionnaire. British Journal of Psychiatry, 1989, 155: 356-362. 5. Gavin N, Gaynes B, Lohr K, Meltzer-Brody S, Gartlehner G, Swinson T: Perinatal Depression. A Systematic Review of Prevalence and Incidence. Obstetrics and Gynecology, 2005, 106 (5): 1071-1083. 6. Lee DTS., Yip ASK., Chan SSM, Tsui MHY, Wong WS, Chung TKH: Postdelivery screening for postpartum depression. Psychosomatic Medicine 2003, 65: 357-361. 7. Garcia-Esteve L, Ascaso C, Ojuel J, Navarro P: Validation of the Edinburgh Postnatal Depression Scale (EPDS) in Spanish mothers. Journal of Affective Disorders, 2003, 75: 71-76. 8. National Screening Committee: www.nelh.nhs.uk/screening. 2001 9. Cox JL, Holden JM, Sagovsky R: Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry1987, 150: 782-786. 10. Thorpe K, Dragonas T, Golding J: The effects of psychosocial factors on the mother’s emotional well-being during early parenthood: A cross-cultural study of Britain and Greece. Journal of Reproductive and Infant Psychology, 1992, 10: 205-217. 11. Cox J, Holden J: Perinatal mental health: a guide to the Edinburgh Postnatal Depression Scale. 2003, Royal College of Psychiatrists. 12. Gonidakis F, Leonardou AA: Maternity blues and Post-partum depression. Findings from Greece. European Psychiatric Review, 2009, 1 (2): 21-23. 13. Boufidou F, Lambrinoudaki I, Argeitis J, Zervas IM, Pliatsika P, Leonardou A, Petropoulos G, Hasiakos D, Papadias K, Nikolaou C: CSF and plasma cytokines at delivery and postpartum mood disturbances. Journal of Affective Disorders, 2009, 115 (1): 287-292. 14. Lambrinoudaki I, Rizos D, Armeni E, Pliatsika P, Leonardou A, Sygelou A, Argeitis J, Spentzou G, Hasiakos D, Zervas I, Papadias C: Thyroid function and postpartum mood disturbances in Greek women. Journal of Affective Disorders, 2009, doi: 10.1016/j.jad.2009.07.001.
Competing interests
The authors declare that they have no competing interests.
Validation of the Edinburgh Postnatal Depression Scale- Previous findings in another sample of Greek women and review of the differences with the newer results.
11 January 2010
Dear Sirs,
In the interesting paper The Edinburgh Postnatal Depression Scale: translation and validation for a Greek sample, the authors mention that “the purpose of this study was to validate the Greek version of the EPDS…”. We would like to inform you that our group has already published a validation of the EPDS based on a different translation and with a different methodology, which has been published in the Journal of Reproductive and Infant Psychology [1]. The fact that our study was not mentioned in the paper by Vivilaki et al. (whereas a study using our translation was referenced therein) [2], might not be significant if the two studies did not produce such different results. We believe that it is important to review here the methodological differences between our study and the study by Vivilaki et al., so that the clinicians and researchers interested in utilising the Greek EPDS can decide on which cut-off they should use in evaluating scores.
Our study was conducted in a sample of new mothers who have given birth in both private and public large obstetric hospitals in Athens. We approached 109 women, of which 95 (87%) agreed to participate in the study. Our initial recruitment of the study subjects took place on the second postpartum day, where initial questionnaires, as well as demographic data were collected. The reassessment of the study subjects took place in the 8th postpartum week. At that time the researcher, blind to the results of the initial questionnaires, conducted a semi-structured interview (SCID), to ascertain the presence of major depression. The diagnosis of minor depression, based on DSM-IV research criteria, was added to the assessment, as it has also been done elsewhere, as 50% of the cases, if unnoticed and untreated, will later suffer from major depression [3]. In the validation study by Vivilaki et al. the recruitment of the study subjects took place between 4 days till 16 weeks postpartum. This poses a major issue regarding the diagnosis of postpartum depression. Newly delivered mothers at 4 days postpartum could be suffering from maternity blues, whereas the women assessed in the 16th week postpartum, could indeed be suffering from postpartum depression. “Postpartum blues”, also known as “5th day blues” is a condition affecting 30-80% of women during the first 3-15 postpartum days, it is significantly influenced by the hormonal fluctuation and it is usually self-limited [4]. Postpartum depression usually manifests itself at 6-8 weeks postpartum. This to our opinion poses an issue on the homogeneity of the group assessed for depressive symptoms in the Vivilaki et al study. Our results showed 5% major and 7.4% minor depression, which is consistent with the percentage found by other authors [5]. The high 31.7% diagnosis of mild depression in the BMC study could be accounted by the inclusion of maternity blues cases.
In our opinion this difference accounts for the rather unexpected difference in the cut-off points in the two Greek validation studies: our validation study shows a cut-off point of 11/12, whereas the newer validation study showed a cut-off point of 8/9, which in our opinion is too low and would include a lot of false positive cases.
