Diagnostic Accuracy of the Stepped Screening Protocol and Its Screening Tools in the Perinatal Period.

This research follows an ongoing study at UZ Gent (registration number B670201836088).

In order to investigate the stepped screening instrument proposed therein for its value in clinical practice, it is important to 1) gain insight into the diagnostic accuracy of its individual screening instruments and 2) examine whether the implementation of the stepped screening protocol leads to better predictability in identifying women at increased risk of a major depressive episode and/or anxiety disorders, with inclusion of posttraumatic stress disorder (PTSD) and obsessive-compulsive disorder (OCD) in the perinatal period.

The diagnostic accuracy of the Edinburgh Postnatal Depression Scale (EPDS) shows good accuracy for major depression in both the ante- and postnatal periods and for any psychiatric disorder postnatally. The Whooley questions also have high diagnostic accuracy for depression and any psychiatric disorder in pregnancy and postnatally . Negative responses to both Whooley questions excluded with acceptable accuracy both perinatal depression and anxiety disorder.

For anxiety instruments, research to date is less unequivocal. The NICE Guidelines recommend the use of the Generalized Anxiety Disorder (GAD-2) (cut-off ≥ 3) for screening for anxiety symptoms based on research in a non-perinatal population (NICE - 2011). However, research shows that the GAD-2 is not specific enough to detect anxiety disorders in general, given that the population prevalence of generalized anxiety disorder (GAD) is only 5% (compared to 17% for all anxiety disorders). As possible good alternatives to the GAD-2, the use of the EDS-3A (a subset of the EPDS) and the GAD-7 are recommended.

This study aims to examine whether the screening protocol using the Whooley (clinical score, score 1) and GAD-2 (clinical score, score 2) as triage instruments followed by the EPDS/EDS-3A and GAD-7 should be updated for detecting a depressive episode, anxiety disorder or another psychiatric disorder ((hypo)manic episode in the context of bipolar disorder, OCD and/or PTSD). To our knowledge, to date, no studies are known concerning the validation of the above-mentioned screening instruments, nor concerning prevalence rates for the above-mentioned disorders for Flanders. This study will contribute to that.

This study will be conducted through hospitals encouraging their pregnant and/or recently delivered mothers to participate in this study by means of flyers. If there is a shortage of participants, this study can also be spread via the website and social media of https://wolkinmijnhoofd.be (initiative of Gents Perinataal Netwerk).

Through a QR code (different for each site) interested parties can log in, after which they receive more information about the study (ICF letter) and, if they wish to participate, are led to the online informed consent. After this, (future) moms are asked to fill in demographic data and answer questionnaires. All questionnaires will be administered from all participants. Based on a positive score on one of the questions of the Whooley and/or GAD-2, a standardized diagnostic interview, SCID-5 (American Psychiatric Association, 2017) will also be administered, to determine whether a depressive episode and/or anxiety disorder or another psychiatric disorder ((hypo)manic episode i.k.a. bipolar disorder, OCD and/or PTSD) is effectively present. The diagnostic interview will be conducted by telephone by a trainee psychiatrist and/or psychologist who will be blinded to the screening results. A sample of those individuals who obtain negative scores on the triage questions (Whooley and GAD-2) will serve as a control group, after which the above-mentioned standardized interview will also be conducted from them. This should allow us to determine the diagnostic accuracy of the screening instruments based on their estimated sensitivity and specificity with respect to the structured interview, the SCID-5 and this for different cut-off scores.

The sample size is based on calculating Wilson score confidence intervals for proportions. A sample size of at least 95 positive scores on the EDS-3A and 95 negative scores on the EDS-3A allows us to estimate the PPV, NPV, and prevalence in each of the groups, respectively, using a Wilson score confidence interval with a half-width of 10% (total width 20%) with a probability of greater than 90%. It is estimated that 10% will score positive on the EDS-3A finally need a sample size of 950 (95* 1/0.2) women, antenatal and postnatal each, or until the beholden number of positive scores on the EDS-3A is reached. Depending on the number of hospitals wishing to participate in this study, the duration of recruitment is estimated to be 24 months.

Data processing Data collection and processing will be through REDCaP. The positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+) and negative likelihood ratio (LR-) of the GAD-2, EDS-3A and GAD-7 will be calculated for anxiety disorders (generalized anxiety, panic disorder, agoraphobia without panic disorder, social phobia, specific phobia). In addition, differential diagnostic also depressive episode (uni- and bipolar), PTSD and OCD will be included in this study. Positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+) and negative likelihood ratio (LR-) will also be calculated for major depression for the Whooley and the EPDS.

A statistician will be used for randomization.