Unpublished data of our study have shown that the EPDS on the second postpartum day has poor psychometric properties, regarding the ‘prediction’ of postpartum depression in the 6-8th week postpartum. ROC analysis showed that at a cut-off point of 8/9 had a sensitivity of 90.00% and a specificity of 53.52%, which means that one of the two women identified by the questionnaire as possibly depressed would be falsly alarmed. So we have found that the use of the EPDS on the first postpartum days is not encouraged, as it has already been shown elsewhere [6].
In contrast to our procedure, the validation study by Vivilaki et al. has validated the questionnaire against another, already validated questionnaire, a procedure as well already adopted by others. However it is well known that using questionnaires to diagnose depression shows a higher prevalence of the disorder [7]. This fact along with the timing of the assessment could account according to our opinion for the cut-off point of the scale being found in this newer validation study a lot lower than in the initial one.
Finally, there is another point to be made, regarding the use of the scale solely as a screening instrument. In the current paper that we comment on it is mentioned that the Greek EPDS could be used by Greek health professional for diagnostic purposes. However as the National Screening Committee [8] has already clearly stated, as well as the constructors of this questionnaire [9], we would like to suggest that it is our responsibility to teach health professional willing to use this screening questionnaire that it is NOT a diagnostic tool, and thus it can only identify possibly depressed subjects, but can not and should not be used for making the diagnosis of depression.
We do not believe that the different translation of the EPDS used by Vivilaki et al. played a significant part in the results reported. The EPDS is a fairly easy scale to translate. However, it might be confusing to users to have many different versions of the same scale, however small differences might be. Our translation came from a previous version used in Greece for over a decade, and was in a study that was published in the same journal our validation study was published [10]. It is also the only one referenced by the originator of the EPDS scale, Prof. J. Cox [11]. Our translation has already been utilized by three other Greek studies [12, 13, 14]. All the above mentioned publications can be found through the Internet when the key words in Google search engine are ‘EPDS, Greek’.
Best Regards
Angeliki A. Leonardou, Ioannis M. Zervas
References
1. Leonardou AA, Zervas YM, Papageorgiou CC, Marks MN, Tsartsara EC, Antsaklis A, Christodoulou GN, Soldatos CR: Validation of the Edinburgh Postnatal Depression Scale and prevalence of postnatal depression at two months postpartum in a sample of Greek mothers. Journal of Reproductive and Infant Psychology, 2009, 27 (1): 28-39.
2. Gonidakis F, Rabavilas AD, Varsou E, Kreatsas G, Christodoulou GN: A 6-month study of postpartum depression and related factors in Athens Greece. Comprehensive Psychiatry 2008, 49: 275-282.
3. Horwath E, Johnson J, Klerman GL: Depressive symptoms as relative and attributable risk factors for first-onset major depression. Archives of General Psychiatry, 1992, 49: 817-823.
4. Kennerley H., Gath D: Maternity blues. I. Detection and measurement by questionnaire. British Journal of Psychiatry, 1989, 155: 356-362.
5. Gavin N, Gaynes B, Lohr K, Meltzer-Brody S, Gartlehner G, Swinson T: Perinatal Depression. A Systematic Review of Prevalence and Incidence. Obstetrics and Gynecology, 2005, 106 (5): 1071-1083.
6. Lee DTS., Yip ASK., Chan SSM, Tsui MHY, Wong WS, Chung TKH: Postdelivery screening for postpartum depression. Psychosomatic Medicine 2003, 65: 357-361.
7. Garcia-Esteve L, Ascaso C, Ojuel J, Navarro P: Validation of the Edinburgh Postnatal Depression Scale (EPDS) in Spanish mothers. Journal of Affective Disorders, 2003, 75: 71-76.
8. National Screening Committee: www.nelh.nhs.uk/screening. 2001
9. Cox JL, Holden JM, Sagovsky R: Detection of postnatal depression. Development of the 10-item Edinburgh Postnatal Depression Scale. British Journal of Psychiatry1987, 150: 782-786.
10. Thorpe K, Dragonas T, Golding J: The effects of psychosocial factors on the mother’s emotional well-being during early parenthood: A cross-cultural study of Britain and Greece. Journal of Reproductive and Infant Psychology, 1992, 10: 205-217.
11. Cox J, Holden J: Perinatal mental health: a guide to the Edinburgh Postnatal Depression Scale. 2003, Royal College of Psychiatrists.
12. Gonidakis F, Leonardou AA: Maternity blues and Post-partum depression. Findings from Greece. European Psychiatric Review, 2009, 1 (2): 21-23.
13. Boufidou F, Lambrinoudaki I, Argeitis J, Zervas IM, Pliatsika P, Leonardou A, Petropoulos G, Hasiakos D, Papadias K, Nikolaou C: CSF and plasma cytokines at delivery and postpartum mood disturbances. Journal of Affective Disorders, 2009, 115 (1): 287-292.
14. Lambrinoudaki I, Rizos D, Armeni E, Pliatsika P, Leonardou A, Sygelou A, Argeitis J, Spentzou G, Hasiakos D, Zervas I, Papadias C: Thyroid function and postpartum mood disturbances in Greek women. Journal of Affective Disorders, 2009, doi: 10.1016/j.jad.2009.07.001.
Competing interests
The authors declare that they have no competing interests.